London Road.Pdf

LIFE AND DEATH IN AROMAN CITY

Excavation of aRoman cemetery with amass grave at 120–122 London Road, Gloucester

By Andrew Simmonds, Nicholas Ma´ rquez-Grant and Louise Loe

with contributions by

Paul Booth, Angela Boyle, Carolyn Chenery, Hilary Cool, Jonny Geber, Tim Haines, Kevin Hayward, Martin Henig, Rebecca Nicholson, Cynthia Poole, Kelly Powell, Ruth Shaffrey, Danielle Schreve, Mike Simms, Alex Smith, Jane Timby, Roger Tomlin and Fay Worley

Illustrations by

Frances Chaloner, Markus Dylewski, Anna Komas, Sarah Lucas, Lucy Martin, Mary Saunders, Georgina Slater and Magdalena Wachnik

Oxford Archaeology Monograph No. 6 2008 The publication of this volume has been generously funded by McCarthy and Stone (Developments) Ltd.

Edited by Alex Smith

q 2008 Oxford Archaeological Unit www.thehumanjourney.net

ISBN 978-0-904220-49-0

Typeset and printed in Europe by the Alden Group, Oxfordshire Contents

List of Figures ...... ix List of Plates ...... xi List of Tables ...... xiii Summary ...... xv Acknowledgements ...... xvi

CHAPTER 1: INTRODUCTION ...... 1 Location,Geologyand Topography...... 1 Archaeological Background ...... 1 The development of Roman Gloucester ...... 1 The cemeteries of Roman Gloucester ...... 6 The Wotton cemetery ...... 6 The hospital of St Mary Magdalen and post-medieval developmentofthe site...... 7 The Excavation...... 7 Previous evaluation of the site ...... 7 The stages of excavation ...... 7 Excavation methodology ...... 7 Location of the archive...... 8

CHAPTER 2: ARCHAEOLOGICAL DESCRIPTION ...... 9 Geology and Pleistocene Faunal Remains...... 9 Activity Pre-dating theCemetery...... 9 The Roman Cemetery...... 9 Pre-Flavianburials...... 9 Late1st–early 2nd century burials...... 12 2nd century or later burials...... 13 3rd–4th century burials ...... 13 Mass grave ...... 14 Other features...... 18 The tombstones...... 18 Gully1358...... 19 Pit1301...... 19 The Burials ...... 19 The cremation burials ...... 19 Urned cremation burials...... 19 Un-urnedcremation burials...... 19 The inhumation burials ...... 20 The grave pits...... 20 The body ...... 21 Discrepantburialrites:prone, decapitated and disarticulated burials...... 21 Cofﬁns and other containers...... 24 Grave goods...... 25 Ceramic grave goods ...... 25 Footwear ...... 25 Jewellery...... 26 Animal remains...... 26 Medieval and Post-medieval Activity ...... 26 Radiocarbon Dating ...... 28

CHAPTER 3: THE HUMAN REMAINS...... 29 Unburnt Human Bone by Nicholas Ma´ rquez-Grantand Louise Loe...... 29 Introduction...... 29 Skeletal biology...... 29

iii Life andDeathinaRoman City

Osteological methodology...... 29 Inventory, conditionand completeness...... 29 Estimation of biological age andsex ...... 29 Metric andnon-metric analysis ...... 30 Taphonomy ...... 30 Background to the Crossbones methodology...... 31 Comparative samples ...... 32 Results...... 32 The discreteinhumations...... 32 Condition and completeness...... 32 Biological age and sex...... 32 Metrical analysis ...... 33 Non-metric traits ...... 33 The mass grave...... 34 Condition and completeness...... 34 Taphonomy ...... 36 Minimum numberofindividuals...... 38 Recovery rate ...... 39 Biological age and sex...... 40 Metrical analysis ...... 40 Non-metric traits ...... 40 Dental health status by Nicholas Ma´ rquez-Grant...... 41 Introduction ...... 41 Materials...... 42 Subadultsample ...... 42 Adult sample...... 42 Methods ...... 42 Results ...... 43 The discreteburial assemblage ...... 43 The adult sample ...... 43 The subadultsample...... 45 The mass graveassemblage ...... 46 Comparison with the discreteskeletons ...... 48 The subadultsample...... 49 Palaeopathology by Louise Loe ...... 49 Methods ...... 49 Results ...... 50 Discrete inhumations...... 50 The mass grave...... 61 The disarticulated material by Nicholas Ma´ rquez-Grant...... 66 Isotope analysis ...... 67 Discussion of theunburnt human remains by Louise Loe, Nicholas Ma´ rquez-Grant and Jonny Geber ...... 67 Somecaveats ...... 67 The under-representation of the youngest...... 69 The mortality proﬁles ...... 69 Taphonomy...... 70 Pathology and physical attributesamong the discreteskeletons and the mass graveskeletons..... 71 Conclusions and future potential ...... 72 The Burnt Human Bone by Nicholas Ma´ rquez-Grant ...... 72 Introduction ...... 72 Methods...... 72 Background: the cremation process ...... 73 Results...... 73 Condition and completeness ...... 73 Weight and skeletal part representation ...... 73 Minimum numberofindividuals, sex and age determination ...... 75 Non-metric traits ...... 76 Fragmentation ...... 76 Colour ...... 76

iv Contents

Palaeopathology...... 76 Periostitis ...... 76 Joint disease ...... 77 Dental conditions ...... 77 Further observations ...... 77 Discussion of thecremation burials ...... 77 Skeletal biology and palaeopathology ...... 77 Efﬁciency of cremationand the funerary rite...... 78 Cremated human bone from non-cremation burial contexts ...... 79

CHAPTER 4: THE GRAVE CATALOGUE ...... 81 Introduction ...... 81 CatalogueofGraves...... 81 Inhumation grave1143...... 81 Inhumation grave1144...... 81 Inhumation grave1145...... 81 Inhumation grave1146...... 83 Inhumation grave1147...... 84 Inhumation grave1148...... 84 Inhumation grave1150...... 84 Inhumation grave1151...... 84 Inhumation grave1167...... 85 Inhumation grave1218...... 86 Inhumation grave1219...... 86 Inhumation grave1228...... 86 Inhumation grave1229...... 86 Inhumation grave1230...... 86 Inhumation grave1234...... 87 Inhumation grave1240...... 87 Inhumation grave1243...... 87 Inhumation grave1246...... 87 Inhumation grave1264...... 87 Inhumation grave1275...... 87 Inhumation grave1283...... 88 Inhumation grave1284...... 88 Inhumation grave1288...... 88 Inhumation grave1315...... 88 Inhumation grave1319...... 88 Inhumation grave1330...... 89 Inhumation grave1334...... 89 Inhumation grave1344...... 89 Inhumation grave1352...... 89 Inhumation grave1353...... 89 Inhumation grave1362...... 90 Inhumation grave1369...... 90 Inhumation grave1370...... 90 Inhumation grave1374...... 90 Inhumation grave1388...... 90 Inhumation grave1398...... 91 Inhumation grave1403...... 91 Inhumation grave1407...... 91 Inhumation grave1422...... 91 Inhumation grave1427...... 91 Inhumation grave1501...... 91 Inhumation grave1503...... 92 Inhumation grave1504...... 92 Inhumation grave1505...... 92 Inhumation grave1507...... 92 Inhumation grave1508...... 92

v Life andDeathinaRoman City

Inhumation grave1720...... 93 Inhumation grave1721...... 93 Inhumation grave1722...... 93 Inhumation grave1723...... 93 Inhumation grave1724...... 93 Inhumation grave1755...... 94 Inhumation grave1756...... 94 Inhumation grave1757...... 94 Inhumation grave1758...... 94 Inhumation grave1759...... 94 Inhumation grave1760...... 94 Inhumation grave1761...... 94 Inhumation grave1762...... 95 Inhumation grave1763...... 95 Inhumation grave1764...... 95 Inhumation grave1765...... 95 Inhumation grave1772...... 95 Inhumation grave2027...... 95 CatalogueofCremations ...... 96 Cremation burial 1196...... 96 Cremation burial 1209...... 96 Cremation burial 1227...... 96 Cremation burial 1266...... 96 Cremation burial 1766...... 97 Cremation burial 1767...... 97 Cremation burial 1768...... 97 Cremation burial 1769...... 97 Cremation burial 1770...... 97

CHAPTER 5: THE FINDSAND ENVIRONMENTAL EVIDENCE ...... 99 The Pottery by Jane Timby...... 99 Introduction ...... 99 Fabrics and forms ...... 99 Pottery from burials ...... 99 Pottery from cremation burials ...... 99 Pottery from inhumations and ‘memorial’ ...... 101 Pottery from grave backfills...... 101 Pottery from the mass grave...... 101 Pottery from buried soils ...... 102 Discussion ...... 102 Catalogueofillustrated sherds...... 103 The Coins by Paul Booth ...... 103 The Small Finds by Hilary Cool...... 104 Introduction ...... 104 The glass vessels...... 104 Small finds from cremation burials ...... 104 Cremation burial 1209...... 104 Cremation burial 1266...... 105 Cremation burial 1766...... 110 Small finds from inhumation burials ...... 110 Inhumation grave 1144...... 110 Inhumation grave 1151...... 110 Inhumation grave 1230...... 110 Inhumation grave 1246...... 111 Inhumation grave 1352...... 111 Inhumation grave 1362...... 111 Inhumation grave 1505...... 112 Small finds from the mass grave ...... 113 Small finds from other contexts...... 115

vi Contents

The Nails by Kelly Powell...... 115 Introduction...... 115 Nailsfrom cremation burials ...... 115 Nailsfrom inhumation burials...... 115 Structural nails...... 115 Hobnails...... 115 Nailsfrom the mass grave ...... 115 Discussion ...... 116 The Sculptural Stone by MartinHenig and Roger Tomlin...... 116 Tombstone1...... 116 Discussion...... 117 Tombstone2...... 117 Discussion...... 118 Other Worked or Utilised Stone by RuthShaffrey...... 118 The Ceramic Building Material by Cynthia Poole...... 118 The Animal Bone by Fay Worley...... 119 Introduction...... 119 Results...... 119 Animal bone from cremation burials ...... 120 Animal bone from inhumation burials ...... 120 Possible inhumation burial or ‘memorial’ 1149 ...... 120 Material from grave back-ﬁlls...... 120 Animal bone from the mass grave ...... 120 Animal bone from other features ...... 121 Discussion ...... 121 The Fish Bone by RebeccaNicholson ...... 122

CHAPTER 6: DISCUSSION OF THE ROMAN CEMETERY ...... 123 Introduction ...... 123 The Development of the Wotton Cemetery ...... 123 The siting of the cemetery ...... 123 The growth of thecemetery ...... 123 The layout and organisation of thecemetery...... 125 Funerary Rites...... 127 The funerary process...... 127 Cremation burials ...... 127 The pyre ...... 127 The grave ...... 128 Inhumation burials ...... 130 Placing the body...... 130 Discrepantburialrites:prone and decapitatedburials ...... 131 Containing the body: cofﬁns, cists and shrouds...... 133 Adorning the body ...... 134 Accompanying the body: food and drink...... 134 Accompanying the body: footwear...... 135 Commemorating the dead...... 135 Grave markers...... 135 Mourning and feasts ...... 136 Cenotaphs, memorials and ritual deposits ...... 137 Funerary rites for children ...... 138 Burial and Belief...... 138 The Mass Grave...... 139 The mass graveand the Antonine Plague...... 140 The PeopleofRoman Gloucester ...... 141 The structure of thepopulation...... 141 Health...... 143 Status and society ...... 143 The peopling of Roman Gloucester...... 144

vii Life andDeathinaRoman City

APPENDICES...... 147 Appendix 1: Dental Health Status Methodology...... 147 Appendix 2: Summary of Disarticulated Human Bone...... 148 Appendix 3: Oxygen and Strontium AnalysisofToothEnamel by Carolyn Chenery...... 150 Appendix 4: Detailed RecordsofCremations ...... 154 Appendix 5: Photomicrographs of the Tombstones by Kevin Hayward ...... 159 Appendix 6: The Pleistocene Vertebrate Remains by Dr DanielleSchreve...... 160

BIBLIOGRAPHY ...... 165

INDEX...... 175

viii List of Figures

CHAPTER 1 Figure 1.1 Site location map...... 2 Figure 1.2 Thegeology and hydrology of the Gloucester area ...... 3 Figure 1.3 TheGloucester area during the Roman period...... 3 Figure 1.4 Plan of ﬁndspots and excavations relatedtothe Wotton cemetery...... 4 Figure 1.5 Plan of the site showing the phases of excavation ...... 8

CHAPTER 2 Figure 2.1 Phased plan of all archaeological features...... 10 Figure 2.2 Theearly phase of the cemetery (1st–early 2nd century)...... 11 Figure 2.3 Thelater phase of the cemetery (2nd–4thcentury) ...... 14 Figure 2.4 Thesequenceofintercutting burials in the south-eastern corner of the excavation...... 15 Figure 2.5 Thedistribution of cremation burials...... 20 Figure 2.6 Thedistribution of inhumation burials...... 22 Figure 2.7 Post-medieval features ...... 27

CHAPTER 3 Figure 3.1 Mortality proﬁle(discrete burials) ...... 33 Figure 3.2 Three dimensional analysis of articulated skeletons with and without lowerlimb bones in the mass grave ...... 37 Figure 3.3 Three dimensional analysis of articulated skeletons with and without upperlimb bones in the mass grave...... 38 Figure 3.4 Mortality proﬁle(mass grave)...... 41 Figure 3.5 Sexdistribution of the adultsample(mass grave)...... 41 Figure 3.6 Periostitis distribution: discrete burials and mass grave (TPR) ...... 62 Figure 3.7 Ante-mortem fracture distribution: cemetery and massgrave (TPR)...... 64

CHAPTER 4 Figure 4.1 Plan of all burials ...... 82 Figure 4.2 Plan of inhumation burial 1145 ...... 83 Figure 4.3 Plan of inhumation burial 1243 ...... 83 Figure 4.4 Plan of inhumation burial 1334 ...... 83 Figure 4.5 Plan of inhumation burial 1352 ...... 83 Figure 4.6 Plan of inhumation burial 1374 ...... 84 Figure 4.7 Plan of inhumation burial 1505 ...... 84 Figure 4.8 Plan of cremation burials 1196, 1209, 1266 and 1767...... 85

CHAPTER 5 Figure 5.1 Romanpottery (1–8) ...... 100 Figure 5.2 Romanpottery (9–15) ...... 102 Figure 5.3 Gaming countersfrom cremationburial 1266...... 106 Figure 5.4 Other objects from cremation burial 1266 ...... 109 Figure 5.5 Iron ring from inhumation burial 1246 ...... 111 Figure 5.6 Unguent bottles from inhumation grave 1352 ...... 112 Figure 5.7 Bracelet from inhumation grave 1362 ...... 112 Figure 5.8 Fingerring from inhumation grave 1505 ...... 113 Figure 5.9 Smallﬁnds from massgrave 1483 ...... 113

ix Life andDeathinaRoman City

CHAPTER 6 Figure 6.1 Thegrowth of the Wotton cemetery ...... 124 Figure 6.2 Orientation of inhumation burials, according to the position of the head (each segment represents an individual inhumation burial) ...... 131 Figure 6.3 Thedistribution of inhumation burialstowhich abiological sex could be assigned...... 142

APPENDICES Figure A3.1 Plotofenamel strontium and oxygen isotope results...... 153

x List of Plates

CHAPTER 2 Plate 2.1 Pit 1149 ...... 12 Plate 2.2 Truncated pit 1306 ...... 12 Plate 2.3 Feature 1352 ...... 12 Plate 2.4a–d The mass grave...... 16 Plate 2.5 Skeleton 1520 in mass grave...... 16 Plate 2.6 Skeleton 1524 in mass grave...... 17 Plate 2.7 Skeleton 1564 in mass grave...... 17 Plate 2.8 Entangled skeletons 1558, 1562 and 1564 within the massgrave ...... 17 Plate 2.9 Skeleton 1553 in mass grave...... 18 Plate 2.10 Tombstone 1...... 18 Plate 2.11 Cremation burial 1766...... 21 Plate 2.12 Cremation burial 1266...... 21 Plate 2.13a–d Supine burials, showing some of the varietyofarm positions encountered...... 23 Plate 2.14 Proneinhumation burial 1145...... 24 Plate 2.15 Proneinhumation burial 1143...... 24 Plate 2.16 Skeleton 1232, grave 1234...... 25 Plate 2.17 Inhumation burial1334 ...... 25 Plate 2.18 Stone cist in inhumation grave1756...... 25 Plate 2.19 Burial 1286 in grave 1288...... 26

CHAPTER 3 Plate 3.1 Permanent teeth from Skeleton 1277 ...... 46 Plate 3.2 Dentition belongingtoSkeleton 1672 ...... 48 Plate 3.3 Non-specific inflammation(highlighted) involving the rightinnominate bone of Skeleton1372...... 53 Plate 3.4 Thorough-going periosteal reaction involving rightand left tibiae of Skeleton1286...... 53 Plate 3.5 Possible old healed fracture involving the right tibia shaft belonging to Skeleton1405...... 55 Plate 3.6 Healed blunt trauma involving the occipital boneofthe skull from Skeleton 1112...... 56 Plate 3.7 Pseudo cut marksobserved on the left femur from Skeleton 1496 ...... 57 Plate 3.8 Possible Legg-Calve´ -Perthes disease involving the left femur belonging to Skeleton1286...... 59 Plate 3.9 Oval, lytic lesions involving the lateral side of the right talus bone of Skeleton1510...... 59 Plate 3.10 Scalloped erosions involving the heads of the right and left greattoes from Skeleton1453 ...... 59 Plate 3.11 Lytic lesion involving the head of the left mandibular condyle. Skeleton 1112 ...... 60 Plate 3.12 Oval lytic lesions involving the necksofrightand left femora. Skeleton 1337...... 61 Plate 3.13 Fractured left humerus belongingtoSkeleton 1524...... 64 Plate 3.14 Destructive lesions involving athoracic vertebrafrom Skeleton 1542...... 66 Plate 3.15 Amenoblastoma involving the left sideofthe mandible from an adultmale (small find 716)...... 66

CHAPTER 5 Plate 5.1 Meltedunguent bottle fromcremation burial 1209 ...... 104 Plate 5.2 Cinerary urn 1252, burial 1266, before excavation of the contents in the laboratory...... 105 Plate 5.3 Gaming counters, dice and melonbead from cremationburial 1266...... 105 Plate 5.4 Meltedunguent bottle fromcremation burial 1766 ...... 110 Plate 5.5 Copperalloybracelet (sf 464) worn on the right forearm of skeleton 1596...... 114 Plate 5.6 Tombstone 1aftercleaning and conservation ...... 116 Plate 5.7 Tombstone 2aftercleaning and conservation ...... 117

xi Life andDeathinaRoman City

CHAPTER 6 Plate 6.1 Burial 1313 in grave 1315...... 133

APPENDICES Plate A5.1 Photomicrograph of tombstone 1...... 159 Plate A5.2 Photomicrograph of tombstone 2...... 159 Plate A5.3 Comparative photomicrograph of Painswick Stone...... 159 Plate A6.1 Partial femoral shaft of Hippopotamusamphibius ...... 160 Plate A6.2 Comminuted upper canine of Hippopotamus amphibius ...... 160 Plate A6.3 Virtually complete ?LM3of Hippopotamusamphibius ...... 160 Plate A6.4 Reconstructed cervical vertebra of Hippopotamus amphibius ...... 161

xii List of Tables

CHAPTER 1 Table 1.1 Known archaeological sitesrelatedtothe Wotton cemetery ...... 5

CHAPTER 2 Table 2.1 Radiocarbon dates from 120–122 London Road, Gloucester ...... 28

CHAPTER 3 Table 3.1 Number of discrete burialsthat could be assigned to finer dateranges ...... 29 Table 3.2 Age categories employed in thisstudy...... 30 Table 3.3 Categories employed in sex determination...... 30 Table 3.4 Condition of the skeletons from the discrete burials...... 32 Table 3.5 Completeness of the skeletons fromthe discrete burials...... 33 Table 3.6 Age-at-death distribution of the discrete burials ...... 33 Table 3.7 Age distribution in the adultsample (N=51) according to biological sex...... 34 Table 3.8 Bones from the discrete burial assemblage employedinthe computation of stature...... 34 Table 3.9 Staturesummary datafrom the discrete burial assemblage...... 34 Table 3.10 Cranial non-metric traitfrequencies in the discrete burial assemblage ...... 35 Table 3.11 Frequency of post-cranial traits...... 35 Table 3.12 Completeness of the articulated skeletons withinthe mass grave...... 36 Table 3.13 Composition of the skeletal small finds assemblage within the mass grave...... 36 Table 3.14 Condition of the articulated skeletons within the massgrave...... 36 Table 3.15 Condition of the small finds withinthe mass grave...... 36 Table 3.16 Condition of the disarticulated elements within the massgrave ...... 36 Table 3.17 MNI calculation for the mass grave assemblage accordingtorepeated anatomical landmarks in different bone elements ...... 39 Table 3.18 MNI calculation of the subadult sample from the mass grave ...... 39 Table 3.19 Skeletal representation andbonerecovery rates for the mass grave assemblage...... 40 Table 3.20 Age-at-death distribution in the massgrave sample ...... 40 Table 3.21 Age distribution in the adultsample (N=76) according to biological sex...... 41 Table 3.22 Cranial non-metric traitfrequencies in the massgrave assemblage...... 42 Table 3.23 Post-cranial non-metric trait frequencies in the mass grave assemblage ...... 42 Table 3.24 Number of teeth present, number lostante-mortem, number lost post-mortem and number generally missing...... 43 Table 3.25 Prevalence of enamel hypoplasia in the discrete burial assemblage ...... 43 Table 3.26 Dental caries prevalence in the discrete burialassemblage ...... 44 Table 3.27 Distribution of pattern of carious lesions in the discrete burial assemblage ...... 44 Table 3.28 AMTL frequencies in the discrete burial population ...... 44 Table 3.29 Dental calculus frequencies in the discrete burial assemblage...... 45 Table 3.30 Periodontal disease frequenciesinthe discrete burial assemblage...... 45 Table 3.31 Periapical cavity frequencies in the discrete burial assemblage ...... 45 Table 3.32 Dental health statusinthe subadult sample from the discrete burial assemblage ...... 46 Table 3.33 Prevalence of enamel hypoplasia in the massgrave assemblage...... 46 Table 3.34 Dental caries prevalence in the mass grave assemblage ...... 47 Table 3.35 Distribution of pattern of carious lesions in the mass grave assemblage ...... 47 Table 3.36 AMTL frequencies in the massgrave assemblage ...... 47 Table 3.37 Dental calculus frequencies in the massgrave assemblage ...... 47 Table 3.38 Periodontal disease frequenciesinthe mass grave assemblage ...... 47 Table 3.39 Periapical cavity frequencies in the mass graveassemblage ...... 48 Table 3.40 Comparison between the discrete burials and the mass grave adultdental assemblages ..... 49 Table 3.41 Comparison between the females and males from the discrete burialand the mass grave assemblage ...... 49 Table 3.42 Dental health statusinthe subadult sample from the mass grave ...... 49

xiii Life andDeathinaRoman City

Table 3.43 Comparison between dental conditions observed among subadults from the discrete burial assemblage and the mass grave ...... 50 Table 3.44 Non-specific inflammation involving the skull...... 50 Table 3.45 Periostitis, absoluteprevalence ...... 51 Table 3.46 Distribution of elements with non-specific inflammation by skeleton...... 52 Table 3.47 Cribra orbitalia, absolute prevalence ...... 54 Table 3.48 Summaryofpossiblefractures...... 54 Table 3.49 Distribution of fractures, absoluteprevalence ...... 55 Table 3.50 Age and sex distribution of skeletons with OA ...... 58 Table 3.51 Frequency of joints affectedwith OA ...... 58 Table 3.52 Age and sex distribution of skeletons with periostitis ...... 61 Table 3.53 Periostitis, absoluteprevalence ...... 62 Table 3.54 Distribution of elements with non-specific inflammation by skeleton (articulated skeletons only) ...... 63 Table 3.55 Cribra orbitalia, absolute prevalence ...... 63 Table 3.56 Distribution of fractures, absoluteprevalence ...... 64 Table 3.57 OA frequency of sites affected...... 65 Table 3.58 Frequency of spines affectedwith Schmorl’s nodes...... 65 Table 3.59 Summaryofthe main osteological findings ...... 68 Table 3.60 Definitions of deposit types ...... 74 Table 3.61 Summaryofcrematedhuman bone...... 74 Table 3.62 Summaryofweights (g) for each of the human cremated bone deposits...... 75 Table 3.63 Summaryofcrematedhuman boneassemblage ...... 76 Table 3.64 Percentage of bonefragmentslarger than 10 mm,between 10–4 mm and smaller than 4mm...... 76 Table 3.65 Percentage of fragments according to weightthat presentedaparticular colour...... 77 Table 3.66 Periostitisdistribution in the human crematedbone assemblage ...... 77 Table 3.67 Human bone from non-cremation burial contexts ...... 78

CHAPTER 5 Table 5.1 Quantification of the pottery assemblage by fabric ...... 101 Table 5.2 The size of the countersinthe gaming set with burial 1266 ...... 106 Table 5.3 Distribution of the material in the urn (burial 1266)...... 106 Table 5.4 Gaming sets from selection of 1st to 3rd century burials...... 107 Table 5.5 The occurrence of intaglios in complete and incomplete simple expanded finger rings ...... 111 Table 5.6 Quantification of Roman ceramic building material by form ...... 118 Table 5.7 The hand collected animal bone assemblage...... 119 Table 5.8 The animal bone assemblage from sieved residues ...... 119 Table 5.9 The cremated animal bone assemblage...... 120 Table 5.10 Fish bones recovered from soil samples taken from burials...... 121

APPENDICES Table A3.1 Details of individuals and teeth samples ...... 150 Table A3.2 Grave and skeletal description...... 151 Table A3.3 Oxygen and strontium isotope results...... 152

xiv Summary

BetweenAugust 2004 and January 2006 Oxford The evidence for funerary rites is described and Archaeology undertook aprogrammeofarchaeolo- the possible beliefs informing them considered. The gical excavationand watching brief in advance of demographic make-up of the population and evi- construction of sheltered apartments on asite dence forstatus based on age and sex have also been formerly occupied by adisusedservice station at examined. Strontium and oxygen isotope analysis 120–122 London Road, Gloucester (NGR SO 843 demonstrated that the population had arange of 189). These investigations, commissioned by CgMs origins,deriving both from the local area and also Consulting on behalf of McCarthy andStone from elsewhere in the province of Britannia,aswell (Developments) Ltd, recorded burials forming part as asmall group probablyfrom the Mediterranean of the Wotton Romancemetery, including arare area and two individuals from another, unidentiﬁed mass grave, along with an assemblage of Pleistocene non-UKsource. vertebrate remains(detailed in Appendix6). The massgrave containedthe remains of at least The partofthe Romancemetery within the area of 91 individuals, thrownhaphazardly intoapit during the excavation comprisedthe remainsofatleast ten the second half of the 2nd centuryAD. The bodies individuals recovered from nine cremationburials showednoevidence of violence and are believedto and 64 inhumation graves. The cemetery was estab- have been the victims of an epidemic. It is tentatively lished during the Neronian period as the burial suggested that thismay have been the Antonine groundfor the fortressatnearby Kingsholm, and Plague, an outbreakofsmallpox that swept across subsequently becameone of the main cemeteries of the RomanEmpire between AD 165 and 189. the colonia that was established at Gloucester follow- After the disuse of the cemetery at the end of the ing the departure of the military. Four crouched Romanperiod the site appears to have been burialsdatingfrom the late 1st–early 2nd centuryare abandoned, with no furtheractivity being recorded interpreted as being the remainsofmembers of the until the area was cultivated during the 11th–15th native populationwho had integrated into the com- centuries. Boundary ditches were dug across the site munityatthe colonia.Inhumation superseded crema- during the 16th–17th centuries, most likely asso- tion as the dominantfunerary riteduring the 2nd ciated with the rebuilding of the adjacent hospital century, after which no further cremation burials of St Mary Magdalen at this time. These features were made. The cemetery continued to be useduntil remained open until theywere levelled when the some time in the 4th century. area was developed during the 19th century.

xv Acknowledgements

Oxford Archaeology would like to thankMcCarthy valuablemanagement support and advice. The and Stone (Developments) Ltd for funding the exca- authors are also grateful to the many otherOAand vationsand post-excavation programme, and Greg non OA staff who worked on the project, both in the Pugh, Steve Weaver and Rob Bourne of CgMs Con- field and at the post-excavation stage. In particular sultingfor acting as the archaeological consultant. we acknowledge the contribution of Peter Hacking RichardSermon, Archaeological Officer for Glouces- to the osteological analysis. ter City Council, also played an importantrole in Oxygen and strontium analysis of tooth enamel facilitating the project. from asample of burialswas carried out by Carolyn The various stages of fieldwork were directed by Chenery at the NERC/BGS Isotope Geoscience Lab- Jonny Geber, Tim Haines, Gerry Thacker, Mike oratory as part of the University of Reading Arch- Simms and Annsofie Witkin, and were managedby aeology Department’sAHRC funded project ‘A AngelaBoyle. Thepost-excavation work was man- Long Way from Home: Diaspora Communities in aged by Alex Smith. Support was provided by Leigh RomanBritain’. Caroline Chenery would like to Allen (finds management), Rebecca Nicholson(en- thankDr. Jane Evans (NIGL/BGS) forproviding vironmental management) and Nicky Scott (archives consultation on strontium interpretation. management). Bob Williams (Director of Operations) Many thanks are due to Dr. John Pearce and Paul and NickShepherd (Head of Fieldwork)gave Booth forreading and commenting on the text.

xvi Chapter 1Introduction

This report describes and analyses the resultsofa along its western boundary. Adepth of up to 1.5 m programme of archaeological excavation carried out of post-medieval and modern deposits was recorded by Oxford Archaeology between August 2004 and during the excavation. At the timeofthe excavation, January 2006 at 120–122 London Road, Gloucester the service station that formerly occupied the site (Fig. 1.1). These investigations were designed to had been demolished and the groundremediated. mitigatethe effectsofthe construction of sheltered This included the removal of its underground fuel apartments on asite formerly occupied by adisused tanks and the back-filling of the resultant pit. service station,asthe site lies within the area of one of the main cemeteries of RomanGloucester. Thework was commissioned by CgMs Consulting on behalf of ARCHAEOLOGICAL BACKGROUND McCarthy and Stone (Developments) Ltd in accor- The developmentofRoman Gloucester dance with acondition attached to the planning permission for the development by Gloucester City At the timeofthe Romaninvasion in AD 43,Glou- Council. cester is believedtohave lain withinthe territory of a tribal group known as the Dobunni. This group has been equatedwith the otherwise unattested ‘Bodun- LOCATION, GEOLOGY AND TOPOGRAPHY ni’, whom the Romanhistorian Dio Cassius records as having surrendered to the invading force following Gloucester lies on the east bank of the River Severn, the defeat of the main resistance at an unnamedriver c 4kmfrom the foot of the Cotswolds escarpment. believedtohave been the Medway (Dio,60.20). The The site is situated1.2 km east of the city centre in Dobunnic territory appears to have passed under the suburb of Wotton, which occupies the north- Romanrule peacefully (or at leastwith very little westernend of agravelridge formedbythe erosion disruption visible in the settlement pattern), anditis of parts of the terrace graveland underlying Lias possible that parts (southern Cotswolds and Upper clay by glacial meltwater flowingwestwardinto the ThamesValley) lay initially within alocally governed Severnfrom the Cotswolds (Fig. 1.2). The ridge is client kingdom (Miles et al. 2007, 385). recorded by the British Geological Survey as partof The first direct intervention in the Gloucester area the fourth (Kidderminster) terrace of the River was the establishmentofafortress at Kingsholm, Severn. It is alignedNW-SE, the ground falling beside aformer channel of the RiverSevern in what away sharply on the north-eastern side intoasmall is nowthe northern part of the city (Fig. 1.3). Tacitus valleycontaining the Wotton Brook, aminor records that in AD 49 the Twentieth Legion was tributary of the Severn, and rather more gradually moved forward from its formerbase at Colchester to to the south-west toward asecond tributary,the subdue the Siluresofsouth Wales (AnnalsXII 32), Twyver, which has been re-alignedand culverted in and this is likely to be the context in whichthe modern times. TheSevernitself has also changedits Kingsholm fortress was built. The precise locationof course since the Roman period, when its eastern the fortress seems to have been selecteddue to the branch is believedtohave followed the course of the likely presence here of anative centre of some size Old Severn, achannel which passed 130 mtothe east and importance, which has only fairly recently been of the present course and which had disappeared by recognised (Hurst 1999b, 119; Timby 1999, 38). The the 19th century(Rowbotham 1978). Beyond the Kingsholm fortress was the destination of the Twyverthe ground rises again onto agravelhillock westernend of Ermin Street, which extended across formerly occupied by the Romancity, and now by southern Britain from Silchester. Abranch road the modern city centre. LondonRoad lies on the line extended south-west from this road along the line of of aRomanroadthat exited the former north gate of the modern LondonRoad towardthe Severn near the Romancity and climbedthe gentle gradient to the modern city centre, where acrossing point was Wottonwhere it joined Ermin Street(Fig. 1.3), the presumably located. The proximity of the fortressto line of which is nowrepresented at this point by this crossing would have provided easy access into Barnwood Road and DenmarkRoad. The site itself is Wales and allowed the army to control traffic mov- located on the south side of London Road, c 80 m ing across and along the river, an additional reason south-west of thisjunction, at NGR SO 8432 1893 for locating it in this area. (Fig. 1.1). It encompassed an areaof c 0.17 ha and The ceramic and numismatic dating evidence was generally fairly level at c 23 mOD, although the indicatesthat the Kingsholm fortresswas abandoned modern topography was clearly the result of during the 60s, at around the same time as anew deliberate levelling as the groundlevel of the site fortresswas established on the site of the modern city was substantially higher than that of alane that ran centre (Hurst 1988, 50). Earliertheories that this

1 Life andDeathinaRoman City

Figure 1.1 Site locationmap.

2 Chapter One

Figure 1.2 The geology and hydrology of the Gloucester area. relocation was prompted by aneed to move to higher was absent from the Gloucester area. The establish- grounddue to ﬂoodingofthe original site have been ment of the city centre fortressformedpartofthe refuted by acloser consideration of the topography of consolidation of the deployment of the British the areaand it is now thought more likely that the garrison afterthe withdrawal from the province of new location was chosen to give abetter command of the Fourteenth Legion in AD 66 (Frere 1991a, 81). It the crossing over the Severn (Hurst 1985, 2–3). This has traditionally been supposed (eg Frere 1991a; was aperiod of frequent alterations to the disposition Webster1993) that the new Gloucester fortress of the legions in Britain, the details of which are still housedthe Second Legion which had moved from only poorly understood, and it is uncertain whether Exeter, with the Twentieth having previously moved the replacement of the Kingsholmfortresswiththat at from Kingsholm to Usk (early 60s?) and then to the city centre represented adirect relocation or Wroxeter ( c AD 66) and Chester(c AD 87). However, whether there was aperiod during which the military Hassall (Hassalland Hurst 1999; Hassall 2000, 63) has

Figure 1.3 The Gloucester area during the Roman period (after Hurst1999, ﬁg. 5).

3 Life andDeathinaRoman City

N 0 500 m

1:25,000

0 200 m

1:5000

Figure 1.4 Plan of findspots and excavations related to the Wotton cemetery. suggested that the Second Legion may have moved (Wacher 1995,150), which would suggest that it was directlyfrom Exeter to Caerleon, with the Twentieth founded during the reignofNerva (AD 96–98). remaining in Gloucester until c AD 75, whereupon it However, an alternative suggestion is that the moved to Wroxeter. conversion of the fortresstocivilian use is repre- The subsequent historyofthe Gloucester fortress sented by amajor programme of rebuilding identi- and the date of its conversion into a colonia,a fied in excavations at anumber of locations and settlement forretired soldiers, is still amatter of dated to shortly after AD 86 (Hurst 1972; 1974). some debate. Atombstone found at Rome of a According to this argument the colonia would have soldierofthe Sixth Legion who originatedfrom been founded under Domitianand renamedfollow- Gloucester indicates that the official name of the ing his assassination and subsequent damnatio colonia was Colonia Nerva (or Nerviana) Glevensium memoriae (Hassall and Hurst 1999, 183–4). It is

4 Chapter One

Table 1.1 Known archaeological sites related to the Wotton cemetery (see Fig. 1.4).

No. Location Comments Reference

150–52 London Road Two inhumation burials, sealed by ametalled surface, were Rawes 1983 found during the digging of service trenches in 1983. 267London Road Disarticulated human remains were found here in 1961. Heighway 1980, 63 383–89 London Road Two cremation burials and two inhumationswere recorded Sermon 2003, 55 during awatching brief at the street frontage. One of the cremations was contained within an early 2nd centuryurn. 493/95 London Road Alarge number of cremations in cinerary urns were found in Fullbrook-Leggatt 1876 on the site of a19th century villa in the vicinity of the 1933, 87 current 93/95 London Road. Most of the cinerary urns from these burials have been lost, but the three that remain are now in Gloucester City Museum. 5StMargaret’s Hospital Cremation burials contained in glass and ceramic vessels placed Heighway 1980, 63–4; in stone cists were found between 1861 and 1864. Rawes 1993, 223–4 Awatching brief undertaken during the digging of service trenches in 1976 recorded five cremation burials of late 1st –2nd century date fringing an extensive metalled surface, and six inhumationsassociated with awall revetting an area of raised ground that may have been part of amausoleum. Asecond watching brief in 1992 uncovered afurther seven inhumations,two cremation burials,part of ametalledsurface and astone structurethat may be asecond mausoleum. 6110 London Road Agroup of three inhumations were discovered in 1886 during Heighway 1980, 63; building work. Afurther burial was recorded in awatching brief Rawes and Wills 1998 in 1998. 7StMary’s Hospital Burials were found in 1856 in the graveyardofthe leper hospital Fullbrook-Leggatt of St Mary Magdalen 1933, 88 83Denmark Road Asingle adult inhumation was found during building work in Atkins and Garrod 1988, 1987. 216; Rawes 1992, 221 Subsequently, agroup of eight inhumations, six cremation burials of 2nd century date, and the stone footing of apossible funerary monument were recorded during awatching brief undertaken on the digging of foundations and service trenches in 1991. 9124–130 London Road During the 19th century agrave containingaflanged bowl and Heighway 1980, 63; coins issued by the emperor Gratian (AD 367–83) was found at Gloucester Archaeology 124 London Road. 1993; Sermon 1996; An archaeological evaluation undertaken in 1993 recorded 17 Foundations Archaeology inhumationsand asubsequent watching brief to the east of the 2003 evaluation area found afurther four inhumations. In 1996 four inhumationburials with coffins were recorded during realignmentofasewer on the street frontage. During 2002 an archaeological excavation in advance of redevelopment of the site uncovered afurther 39 inhumations and 19 cremations. 10 St. Catherine’s Church, Acremation within astone cist was found during the building of Heighway 1980, 64; London Road the church in 1912. Wills 2000, 226 Awatching brief here undertakenduring the digging of drainage trenches in 1999 recorded alate 1st century cremation burial, the inhumation of achild, agravel surface, and apossible monument foundation.

possible that the fortress may have been abandoned defensive circuit, had passed wholesaleinto civilian during the 70s, when the Second Legion moved to hands before the end of the 1st century. Alarge forum Caerleon, and stood empty for some time before was established on the formersite of the legionary being given over to civilian use. principia ,presumably during the reign of Trajan Whichever date is accepted for the foundation of the (AD 98–117),and the areas of formerbarracks were colonia,itiscertain that the site, complete with its given over to civilian housing (Wacher 1995, 141).

5 Life andDeathinaRoman City

The traditional view of the subsequent development period, and the group of burialsexcavatedatSt of the colonia has been one of relative failure. This Oswald’s Priory were likewise all inhumations, and opinion was firmly established by Richmond, based are likely to indicate asimilarly late date. During the on the belief that the town failed to grow beyond the later partofthe Romanperiod the cemeteries at limits of the former fortressdefences (Richmond 1946, Kingsholm andWotton expanded southward to- 83), and has beenmaintained by unfavourable ward the north gate of the city, possiblydue to a comparisons of Gloucester with the neighbouring city contraction in the area of extramural settlement. at Cirencester (eg Wacher1995, 165). Recent research (eg Hurst 2005), however, has indicatedthat The Wotton cemetery Gloucester possessed sizeable areas of extramural settlement, particularlytothe north and north-east The present site lies within the Wotton cemetery, and to the west between the fortressand the river which is known to have extended for at least500 m Severn(Hurst1999b,120–1), giving it atotal area on a along the road connecting the colonia to Ermin Street, par with more successful civitas capitals such as between 50–52 London Road and St Catherine’s Silchester and Verulamium. Clarke (1996) meanwhile Church at the junction with DenmarkRoad. The has argued from arank-order analysis of settlements majority of the cremations known from this cemetery in the areathat Gloucester was better integrated are concentrated at the east end, suggesting that it with its hinterland than was Cirencester, with awell- began as acremationcemetery locatedatthe junction developed trading infrastructure.Gloucester is not of the two roads during the 1st century AD. Inhuma- thought to have had amajor administrative role, tions extending along the road to the south-west attest becoming neither a civitas nor aprovincial capital, to the subsequent expansion of the cemetery toward although Reece (1999, 77–8)has argued that Glouce- the city during the later Romanperiod. ster rather than Cirencester may have become capital Much of whatisknown about the cemetery of the late Romanprovince of Britannia Prima. derivesfromindividual finds of burialsknown from Neverthelessthe colonia is likely to have been an observations made during the developmentof importantnodeinthe trading and communications the area during the 19th century, and from small network of the south-western part of Britannia, on scale investigations of more recent date (Fig. 1.4 accountofits location at the meeting point of water- and Table 1.1). The first recorded evidence for the borne traffic arriving from the Bristol Channel and the cemetery was the discoveryin1827 of two military main land route between south-eastern Englandand tombstones (Fullbrook-Leggatt 1933, 89). One of southern Wales. these, now lost, was dedicated to asoldierof Twentieth Legion (RIB 122) and is oneofthe main pieces of materialevidence associating that legion The cemeteries of Roman Gloucester with the Gloucester area.The other tombstone (RIB Romanlaw required burialsaturbansitestobe 121) commemorates acavalryman of the Sixth located beyond the official limitsofthe city. Most Cohort of Thracians. Other antiquarian discoveries Romancities were consequently surroundedbya includeanunspecified number of burials in the ring of cemeteries, usually located along the main graveyard of the church of the Hospital of St Mary approach roads. At Gloucester cemeteries have been Magdalen, on the opposite side of London Road identified along the roadsapproaching the south and from the site (Fullbrook-Leggatt 1933, 88), and stone east gates, to the west near the old course of the cists containing cremated remainsinpottery and River SevernatStOswald’s Priory, and beyond the glass vessels found at St Margaret’s Hospital (Heigh- extramural settlementtothe north and north-east at way 1980, 63–4). Alarge group of cremation burials Kingsholm and Wotton (Fig. 1.3, Heighway 1980). was foundin1876 on the site of a19th century The cremation cemeteries at Kingsholm and Wotton villa in the vicinity of the current93/95 London appear to have been the earliest, having perhaps Road, but most of the cinerary urns have been lost, been originally established to serve the fortress at with the exception of threenow in Gloucester City Kingsholm, while the discoveryofpotentially early Museum. cremations to the south of the defencesduring the Many of the early observations were not accu- construction of parts of the Dock basin in the 19th rately recorded, and Fullbrook-Leggatt (1933, 88–90) centurymay indicate that the south gate cemetery records numerous finds which have now beenlost, was also established during the early partofthe and forwhich no precise findspot is known. These Romanperiod (Hurst 1988, 62). Themajority of the includeanassemblage of 110 urned cremations burialsfound at Gloucester are inhumations and are foundonlandbelongingtoaMr Niblett at Wotton thus likely to datetoafter the 2nd century, when in 1870, and a‘large quantity’ of human boneand inhumation replacedcremation as the predominant finds including abrass handle, partofasurgical form of burial in the Romanworld (Jones 1987, instrument, coins, tile and pottery foundinafield 815–7), although the London Road excavationhas adjoining London Road in 1827. demonstrated that earlier inhumations can occur During the past three decades anumber of more (see Chapter2). The absence of cremations from the controlled archaeological investigations within the east gate cemetery suggeststhat it maynot have area of the cemetery have addedfurther information. been established until the later part of the Roman Most of these investigations have taken the form of

6 Chapter One watching briefs during the digging of foundation or the north-eastern end of London Road was diverted service trenchesand have consequently uncovered from its original course to run to the south of only asmall number of burials. Anumber of the Chapel(Spry 1971, 4). During the 20th century observations relating to the Wotton cemetery have the house was replaced by aservice station, been made at 124–130 London Road, adjacent to the which was demolished shortlybefore the excavation currentsite.The earliest of these was the discovery took place. during the 19th century of agrave containing a flanged bowl and coins issued by the emperor Gratian (AD 367–83).More recently an archaeologi- THE EXCAVATION cal evaluation undertaken in 1993 recorded 17 Previous evaluation of the site inhumations(GloucesterArchaeology 1993) and a subsequent watching brief to the east of the evalua- An archaeological field evaluation of the site, com- tion area foundafurther four inhumations, includ- prisingthe excavation of three trenches,was carried ing an infant and achild. In 1996 four inhumation out during 2004 by Gloucester Archaeology Unit burialswith coffins were recorded during realign- (Cook 2004). Asequenceofdeposits dating fromthe ment of asewer on the street frontage, all buried on Roman, medieval and post-medieval periods was north-south or east-west alignments (Sermon 1996), foundtosurvive in the southern and eastern parts of and during 2002 an archaeological excavation in the site, including an urned cremation near the south- advance of construction of new buildings and an east corner. The north-western part of the site and the access road uncovered an additional 39 inhumations street frontage had beensignificantly disturbed by the and 19 cremations, as well as anumber of ditches, construction of the service station that formerly pits and postholes (Foundations Archaeology 2003). occupied the site and the digging of apit for its fuel tanks, destroying any archaeological remainsthat may have been located in this area. The hospital of St Mary Magdalen and post-medieval developmentofthe site The stagesofexcavation Gloucester appears to have been more or less The excavation strategy originally adoptedtomiti- abandoned as an urban settlement after the Roman gate the effects of the development consistedofthe period, although it retained some significance as an full excavation of alimited areameasuring 6 · 3m administrative and religiouscentre forthe Anglo- (Fig. 1.5, Area A) and comprising the part of the Saxon kingdom of the Hwicce (Herbert 1988, 2). It footprint of the new building that would impact on was not until the 10th century that the town emerged the rearofthe site, where the previous evaluation again as acommercial centre,and becameaprosper- had indicated that archaeologicalremains survived ous manufacturing and market centre during the intact. The rest of the rearofthe property (Fig. 1.5, medieval period. During this time Wotton would Area B) lay outside the intended area of impact and have lain some distancefrom the built up area, and it any archaeological remains in thisarea were to be is for thisreason that it was chosen as the locationof preserved in situ.Inthe area towardthe street the leper hospital of St Mary Magdalen, established frontage, where the impact of the construction of the by Llanthony Priory in the early 12th century.The formerservice station was greatest, awatching brief site of the excavation lay within the grounds of the was to be maintainedduring ground reduction and hospital, the main buildings of which lay ashort the digging of foundation trenches (Fig. 1.5, Area C). distancetothe east, where awallfooting and However, following the excavation of Area A associated drain and boundaryditches were found during August 2004, groundreduction within Area during arecent excavation at 124–130London Road Bexposed anumber of burialsand consequently it (Foundations Archaeology 2003). Followingthe Dis- was decided, in consultationwith the Gloucester solution the hospital fell intoruins but it was rebuilt City Archaeologist, that the whole of thisarea during the 17th century, finally being demolished shouldbesubject to full excavation. The excavation when the area was developed as residential housing of this areawas carried out in two phases, running in the mid-19th century. The only partofthe hospital from October 2004 to March 2005 and May to to survive was the chancel of the church, which is September 2005. During the latter phase of work now known as St Mary Magdalene’s Chapel and part of asubstantial square pit containing the standsinHillfield Gardens,acrossthe road from the remainsofalarge number of individuals was site of the excavation. ArthurCauston’s map of 1843 discovered, the larger part of which was subse- shows that the Victorian development of the area quentlyexcavated during an additional phase of included the construction of alarge rectangular excavation in November and December 2005. building on the site of the excavation. The building, which was presumably asubstantial villa similar to others built along London Road during the mid-19th Excavation methodology century, was situated in the north-western partof the site,fronting onto London Road and was In Area Athe post-medieval and modern over- surrounded by gardens. It was also at this time that burdenwas removed by amechanical excavator

7 Life andDeathinaRoman City

in the soil. It was therefore necessary to adopt a strategy of excavating the area in spits by hand in order to locate further burials, and consequentlyit was not possible to establish the height from which grave pits had been cut. Awatchingbrief was maintained during ground reduction and the digging of foundation trenches in the northern part of the site but only asmall number of burialswere identified. It is likely that this area had experienced some degreeoftruncation during the construction of the Victorian villa and 20th centuryservice station that had formerlyoccupied this partofthe site and that this had resulted in the loss of anumber of features. In the areatoward the London Road frontagethe archaeological horizon had been completely removed, and consequently no burialssurvivedhere. Similarlynoburials were identified in the western partofthe site despite this area being subject to carefulexcavationinspits and detailed cleaning, and this absence is attributed to truncation of the area associatedwith the lane that borderedthe western edge of the site. Once each inhumation grave had been identified it was fullyexcavated by hand and recorded in accordance with the recommendations of IFA Tech- nical Paper No. 13 (McKinley and Roberts 1993) and the OA Field Manual (OA 1992). Afull graphic and photographic recordwas made of each burial and any associatedartefacts were located three- Figure 1.5 Plan of the site showing the phases of dimensionally by meansofanEDM. In the later excavation. phases of work, from October 2004, hand drawing of graveswas largelyreplaced by geo-referenced under archaeological supervision,revealing amed- photography, from which digital plans could be ieval soil layer (1025) into which medieval and later generated. features were cut. Following hand excavationof All cremationdeposits were subjected to full these features the layerwas removed by hand to recovery for wet sievingand flotation to maximise exposeasoil layer of Roman date (1020).Anumber the retrieval of cremated bone, charred plant remains of inhumations andcremations were cut into this and small artefacts. The cremation pit, urn and any layer, as well as apit containing the fragmentsofa othercontents were planned at ascale of 1:10. The sculpted tombstone (see Chapter 5). These features majority of the cremation urns were found to be in a were excavated and layer1020 removed by hand to fragmentary condition, but two (1061, 1251) were exposethe natural geology beneath,into which sufficiently well preserved to be lifted intact with the furtherburialswere cut. cremation deposit in situ for laboratoryexcavation by Across the rest of the site the overburden was an osteologist. All non-burial discrete features were removed by machine underwatching brief condi- half-sectioned and planned at ascale of 1:10. Linear tions, exposing burials in Area Band resulting in the features were sampled by excavation of sections decisiontocarry out afull excavationofthis area. regularly spacedalong the length of the feature The exposed surface was cleaned in order to define sufficient to characterise it, to establish any strati- the archaeological features, and those features that graphic relationships with other features, andto could be identified at this level were excavated. It retrieve artefactual material. was found that archaeological features were very difficult to identify due to the similarity of their Location of the archive back-fills to the surrounding geology, and graves and cremationswere generally recognisedinitially The finds, paper records and digital archive are to be only by the presence of coffinnails, skeletal material deposited at Gloucester City Museum under acces- or grave goods rather than by any visible difference sion code GLRCM 2004/52.

8 Chapter 2Archaeological Description

The total areaexcavatedmeasured 37 mnorth-south THE ROMAN CEMETERY by 32 meast-westand encompassed an areaof0.1 ha The excavationrevealedpart of acemetery dating (Fig. 2.1). The northern partofthe site,fronting onto from the 1st–4th centuries that contained both London Road, had beencompletely truncated by the cremation and inhumation burials, as well as amass construction of the 19th century villa and 20th centurygarage that had been sitedhere, resulting grave.Aburiedsoil associated with the use of the cemetery was recorded across most of the site, except in the lossofany archaeological remainsinthis area. in the north-western and south-westernparts where it Similarlynofeatures were identified in the western had been truncated by medieval ploughing. The soil part of the site,and this is attributed to truncation layer (1020/1106/1440) was 0.15–0.26 mthick and associated with the lanethat ran along this edge of was composed of brownish yellow/orange silty clay, the site.Inaddition to these areas of significant with inclusions of limestone gravelthat varied in truncation anumber of modern intrusions pene- quantity according to variations in the underlying trated intothe natural geology and in some instances geology. Thesimilarity of this material to the deposits were dug partly through graves. with which the graveswere back-filled made it impossible to establish its stratigraphicrelationship with all but ahandfuloffeatures, which were all dug GEOLOGY AND PLEISTOCENEFAUNAL into the soil layer. These included cremation burial REMAINS 1767, which was one of the earliest burials, suggesting that this soil layer was the original ground surface The stiff grey Lias clay outcropped along the eastern into which the graveswere dug. The layer yieldedan side of the excavation and sloped down to the west. assemblage of pottery spanning the 1st and 2nd Across most of the site it was overlainbyalayerof centuries, which was presumably introduced into it orange brown silty gravelthat increased in depth by disturbance resulting from the use of the cemetery. westwardtoamaximumof0.5 matthe western The partofthe cemetery within the excavation edge of the excavation. area comprised atotal of ninecremation burialsand During the watching brief stage of the investiga- 64 inhumation graves, including threeprobable tion poorly preserved faunal remains within the gravesinwhich no skeletal remainswere present gravelwere exposed in asection in the north central (Fig. 2.1). Three disturbed inhumations for which no part of the site. The section was subsequentlyredu- grave could be identified were also recorded, ced by machine to enable the remainstobecleaned, resulting in atotal of 64 bodies from inhumations recorded and excavatedbyhand. Where possible (thoughone was unable to be examined, see Chapter the bones were liftedinblocks of soilfor detailed 3). The cremation burialscomprised six urned excavation by aPleistocenespecialist (see Schreve cremations and three deposits of burntboneplaced Appendix 6). The remainswereidentified as those of unaccompanied in shallowpits, and were generally part of asingle adulthippopotamus,with additional located in the south central partofthe excavation. material from asingle individual each of probable The inhumation graves were more widely,though bison and elephant. They appeared to lie on the unevenly, distributed and occurred as bothdiscrete surface of the Lias clay, withinthe lowerpart of the gravesand small clusters of intercutting burials. gravel, although some of the fragments to the south Ceramic and stratigraphic evidence provided overlay an area of sub-angularcobbles. dates for asmall number of graves, and radiocarbon determinations were obtained forseven, but the majority could notbeascribed absolute dates. The sequence of the dated gravesisdiscussedinmore ACTIVITY PRE-DATING THE CEMETERY detail below. Agroup of plough scars were observed in the eastern part of the site,where the plough had dragged bands Pre-Flavian burials of limestone gravelinto areas of Lias clay (Fig. 2.1). Parts of at least twelve such plough scars were The earliest use of the cemetery comprisedthree observed, measuring up to 2.55mand all lying on the cremation burials(1766, 1767, 1770) and three pits same ENE-WSW orientation. No datable material (1149, 1074, 1306)containing pottery dating from the was recovered fromthese features, but theywerecut Neronian period (AD 54–68, Fig 2.2). Pit 1149 lay by at least onegrave (1151) and are likely to represent towardthe eastern edgeofthe excavationbut the prehistoric or early Roman cultivation of the area otherfive features were locatedclosetogether in the pre-dating the establishmentofthe cemetery. central part of the site, the two urned cremations

9 Life andDeathinaRoman City

Figure 2.1 Phased plan of all archaeological features.

(1766 and 1767) being only 2.5 mapart. Both these half of cremation burial1766 had been removed by cremation burials had been placedinpits apparently amodern geotechncal pit, and cremation burial dug for thispurpose. In bothcases the cremated re- 1767 contained the remainsofatleastthree adults mains were contained in agreyware jar. The northern of indeterminate sex and was accompanied by a

10 Chapter Two

Figure 2.2 The early phase of the cemetery (1st–early2nd century). collaredrim ﬂagon of Hofheim type (1063) that had Pit 1149 (Plate 2.1) was rectangularinplan, toppled onto its side andlay against the eastern side measuring 0.6 mx0.4 mand 0.2 mdeep. The upper of the cinerary urn. Thecrematedremains (1324) part of aﬂagon of Neronian-earlyFlavian date (1141, from un-urned cremation burial 1770 further south see Timby, Chapter5), which may have been placed yieldedaradiocarbon date of 50 BC–AD70(OxA- deliberately, lay inverted near the centre of the pit, 16792 cal 2sigma). and was accompanied by asmall assemblage of

11 Life andDeathinaRoman City

inhumation (1352),and apit (1196) were dated to the late 1st-early2nd century (Fig. 2.2). The cremation burialswere located in the same general area of the site as thoseofthe pre-Flavian period. The four urned cremation burialsall contained pottery usedascinerary urns or placedas grave goods that was of Flavian-Trajanic date, and burial 1266 was particularlynotable for the presence of agaming set within the cinerary urn (see Cool, Chapter5). The crematedremains (1187) from burial 1769 produced aradiocarbon determination of AD 60–220 (OxA-16811 cal 2sigma), and it is likely that cremation burial 1768, which was not dated by radiocarbon,isofsimilar date. Inhumation burials 1145,1243 and 1334 were each Plate 2.1Pit 1149, with flagon (1141) placed in an accompanied by avessel of Flavian-Trajanic date. inverted position near its centre. Burial 1334 was placedinacrouched position, burial 1243 was partly crouched, placedonthe rightside with legspartially flexed,and burial 1145 was a animalboneincluding aright forelimb pork joint. A proneburial. The dating of burials 1243 and 1334 pottery vessel had also been deliberatelyplacedin were confirmed by radiocarbon determinations from pit 1306. Thepit (Plate 2.2) was circular in plan, samples of the skeletons,which yieldeddates of AD measuring 0.5 mindiameter and only 0.1 mdeep. A 1–130 (NZA 27005 cal 2sigma) and AD 50–230 (NZA buff ware flagon appearedtohave been deliberately 27004 cal 2sigma) respectively.Itislikely that two placedupright in the bottom of the pit, although othercrouched burials (1219, 1240) that lacked any only the base of the vessel had survived the intrinsic datingevidence also date from thisphase. subsequent truncation of the feature. Foursherds Pit 1352 (Plate 2.3) contained several items that of agrey ware beakerwerealso recovered,all the appeared to have beenplacedasgrave goodsand pottery being of Neroniandate. Pit 1074 contained ahandful of sherds of 1st centurypottery and part of acopper alloypin (sf 11) and was cut by cremation burial 1766. The pit appeared to be roughly circular and measured c 0.65 mindiameter, although its northern half had been removed by the sametest pit that had truncated the cremation burial.

Late 1st-early 2nd century burials Atotal of four urned cremationburials(1196, 1209, 1227, 1266), at leastone (and probablytwo) unurned cremation burial (1768, 1769), four crouched inhumations(1145, 1240, 1243, 1334), apossible

Plate 2.3Feature 1352, facing south-west. Two ﬂagons and some animal bone can be seen at the north-eastern end Plate 2.2Truncated pit 1306, with base of aﬂagon of the pit, with two glass unguent bottles near the north- preserved in situ. west edge.

12 Chapter Two may have beenaninhumation grave,although no The densest and most complex of the clusters of bone survived. Two flagonsand some animal bone late burials was asequence of sevengravesdug had been placedatthe north-eastern end of the during the 3rd–4th centuries into the upper part of feature, and apair of glass unguent bottles lay near the back-fill of the mass grave (Fig. 2.4). Theearliest its north-western side. An assemblage of seven nails of these burialsweretwo graves (1505, 1764) on recoveredfrom the grave may be evidence for a parallel NE-SW alignments, both of which were wooden box or coffin. Asingleoutlier of this group truncated by the diggingofsubsequent graves with of nails was locatedatthe north-eastern end of the the result that only partial skeletons survived in each grave and it is possible that this nailhad been feature. The burial in grave 1505, an elderly male, disturbed and that the box/coffin occupied the was accompanied by aDorset black burnished ware south-westernhalf of the feature, with the other jar dating from the 3rd–4th century(1489) that items placed outside it. If thiswas the case, the box/ provided a terminus post quem for the sequenceof coffincan have measurednomorethan 0.6 mx0.3 burials. The north-eastern half of grave 1764 was m, but this may have been large enough to contain subsequently truncated by the diggingofagrave the body of achild. orientedroughly east-west (1765),which appears to represent the adoption of the practice in this partof 2nd century or later burials the cemetery of aligning gravesonmore cardinal orientations. Grave 1503 was then dug, cutting Afew burials withinstratigraphic sequences pro- slightlyintothe western end of grave 1765and also duced radiocarbon dates suggestive of a2nd/early truncating the end of grave 1505. Graves1501 and 3rd century AD date, thoughlater 1st century is also 1508, which bothlay on the same east-west align- possible (Fig. 2.3). The mass grave is dated to the ment, were also later than grave 1765, but it is later 2nd century. uncertain whether theywere contemporarywith Asample from skeleton1286, an adultmale 1503. If they were, then the final phase of burial in buriedingrave 1288, yieldedaradiocarbon deter- this areaisrepresented by grave 1504, but it is also mination of AD 60–240 (NZA 27007 cal2sigma). possible that graves1501 and 1508 were later than This grave was partofagroup of three intercutting burial 1503 and form part of the final phase with graves, and was later in thissequencethan grave 1504. An eighth grave (1507) was located adjacent to 1243 (see above), which was radiocarbon dated to this complexoffeatures and is likely to have been AD 1–130 (NZA 27005 cal 2sigma), and earlier than contemporary with it, but where it fitted intothe grave 1264, which has been assigned a3rd–4th sequence is uncertain as it did not intersect with any centurydate on the basis of its apparent spatial of the othergraves. The orientation of thisgrave, association with grave 1370. Similarly, asample from however, suggests that it is most likely to be contem- skeleton 1184, which was buried in grave 1228, the porarywith eithergrave 1503 or graves1501, 1504 earliest in this sequence, produced aradiocarbon and 1508. determination of AD 70–240 (NZA 27006 cal 2 The second clusterof3rd–4th century graves sigma). The burial was cut by grave 1229 and this consisted of agroup of burials locatedtowardthe was in turn cut by 1230, which produced alate 3rd westernedgeofthe distribution of graves(Fig. 2.3), centurycoin (see below). and included asequence of intercutting graves(1374, The remainsofayoung child (1277) aged 2–5 1369, 1370). Grave 1374 was the earliest grave in this yearsburiedinisolated grave 1275 produced a sequence, and contained the burial of an adultof radiocarbon determinationofAD50–220 (NZA indeterminate sex accompanied by the substantial 29423 cal 2sigma) and so could be either1st or part of aflared-rim grey ware jar. This grave was cut 2nd/early 3rd century AD in date. by two subsequent graves (1369 and 1370) which Pottery dating from the 2nd century was recov- must therefore also date fromthislater phase of the ered from the back-fills of graves 1388, 1403 and cemetery. Graves 1264 and 1353 were located in 1720, the latter including afragment of amphora, close proximity to this group and shared acommon whileafragment of vessel glassdatingfrom the alignment with grave 1370, possiblyindicating that 2nd–3rd centuries was recoveredfrom grave 1144. they are contemporary. Although this material does notprovidedefinite Grave 1230 also belongstothis latestphase of dates for these burials it doesenable a terminus post burials, as acoin datingfrom AD 270–95 was quem in the 2ndcentury to be assigned to each of recoveredfrom the back-fill. This burialwas the them. latest in asequenceofthreegraves dug in approximately the samelocation, albeit with slightly varying orientations (see above). 3rd–4th century burials Asingle discrete burial was also assigned to the Atotal of five graves containedartefactsdatingfrom later partofthe Romanperiod. Grave 1362 contained the 3rd to 4th century, and anumber of further the remains of ayoung adult of indeterminate sex graveswith stratigraphic relationships with these buriedwearing acopper alloy ring(sf 219) and burialsare also likely to datetothis period (Fig. 2.3). bracelet (sf 216). Thering is of a3rd–4th century Most of these late burialswere locatedwithin two form and the presence of the bracelet also probably distinct clusters. indicatesa4th century date forthis burial.

13 Life andDeathinaRoman City

Figure 2.3 The later phase of the cemetery (2nd–4thcentury).

Mass grave plan and alittle over 3.5 macross, with verticalsides and adepth of c 0.85 m. The north-western corner of During the late 2nd or early 3rd century multiple the grave had been truncated by post-medieval ditch individuals were deposited in amass grave (1483) in 1055, while agroup of burials datingfrom the later the south-eastern partofthe site (Fig. 2.3; Plate 2.4), in phase of the cemetery had beendug intoits back-ﬁll an area that does not appear to have previously been (see above). However, the latter features had only used forburials. The grave was roughly square in been dug to arelatively shallowdepth and did not

14 Chapter Two

Figure 2.4 The sequence of intercutting burials in the south-eastern cornerofthe excavation.

penetrateasdeepasthe mass burial deposit at the skeletal parts (Plates 2.5–2.7). Thedegreeofentan- base of the grave,although disturbance associated glement certainly suggests that bodies were depos- with themhas resulted in the introductionofsome ited in asingle event rather than individually (see later pottery intothe upper grave ﬁll. Chapter6for full discussion). Skeleton1558, for The lower part of the grave containedadeposit of example, lay on top of the legs of skeleton 1562 and humanremainsthat appeared to have been uncer- the skull of skeleton 1564, but the arms of 1558 were emoniously dumped into it withoutany degree of beneath the pelvis of 1562 and the left arm of 1564 care, resulting in acomplicated array of entangled (Plate 2.8). Such acircumstance would clearly not

15 Life andDeathinaRoman City

(a) (b)

Plate 2.4a–d The mass grave. (a) The mass grave as ﬁrstdiscovered. (b) General view of the mass grave during excavation. (c) The mass grave under excavation.(d) Detail of the northern part of the mass grave showing its depth and complexity.

have arisen had the bodies beendepositedone at a time, which would have resulted in amuch more clearlystratified deposit, and it suggests bodies were dumped into the pit together,perhaps from acart or wheelbarrow. The complexity of the deposithad been exacer- bated by the post-depositional movement of elements resulting from the settling of individual bones under the influence of gravity as the soft tissue connecting them and beneath them decomposed, disassociating themfrom the skeleton to which they belonged (see Chapter 3). In some cases it was clear from the proximity of elements that they had originally been associated, and had presumably been deposited as acomplete, articulated corpse (Plate 2.9), but many individual bones or articulating elements could not be confidently associated with a Plate 2.5Skeleton 1520 in mass grave lying in aprone specific skeleton at the timeofrecovery. position with the upper body turned slightlytothe left Due to the complexity of the deposititwas not and the right arm pinned beneath it. possible to apply normal excavation procedures.

16 Chapter Two

Plate 2.6Skeleton 1524 in mass grave lying in asupine position with the back arched and twisted to the right, with the right arm thrownbackover the head. Plate 2.8Entangled skeletons 1558, 1562 and1564 within the mass grave.

more pragmatic approach to the excavation and recording of the deposit: each complete or substantially complete articulated skeleton was allocated a uniqueskeleton number, prefixedwith‘sk’, and was recorded on a pro forma sheet, including asketch plan recording its location and posture, and each individual bone or articulating elementthat couldnot be associated with aspecific skeleton at the time of recording was allocatedasmall find number, prefixedwith ‘sf’. Both skeletons and small finds were recorded in situ by rectifieddigital photography, which recorded their exact position so that spatial relationships and associations couldbe explored during post-excavation analysis. Many of the remains to which skeleton numbers were allocated were partial skeletons and many of thoseto Plate 2.7Skeleton 1564 in mass grave lying on the right which small find numberswere allocatedwere side with the head tilted sharply backand the right arm articulated elements, or even complete limbs,and extended behind the body. consequentlythere was acertain amount of sub- jectivity to the allocation of individual remainsto either group, which were not absolute categories. Normal practice would be to record and lift complete This procedureresulted in the recording of 201 individual skeletons in their correct stratigraphic articulated skeletons and 419 elements recorded as sequence, but the nature of the deposit was such that small finds, although it was appreciated at the time it was often notpossible to isolate individual of recording that the true number of individuals skeletons or to understandthe sequence of deposi- buriedinthe pit was likely to be smaller than this. As tion in the field. It was therefore necessary to adopt a the majority of the remains to which skeleton

17 Life andDeathinaRoman City

iron penannular brooch (sf 345) that was adhered to askull (sf 344) but was not necessarily originally associated with that individual. The remains of acomplete chicken and the left hind leg of asheep or goat also lay among the humanremains (Worley, Chapter5), along with an assemblage of 168 sherds of pottery weighing 1341 g, including the substantially complete remains of at least six south-west black burnished ware jars dating from the late 2nd to 3rd century that provide a terminus post quem forthe deposit (Timby, Chapter 5). The brooches, bracelets and ring are also types consistent with such adate (Cool, Chapter 5). A radiocarbon determinationofAD70–240 (NZA 27008 cal 2sigma) was obtained for asample of bone from skeleton 1630, oneofthe bodies toward the base of the deposit, but samples from afurther three skeletons failed to produce adatedue to poor collagen preservation (see below). Thepit had been back-ﬁlled with adeposit of orange and blue clay that is likely to have been derived from its original excavation.

Other features The tombstones Two tombstones were recoveredinthe course of the Plate 2.9Skeleton 1553 in mass grave in aprone excavation, though sadly neither was in situ and position with the legs drawnupunder the torso. The consequentlyitwas not possible to associate them lumbar vertebraeand pelvis had become disarticulated with specific graves(see Henig and Tomlin, Chapter from the upper part of the body duetosettling as the soft 5for full analysis). tissue decomposed. Tombstone1had been buried in apit (1003) cut into the soil layer associated with the use of the cemetery (Plate 2.10, Fig. 2.2). The tombstone was of numbers were allocatedwere only partial skeletons Aalenian Lower Jurassic limestone (see Appendix5) it was acceptedthat in some instances more than one and survived in three main pieces and anumber of skeleton number may have been allocated to the smaller fragments. The top and bottomwere both remainsofasingleindividual,particularlywhere incomplete. The tombstone boreaninscription parts of the skeleton were notrecognised as dedicated to aslave boy, ‘Martialis, slave of C belonging to asingle individual due to the entangled ..lonus’, above which was arecessed panel contain- nature of the deposit or where they had become ing arelief sculpture. The tombstone had clearly disassociated due to post-depositional effects. been removed from the location at which it had been There was evidence that at least some of the bodies in the mass grave were clothed at the time of burial (see Cool, Chapter 5). An iron buckle(sf 386) was foundbeneath the pelvis of skeleton 1544, where it may have fastened abelt. Two fragmentsofaniron bar (sfs 399, 408) found over the left hip of skeleton 1542 may also be some form of belt equipment. Skeleton 1660 was wearing apair of shoes or boots with hobnail soles (sfs 676, 677) and four other such items of footwear (sfs 516, 517, 2004, 2018) were recoveredfrom the pit but could not be attributed to specific individuals. Skeleton 1596, afemale aged 26– 35 years, had acopper alloy bracelet (sf 464) on the right forearm. In addition to these items, the deposit contained three copper alloyrings(sfs 378, 383, 2027), aWroxeter brooch (sf 406) acopper alloy knee brooch (sf 437), two bone pins (sfs 341, 711) and two unidentified iron objects (sfs 409, 541) that couldnot Plate 2.10 Tombstone1in pit 1003, as initially exposed be associatedwith individual skeletons, as well as an by machining.

18 Chapter Two originally set up and so could not be associated with 0.70 x0.50 m(1196),and one(1767) lay in asquare aspecific grave. No other material was recovered pit 0.45 mwide. Theremainingtwo burials (1266, from the pit that might aid in establishing the date of 1766) were roughly circular in shape, measuring 0.44 the disposal of the tombstone, although it is possible m(1766) and 0.60 m(1266) in diameter. All six had that it was buriedduring the clearing of the site for been truncated to agreater or lesserdegree by the cultivation during the medieval period. effectsoflater ploughing, and the northern partof Asecond tombstone (Tombstone 2), also of Aale- burial 1766 had been removed by ageotechnical nian Lower Jurassic limestone was uncovered lying investigation pit (Plate 2.11). Only in burials1266 face downinthe upper surface of the soillayer, c 5m and possibly 1767 had the cinerary urn and its from the first tombstone (Fig. 2.2). This monument contents survivedintact, although in the latter case was largely complete, albeit in alarge number of the urn was still veryfragmented. Burial 1266 com- fragments. It was dedicated to Lucius Octavius Mar- prised aciruculargrave pit (1253) with agrey ware tialis, asoldierofthe Twentieth Legion. The lower cinerary urn and an ancillary vessel (ring necked part of alocally madevessel of Flavianic-Trajanic flagon), both of which had toppled over presumably date (1101) lay crushed beneath the stone, which during the funeral or back-filling, with the result appears to have toppled over and to have been left that theylay below the level to which the feature where it fell. had been truncated (Plate 2.12). This burial, which contained the crematedremainsoftwo children aged 5–10 yearsand 10–15 years, was particularly Gully1358 (Fig. 2.2) notablefor aset of 33 bone and glass gaming Part of ashallow curvinggully (1358) recorded counters (sfs 249–252, 256–269, 271–277, 279–286), towardthe western limit of the distribution of graves two bonedice (sfs 253, 254) and afrit melonbead was dated to the 1st century on the basis of part of (sf 248), which appeared to have been placed the rim of awhite ware flask recoveredfrom its fill. within the urn in astack against one side. None of The gully was difficult to distinguish as its fill was these items showed any evidence of having been very similar to the surrounding natural clay, and burnt. could be traced for atotal length of only 3.4 mfrom a The truncation of the remaining four urned burials possible terminus at its southern end. To the north it had removed the upper parts of the urns,and peteredout before reaching agroup of later burials, potentially causedthe loss of partofthe cremation whiletothe south it was truncated by inhumation deposit. Burial 1227 was particularly badly affected, grave 1422. with only the bottom 0.05 mofthe featuresurviving, along with the base of aflagon that may have been the cinerary urn and amere 1.5 gofcremated bone. Pit 1301 (Fig. 2.3) The cinerary urns containing five of the cremations Alarge circular pit (1301) was recorded in the south- were grey ware jars, four of which were accompa- central part of the excavation. It measured c 2.9 min nied by ancillary vessels in the form of flagonsin diameter and 0.34 mdeep and is likely to have been oxidised fabrics. Only in burial 1227 was there a dug as aquarry pit to extract gravel. Grave 1427 was variation from this pattern, with an oxidised flagon discovered cut intothe natural beneath the base of apparently being used as the cinerary urn, although the pit but it is uncertain which feature was earlier. the disturbance of this feature due to later ploughing As the fills were very similar, it was notpossible to was such that it is possiblethat this was an ancillary establish whether the grave had been dug through vessel and that the cinerary urn had been completely the fill of the pit or the pit had truncated the upper destroyed. part of the grave. Two coins datingfrom the 4th centurywere recovered fromthe fill of the pit but it Un-urned cremation burials is possiblethat they are residual and that the feature is later in date. The three un-urned cremationburials(1768, 1769, 1770) each comprised ashallow pit in which the cremated remains of asingleadult of indeterminate THE BURIALS sex had beendeposited with no accompanying The cremation burials (Fig. 2.5) artefacts. These features had been particularly severely truncated by later ploughing and survived Nine cremation burials were identified withinthe to depths of only 0.03–0.07 m. Burial 1769 comprised area of the excavation, six of which (1196, 1209, 1227, arectangular pitmeasuring 0.43mx0.28 minto 1266, 1766, 1767) were placed in ceramic cinerary which adepositofcharcoal and calcinedbone from urns,whilethree were un-urned (1768, 1769, 1770). the pyre had been deposited, whileburials 1768 and 1770 had been placedincircular pits measuring Urned cremationburials c 0.25 mindiameter. The deposits in these two burials consisted of re-deposited natural clay containing Of the six urned cremationburials, three (1196, 1209, small quantities of calcined bone and very little 1227) had been placedinsub-rectangular pits, with charcoal,the latter suggesting that the bone had been dimensions ranging from 0.55 x0.40 m(1209) to picked very carefully out of the pyre debris.

19 Life andDeathinaRoman City

Figure 2.5 The distribution of cremation burials.

The inhumation burials (Fig. 2.6) full dimensions couldbemeasured varied from 1.72–2.17 minlength and 0.38–0.77 minwidth. Due The grave pits to the nature of the excavation, with material being Grave pits were generallyrectangularorsub- removed in spits until each burial becameapparent, rectangular in plan and of asize to accommodate it was not possible to establish the level from which the body of the deceased. Many burials had been each grave was cut and so the depths of the grave truncated by modern features or the digging of pits could not be established. Afew smaller graves subsequent graves, but those adultgravesfor which had been provided for burials of children. These

20 Chapter Two

adult burials that were sufficiently well preserved for the posture to be established 40 were supine, the remainder comprising six prone burials, four crouched burials(Fig. 2.6) and asingle individual (1288) who appeared to have been placedinthe grave lying on his left side in an extended rather than crouched posture, and to have subsequently slumpedsomewhat onto his front. Althoughitis possible that in some instancesthe position that survivedtoberecorded in the excavationmay have been the result of disturbance during burial or the movement of bones during the course of decomposition, there were no burials in which thiswas demonstrably the case, and it is assumedthat the postures that were recorded preservethe posture chosen for thematthe timeofburial. Plate 2.11 Cremation burial 1766 exposedinsection in The arrangement of the arms was particularly the side of amodern geo-technicalpit. Scale=0.5 m. variable (Plate 2.13). Four principle arrangements were identified: extended beside the body; flexed with the hand lying over the pelvis; bent at aright anglewith the hand lying across the stomach; and tightly flexed with the hand on the upper torso or shoulder. Atotal of 20 adultburials was sufficiently well preserved for the positions of botharms to be established, and among these eight of the 16 possible combinations of arm positions were represented. No single arrangement was dominant, the most fre- quentlyencountered being with both hands resting on the pelvis (five instances), bothbeside the body (four instances) and with the left arm beside the body and the rightflexed with the hand lying on the pelvis (four instances). No othercombination was represented by more than two examples. Asymme- trical arrangementswere slightly more common than symmetrical ones, with 11 instances. Where one arm was placedacross the waist or chest the otherwas usuallybythe side. Of the 32 supine burials where both legs survived Plate 2.12 Cremation burial 1266, showing cinerary 22 lay with the legsextended andparallel and eight urn and complete ring-necked flagon. with the legs together, with two in more irregular positions. were the same shape as thosefor adults, but Discrepant burial rites: prone, decapitated and correspondingly smaller. Theonly gravesthat varied disarticulated burials from thisshape were the sub-circular pits dug for crouched burials1219 and 1234, and arather oval The burials of six individuals had been placed face grave for crouched burial 1344. down in aprone position (1143, 1145, 1150, 1230, Rectangular gravesnaturally lend themselvesto 1234, 1765). All but one of these burials were alignment on deliberately chosen orientations. A distributed in the southern part of the excavation, total of 31 of the 61 rectangulargraveswith skeletal the only outlier being burial 1145 which was located material presentwithin the area of the excavation towardthe northern edge of the distribution of had cardinal orientations, divided almost equally burials(Fig. 2.6). The pits dug forthese burialsdid between east-west (26.2%)and north-south (24.6%). not differ in any way from those containing supine Of the remainder, 11 (18%) were oriented NE-SW,10 burials. All six were sub-rectangular in shape and (16.4%)NW-SEand nine(14.8%)couldnot be the three that had not beentruncated by subsequent assignedtoany of these orientations. burialsmeasured 1.88–2.06 mlong and 0.41–0.69 m wide. Both sexes were represented, and they ranged The body in agefrom an adolescent male (1112, grave 1143) to amature female (1103, grave 1150). In burial 1145, a Aconsiderable amount of variation was observed in flagon had beendeposited beside the skull (Plate the details of the laying out of the corpse,although 2.14). The postures of the bodies themselves varied most of the postures were very similar. Of the 51 somewhat. Allsix appeared to have been placedin

21 Life andDeathinaRoman City

Figure 2.6 The distribution of inhumation burials. the grave fairly carefully rather than throwninlike the left arm was straight beside the body and the the bodies in the mass grave,although body 1127 in right flexed with the hand beneath the pelvis. The grave 1145 was slightlyaskew.Three burials had bodies in two burials (1143, 1765) were buriedwith arm postures consistent with the range of positions their arms behind their backs, although the lower foundamong the supine burials: in burial1145 the parts of the arms of the former had been removed by left arm was straightand the right flexed beneath the truncation (Plate 2.15). Grave 1234 had been trun- chest,inburial 1150 the arms were slightlyflexed cated by amodern geotechnical test pit, leaving with the hands beneath the pelvis, and in burial 1230 only the legs, which lay in aproneposture with a

22 Chapter Two

(a)

(b) (c) (d)

Plate 2.13a-d Supine burials, showing some of the variety of arm positions encountered. (a) Skeleton 1165, grave 1167, with both hands resting over the pelvis (b) Skeleton 1337, grave 1353, with left arm lying beside body and right flexed across pelvis (c) Skeleton 1372, grave 1374, with the right arm straight and left flexed across the stomach (d) Skeleton 1496, grave 1507, with right arm straight and left flexed tightly across the chest. decapitated skull, presumably from the samein- (1238, grave 1240) and two adult individuals of dividual,placedonthe backofthe knees (Plate 2.16). undetermed sex (1211, grave 1243; 1332, grave 1334) Four adult individuals were buriedinacrouched lying on their left sides. Burial 1243 was accompa- position, lying on one side with the arms and knees nied by the upper partofaring-necked flagon, flexed (Fig. 2.6). These burials comprisedthe remains which had toppled over and lay on the individual’s of afemale aged 18–25 yearslying on her right side feet, and burial 1334 was accompanied by agrey (1206, grave 1219), and amale aged over 45 years ware jar, also placed near the feet (Plate 2.17).

23 Life andDeathinaRoman City

Plate 2.15 Prone inhumation burial 1143 with arms Plate 2.14 Prone inhumation burial 1145, with ﬂagon positioned behind back. placed by the skull.

In additiontothe articulated burials, asingle The groups of nails from most of the graves were deliberate burial of disarticulated remains was found small,and none containedevidence for an upper (1757).This comprisedarectangular pit into which layer of nails usedtosecure the lid. This could the disarticulated bones of an adult male (1089) had suggest that some or all of the assemblages represent been placed, in atightly packed clusterasthough the remains of biers rather than coffins, or that bound or held withinanorganiccontainersuch as a coffins were buried withoutlids. However, it could bag or sack that has since decomposed. also be explained by acombination of poor survival and the verticalmovement of upper nails resulting Coffins and other containers from the collapse of the coffinunder the weightof overburden. The method of excavation, in which soil No evidence for coffins survived in the form of was removed in spits across the area until evidence timber or coffinstains, but groups of nails indicative for aburial was identified, may also have resulted in of the presence of coffins were recorded in situ in a upper nails from some gravesbeing lost before the total of 19 graves. Smaller numbersofnails that may features were recognised. The only coffined burials also indicate the former presence of coffins were with grave goods were burial 1374, which was recoveredfrom afurther 21 burials, but some of accompanied by apottery vessel that mayhave lain these assemblages may comprise material incorpo- within the coffin, and skeleton 1360 in grave 1362, rated into the back-fills of graves accidentally and it which was wearing abracelet and ring. is thus not possible to establish the exact number of Asingleburial (1756) had beenprovided with a gravesthat contained coffins. Nevertheless, even if stone cist (Plate 2.18). Thecist was formed from five some instancesdonot represent the remains of slabs of limestone set on edge to form atrapezoidal coffins, the presence of nails in 40 of the 66 box (1081) measuring 0.5 mx0.3 mand 0.2 m. Aflat inhumation graveswithin the area of the excavation stone at the wider, south-western end of the grave indicatesthat asubstantial proportion of the graves may have been provided for the headtorest on, contained burialsincoffins secured by nails (see although this cannot be proven as the only surviving Powell, Chapter 5). skeletal material was the left femur of aperinatal

24 Chapter Two

Plate 2.16 Skeleton 1232, grave 1234, mostly truncated with just the legs (in prone position) and decapitated skull remaining. Plate 2.17 Inhumation burial 1334, with crouched skeleton and grey warejar placed by feet. individual. No capstones were found, but these may have been removed by subsequent truncation or deliberate robbing forbuilding material.

Grave goods Atotal of 17 burials, or 26.6% of the graveswithin the area of the excavation, were provided with grave goods.The artefactsplacedwith the burialscon- sistedofceramic vessels, hobnailed footwear, jewellery and animal remains, although only one burial (1505) was accompanied by more than one item.

Ceramic grave goods Pottery vessels were placedwith six burials (1145, 1243, 1334, 1374, 1352, 1505) including possible Plate 2.18 Stone cist in inhumation grave 1756. burial 1352 which may have been acenotaph or memorial rather than agrave (see Timby, Chapter 5). This feature containedtwo vessels,but otherwiseno Footwear grave contained more than onevessel. The vessels were normally placednear the feet (in burials1243, The burials in six graves(1234, 1330, 1344, 1353, 1374, 1505 and 1334). The only exception was prone 1369, 1764) were recorded as wearing footwear burial 1145, where the flagon lay beside the head. represented by groups of hobnails associated with The black burnished ware jar placedingrave 1505 the feet (see Powell, Chapter 5). Grave 1167 con- contained the partialskeleton of adomesticfowl, but tained agroup of hobnails that appear to represent no evidence survivedfor the contents of the vessels footwear placedatthe foot of the grave,and possibly in the other burials. outside the coffin, rather than worn. Afurther five

25 Life andDeathinaRoman City graves(1283, 1284, 1315, 1508 and 1765) yielded from grave 1246, the burial of achild aged 5–12 sufficiently large assemblages to indicate the prob- years. The latter item was not worn at the timeof able presence of footwear, although these were not burial but may have been deliberatelyplaced as a recorded in situ and it is uncertain whether the items grave good. were worn or were placedinthe grave separately. Variations in the numbersofhobnails recovered Animal remains from different gravespresumably represent different types of footwear. The largest assemblages, compris- In addition to the domesticfowlcontainedwithin ing 178 hobnails fromgrave 1344 and104 hobnails the jar in burial 1505, parts of similar birds had been from grave 1369, may representheavily shod boots, placedinburials1374 and 1759, and some mammal but were notaccurately locatedand may be the bone had been placed behind the legsofamale aged remainsofmorethan one pair of shoes. In the burials 18–25 yearsburied on his left sideingrave 1288 where it was possibletoattribute hobnails to (Plate 2.19; see Worley, Chapter5). The placing of the individual feet, the surviving leg of burial 1510 in individual with his legsinaslightlyflexedposture grave 1764 wore aboot or shoe with 42 hobnails, and may have been deliberate,intended to create aspace in burial 1232 in grave 1234 32 hobnails were asso- for placing of the animal. The partialskeletonofa ciated with the left foot and 37 with the right. Grave third domestic fowl was recovered from possible 1330 produced asimilar total,with 13 hobnails cenotaph 1759. associated with the left foot, 22 with the rightfoot and 28 unlocated. Four other gravescontained between 5 MEDIEVAL AND POST-MEDIEVAL ACTIVITY and 11 hobnails, but it is impossible to be certain whether these assemblages represent items of foot- The burialsofthe Roman cemetery were sealed by a wear, with the number of hobnails perhaps dimin- soil layer (1025/1047/1052) containing an assem- ished by truncation or differential preservation, or blage of pottery datingfrom between the 11th and else accidentalinclusions. 15th centuries, as well as some residual Romano- British material. Jewellery During the early part of the post-medieval period anumber of boundaryditches were dug across the Threeindividuals (1246, 1362, 1505) were buried area of the excavationand cut through the soil layer wearing or accompanied by items of jewellery. (Fig. 2.7). Ditch 1725 defined aboundaryextending Skeleton 1360 in grave 1362 wore acopper alloy on aslightly curving north-south orientation bracelet (sf 216) on the left arm. The sex of the through the middleofthe site,cutting through individual could notbeestablished from osteological graves1501, 1504 and 1758 and the north-western evidence, but the wearing of bracelets is normally cornerofthe mass grave.Pottery from the fill of the associated with female burials (Cool, Chapter 5). ditch indicated that it was probablydug during the This burial also contained aring or earring (sf 219), 16th–17th centuries, although it also yielded a although it is uncertain from the excavationrecords substantial quantity of residual material of Roman whether this item was worn at the time of burial or and medieval date. The ditch had asinglere-cut simply placedwith the corpse.Acopper alloy ring (1055) which contained four sherds of earthenware (sf 319) was worn on the third or fourth finger of the datingfromthe 17th–19th century. left hand of the individual buriedingrave 1505, and Boundary ditch 1461 extended on an approximate an iron ring(sf 88), minus its intaglio, was recovered east-west orientation across the southern edge of the

Plate 2.19 Burial 1286 in grave 1288, amaleaged 18-25 buried on his leftside with legs tightly ﬂexed. Animal bones had been placed behind the legs.

26 Chapter Two

Figure 2.7 Post-medieval features.

site and was subsequently re-cut on astraighter ship between this boundary and that represented by alignment as ditch 1729, which partly truncated ditch 1725/1055 were frustratedbythe presence of a graves1505, 1720 and1723. No artefactual material modern pit at the point where the ditches inter- was recovered from the earlier phase of the ditch, sected. and the later phase contained only Romanpottery, Aditch terminus (1412) extended for1.0 minto the but its relationship with the burials suggests that it is edge of the excavation from the eastern side. The post-Roman in date. Attempts to clarify the relation- ditch contained two sherds of 12th–13th century

27 Life andDeathinaRoman City

Table 2.1 Radiocarbon dates from 120–122 London Road, Gloucester.

Lab ID Sample ID Grave group Material Radiocarbon Calibrateddate age (BP) (95% conﬁdence)

OxA-16792 GLLR04 SK1324 ID10 1770 Cremated human bone 1993+ 27 50 BC–AD 70 OxA-16811 GLLR04 SK1187 ID11 1769 Cremated human bone 1884+ 28 AD 60–220 NZA 29423 1277 GLLR04 1275 Left petrous temporal, Occipital 1893+ 30 AD 50–220 crest fragments, parietal fragments NZA 27004 GLLR04 SK1332 1334 Left femur, mid-shaft 1885+ 30 AD 50–230 NZA 27005 GLLR04 SK1211 1243 Left femur, proximal shaft 1932+ 30 AD 1–130 NZA 27006 GLLR04 SK1184 1228 Right femur, mid-shaft 1867+ 30 AD 70–240 NZA 27007 GLL04 SK1286 1288 Left femur, mid-shaft 1869+ 30 AD 60–240 NZA 27008 GLLR04 SK 1630 Mass grave Left femur, shaft fragments 1856+ 30 AD 70–240 pottery, but as with the medieval pottery recovered to CO2. Theresultant gas was converted to from ditch 1729 this may be residual. elemental carbongraphite and the proportion of Pit 1041 and posthole1016 both contained small C 14 measured using conventional mass spectro- quantities of 16th–17th century pottery, but the metry. function of these apparently isolated features could The cremated bone sampleswere preparedand not be ascertained. measured following procedures outlined by Bronk Ramsey et al. 2002, BronkRamsey et al. 2004a and Bronk Ramsey et al. 2004b. RADIOCARBONDATING The dates of the successful samplesare shown in Atotal of eleven sampleswas submitted for radio- Table 2.1. They are presented as conventional radiocarbondating. Human bonefrom nine burials was carbonages(Stuiver and Polach 1977), quoted submitted to Rafter RadiocarbonLaboratory, New according to the international standard known as Zealand and two samplesofcrematedbone were the Trondheim convention (Stuiver and Kra 1986). submitted to the Oxford Radiocarbon Accelerator The calibrations of these results, which relate the Unit, Oxford University. Threeofthe samplescould radiocarbon measurements directly to the calendrical not be dated due to the poor preservation of the time scale, were calculated using the atmospheric collagen in the bone, which resulted in no collagen data publishedinReimer et al. (2004) and the Oxcal being extractedfollowing the demineralisation and (v3.10) computer program(Bronk Ramsey 1995; gelatinisation processes. 1998; 2001) using the ‘INTCAL04’ dataset (Reimer The samples of human bone were pre-cleaned, et al. 2004). The calibrated dateranges cited within crushed and demineralised to obtain collagen, which the text are thosefor the 95% conﬁdence level (2 was gelatinised and combustedtoconvert the carbon sigma).

28 Chapter 3The Human Remains

UNBURNT HUMAN BONE Skeletal biology by Nicholas Ma´ rquez-Grant and Louise Loe Osteological methodology Standard osteological analysis of all remainswas Introduction undertaken by following the recommendationsset out by Brickley and McKinley (2004) and Buikstra Atotal of 63 discrete skeletons and at least 91 com- and Ubelaker(1994).Human boneidentification was mingled skeletons fromthe 1st to 4th century cemetery were subjected to standard osteological facilitatedbyreferencetoWhite and Folkens(1991), Bass (1987),and Scheuer and Black (2000).Tooth examination.This is in addition to asmall quantity of identification benefited fromWhite and Folkens disarticulated bone, recovered from various contexts within the cemetery. Afurtherpartialskeleton (1420, (1991) and Hillson (1996b).All data were recorded on askeletal recording form and entered onto a grave 1422) was notexamined due to its very poor Microsoft Accessdatabase. preservation. The remains represented amixed population of males and females who ranged in agefrom perinatal to over 45 years. In terms of nutrition,health Inventory, conditionand completeness and lifestyle, biological and pathological indicators suggest that they are broadly similar to other popu- Adetailed record of bonepresence and absence, lations that areofsimilar date and type. These trends condition and completeness was madefor all did not vary greatly between the discrete inhuma- remains. Presence or absence was scored according tions and the commingled inhumations. However, to bone segment. For example,for the long bones, more non-specific infection was indicated among the this included the proximal epiphysis, proximal third London Road sample compared to otherRomano- of the shaft, middle third of the shaft, distal third of British populations and among the discrete skeletons the shaft and the distal epiphysis. This allowed compared to the commingled skeletons. calculation of the trueprevalence rate (TPR), as The assemblage was most notablefor the context in based on the number of observableelements or teeth, which the commingledremainswere found, which in addition to the crude prevalence rate (CPR), as suggests they had been deposited in asingle episode based on the number of skeletons, for palaeopatho- as aresult of acatastrophic event. This deposit (the logical and dental analyses, described below. ‘mass grave’ hereafter) has been dated to around the Bone condition was scored by employing the later 2nd century AD. No such deposits of this date grades set out by McKinley (2004),whereby 0means have beenosteologically analysed from Britain before. excellent and 5meanspoor, depending on the level Data for age, sex and pathology were examined to of erosion of the bone surface. Thecompleteness of explorewhat events might have led to the deposition discrete skeletons was indicated by assigning them of these individuals in asingle grave.Inaddition, a to one of the following categories: basic examination of taphonomic changes was also 1=>75% undertaken to explore mode of disposal.Awider dis- 2=75–50% cussion of the massgrave can be foundinChapter 6. 3=50–25% Few of the discrete burials couldbeassigned to 4=<25% finer date ranges other than 1st to 4thcenturies AD (Table 3.1). Thus, there has been alimited attempt to examinethe remains by phase. Discrete inhumations Estimation of biological age andsex and the massgrave are reported upon separately here, but compared when relevant. Subadult ageestimationwas based primarily on observations relating to the stage of dental development (Moorrees et al. 1963; Smith 1991, Ubelaker 1989) and epiphyseal fusion(Scheuerand Black 2000). Diaphyseal long bonelengthswere usedas Table 3.1 Number of discrete burials that couldbe the basis for ageing perinates and neonates, as based assigned to finer date ranges. on the methodsdeveloped by Fazekas and Ko´ sa (as adapted in Scheuer and Black 2000) and, for older Date range Number of inhumations children, Maresh(1972). Wherever possible, acom- 1st–early 2nd century 4 bination of methods was used to estimate age-at- 3rd–4th century 15 death for adult individuals, classified here as those over 18 years of age. This involvedobservations

29 Life andDeathinaRoman City relating to the degeneration of the auricular surface Metric andnon-metric analysis of the pelvis (Buckberry and Chamberlain 2002)and the pubic symphysis (Brooks and Suchey1990), late Astandard set of cranialand post-cranial measure- fusing epiphyses (Scheuer and Black 2000) and ments were taken whereverpossible, as recom- dental attrition (Miles 1962; 2001). General observa- mended by Brothwell and Zakrzewski(2004). tions relating to cranial suture closure(eg Meindl These were primarily taken as ameansofproviding and Lovejoy 1985) were only used as acomplemen- data for estimating biological sex and for calculation tary aid, since this method is not considered to be of stature. Spreading and sliding callipers were used very reliable (Cox 2000; Key et al. 1994). The spheno- for these measurements and an osteometric board occipital synchondrosis was employed to assigna was employed to measure the lengths of the long minimumadultage (Scheuer and Black 2000). bones. Where possible, measurements were taken The highly fragmentary nature of the assemblage using bones from the left sideinkeeping with precludedthe application of othermethods (for standard osteologicalpractice. example, observing of sternal rib ends) that may be Staturewas estimated by taking the maximum employed to estimate age. Skeletons were assigned lengthsofany complete major long bones and to one of the age categories in Table 3.2. applying them to the relevant regression equations Sexually dimorphic features of the pelvis and devised by Trotter and Gleser (1952; 1958) and craniumwere employed to estimate the sex of modified by Trotter(1970).The maximumlength of skeletons (Buikstra and Ubelaker 1994). Osteometrics the femur was used preferentially, as the femur is were employed as secondary indicators (Brickley considered to be the most accurate bone for estimat- and McKinley2004). Skeletons were assigned to one ing stature forBritish populations (Trotter and of the categories listed in Table 3.3. This does not Gleser 1952; Waldron 1998). If the femur was not applytoimmature (subadult) skeletons for whom available, the tibia was the next boneofchoice. The there are currently no acceptedmethods available major bones of the upper limb were only usedifno (ibid.). For the purposes of analysis, all possible fema- lower limb bones were available. les have been treated as females,all possible males Where appropriate, adultskeletons were exam- have beentreated as males, and all ambiguous cases ined for non-metric traits by following the guidelines have been categorised as ‘sex unknown’.Thisis set out by Berry and Berry (1967),Finnegan (1978), unless otherwise stated. Hauser and De Stefano(1989) and Buikstra and Ubelaker(1994).Non-metric traits are minor variants in the skeletonand are of no pathological significance.Early studies of these traits placedgreat Table 3.2 Age categories employed in this study. emphasisonexamining them to explore relatedness between individuals (Berryand Berry 1967). How- Age category Age ever, experimental studies have shown that the extenttowhich they are genetically controlled is Fetus Third month to birth far fromclear (Mays 1998). Some are believed to be Perinate Around the time of birth indicative of mechanical or occupational stress Neonate birth–1 month (Kennedy 1989; Tyrell 2000). For example,accessory Infant >1 month–2 years facets on the ilium of the pelvis may be related to Young child >2–5 years weightbearing as aresult of carrying heavy loads Older Child >5–12 years (Kennedy 1989, 146). Cranial traits may have a Adolescent 13–17 years stronger genetic component in their manifestation Young Adult 18–25 years than post-cranialtraits (Hauser and De Stefano Middle Adult 26–35 years 1989). However, thiswas not borne out by astudy Mature Adult 36–45 years by Mollesonand Cox (1993) on adocumented post- Older Adult >45 years medieval assemblage from Spitalfields,London. Subadult (age unknown) <18 years Adult (age unknown)>18 years Unknown Unknown Taphonomy Taphonomyrefers here to the processes that are responsible for the organisation, appearance and Table 3.3 Categories employed in sex determination. composition of aburial assemblage from its deposition to recovery andanalysis (Haglund and Sorg Sex category Definition 1997). It primarilyconcerns burial position, distribution of elements and alteration to bone surfaces. FFemale Standard taphonomicanalysis of the skeletons F? Possible female from the mass grave was undertaken to explorethe Ambiguous Ambiguous traits treatment of the individualsfollowing death,mode M? Possible male MMale of burial and whether burial of the individuals had ?Unknown taken place at once or over several episodes. More specifically, they were examined to explorewhy

30 Chapter Three disarticulation had occurred, whetherthe remains appreciation of the spatial relationships (from all weredisarticulatedprior to their burial (eitherasa angles) between them, otherthan indicated in two result of natural decomposition or deliberatedismem- dimensional photographs, sketches and plans. Ana- berment), or whetherthe disarticulation was aresult lysis of the deposit from all angles was essential in of naturalprocesses followingdeposition. These were order to explore whether individuals had been exploredbyanalysing bone surfacesfor ancient disposed of in aorderlyfashion, whether the deposit modifications and by undertaking elementorskeletal represented asingleevent or several burial episodes, part matching exercises,whilsttakingintoaccount and whether complete or incomplete, fullydecom- archaeological contextinformation (seeChapter 2). posed or partially decomposed bodies had been Analysis of ancient modifications involvedmacro- deposited. scopic observation of bonesurfaces formarks that Although several different approaches have been may be indicative of deliberate and/or natural employed to record and analyse previously exca- disarticulation. For example, cut marks in areas of vated mass deposits of human remains, these were muscleattachment andinthe region of articulations not considered to be ideal forthe presentstudy. may indicate deliberate disarticulation (Hurlburt Two-dimensional recording was employed during 2000), while animal gnawing and bleaching may the excavation of mass deposits from the north and suggest exposure (ibid.). All ancient modifications south burial chambers at Hazelton North, Somerset, were recorded with reference to published criteria England(Saville 1990). In additiontophotography, (McKinley 2004; Loe and Cox 2005). this involved drawing bones in plan on overlays, The visual matching of elements or skeletal parts assigning each boneaunique number and recording was undertaken based on overallmorphology, the absoluteheight of each bone in situ,asavertical whilst takinginto accountarticulations, conjoining control (ibid., 81). This information was used to fragments, biological parameters(eg age and sex), create aseries of plans, in sequence, which were pathology, condition, ancientmodifications, bone amalgamated to give an overalltwo dimensional dimensions and stratigraphic relationships (Adams impression of the density of the deposits in plan. and Byrd 2006; Duday 2006; Stewart 1979; White Combined with an element matching exercise and 2003; Wright 2003). Thethree dimensional relation- creatingplots of particularskeletal parts, the distri- ship between elements and/or skeletal parts with bution of elements, both horizontally and vertically, each other and articulated skeletons was explored by was analysed to explore patterning. Larger,more the application of Crossbones,acomputer software recognisableelements were represented naturalisti- programme specifically designed for the present cally, while symbolswereemployed to represent analysis (Isaksen et al.forthcoming). The background smaller bones. to this methodologyisdescribed more fully below. This approach was timeconsuming and provided Time constraints did not allowfor anything but no opportunityfor on-site interpretation of the minimalexamination of the assemblage for modifi- depositsothat excavation strategy could be in- cationsand matches. Thus, fragmented elements formed. Some bones were overlainand thusnot were not reconstructed. This has precludedthe represented in the overallamalgamated drawing. employment of conjoining exercises, except in the Further, the drawing, as the authors state, givesa most limited manner, as described by White (1992, rather false impressionofthe deposits in that they 67–92). Further, there was no microscopic analysis of ‘ ...compress the verticaldimension...[and]...record bone surfaces to facilitate the identification of ancient only the bones and not anyofthe accompanying modifications. More detailed examination would be stones...... ’Thus, relationships to structures and beneficial in light of the results of the preliminary layers cannot be fully appreciated. observations presented below. An approach that involved superimposing rectified photographsover surveyed points was employed by Sutherland (2000) to amass grave of Background to the Crossbones methodology individuals who mettheir fate at the battle of Crossbones is asoftware package that was developed Towton(1461) in Yorkshire, England. Vertical to facilitate the visualisation of relationships between photographs, copied to scale, were traced, super- the skeletons in the mass grave (Isaksen et al. imposed over the surveyed points of each skeleton forthcoming).Itworks in conjunction with the and digitally scanned to create acomposite of the X-Bones surveymethodologywhich createstotal skeletons superimposed over each other(ibid., 38). station data using ID points that are assignedto This resulted in ‘ ... arepresentation of the original locations on each major element of the skeleton. appearance of the grave,asifitwere in an x-ray.’ These methods were developed because the ‘three- (ibid.). This processallowed stratigraphicrelation- dimensional chaos’ (Wright 2003, 1) presented by the ships within the grave to be explored, elements of mass grave meant that traditional methods, for skeletal parts to be matched, and the calculation of example, two dimensional plans and photographs, the minimum number of individuals to be clarified were an inadequate meansofrecording and inter- (the Harris matrix suggested adifferent number to preting the deposit. During excavation, skeletons that calculatedduring osteologicalanalysis). and/or bones often had to be removed before a Furthermore,hypothesesraised during the excava- sufficient amounthad been excavated to allowfull tion could be tested.This included purported

31 Life andDeathinaRoman City evidence forbinding based on the posture of one of undertaken to explore whether the LondonRoad the skeletons, ahypothesis that was disproved when population differed significantlyfrom other Roma- the rectified images were scrutinised and, combined no-British populations in terms of its health and with osteological analysis,suggested an alternative physical attributes. Selection of the comparative explanation. Lastly,the images revealedthe manner populations was largelydeterminedbyavailability in which the pit had been filled with bodies, in this of data, but prioritised samplesfrom Gloucester- case by laying (rather than throwing) them down in shire, as well as acombination of urban and rural away that aimed to fit as many individuals in as populations. Thesamplesthat were employed from possible. Gloucestershire were 124–130London Road (Clough Similar methodsthat employ digital rectification 2003), Cirencester (Wells 1982aand b), and King- have been employed elsewhere (for example, see sholm (Robertsand Cox 2003). Other samples are Jacobi et al. 2006 on an Iron Age mass grave from referenced, where appropriate, in the text. Westerhausen, Germany). These provideamore rapid meansofrepresenting mass deposits than hand Results drawing. However, the method is neverthelessstill time consuming and relies on good quality photo- The discreteinhumations graphs taken in ahorizontal plane, requirements that Condition and completeness may not always be metduring some excavations where access is inhibited, timeispressured and the Most skeletons were in avery good to fair condition work undertaken in poorweather conditions. Further, (Table 3.4). This meansthat most of the assemblage the resultant images, which show naturalistic repre- had undergone minimalormoderatesurface ero- sentations of bones superimposed uponone another, sion. However, fragmentation was considerable are un-necessarily complex for exploring spatial acrossall condition categories. Few epiphyseshad relationships. survivedand there were no skulls that had survived Asolution to these problems was recently provided completely intact. This has significantly limited by Wright (2003) who produced athree-dimensional metrical analysis and the estimation of age and sex. programcalled ‘BODROT’. Aimed at massgraves that The fragmentary nature of most pelves has meant are encountered in forensic contexts, the program that, for the vast majority of skeletons, observations createsthree-dimensional stick figures based on relating to sex have had to rely on traits of the skull survey datataken at certain anatomical landmarks. and for age, wear patterns on the molar teeth. The figures can be rotatedsothat they maybeviewed Almost half of the skeletons were less than 25% from all angles to explorespatial relationships. complete (Table 3.5). This is notsurprising given the Primarily designed for awide audience comprising high level of truncation that had occurred across lawyers, archaeologists, magistrates and courtsin the site caused by the later diggingofditches general,the programwas made freely available and (five skeletons),modern features (21 skeletons) and was designed to be veryaccessible. Unfortunately, the Romangraves(nine skeletons). linear representation of bones makes it hard to ascertain bone type or directionality, and the software Biological age and sex is not easy to integrate with otherdata sources. Crossbones works on the same principle as BOD- The assemblage comprised 51 adults and nine ROT: it usesabstractionand schematisation in order subadults. Threeskeletons were of unknown age, to facilitate the visualisation of relationships between but all had dimensions that indicatedthat they were skeletons and elements (Isaksen et al.forthcoming). probablyover 10 years old when they died. However, unlike BODROT,itgenerates pyramids Adult deaths were highest between the ages of 18 which allowthe direction of the bones, and orienta- and 35 years(Table 3.6; Fig. 3.1). Seven individuals tion of the skeletons, to be observed(see Figs 3.2–3 had survivedinto the older age category, set hereas below). When integrated with osteological and arch- being over the age of 45 years. Subadults comprised aeological databases skeletons can be sorted visually (by colour or line)according to different variables (for example, sex and location). For the purposes of the present report, data were sorted by colour accor- Table 3.4 Condition of the skeletons from the discrete ding to bone type and/or skeletal component (eg burials. cranium, mandible, upper limb or lower limb) and Condition Number of skeletons % their location within the pit in an attempttomatch elements and/or components and analyse spatial Excellent (grade 0) 00 relationships between elements and/or components. Very good (grade 1) 16 25.3 Good (grade 2) 15 23.8 Comparative samples Fair (grade 3) 20 31.7 Poor (grade 4) 12 19.0 Osteological data were compared with those that Destroyed (grade 5) 00 have been reported for several Romano-British TOTAL 63 assemblages, where relevant. This was primarily

32 Chapter Three

Table 3.5 Completeness of the skeletons from the There were 11 males, 12 possible males, four discreteburials. femalesand sevenpossible females(Table 3.7). In addition, one late adolescent skeleton was estimated Completeness Number of % to have been male (Skeleton1112), based on cranial skeletons and pelvic traits. This skeletonhas been included in >75% (grade 1) 711.1 all calculations involving male skeletons. Sex could 75–50% (grade 2) 14 22.2 not be estimatedfor 17 skeletons. 50–25% (grade 3) 12 19.0 There were more females(36.3%; 4/11) below the <25% (grade 4) 30 47.6 age of 25 years(young adult age category)compared TOTAL 63 to males(26%;6/23). However, more males (21.7%; 5/23) were presentthan females (9%; 1/11) in the over 45 year age category(Table 3.7). Numbers were too small to observe any pattern in the age and sex Table 3.6 Age-at-death distribution of the discrete distribution of skeletons by phase. burials. Metrical analysis Age category Number of % skeletons As aresult of fragmentation,only alimited number of cranial and post-cranial measurements could be Fetus 00taken.Only those relating to stature estimationare Perinate 11.5 presented here. The remainder have been retained Neonate 00for the archive.Itwas possibletocalculate the Infant 00stature of eight skeletons (Table 3.8), including four Young child 23.1 femalesand four males. One of these, Skeleton1181, Older Child 23.1 dated from the 3rd to 4th centuries AD and was Adolescent 34.7 estimatedtohave been 162 cm (50 3.100)tall. Four Subadult (age unknown) 11.5 skeletons provided measurements from the most Young adult 12 19.0 accurate element, the femur.Threeskeletons pro- Middle Adult 11 17.4 vided measurements from the humerus, the least Mature Adult 34.7 accurate bonefor estimating stature. Older Adult 711.1 The average female height was 160 cm (50 2.400) Adult 18 28.5 and the averagemale was 169 cm (50 5.400)(Table 3.9). Unknown 34.7 Although the average male stature is skewed by Total 63 aheight of 156 cm (50 1.100)for Skeleton 1393, the three remaining males measured170 cm (50 5.700) and over. 15% (9/60) of the total aged sample (Table 3.6). Deaths among this group were highest in adoles- Non-metric traits cence, followed by young and old childhood. Subadults, including infants and younger, were Atotal of 46 adult individuals provided information under-represented. Only oneperinate was identiﬁed on 20 cranial and 18 post-cranialnon-metric traits based on the presence of one rightfemur (Skeleton (Tables3.10 and 3.11). The small samplesize has 1073). precludedanalysing the data by sex or age.

Figure 3.1 Mortality proﬁle(discrete burials).

33 Life andDeathinaRoman City

Table 3.7 Age distribution in the adult sample (N=51) according to biological sex.

Age category Female ?Female Unknown ?Male Male %Females (n) %Males (n)

Young adult 04 2517.8 (4) 11.7 (6) Middle adult 22 2417.8 (4) 9.8 (5) Mature adult 10 1101.9 (1) 1.9 (1) Older adult 01 1141.9 (1) 9.8 (5) Adult 10 11 15 1.9 (1) 11.7 (6) Total 47 17 12 11 21.5 (11) 45.0 (23)

Table 3.8 Bones from the discrete burial assemblage employed in the computation of stature*.

Skeleton Sex Element employed Side Maximum Stature Feet/inches in Stature calculation Length (Trotter 1970)

1103 FFemur L41.3 cm 156 cm 6 3.72 cm 5’ 1’’ 1165 FHumerus R31.4 cm 163 cm 6 4.45 cm 5’ 3’’ 1181 FFemur L43.9 cm 162 cm 6 3.72 5’ 3’’ 1390 FFemur L43.5 cm 161 cm 6 3.72 5’ 2’’ 1109 MHumerus L34.5 cm 176 cm 6 4.05 cm 5’ 7’’ 1153 M? Humerus L32.4 cm 170 cm 6 4.05 5’ 5’’ 1313 MRadius R25.4 cm 175 cm 6 4.32 5’ 7’’ 1393 M? Femur L40.1 cm 156 cm 6 3.27 cm 5’ 1’’

*F =female; F? =possibly female; M=Male; M? =possibly male; L=Left; R=Right

The most frequent cranial non-metric traitwas the surfaces in particular may be under greater mechan- supraorbital notch (93.7%), followed by the mastoid ical influence than geneticinfluence. For example, Os foramen(53.8%). The supraorbital notch is mani- acromiale, represented as non-fusion of the acromial fested early on in skeletal development and therefore processofthe scapula,has beenlinked to activity- may be under strongergenetic influence than other inducedtrauma occurring at ayoung age (Stirland traits that are not (Hauser and De Stefano1989). The 1998). In the presentsample,nosuch trait was mastoid foramen on the otherhand,isbelievedtobe observedand, overall,traits involving joint surfaces under alow genetic influence compared to other were not very frequent. cranial traits (Sjøvold 1984). In fact, this trait is Comparative analysis of non-metric traits is frequently observed in skeletal remains (Hauser and difficult owing to the varied manner in which data De Stefano1989). Preliminaryanalysis of the present for these have beencollected and published for data suggests no obvious association between various populations. However, metopism would individuals with the samecranial traits and their seem to have been more frequent among the London locationinthe cemetery.Cranial non-metric data, Road individuals (TPR 25%) compared to those from therefore, would not seemtoindicate anygenetic 124–130 LondonRoad (TPR 5.4%), Kempston,Bed- relationships between individuals. fordshire (TPR 12%) and Cirencester (TPR 8.2%) The most frequent post-cranialnon-metric traits (Boylston and Roberts 1996). Also more frequent was were atlas double facets and lateral tibial squatting the septal aperture which had aTPR of 25% at facets (Table 3.11). These were both presentonover London Road compared to aTPR of 8.4% for 66.6% of the skeletons that had these elements Kempston and aTPR of 4.5% for Cirencester, and availablefor examination. Traits that involve joint the vastus notch which had aTPR of 28.5% at London Road, aTPR of 5.7% at Kempstonand aTPR of 11.8% at Cirencester. Table 3.9 Stature summary data from the discrete burial assemblage*. The mass grave Sex NMethod Average Minimum Maximum S.D. Condition and completeness stature Articulated skeletons within the mass grave were very F4Trotter 160 cm 156 cm 163 cm 2.69 incomplete (Table 3.12). The majority (65%) were M4(1970) 169 cm 156 cm 176 cm 7.98 represented by less than 25%and mainly comprised articulated limbs. Only 17 skeletons (8.4%) were over *F=female, M=male; N=number of individuals in the sample; 75% complete. All of the small finds (‘sf’ number S.D. =standard deviation allocated to each individual boneorarticulating

34 Chapter Three

Table 3.10 Cranial non-metric traitfrequencies in the discreteburial assemblage.

Trait Left side Right side Unilateraltraits Cranial prevalence %(n/N) %(n/N) %(n/N) %(n/N)

Highest nuchal line ––14.2 (2/14) 14.2 (2/14) Ossicle at lambda –– 0(0/13) 0(0/13) Bregmatic bone –– 0(0/13) 0(0/13) Palatine torus ––25.0 (2/8) 25.0 (2/8) Metopism ––19.0 (4/21) 19.0 (4/21) Lambdoid ossicle 42.8 (3/7) 66.6 (4/6) –55.5 (5/9) Coronal ossicle 0(0/7) 0(0/7) –0(0/9) Epipteric bone 0(0/1) (left?) –– 0(0/1) Asterion ossicle 0(0/2) 33.3 (1/3) –33.3 (1/3) Parietal notch 0(0/1) (left?) –– 0(0/1) Fronto-temporal articulation 0(0/1) (left?) –– 0(0/1) Parietal foramina 54.5 (6/11) 45.4 (5/11) –64.2 (9/14) Access infraorbitalforamina0(0/2) –– 0(0/2) Multizygomatic foramina 46.6 (7/15) 38.4 (5/13) –47.0 (8/17) Auditory torus 0(0/22) 5.2 (1/19) –4.5 (1/22) Mandibulartorus 20.0 (4/20) 19.0 (4/21) –19.0 (4/21) Maxillatry torus 33.3 (3/9) 33.3 (3/9) –33.3 (3/9) Supraorbital foramen 41.6 (5/12) 21.4 (3/14) –40.0 (6/15) Supraorbital notch 81.8 (9/11) 93.3 (14/15) –93.7 (15/16) Mastoid foramen 38.4 (5/13) 40.0 (4/10) –53.8 (7/13)

Table 3.11 Frequency of post-cranial traits.

Trait Side prevalenceSkeletal prevalence %(n/N) Left %(n/N) Right %(n/N)

Septal aperture 7.6 (1/13) 25.0 (4/16) 25.0 (4/16) Supracondyloid process 0(0/13) 0(0/16) 0(0/17) Atlas double facet 55.5 (5/9) 66.6 (6/9) 66.6 (6/9) Atlas lateral bridge 0(0/7) 0(0/7) 0(0/9) Atlas posterior bridge 12.5 (1/8) 14.2 (1/7) 12.5 (1/8) Atlas bipartite foramen 0(0/5) 0(0/4) 0(0/6) Allen’s fossa 0(0/11) 0(0/9) 0(0/12) Poirier’s facet 10.0 (1/10) 14.2 (1/7) 18.1 (2/11) Plaque 20.0 (2/10) 16.6 (1/6) 20.0 (2/10) 3rd trochanter 8.3 (2/24) 28.5 (6/21) 23.0 (6/26) Hypotrochanteric fossa 31.0 (9/29) 31.8 (7/22) 33.3 (10/30) Exostosis in intertrochanteric fossa 42.8 (6/14) 33.3 (3/9) 46.6 (7/15) Vastus notch 20.0 (1/5) 33.3 (2/6) 28.5 (2/7) Vastus fossa 0(0/4) 0(0/6) 0(0/6) Emarginate patella 0(0/5) 0(0/7) 0(0/8) Tibia lateral squatting facet 60.0 (3/5) 60.0 (3/5) 66.6 (4/6) Tibia medial squatting facet 0(0/5) 0(0/4) 0(0/5) element that couldnot be associatedwith aspeciﬁc bone surface erosion was minimalorslight and skeleton at the timeofrecording; See Chapter2) bones generally had afresh appearance and clearly represented less than 25% of acomplete skeleton. visible surface morphologies (McKinley 2004, 16). Most small ﬁnds comprised crania (8.3%;35/419); Fragmentation was considerable acrossall condition mandibles (7.3%;31/419); ribs (21.7%; 91/419); and categories. major long limb bones (32.9%;138/419). Hand and Nine skeletons (4.4%)were in an excellent condition foot bones, sternums and vertebrae were infrequent and, of these, all but one (1518) were less than 25% (Table 3.13). complete. Interestingly, amongst these nine skeletons, Most of the articulated skeletons were in agood or were three subadults. The excellent preservation of very good condition (Table 3.14). This means that subadult skeletons is unusual because their small,

35 Life andDeathinaRoman City

Table 3.12 Completeness of the articulated skeletons Table 3.15 Conditionofthe skeletal small finds within within the mass grave. the mass grave.

Completeness Number of %ofthe assemblage Category Number of skeletons % skeletons (N ¼ 201 skeletons) Excellent 81.9 >75% 17 8.4 Very good 141 33.6 75–50% 23 11.4 Good 148 35.3 50–25% 29 14.4 Fair 94 22.4 <25% 132 65.6 Poor 26 6.2 Total 201 Destroyed 20.4 Total 419

Table 3.13 Composition of the skeletal small finds The condition of the small finds and disarticulated assemblage within the mass grave. elements (Tables 3.15 and 3.16) was similar to that of the articulated skeletons. Thus, regardless of location Element Number and human remains category (articulated skeleton, (% of 419 elements) small findordisarticulated element), all bones in the Skull 11 (2.6) mass grave were fairly homogenousinterms of their Cranium 35 (8.3) condition. Mandible 31 (7.3) Small finds that are only one 8(1.9) Taphonomy loose tooth Sternum 1(0.2) Therewere no obvious cases where fragments of a Vertebrae 10 (2.3) single bonefound in different locations within the pit Ribs 91 (21.7) conjoined. This may be because no formal conjoining Scapula 18 (4.2) exercises (see White 1992, 65–77) were undertaken Clavicles 21 (5.0) and thus such informationhas been missed. Alter- Pelvis 11 (2.6) natively, this may indicate that fragmentsofthe Long bones (Humerus, Radius, 138 (32.9) same bone were notdispersed within the pit, just Ulna, Femur, Tibia, Fibula) bones from the sameskeleton. Patella 19 (4.5) Attempts at matching met with limited success. Hand bones (carpals, metacarpals, 11 (2.6) During excavation, several records were madethat phalanges) suggested possible matches between elements, and Foot bones (carpals, metacarpals, 14 (3.3) these were confirmed during macroscopic examina- phalanges) tion and by application of Crossbones.Inall of these Total 419 cases, the matches occurred between elements that were in approximate anatomical alignment with one another. Thus, for example, Skeleton1569, represented by aleft leg was matched with the remains Table 3.14 Conditionofthe articulated skeletons within comprising Skeleton 1560, and the remainsthat the mass grave. comprised Skeleton 1521 were matched to the bones recorded as Skeleton 1591 and small find390. Category Number of skeletons % Based on the recommendations of Wright (2003), Excellent 94.4 measurements of the humerus and femur were Very good 72 35.8 employed to match dispersed elements and bone Good 82 40.7 units. However, the incompleteness of these elements Fair 29 14.4 meantthat the full set of requiredmeasurements for Poor 94.4 Destroyed 00Table 3.16 Conditionofthe disarticulated elements Total 201 within the mass grave.

Category Number of Elements % fragile bones do not usually survive as well as thoseof Excellent 10.3 adult skeletons (Chamberlain 2000). The subadults Very good 41 13.0 from the mass grave may be an exception because the Good 85 27.1 burial of many individuals together may have Fair 151 48.2 provided amoreprotectedenvironmentthan asingle Poor 33 10.5 grave.The overall better condition of the skeletons Destroyed 20.6 from the mass grave compared to the discrete Total 313 skeletons,may also be explained by this.

36 Chapter Three each bone (seven for each)could not be taken in all separate queries that were run using Crossbones. but onecase(ahumerus from Skeleton 1668). Thus, The first explored horizontal and/or vertical rela- the remains did not provideany metrical data that tionships between articulated skeletons with lower could be employed to match elements. limb bones and articulated skeletons that lacked There was no evidence that any of the remains had lower limb bones. The second explored horizontal been modified by scavengers, exposureindaylight and/or vertical associations between articulated or deliberate disarticulation with ablade or other skeletons with upperlimb bones and articulated sharp tool. This suggests that corpses had not skeletons that were withoutupper limb bones. No skeletonised or partially decomposedprior to their obvious relationships were observedinthe vertical deposition in the pit. This is supported by a plane for bothanalyses. Patterns were,however, preliminary analysis of the occurrence of articulating observedinthe horizontal plane for bothanalyses. joints in the entireassemblage (based on sketches In general, this showedthat upper limb bones made of the boneinthe field). This indicatesthat were more dispersed than lower limb bones (Figs 3.2 there were at least 35 articulated crania and and 3.3). The latter were concentrated in the western mandibles, 27 righthips,24left hips, 16 rightknees side of the pit in the sameareaasskeletons that and 14 left knees.Shoulders, elbows, wrists and lacked lower limbs. The formerwere located ankles scored between 12 and three occurrences. comparatively further from skeletons that lacked Three dimensional analysis of the assemblage did upper limb bones. Some skeletons without upper not identify anymatches between elements but limbs were locatedinthe middleofthe pit, some provided informationregarding the spatial relation- distancefrom the western half where the majority of ship of elements. This primarily relates to two the small finds were located.

Figure 3.2 Three dimensional analysis of articulated skeletons with andwithout lower limb bones in the mass grave.

37 Life andDeathinaRoman City

Figure 3.3 Three dimensional analysis of articulated skeletons with and withoutupper limb bones in the mass grave.

Despite the limited successatmatching elements of specific anatomical landmarks such as the glabella, and skeletal units from the pit, thisbasic taphonomic the external occipital protuberance,the radial notch analysis of the remains suggests that the individuals of the left ulna and the fovea capitis of the left femur. had beendepositedinthe pit as articulated fleshed, The highest MNI achieved by the most frequent fresh, corpses. The disassociation of elements is, anatomical landmark or bone segment, the mid-shaft most likely, the result of post-depositional processes of the right femur, was 87 individuals (Table 3.17) relating to body decomposition (Duday 2006). This is and was based on femoral mid-shafts that were over aplausible explanation for the concentration of 50% complete. Of these, 73 belonged to articulated partialskeletons and skeletal elements in the same skeletons,13were small finds and one was disarti- area of the pit. culated. The mid-shaft of the femur, which comprises athick cylinder of compactbone (White 2003), is the strongest boneinthe skeleton. It is therefore Minimum numberofindividuals not surprising that this element shouldyield the high- To avoid duplication, the minimum number of est MNI count in the present sample. This MNI was individuals (MNI) was calculated for the entire followed by alower count for the left femur (n=84) assemblage, including articulated skeletons, small and the mental eminence of the mandible (n=78). finds and disarticulated bones. This was based on the Based on the number of right femora(87) there was number of repeatedelements, whilst taking into aminimum of 76 adults and sevensubadults (four accountage, sex and bone morphology. Owing to the were of unknown age). Thenumber of subadults was, fragmented nature of the sample, it was importantto however, higher than this, as indicatedbyrepeated ensure that abone was not counted more than once. countsofthe mental eminence.This suggests a This analysis, therefore, has employed the calculation minimumnumber of 10 subadults. Although other

38 Chapter Three

Table 3.17 MNI calculation for the mass grave assemblage according to repeated anatomical landmarks in different bone elements.

Speciﬁc skeletal region Number of repeatedelements MNI Total

Subadult Adult Age unknown

Right femur mid-shaft (>50% complete) 7764 87 Left femur mid-shaft (>50% complete) 9714 84 Mental eminence (mandible) 10 68 078 Right Mastoid process (temporal bone) 9614 74 External protuberance (occipital bone) 8640 72 Left mastoid process (temporal bone) 7640 71 Glabella (frontal bone) 8510 59 Left talus (>50% complete) 3395 47 Right talus (>50% complete) 5334 42 Radial notch (left ulna) 3341 38 Radial notch (right ulna) 5311 37 Fovea capitis (right femur) 5311 37 Fovea capitis (left femur) 3330 36 Nutrient foramen (left tibia) 1302 33 Olecranon fossa (right humerus) 3251 29 Radial tuberosity (right radius) 5222 29 Nutrient foramen (right tibia) 2212 25 Tibial tuberosity (left tibia) 0232 25 Radial tuberosity (left radius) 3180 21 Tibial tuberosity (right tibia) 1182 21 Olecranon fossa (left humerus) 3170 20

subadult bone fragments, such as those from the left and six in the adolescent agegroup. The remainder temporal and left parietal, were repeatedamong 19 were aged based on skeletal maturity and/or long skeletons and 20 small finds numbers, these were less bone lengths. Addingthese 15 subadult individuals than 50% complete and aretherefore unreliable to the 76 adult individuals, as inferred from the right for estimating the minimum number of subadults femur middle shaft, the total minimum number of (Dobneyand Reilly 1988). individuals from the mass grave is 91. The minimum number of subadults is increased furtherto15individuals based on age indicators Recovery rate rather than repeated boneelements. There were no perinatal or neonatal remains, but an MNI for all Based on an MNI of 91 individuals, the recovery othersubadult age groups couldbeestablished rate of bones from the grave pit was calculated (Table 3.18). Estimations were based on dental (Table 3.19) to explore whether partially complete development for oneskeleton in the young child skeletons or corpses had been deposited. This was age category, three in the older child agecategory undertaken by dividing the number of elements presentbythe total number that would be expected for 91 individuals. This calculation assumesthat all Table 3.18 MNI calculation of the subadultsample from individuals were complete when they were buried. the mass grave. Only the long bonesegments that provided the highest count were included. Age category MNI (based on Elements employed in Recovery rates were highest for femora and skull most frequently the MNI calculation bones. Allother elements had arecovery rate of repeatedelements) around50%. This is with the exception of verysmall Infant 1Not applicable bones, such as foot phalanges, which had alow Young child 3Repeated left femora recovery rate. Some bias may have beenintroduced Older Child 4Several repeatedelements into these resultsbecauseelements that were 5 50% including the mandible, complete were not counted. Further, bone fragments the atlas and axis vertebrae from some parts of the skeleton, for example, the Adolescent 7Several repeatedskull skull, are more easily identified by the Osteologist elements, the atlas and than others such as hand and foot bones. axis vertebrae Despite these caveats,overall patterns of reco- Total 15 very would seem to be reflective of characteristic preservation patterns of the skeleton resultingfrom

39 Life andDeathinaRoman City

Table 3.19 Skeletal representation and bone recovery Table 3.20 Age-at-death distribution in the mass grave rates for the mass grave assemblage. sample.

Bone element/anatomical Number recovered/ % Age category Number of skeletons %ofindividuals landmark number expected (N=91)

Femora (middle shaft >50%) 171/182 93.9 Foetus 0 mental eminence (mandible) 78/91 85.7 Perinate 0 Temporal bone (mastoid process) 145/182 79.6 Neonate 0 Occipital bone (Nuchal crest area) 72/91 79.1 Infant 11.0 Clavicles (shafts) 144/182 79.1 Young child 33.2 Humeri (distal shaft >50%) 128/182 70.3 Older Child 44.3 Frontal bone (glabella) 59/91 64.8 Adolescent 77.6 Zygomatics 117/182 64.2 Young Adult 36 39.5 Patellae 109/182 59.8 Middle Adult 13 14.2 Vertebrae 1274/2184 58.3 Mature Adult 77.6 Ulnae (proximal shaft >50%) 104/182 57.1 Older Adult 55.4 Tibiae (middle shaft >50%) 101/182 55.4 Adult 15 16.4 Radii (middle shaft >50%) 99/182 54.3 Total 91 Fibulae (middle shaft >50%) 98/182 53.8 Scapulae (glenoid fossa) 81/182 44.5 Metacarpals 371/910 40.7 All subadults were assigned to an age category as Ribs 850/2184 38.9 described above. Metatarsals 340/910 37.3 Almost 40% (N=36) of the skeletons were assigned Tarsals 449/1274 35.2 to the young adult age category (Table 3.20 and Hand phalanges 680/2548 26.6 Fig. 3.4). Thenumber of skeletons assignedtothe Pubis 49/182 26.9 infant and youngerchild agecategories was lower Carpals 356/1456 24.4 than expected(N=1)and only five skeletons were Foot phalanges 184/2548 7.2 estimatedtohave beenolder adults whenthey died. Nasal bones 10/182 5.4 Sex estimation was based on acombination of pelvic Vomer 4/182 2.1 and skull indicators forthe more complete skeletons, Palatines 4/182 2.1 but relied on alimited number of skull indicators Hyoid 2/91 2.1 for the small finds and disarticulated elements Sternum 2/91 2.1 (Table 3.21, Fig. 3.5). Most of the individuals were Lacrimal bones 0/182 0 male, although 35.5% of the adult skeletons were of undeterminedsex. Deaths were highest for bothmales and females in the inherent structural properties of bone (Bello and the young adult age category. Out of atotal of 14 Andrews2006, 10). Thus, large, dense bones, such as femalesand 35 males,seven females (50%)and 19 the femur and skull, have ahigher survival rate than males (54.2%)were assigned to this agecategory. do less dense,smaller bones, such as thosefrom the hands and feet (ibid.). These patterns aresimilar to Metrical analysis recovery ratesthat are observed among discrete medieval skeletons that were interred as complete Alimited number of measurements could be taken corpsesand were not subsequentlyremoved or from skulls and long bones and have been recorded deliberately disturbed until archaeological excava- for the archive.Nostatures could be estimated.This tion (ibid., 9). These findings therefore suggest that is with the exception of skeleton 1668 whichhad a corpseshad probablynot skeletonised before they complete left humerus that provided amaximum were interred in the mass grave and that no selection length of 30.5 cm. The absence of askull and pelvis of elements for burial had taken place. meantthat it was not possible to estimate the sex of this skeleton. Applying the humeral length to regression equations for both females and males suggests a Biological age and sex female height of 160 cm (50 2.400)and amale height of The ages of adult skeletons were primarily estimated 164 cm (50 3.800). by employing the most frequently occurring skeletal indicator; wear patterns on the molar teeth combined Non-metric traits with ante-mortem tooth loss. This is with the exception of 17 articulated skeletons forwhich ages could Limiteddata on non-metric traits couldbeobtained be estimated based on acombination of indicators (Tables3.22 and 3.23). The small size of the sample including dentitions and pelves. Only dentitions precludedanalysis of the data by sex and age. that had the mental eminence presentwere counted Several traits including the coronal ossicle, epipte- for thisanalysis, so as to avoid anyduplication. ric bone, asterion ossicle, parietal notch and the

40 Chapter Three

Figure 3.4 Mortality proﬁle(massgrave). fronto-temporal articulationcouldnot be scored Dental health status owing to the incompleteness of the remains. by Nicholas Ma´ rquez-Grant The most frequently occurring cranial traits were the parietal foramina, multizygomatic foramina, Introduction access infraorbital foramina and the lambdoid ossicle. Examinationofconditions such as dental caries, ante- However, for the latter two, only one bone was mortem tooth loss and dentalcalculus can provide availablefor each observation. Themost frequently information on the dietofthe populationaswell as occurring post-cranialtraits were the hypotrochan- the levels of oral hygiene.Moreover, lesions called teric fossa, the tibial squatting facet and the vastus hypoplasias, which can be present on tooth enamel, notch. Again, for the latter,observations were biased may provideaninsight intoliving conditions and by the small number of skeletal elements that were levels of nutrition, infection and hygiene during the availablefor examination. childhood years. In addition, data on oral pathology Overall,the range and frequency of cranial and provideinformationonthe social, cultural and econo- post-cranial non-metric traits is broadly similar to mic context in which apopulation lived and may be that which has beenrecorded for the discrete ameans of inferringcultural change (Emery1963). skeletons.This suggests that both populations The examination of the dentitions that belonged to were exposed to similar genetic and environmental the discrete skeletons and the mass grave skeletons inﬂuences. focusedonenamel hypoplasia, dental caries, ante-

Table 3.21 Age distribution in the adult sample (N=76) according to biological sex.

Age category Female ?Female Unknown ?Male Male %Females(n) %Males (n)

Young Adult 52 10 11 89.2 (7) 25.0 (19) Middle Adult 31 2435.2 (4) 9.2 (7) Mature Adult 02 3202.6 (2) 2.6 (2) Older Adult 01 1301.3 (1) 3.9 (3) Adult 00 11 13 0(0) 5.2 (4) Total 86 27 21 14 18.4 (14) 46.0 (35)

Figure 3.5 Sex distribution of the adult sample (massgrave).

41 Life andDeathinaRoman City

Table 3.22 Cranial non-metric traitfrequencies in the mass grave assemblage.

Trait Side prevalence Unilateral trait prevalence%(n/N) Left %(n/N) Right %(n/N)

Highest nuchal line 0(0/2) Ossicle at lambda 0(0/1) Bregmatic bone 0(0/7) Palatine torus 25.0 (1/4) Metopism 9.3 (4/43) Lambdoid ossicle 100 (1/1) – Parietal foramina 80.0 (4/5) 100 (9/9) Access infraorbitalforamina100 (1/1) – Multizygomatic foramina 68.0 (17/25) 72.7 (16/22) Auditory torus 0(0/32) 0(0/30) Mandibulartorus 40.0 (2/5) 33.3 (1/3) Maxillary torus –– 60.0 (3/5) Supraorbital foramen 53.8 (7/13) 46.6 (7/15) Supraorbital notch 69.5 (16/23) 75.0 (15/20) Mastoid foramen 0(0/3) 50.0 (2/4)

mortemtooth loss, dental calculus, periodontitis, 1317) provided dental information, since in the other periapical cavities,dental attrition and dental skeletons the dentitions were not preserved. In the anomalies such as toothagenesisorchipping. Data mass grave, dental information came from small are presented separately for subadult and adult ﬁnds 540, 564,588 and 6873, and skeletons 1526, individuals, and forthe discrete burial assemblage 1547, 1559, 1570, 1584, 1595, 1608, 1615, 1623, 1642, and the mass grave assemblage. 1672 and 1685. Atotal of 38 deciduous and 63 (21 unerupted) permanent teeth from the discrete burial assemblage Materials were availablefor examination.The total sample Subadultsample examined for the subadult individuals from the mass grave consisted of 79 deciduousteeth and197 (three Of the ninesubadult skeletons in the discrete burial unerupted) permanent teeth. assemblage, only four (skeletons 1066, 1086, 1277,

Adult sample Table 3.23 Post-cranial non-metric traitfrequencies in the mass grave assemblage. Therewere51adult individuals in the discrete burial assemblage and 76 adult individuals from the mass Trait Side prevalence grave sample.Ofthese, 36 from discrete samples and an MNI of 68 from the mass grave provided Left %(n/N) Right %(n/N) information regarding dental health. Septal aperture 0(0/10) 0(0/18) Table 3.24 provides abreakdownofthe number Supracondyloid process 6.2 (1/16) 5.2 (1/19) of teeth presentinthese adultsamples. The number Atlas double facet 18.7 (3/16) 21.4 (3/14) of teeth expected in this table has been determined Atlas lateral bridge 0(0/14) 0(0/11) by counting the number of common teeth in a Atlas posterior bridge 0(0/14) 0(0/11) humanbeing (32 permanentteeth)and multiply- Atlas bipartite foramen 0(0/4) 0(0/2) ing this by the number of adult individualsinthe Allen’s fossa 14.2 (1/7) 11.1 (1/9) sample. Poirier’s facet 0(0/7) 0(0/8) In total,557 and 1913 adult teeth were examined Plaque 14.2 (1/7) 0(0/9) from the discrete burial and the massgrave assem- 3rd trochanter 3.5 (1/28) 6.6 (2/30) blagesrespectively. Hypotrochanteric fossa 20.4 (10/49) 24.0 (12/50) Exostosis in 0(0/3) 0(0/3) Methods intertrochanteric fossa Vastus notch 18.1 (2/11) 23.0 (6/26) All observations were made macroscopically and Vastus fossa 0(0/7) 0(0/25) employed standard published criteria (Brickley Emarginate patella 0(0/25) 0(0/36) and McKinley 2004). For subadults, deciduous and Tibia lateral squatting facet 100 (1/1) 100 (2/2) permanent teeth were analysed separately. Afull Tibia medial squatting facet 0(0/1) 0(0/1) description of the methodsemployed is presentedin Appendix 1.

42 Chapter Three

Table 3.24 Number of teeth present, numberlost ante-mortem, number lost post-mortem and numbergenerally missing (no data).

Discrete cemetery Mass grave (51 adults) n(%) (76 adults) n(%)

Total teeth expected for 100% preservation and recovery 1632 2432 (n =32teeth per adult dentition) Teeth present 557 (34.1) 1913 (78.6) Teeth lost ante-mortem 42 (2.5) 107 (4.3) Teeth missing post-mortem 120 (7.3) –* Agenesis (congenitally absent/unerupted) 15 (0.9) 22 (0.9) Tooth status unknown(missing data) 898 (55.0) 441 (18.1)**

*teeth missing post- mortem in the mass grave. Since they are commingled, the number of loose teeth is greater than the number of alveoli that were observed as having teeth lost post- mortem (108 alveoli, 4.4%). Since it cannot be determined whether the teeth that are missing were or were not among the loose teeth in the assemblage,this ﬁgure cannot be calculated. **This calculation has been obtained by totalling the number of teeth present with the number lost ante-mortem and the number of examples of agenesis and subtracting this from the total number of expected teeth. The number of teeth lost post- mortem was excluded from this calculation.

Results geneticbackground (Pindborg 1982; King et al. 2002). Generally speaking, the prevalence of these All samplescontained across-representation of adult defects in apopulation is likely to result from a ages and bothsexes.Samples fromthe discrete burial complex interaction of factors. assemblage and the mass grave are presented along- The sample examined for enamel hypoplasia was side each othertofacilitate comparisons and discus- 279 teeth.This does notinclude teeth that were too sions. Adult ages have been pooled as well as left and poorlypreserved, had surfaces that were obscured right teeth,and maxillary and mandibular teeth. by calculus, or were too worn. The total prevalence Data for the discrete burial assemblage are rate was 35.1% (Table 3.25). In the majority of cases presented first, followed by thosefor the mass (>90%) the defects were linear, but the pitting form grave.These results are then compared. Adiscussion was also observed. and interpretation of the resultsfollows after. The total prevalence for London Road is higher than otherRomano-British populations (Robertsand The discreteburial assemblage Cox 2003, 140), including the cemetery of King- sholm, Gloucester where 10.4% of all teeth were The adult sample affected(Roberts1989). Enamelhypoplasia (EH) Dental caries Enamelhypoplasia is adental enamel defect caused Dental caries involvesthe destruction of the enamel by generalised disruptions to the enamel matrix surface, the dentine (internal partofthe tooth) and formation (Hillson 1996b). The prevalence of these the cement (outer layerofthe roots). This is caused lesions has been used to infer the level of physiological by the acid produced by bacteria present in dental stress in apopulation (Skinnerand Goodman 1992). plaque(Hillson 1996b, 269). Classified as an in- Since these disruptions occur during toothformation, fectiousdisease, it usually progresses gradually. these lesions provide an insight into health status The prevalence of dental caries is strongly influ- during childhood. enced by diet and subsistence technology (Larsen There are several types of enamel defects, but pits 1997). In particular, the consumption of carbohy- and linear defects are thosemostcommonin drates, especially sugar, is awell established cause of archaeological populations. The aetiology of hypo- cavitations (Hillson 1996b; Moyniham 1998). Other plasias is multifactorial (Pindborg 1982). Amongst factors, such as tooth crown shape, food texture, the factors contributing to the presence of these dis- attrition, frequency of consumption of food and ruptions are malnutrition, infectious disease,genetic inheritance may also play arole in the manifestation anomalies, diabetes, allergies, trauma, infections, of this condition (Powell 1985; Larsen et al. 1991). neonatal disturbances, psychological stress and

Table 3.25 Prevalence of enamel hypoplasia in the discrete burial assemblage.

Females %(n/N) Males %(n/N) Unsexed%(n/N) Total %(n/N)

All anterior teeth 54.3 (25/46) 79.4 (31/39) 54.2 (13/24) 63.3 (69/109) All posterior teeth 17.7 (14/79) 15.0 (9/60) 19.3 (6/31) 17.0 (29/170) Total 31.2 (39/125) 40.4 (40/99) 34.5 (19/55) 35.1 (98/279)

43 Life andDeathinaRoman City

Table 3.26 Dental caries prevalence in the discreteburial assemblage.

Females %(n/N) Males %(n/N) Unsexed %(n/N) Total %(n/N)

All anterior teeth 1.4 (1/70) 4.1 (4/96) 3.0 (1/33) 3.0 (6/199) All posterior teeth 12.5 (17/135) 18.9 (30/158) 16.7 (9/54) 16.1 (56/347) Total 8.8 (18/205) 13.4 (34/254) 11.5 (10/87) 11.4 (62/546)

(These are uncorrected rates which means that ante-mortem and post-mortem tooth loss have not been taken into account.)

The total prevalence fordental caries for the Atotal of 42 alveoli showedcomplete regeneration discrete skeletons was 11.3% (Table 3.26) which or initial signs of regeneration (teeth lostinarela- placesLondon Road among the highest rates to have tively short timebefore death). The total prevalence been recorded for Romano-British material (Roberts was 5.8% (42/719). Femaleswereless affectedthan and Cox2003, 131). Femaleshad alower frequency males (Table 3.28). than males, although ten teeth with cavities be- Dental calculus longedtoindividualsofundeterminedsex. Most Dental calculus consists of mineralised plaque cavitieswere present on the coronal surfaces of teeth, composed of micro-organisms that accumulate in or at the cemento-enamel junction (Table 3.27). the mouth and become imbedded in amatrix of ‘Gross’referstothe involvement of the entire surface proteinand saliva. Sugar in the dietaccelerates this of the tooth. process(Hillson 1996b, 254–55). Ante-mortem toothloss Atotal of 66.7% of the teeth from the discrete The loss of atooth before death is the end result of assemblage had deposits of calculus (Table 3.29). several disease processes. Calculus deposits can irri- This frequency is among the highest to be reported tate the soft tissue and the underlying bone, which for Romano-British material (see Roberts and Cox may lead to the reduction of the bone(periodontal 2003, 132). This is particularlyhigh when takinginto disease)and ante-mortem tooth loss (AMTL) accountthe fact that the prevalence of calculus in (Robertsand Manchester1995, 45). Teeth mayalso once living populations is underestimatedinarchae- be lost as aresult of peri-apical abscesses formed ologicalmaterial because it is easily flakedoff during through the exposure of the pulp cavity, causedby excavation and post-excavation handling. Calculus caries or excessive attritioncoupled with localised was judged to have been considerable for 3.2% of all resorption of the alveolarmargin. AMTL is regarded teeth.Rates were similar for bothmales and females. as adegenerative disease where the main contribu- Periodontal disease tory factors are old age and poor oral hygene. The principal predisposing factor in periodontal AMTL was identified as partial or complete disease is the accumulationofcalculus in dental remodelling of the alveolarboneand tooth socket, pockets. The disease begins as gingivitis (an inflam- which occurs after atoothislost (Costa 1980, 580). mationofthe soft tissues), which is transmitted to the The calculation of AMTL prevalence rates usually jaw itself. Resorption of the bone commences, follow- involvesobserving the number of regenerated alveoli ed by tooth loss. Thereisastrong link between in the assemblage and dividing it by the number of increase in ageand the increase in the prevalence alveoli present, the former indicating that the teeth were lost before death.However, in the present assemblage, the alveoli were extremely poor. Thus, there were discrete skeletons that did not have any Table 3.28 AMTL frequencies in the discreteburial alveoli preservedbut had all of their teeth present. population. The calculation of AMTLwas therefore determined Females % Males % Unsexed % Total % by dividing the number of regenerated alveoli by the (n/N) (n/N) (n/N) (n/N) total number of teeth present, added to the total of number of teeth lost post-mortem and number of All teeth 2.4 (6/242) 8.6 (30/346) 4.6 6/131 5.8 (42/719) alveoli with AMTL in the sample.

Table 3.27 Distribution of pattern of carious lesions in the discrete burial assemblage.

Lesion distribution Occlusal n/N (%) Coronal n/N (%) CEJ n/N (%) Root n/N (%) Gross n/N (%)

Anterior teeth (6 teeth; 6lesions) –4/6 (66.6) 1/6 (16.6) 0/6 (0) 1/6 (16.6) Posterior teeth (56 teeth; 59 lesions) 11/59 (18.6) 13/59 (22.0) 19/59 (32.2) 1/59 (1.6) 15/56 (26.7)

(Frequencies have been calculated by dividing the number of observed dental surfaces by the number of carious lesions present. Gross lesion frequencies were based on the number of teeth observed rather than the number of lesions observed since there was no indication as to the initiation site of the lesions.)

44 Chapter Three

Table 3.29 Dental calculus frequencies in the discrete burial assemblage.

Females %(n/N) Males %(n/N) Unsexed %(n/N) Total %(n/N)

All teeth 68.8 (139/202) 73.0 (165/226)45.3 (39/86) 66.7 (343/514) Considerable cases (considerable/ 2.9 (6/202) 2.2 (5/226) 0(0/39) 3.2 (11/343) teeth with calculus) of periodontal disease in modern populations, and Roberts and Cox (2003, 136). This is with the this is also the case with past populations. However, exception of Kingsholm, Gloucester which has a the aetiology is multifactorial with genetic predis- total prevalence of 0.6% (ibid.). position, environment, diet and hygiene all being Dental anomalies factors influencing the development of the disease. Dental anomalies may be inherited but can also be Most jaws in the present sample were very culturally induced. At leastseven skeletons revealed fragmented and this meant that the number that agenesis (congenital absence)ofthe third molar.Due could be examined for periodontal disease was to fragmentation and poor preservation, it can only be limited. The total prevalence was high (Table 3.30), statedthat at leastone tooth was unerupted in the although advanced recession (considerable cases) adult dentitions of skeletons 1057 (male),1131 was of relatively low prevalence. There was no (female), 1206(female) and 1381 (not sexed). Deter- markeddifference between males and females. mination of whether these teeth had not developed or Periapical cavities had developed but were unerupted would require Periapical cavities have often, in anthropology, been radiology. Three third molarswere absent in skeleton erroneously referred as abscess cavities,since it is 1165 (female) and the four molars were evidently unlikely that abscesses form these cavities in absent in skeletons 1337 (female) and 1360 (not sexed). the majorityofcases (Dias andTayles1997). They The total number of teeth,therefore, that had not are identified as openings or holes in the periapical eruptedwas 15. Agenesis of third molars may be bone of the mandible or maxilla at the apex of the inherited (eg Lasker1951; Bermu´ dez de Castro 1989). tooth root.They arise as aresult of inflammation of the dental pulp that can occur as aresult of trauma, The subadultsample caries or attrition. Depending on severity, these cavitiesmay contain granulation tissue (a ‘granulo- Data are presented for dental caries and enamel ma’),afluid filled sac (a ‘periapicalcyst’) or apus hypoplasia only (Table 3.32). No carious cavities filled sac (an ‘abscess’). Granulomata and periapical were identified among the deciduousteeth and only cysts are usually asymptomatic. Abscesses may result one example was observed among the permanent in apersistent fever, ageneral feeling of being unwell teeth.The veryhigh prevalence of hypoplasia has and, whenthey burst and discharge their contents, been biased by the dentition belongingtoskeleton halitosis. Acute abscesses maylead to osteomyelitis 1277 which had severe EH involving all 13 (eight (bone infection) which canbefatal if, for example,it anterior and four posterior) un-erupted permanent leads to septicaemia. teeth (Plate 3.1). The deciduous teeth from this The fragmentation of maxillae and mandibles individual also displayed EH (2/4anterior dentition; limited the number of observations. Only 0.7% 3/5 posterior dentition). Severe enamel hypoplastic of the alveoli observed had aperiapical cavity defects may be observed among individuals with (Table 3.31) and involved just three skeletons (1103, congenital syphilis (Hillson et al. 1998; Aufderheide 1153 and 1328). This prevalence is the lowest of and Rodrı´ guez-Martı´ n1998), but the nature of the those presented for Romano-British populations by lesions,the incompleteness of other diagnostic

Table 3.30 Periodontal disease frequencies in the discreteburial assemblage.

Females %(n/N) Males %(n/N) Unsexed %(n/N) Total %(n/N)

All alveoli with teeth present post-mortem 56.2 (63/112) 55.8 (38/68) 13.2 (5/38) 48.6 (106/218) Considerable cases (considerable/alveoli 7.9 (5/63) 5.2 (2/38) 0(0/5) 6.6 (7/106) with periodontitis)

Table 3.31 Periapical cavity frequencies in the discreteburial assemblage.

Females %(n/N) Males %(n/N) Unsexed%(n/N) Total %(n/N)

All alveoli observed 0.5 (1/177) 1.4 (2/139) 0(0/107) 0.7 (3/423)

45 Life andDeathinaRoman City

Table 3.32 Dental health status in the subadult sample from the discrete burial assemblage.

Deciduous Anterior dentition Posterior dentition

Age range Dental caries Enamel Hypoplasia Dental caries Hypoplasia %(n/N) %(n/N) %(n/N) %(n/N)

Young Child 0(0/15) 13.3 (2/15) 0(0/16) 23.0 (3/13) Older Child 0(0/4) 0(0/4) 0(0/3) 0(0/3) Total 0(0/19) 10.5 (2/19) 0(0/19) 18.7 (3/16)

Permanent Dental caries Enamel Hypoplasia Dental caries Hypoplasia Young Child –91.6 (11/12) –66.6 (4/6) Older Child –100 (1/1) 0(0/5) 16.6 (1/6) Adolescent 0(0/10) 100 (10/10) 8.3 (1/12) 91.6 (11/12) Total 0(0/10) 95.6 (22/23) 5.8 (1/17) 66.6 (16/24)

features of the examined skeleton and the need for 2003). The majority (>90%) of the cases were futurework cannot, at present, pinpoint aspecific classifiedaslinear enamel hypoplasias, the remain- cause forthese defects. Skeleton 1277 was less than ing being pit defects of the enamel. Females had 25% complete and was represented by the skull, higher frequencies than males. There were also 34 vertebrae, and ahumerus and femur shaft. No teeth belongingtoskeletons of unknown age(ie pathological lesions were identified on these remains adult or subadult). The frequency of hypoplasia for apart from cribra orbitalia. the anterior sample was 45.4% (5/11) and forpost- erior teeth 4.3% (1/23). Dental caries The mass grave assemblage The total prevalence rate for dental caries was 7.9% Enamelhypoplasia (Table 3.34). This prevalence is higher than King- Discounting poorly preservedteeth and those that sholm (5.3%), Cirencester north (4.8%), Cirencester were too worn or had surfaces that were obscuredby south (5.1%), and is average for thisperiod (Roberts calculus, 1222 teeth couldbeexamined forenamel and Cox 2003, 131). Teeth from skeletons of hypoplasia. The total prevalence was 28.5% undeterminedage (ie adult or subadult) had no (0/ (Table 3.33). Like the discrete burials, this is high 15) carious teeth in the anterior dentition and 7.4% for aRomano-British population (Roberts and Cox (2/27) in the posterior teeth.

Plate 3.1Permanent teethfrom Skeleton 1277, atwo to ﬁve year old displaying considerable enameldefects.

Table 3.33 Prevalence of enamel hypoplasia in the mass grave assemblage.

Females %(n/N) Males %(n/N) Unsexed %(n/N) Total %(n/N)

All anterior 50.5 (47/93) 42.6 (96/225) 33.7 (61/181) 40.8 (204/499) All posterior 32.6 (51/156) 20.0 (67/334) 11.6 (27/233) 20.0 (145/723) Total 39.4 (98/249) 29.2 (163/559) 21.3 (88/414) 28.5 (349/1,222)

46 Chapter Three

Table 3.34 Dental caries prevalence in the mass grave assemblage.

Females %(n/N) Males %(n/N) Unsexed %(n/N) Total %(n/N)

All anterior 0(0/137) 0.3 (1/331) 3.9 (10/225) 1.5 (11/693) All posterior 7.7 (17/218) 11.7 (66/563) 14.7 (51/347) 11.8 (134/1128) Total 4.8 (17/355) 0.8 (7/894) 10.7 (61/572) 7.9 (145/1821)

Carious lesions were most common at the cemen- subadults) yielded aprevalence of 39.4% (15/38). to-enamel junction (CEJ) in the anterior teeth,whilst Only one toothhad aconsiderable amount (6.6% of cavitiesonthe coronal surfaces as well as gross the teeth with calculus). lesions were more abundant in the posterior teeth Periodontal disease (Table 3.35). The prevalence of periodontal disease was relatively Ante-mortem toothloss high (Table 3.38). Males had ahigher prevalence AMTL was calculated in the sameway as described than females. Five alveoli from the age unknown for the discrete skeletons (Table 3.36). Atotal group had periodontitis. Since these are likely to prevalence of 5.3% was observed which is low for belongtoadult individuals, this was added to the aRomano-British population. unsexedgroup and the total count. There were no considerable cases. Dental calculus Like the discrete skeletons,rates for calculus were Periapical cavities high for this period with 59.4% of the teeth identiﬁed In total 1.6% of the alveoli that could be observed as having deposits (Table 3.37). Only 3.2% of the had aperiapical cavity (Table 3.39). This ﬁgure is teeth in the assemblage had deposits that were similar to that in other Romanpopulations from recorded as considerable. Teeth that belonged to Gloucestershire (Kingsholmand Cirencester south), individuals of unknown age (ie theywereadults or but low compared to samples from the rest of

Table 3.35 Distribution of pattern of carious lesions in the mass grave assemblage.

Lesion distribution Occlusal n/N (%) Coronal n/N (%) CEJ n/N (%) Root n/N (%) Gross n/N (%)

Anterior teeth (11 teeth, 12 lesions) –3/12 (25.0) 6/12 (50.0) 0/12 (0) 3/11 (27.2) Posterior teeth (134 teeth, 156 lesions) 35/156 (22.4) 38/156 (24.3) 32/156 (20.5) 2/156 (1.2) 49/134 (36.5)

Table 3.36 AMTL frequencies in the mass grave assemblage.

Females %(n/N) Males %(n/N) Unsexed %(n/N) Total %(n/N)

All teeth 2.9 (11/372) 3.7 (37/994) 7.1 (44/618) 5.3 (107/1984)

Table 3.37 Dental calculus frequencies in the mass grave assemblage.

Females %(n/N) Males %(n/N) Unsexed %(n/N) Total %(n/N)

All teeth 66.5 (231/347) 53.4 (636/1191) 68.3 (372/545) 59.4 (1239/2083) Considerable cases (considerable/ 1.7 (4/231) 5.5 (35/636) 4.6 (17/372) 4.5 (56/1239) alveoli with calculus)

Table 3.38 Periodontal disease frequencies in the mass grave assemblage.

Females %(n/N) Males %(n/N) Unsexed %(n/N) Total %(n/N)

All alveoli with teeth present post-mortem 19.1 (9/47) 78.3 (116/148) 70.4 (81/115) 66.4 (206/310) Considerable cases (considerable/alveoli 22.2 (2/9) 13.7 (16/116) 0(0/81) 8.7 (18/206) with periodontitis)

47 Life andDeathinaRoman City

Table 3.39 Periapical cavity frequencies in the mass grave assemblage.

Females %(n/N) Males %(n/N) Unsexed %(n/N) Total %(n/N)

All alveoli observed 0(0/134) 1.6 (5/300) 3.8 (4/104) 1.6 (9/538)

RomanBritain (Robertsand Cox 2003,136). was buriedinthe north-west quadrant of the pit and Amongstthe alveoli from the unaged individuals, in the lower layer. Histological analysis would be two out of three (66.6%)had periapical cavities. required to explore this diagnosis further. Dental anomalies Twenty-two teeth had not erupted and might have Comparison with the discrete skeletons been congenitallyabsent.These are third molars from adultdentitions. Thepresence of abnormalities Examinationofthe total sample frequencies calcu- that involved the enamel surfaces of the anterior and lated for adult dentitions from the discrete burial posterior teeth belongingtoskeleton 1672 (Plate 3.2) assemblage compared to the massgrave indicates was an exceptional ﬁnding. The abnormalities were that the discrete burialshad higher prevalences for reminiscent of the notched incisors and mulberry all conditions, except for periodontitis and periapical molarsthat appear in congenital syphilis (Langsjoen lesions (Table 3.40). The females from the massgrave 1998, 405–6, ﬁgs 14.10 and 14.11; Hillson et al. 1998). had higher rates of enamel hypoplasia on their This skeleton, a15–18 year old,was represented by posterior teeth but less periodontitis than the females its skull, vertebrae and aleft tarsal, all of which from the discrete assemblage (Table 3.41). With displayed no pathological lesions. Theindividual regard to males, pooreroral health was indicated

Plate 3.2Dentition belonging to Skeleton 1672, a15to18year old with enamel defects. Of particular note are the molars and central incisors (detail).

48 Chapter Three

Table 3.40 Comparison betweenthe discrete burials and the mass grave adult dental assemblages.

Oral condition Total adult sample

Discrete burials %(n/N) Mass grave %(n/N)

Enamel hypoplasia Anterior 63.3 (69/109) 40.8 (204/499) Posterior 17.0 (29/170) 20.0 (145/723) Dental caries Anterior 3.0 (6/199) 1.5 (11/693) Posterior 16.1 (56/347) 11.8 (134/1128) Ante-mortem tooth loss 5.8 (42/719) 5.3 (107/1984) Dental calculus 66.7 (343/514)59.4 (1239/2083) Periodontitis 48.6 (106/218)66.4 (206/310) Periapical cavities 0.7 (3/423) 1.6 (9/538)

Table 3.41 Comparison betweenthe femalesand males from the discreteburial and the mass grave assemblage.

Females Males

Oral condition Discrete burials Mass grave Discrete burials Mass grave %(n/N) %(n/N) %(n/N) %(n/N)

Enamel hypoplasia Anterior54.3 (25/46) 50.5 (47/93) 79.4 (31/39) 42.6 (96/225) Posterior17.7 (14/79) 32.6 (51/156) 15.0 (9/60) 20.0 (67/334) Dental caries Anterior1.4 (1/170) 0(0/137) 4.1 (4/96) 0.3 (1/331) Posterior12.5 (17/135) 7.7 (17/218) 18.9 (30/158) 11.7 (66/563) Ante-mortem tooth loss 2.4 (6/242) 2.9 (11/372) 8.6 (30/346) 3.7 (37/994) Dental calculus 68.8 (139/202) 66.5 (231/347)73.0 (165/226) 33.4 (636/1191) Periodontitis 56.2 (63/112) 19.1 (9/47) 55.8 (38/68) 78.3 (116/148) Periapical cavities 0.5 (1/177) 0(0/134) 1.4 (2/139) 1.6 (5/300) among the discrete burials, with the exception of Palaeopathology periodontitis (Table 3.41). by Louise Loe Methods The subadultsample All skeletons were examined for pathology or bony In the massgrave assemblage caries affectedtwo abnormality based on the recommendations of Ro- deciduous teeth (5.4%)and five permanent teeth berts and Connell(2004).Whenpresent, lesions were (4.4%). The prevalence of caries affecting deciduous described and diagnosedbyemploying standard teeth was higher than that calculated for the discrete definitions (eg Aufderheide and Rodrı´ guez-Martı´ n skeletons (Tables 3.42 and 3.43). Enamelhypoplasia 1998; Ortner and Putschar 1985; Roberts and Connell was confined to permanent teeth only and was less 2004). Conditions were broadly classified thus: infec- prevalent than in the discrete assemblage. tion, metabolic disorders, trauma, congenital and

Table 3.42 Dental health status in the subadult sample from the mass grave.

Deciduous Anterior dentition Posterior dentition

Age Dental caries Enamel Hypoplasia Dental caries Enamel Hypoplasia %(n/N) %(n/N) %(n/N) %(n/N)

Infant 0(0/7) ––– Young Child 0(0/15) –0(0/13) – Older Child 0(0/18) –8.3 (2/24) 0(0/4) Total 0(0/40) –5.4 (2/37) 0(0/4)

Permanent Dental caries Enamel Hypoplasia Dental caries Hypoplasia Young Child –––– Older Child 0(0/30) 22.2 (2/9) 0(0/19) 0(0/10) Adolescent 0(0/60) –5.8 (5/93) – Total 0(0/90) 22.2 (2/9) 4.4 (5/112) 0(0/10)

49 Life andDeathinaRoman City

Table 3.43 Comparison betweendental conditions observed among subadults from the discrete burial assemblage and the mass grave.

Oral condition Deciduousteeth Permanent teeth

Discrete burials Mass grave Discrete burials Mass grave %(n/N) %(n/N) %(n/N) %(n/N)

Enamel hypoplasia Anterior10.5 (2/29) –95.6 (22/23) 22.2 (2/9) Posterior18.7 (3/16) 0(0/4) 66.6 (16/24) 0(0/10) Dental caries Anterior0(0/19) 0(0/40) 0(0/10) 0(0/90) Posterior0(0/19) 5.4 (2/37) 5.8 (1/17) 4.4 (5/112)

(While enamel hypoplasia and dental caries have been calculated based on the number of observed teeth. Calculus, periodontitis and periapical cavities have been calculated by counting the number of alveoli available for examination.) developmental conditions, joint disease, circulatory mal-adaptation to environmental conditions, or disorders, neoplastic disease, miscellaneous condi- more specifically, malnutrition, poor sanitation and tions and unclassified pathological changes. generalised health stress (Robertsand Manchester 1995). Non-specific inflammationinvolving the skull: Three Results skeletons,one 18 to 25 year old male (Skeleton1386), Discrete inhumations one adolescent (Skeleton1317) of undetermined sex and one 25 to 35 year old female (Skeleton 1181), had Infection inflammatory changes involving ectocranial and/or Infectionmanifestsonboneinthe form of inflam- endocranialsurfaces of the skull vault (Table 3.44). mationand may involve the entirebone(‘osteomye- The female had spiculatedand porous new boneon litis’), the cortical bone(‘osteitis’) or the fibrous both the endocranial and the ectocranial surfaces of sheaththat covers bone, the periosteum (‘periosti- the parsbasilaris and the ectocranial surface of the tis’). These changesmay be observed as aresult of left pars lateralisofthe occipital bone. The changes tuberculosis, leprosy and syphilis (among others)or, on the male and subadult involvedthe endocranial where the pattern of change is non-diagnostic and surfaces of their parietal bones and were in the form the pathogen is unknown, non-specific infection. of capillary-like new bone. Inflammation involving Infectionmay arise as aresult of pathogens spread- the cranial vault may be the result of anumber of ing from an adjacent lesion via the blood stream(for different conditions including scurvy, chronicme- example, as seenintrauma, chronic skin ulceration, ningitis, trauma, anaemia, neoplasia, venous drai- paranasal sinusitis,middleear cavityinfection, a nage disorders and tuberculosis (Lewis 2004; Ortner dental abscess and visceral rib surface inflamma- and Erikson1997). There were no changesonother tion), or as aresult of direct implantation intobone parts of these skeletons that suggest which of these (for example, as seen in puncture and penetrating diagnoses is more likely here. injuries). Periostitis is the most commonlyobserved Skeleton 1181 also had inflammation affecting the lesion in archaeological populations. This surface alveolar bonearound the third mandibularmolars. inflammationmay be identified on dry bone as fine Both molars had been lost post-mortem and the pitting, longitudinal striations, swelling and/or empty sockets did notdisplay any pathological plaque-like new bone formation on the original bone changes. Inflammatory changes were also observed surface. Surface inflammationmay also affect the on the mandibles of Skeleton 1232, an 18–25 year old craniumbut, owingtoadifference in anatomy, the female,and Skeleton1328, a25–35 year old male. term periostitis does not applyhere. These changes Skeleton 1232 had active and healed new boneonthe may occur as the result of infection, or may accom- internal and external surfaces of the right ascending pany otherconditions of ametabolic, neoplastic or ramus boneand the damaged socketofthe right traumaticnature (Resnick and Niwayama1995). The third molar.Nodental conditions were observedin level of non-specific infection in apopulation is association with these lesions, but the bonewas generally regarded as an indicator of adaptation or incomplete. On Skeleton 1328, active new bone was

Table 3.44 Non-specific inflammation involving the skull.

Element involved Total %Unsexedadults Male %Female %Subadults %

Ectocranium 2/48 4.2 0/8 0/24 1/11 (9.1) 1/5 Endocranium 1/48 2.1 0/8 0/24 1/11 (9.1) 0/5 Mandible 5/77 6.5 R: 0/6 L: 0/8 R: 1/18 (5.6) L: 1/17(5.9) R: 2/9 (22.2) L: 1/10 (10) R: 0/4 L: 0/5

50 Chapter Three presentonthe alveolar bone(buccal aspect) below observations that couldnot be made owing to the sockets for the left first molar and the rightfirst preservation. This therefore suggests that something and second molars. Therewas aperiapical cavity otherthan just mild traumawas causing periosteal associated with the right first molar and all teeth reactionsinthe present sample. Multiple element were associatedwithcarious cavities. An erupting involvement is indicativeofsystemicdisease, tooth, dental disease and scurvy are among the although the patterns of involvement in the present conditions that maycause inflammatory changes skeletons are not exclusive to any oneparticular involving the mandible. disease process. Sinusitis: Skeleton 1181 also presented changes The affectedinnominate bones (Plate 3.3) were that were consistent with maxillary sinusitis.Thisis from one individual,Skeleton 1372, an adult male. in addition to Skeleton 1381, a25to35year old of This individual also had periosteal new bone undeterminedsex and Skeleton 1328, a25to35year deposits on all of his lower major limbbones and old male. Maxillary sinusitis is diagnosed based on most of his metacarpalsand metatarsals. On all the presence of new bone in the nasal sinuses. Upper bones the changes were typical of inflammation of respiratory tract infections, poor living conditions, the periosteum (ie striae, pitting and lamellar bone environmental pollution, congenital abnormalities, formation), with the exception of the innominate dental disease and specific infectiousdiseasessuch bones where the changes,inthe regionofthe as tuberculosis and leprosy are among the aetiolo- attachment site forthe ischiofemoral ligament, gical factors associated with this condition (Lewis included capillary-likenew bone formation. Because 2002, 21). In additiontonew boneonafragment of of the location of these changes, it is just possible that nasal sinus,the same changeswere also presentona they arethe result of ligament trauma at this site. poorlypreserved fragment of skull bonebelonging However, owingtothe fact that the changes to Skeleton 1328. The fragment could not be pre- involvedbothsides, non-specificsystemic disease cisely identified. It was eitheranadditional piece of is the preferred diagnosis. sinus bone, or it may have been apiece of sphenoid Comparison with other Romano-British popula- bone. tions indicatesthat London Road has ahigh Periostitis: Periostitis was observed on 28 out of 63 prevalence of periostitisfor thisperiod. For example, skeletons (44.4%), 25 adults (49%) and three sub- aprevalence of 24.2% (TPR) has beenreported for adults (33.3%). It was encountered on more females tibiae and 17.2% (TPR) forfibulae from Kingsholm (seven out of elevenfemales, or 63.6%)than males and 25% (TPR) has been reportedfor tibiae and (12 out of 23 males, or 50%),although six adults were 12.2% (TPR) for fibulae from Kempston (Robertsand of unknown sex. The most frequently involved Cox 2003). Wells (1982a,182) reports that between 10 elements were the lower long limb bones and the and 12 percent of alladult tibiae in the Cirencester bones of the feet (Table 3.45). south assemblage had periostitis and that there was The tibiawas the most frequently affectedelement little difference between males and females. These (TPR 28.9%). This finding is typical for most figuresare compared to aTPR of 28.9% for tibiae and archaeological skeletal assemblages and is explained 16% for fibulae from the present site. TheTPR for by the fact that the tibia is more easily affected by 124–130 Londonroad was not available but aCPR of mild trauma compared to otherbones in the skeleton 7.1% has been reported(Clough 2003). Overall, out (Robertsand Manchester 1995, 130). However, in the of 5,716 skeletons from 52 different Romano-British presentsample, few tibiae were affected unilaterally. sites,381 (CPR 6.7%) had non-specific infection Further, the plotted distribution of element involve- (Robertsand Cox 2003, 124). These figures are far ment indicates that at least16ofthe affected lower than the CPR of 44.4% for London Road. individuals had changes that involved more than Osteomyelitis: Therewere no definite examplesof one element(Table 3.46), thisbeing aminimum osteomyelitis (non-specific bone infection).However, number becausethe plot does not takeinto account it is possible that some of the more thorough-going

Table 3.45 Periostitis, absolute prevalence.

Element involved Total % Unsexed adults Males Females Subadults

Right Left Right %Left %Right %Left %Right Left

Ribs 5/316 1.6 0/10 0/8 1/62 (1.6) 0/87 1/60 (1.7) 1/47 (2.1) 1/16 1/26 Innominate 2/61 3.3 0/4 0/4 1/13 (7.7) 1/15 (6.7) 0/9 0/9 0/4 0/3 Femur 9/86 10.5 1/9 1/9 1/17 (5.9) 1/18 (5.6) 1/10 (10) 1/11 (9.1) 2/6 1/6 Tibia 24/83 28.9 4/13 2/9 1/18 (33.3) 6/16 (37.5) 3/10 (30) 2/11 (18.2) 0/3 0/3 Fibula 12/75 16.0 3/11 1/7 4/17 (23.5) 4/14 (28.6) 0/10 0/10 0/3 0/3 Metacarpals 2/172 1.2 0/13 0/5 1/34 (2.9) 1/52 (1.9) 0/27 0/27 0/6 0/8 Metatarsals 8/162 4.9 0/22 1/15 3/45 (6.7) 4/39 (10.3) 0/16 0/25 00 Calcaneus3/41 7.3 0/6 0/5 0/11 0/10 1/4 (25.0) 2/5 (40.0) 00

51 Life andDeathinaRoman City

Table 3.46 Distribution of elements with non-specific inflammation by skeleton. mber nu us ial erus ble phals. la it ur ius ls a a om. phals. ton Cs Ts Cs Ts rbit cranial rb lna umer ln um adius ad emur ibia ibula nnom. em ibi ibu nn RM LM Hand Rf Lf Rt Lt Rf Lf RM LM Foot Tarsa Lh Rr Lr Ru Lu Age Sex Skele En.- Ec.-cran Mandi RO LO Ribs CV TV LV Ri Li Rh

5–12 u/k 1086 X 46+ M1089 X 35–45 F1103 X 25–35 M1109 XX 25–35 F1131 XX X 18–25 u/k 1138 X 18–25 F1165 X 25–35 F1181 XXX X 46+ M1216 XXX 18–25 F1232 X 25–35 M1262 XXXXX Adult u/k 1279 X 111 M1286 XX X 108 u/k 1317 XXX 25–35 M1328 X 25–35 F1337 XX Adult M1372 XX XXXX 25–35 F1390 XX Adult M1393 XX 12–18 u/k 1396 X Adult u/k 1401 XXXXXX Adult u/k 1405 XX Adult M1453 XXX 46+ M1459 X 25–35 M1496 XX Adult M1499 X Adult u/k 1510 X Adult u/k 1709 XX

examples of periostitis were caused by this. In example, the scapula and the sphenoid bone) that are particular, this concerns Skeleton1286, a25–35 year affectedinthis disease (Ortner and Erikson 1997). old male (Plate 3.4). This skeleton displayedcon- However, the inflammatory changesthat were siderable, long-standing, new boneformation, sug- identified on the skulls,asdescribed above, are seen gestiveofmarrow involvement. However, there was in this condition. On their own, these changescannot no evidence foradraining sinus(acloaca),orsequ- be considered diagnosticofthe disease. estra (areas of dead bone) and there was no involu- Cribra orbitalia andporotic hyperostosis: Cribra crum (proliferative new bone) aroundthe affected orbitalia and porotic hyperostosis refer to small bones. Radiographytodemonstrate the extentofthe porosities or large interconnectedtrabeculae on the changeswould be required to explore thisfurther. roof of the orbits (cribra orbitalia) and the frontal, Metabolic disorders parietal and occipital bones of the cranial vault These are triggered by either an excess or a (porotic hyperostosis) (Stuart-Macadam 1991). These deficiencyinthe body’sdietary requirements and changesare believed to represent the skeletal hormones andresult in specific changestothe manifestation of iron deficiencyanaemiawhich skeleton, such as in the form of increased or may arise as aresult of dietary deficiency, malab- decreased bone turn over. Included in this category sorption (due to gastro-intestinalinfection or para- are scurvy (vitamin Cdeficiency) and cribra orbita- sites), blood loss and chronicdisease (Robertsand lia. No definite cases of scurvy were identified, but Cox, 2003, 234). Iron deficiencyanaemiamay also be the lesions may be fairly subtle in archaeological inherited, but this form in rare in Britishmaterial bone and,therefore, difficult to detect(Robertsand (Boylston et al. 1998). In the present sample, cribra Manchester 1995). Therewas no evidence for orbitalia was idenfitied on one or both of the orbital inflammationonthose parts of the skeleton (for bones of 11/51 adults (21.6%)and 2/9 subadults

52 Chapter Three

0 100 mm

1:2

Plate 3.3Non-speciﬁcinﬂammation(highlighted)involving the right innominate bone of Skeleton 1372, an adult male.

(22.2%)(Table 3.47). There were no examplesof porotichyperostosis. The youngest individual with cribra orbitalia was approximately two to five yearsold and the oldest, over 45 yearsold. The higher prevalence among subadults is to be expected because cribra orbitalia does not develop in adultbone (Stuart Macadam 1985). The changes,therefore,all represent achild- 0 100 mm hood condition. 1:2 In five cases bothorbits were involved and in four cases only the right orbit was involved. Therewere Plate 3.4Thorough-going periostealreaction involving no cases where only the left orbit was involved. right and left tibiae of Skeleton 1286, a25–35 year old male. However, two skeletons only had their rightorbits availablefor examination and oneonlyhad its left are reported to have had the condition (CPR 7.1%) orbit available. An additional skeleton only had one (Clough 2003), which is drastically lower than the orbit present, but this was unsided. CPR for the present sample (20.6%). These differ- The prevalence rates for cribra orbitalia were ences may be explained by the fact that there were similar to thosereportedfor Poundbury(Stuart- more older adults among the populations from Macadam 1991) which were 24.4% (TPR) for males, Kempston and 124–130 London Road. Lesions heal 27.7% (TPR) for females and 36.4% (TPR) for with age and are thus less common among the older subadults (ibid., 102). However, rates were higher adult age categories. than those reportedfor Kempston, Bedfordshire Trauma (Boylston and Roberts 1996) where the TPR for Traumareferstoany injury or wound to the body males was 18.3%,for females 10.5% and for that mayaffect the bone and/or soft tissues (Roberts subadults 26.7%. Males, females and subadults from 1991, 226). Dislocations, ligament trauma in the form Cirencester (Wells 1982a, 186) also had lower rates. of new boneformationand fractures are some of the The trueprevalence rates observed among these types of change in this category. The most common groupswere 19.9%,13.3% and 35.1% respectively. of these are fractures and these representthe main Four out of 56 skeletons from 124–130 LondonRoad type of trauma identified in the present sample.

53 Life andDeathinaRoman City

Table 3.47 Cribra orbitalia, absolute prevalence.

Orbit Male %Female %Unsexed %Unsexedand %Subadult % adult unaged

Right 3/9 33.3 4/8 50 3/3 100 0/1 01/2 50 Left 2/9 22.2 0/8 03/4 100 0/0 01/4 25 Total 5/18 27.8 4/16 25 6/7 100 0/1 03*/6 50

*includes one orbit that could not be sided.

Afracture is definedasacomplete or partial break Table 3.48 Summary of possible fractures. in the continuity of bone(ibid.). Fractures may result from underlyingpathology, repeatedstress or acute Skeleton Sex Age Element(s) injury (Robertsand Manchester 1995). Fractures Number occurring around the time of death,when the organicmatrix of the bone is still present, are termed 1118 Male Adult Third and fourth left peri-mortem, whilethoseoccurring afterdeath, when metacarpals the organic matrix has decomposed, are termedpost- 1281 Unsexed Adult One unsided hand mortem. Ante-mortemfractures occur before death phalanx and may be identified based on evidence for healing. 1286 Male 25–35 years Right tibia The identification of these different types of fracture 1405 Unsexed Adult Right tibia in archaeological humanbone provides enormous scope forfurthering knowledge of the lives of past populations including, their social interactions, acti- long time before death. There was also no evidence vities, socio-economic status and treatment of the for associated thorough-goingbone infection, dead. For example, certain types of ante- or peri- although two tibiae (Skeletons 1286 and 1405) and mortemfracture are indicative of inter-personal two fibulae (Skeletons 1089 and 1453) had associated violence, while others couldbefrom accidents (such periostitis. Only one skeletonwith apossible as afall from aheight)(Crawford Adams 1983; fractured hand phalanx (Skeleton 1281) had osteoar- Galloway 1999; Walker 1997). The alignment of an thritis in an associatedjoint. There were probably ante-mortem breakand evidence for secondary path- more cases like this, however, because associated ology (among other changes)may indicate quality joint surfaces were frequently missing. Fracturesthat of diet and treatment (Grauer and Roberts 1996) become infected may result in septicaemia and while, in certain burial contexts (for example,amass ultimately in death. They may also disrupt the grave), some types of peri-mortem breakmay be mechanics of ajoint and result in arthritis, leading associated with post-mortem dismemberment (Villa to deformity and debilitation. There is no evidence in and Mahieu 1991; White 2003). the present sample that the secondary changes had a Ante-mortem fractures: Healed ante-mortem frac- significant impact on the population. tures were observed on the bones of nineskeletons The most frequently affected elements were the (14.1% of all individuals); sevenout of 24 males fibula(2.9%) followed by foot phalanges (2.6%)and (29.2%)including three adults, oneadolescent,one old the tibia (2.6%)(Table 3.49). The prevalence of upper adult and three middleadults.Two skeletons were limb bone (humerus, radius, ulna and clavicle) unsexedindividuals (both adults). Fractures were not fractures was 0.4% (1/246), compared to 2.3% identified on the skeletons of any females. This is (5/220) for the lower limb bones (femur, tibia and despite the fact that the survival of the different female fibula). Fibulae and tibiae were also the most skeletal elements was proportional to that of male frequently fractured elements among the skeletons skeletal elements. Fractures were less frequent among from Cirencester south (Wells 1982a, 168). Here, 4.7% adult females from Cirencester (south)(CPR 6.6%) (13/274) of all fibulae and 2.4% (8/334) of all tibiae than they were among adultmales (CPR 26.7%)(Wells had sustained ante-mortem fractures.This trend is 1982a, 167). At 124–130 London Road 30% of all males also seen among other Romano-British assemblages had sustained afracture compared to 14.3% of all (Robertsand Cox 2003, 157). females(Clough 2003). Foot phalanges are susceptible to fracturing being Atotal of 14 elements displayed macroscopic ‘ ...ideally positionedtocollide with avarietyof evidence for afracture, identified as abreakinthe objects’(Galloway 1999, 222). It is unlikely that the continuity of the alignment of the bone. However, of fractures involving feet would have greatly incapa- these, five (see Table 3.48) require radiography to citated the individuals that they involved(ibid.). confirm their diagnosis. For the purposes of the Both of the tibia fractures (skeleton 1405 and presentanalysis, all of these elements have been skeleton 1286) require confirmation radiologically counted as definite fractures. (see Plate 3.5). Both presented swelling that was All fractures were well healed and showed good located on the middle portion of the shafts. A alignment suggesting that they had been sustaineda subperiostealhaematoma or boneinfection are

54 Chapter Three

Table 3.49 Distribution of fractures, absolute prevalence.

Skeletal element Total % Males Females Unsexedadults n/N Right %Left %Right %Left %Right %Left % n/N n/N n/N n/N n/N n/N

Clavicle 1/61 1.6 1/17 5.9 0/17 0/8 0/8 1/6 16.7 0/5 Ribs 2/274 0.7 1/62 1.6 1/87 1.1 0/60 0/47 0/10 0/8 Humerus 0/75 0/20 0/20 0/8 0/9 0/10 0/8 Radius 0/56 0/14 0/18 0/8 0/9 0/4 0/3 Ulna 0/54 0/13 0/18 0/8 0/7 0/3 0/5 Mcs 4/158 2.5 0/34 2/52 3.8 0/27 0/27 0/13 2/5 40 Hand phalanges 1/254 0.4 0/113 0/79 1/62 1.6 Femur 0/74 0/17 0/18 0/10 0/11 0/9 0/9 Tibia 2/77 2.6 1/18 5.6 0/16 0/10 0/11 1/13 7.7 0/9 Fibula 2/69 2.9 0/17 1/14 7.1 0/10 0/10 0/11 1/7 14.3 MTs 0/162 0/45 0/39 0/16 0/25 0/22 0/15 Foot phalanges 1/38 2.6 0/20 0/9 1/9 11.1

MCs =metacarpals; MTs metacarpals otherconditions that may result in bony swelling, skeletons,itisimpossible to say what may have but the preferred diagnosis is old trauma. The shaft causedthem becauselower limb boneelements are is the most common partofthe tibia to be fractured missingand thus the pattern of elementinvolvement and the middle or distal portions are usually (for example,whether theywere isolated fractures or involved(ibid., 193). whether otherassociated bones were also affected) is Fracturesthat involve the fibula usually show little not known. displacement and therefore the good alignment Rib fractures were generally infrequent and in- shown by the examplesinthe present sample is volvedtwo individuals out of 63 skeletons (CPR not surprising. Further, where the tibia was not 3.2%),ortwo ribs out of atotal of 274 (TPR 0.7%). This involved, it is unlikely that the fibula fractures is lower than the percentages calculatedfor skeletons causedthe individuals to experience significant from 124–130 London Road (CPR7.1%) (Clough mobility problems. This is because the un-fractured 2003); Kempston, Bedfordshire (TPR 10%) (Boylston tibia would have acted as anatural splint, providing and Roberts 1996); Kingsholm,Gloucester (CPR support and stability to the broken fibula. In modern 12.5%); Cirencester (north)(CPR 4.4%) and Cirence- clinicalsettings,fractures that involve the fibula ster (south)(CPR 7.7%) (Robertsand Cox2003). The often go undetected by the patient (Crawford Adams rate was, however, higher than that reported for 1983). It is possible that this was also the case for Poundbury (CPR 2.0%) (Farwell and Molleson 1993), individuals from London Road. Fibula fractures may Eastern Cemetery, London(CPR1.3%) (Conheeney result from direct force trauma such as ablow, if 2000), Cassington (CPR 1.4%) and Ancaster (CPR they occur withoutthe involvement of other ele- 2.1%) (Roberts and Cox 2003). The rib fractures in the ments (ibid.). Rotation of the ankle round an axis presentsample involved the lateral portion of the provided by the stationary tibia is another mechan- shaft of oneunidentified right rib (Skeleton1109) and ism, such as may occur accidentally whilst walking the neck, or vertebral end, of one unidentified left rib on rough ground, or in amodern setting, skiing (Skeleton1089). Fractures that involve the shaft (Boylston and Roberts 1996). For the present (rather than the neck) of the rib are likely to be the result of interpersonal violence (Wells 1982a). The absence of radial and ulnar (forearm) fractures is unusual. Clough (2003) reports aCPR of 1.8% (1/56) for the assemblage from the 124–130 London Road cemetery. For Kempston, Bedfordshire, the TPR was 2.8% and 2.6% for radial and ulnar fractures respectively (Boylston 1996). Similarly, Wells (1982a) reports that 2.9% of all ulnae and 1.7% of all radii were fractured among the elements examined from Ciren- cester (south). Thecrude prevalence rate forulnar and radial fractures observed in Romano-British populations is between approximately one and three percent Plate 3.5Possible old healed fracture involving the right (Robertsand Cox 2003, 154–55).Generally,forearm tibia shaft belonging to Skeleton 1405, an adult of fractures were infrequent during the Romano-British undetermined sex. period compared to other periods (Wells 1982a).

55 Life andDeathinaRoman City

Defence injuries and accidental falls onto an out- prevalence rate of 1.6% for the population overall. stretched hand are the most commoncausesof This is higher than the total overallCPR (0.9%) forearm fractures (CrawfordAdams 1983). calculated for the Romano-British period by Roberts Two skeletons had more than one fractured and Cox (2003). element. Skeleton1089, amale aged over 45 years, Peri-mortem trauma: The in situ locationofthe skull had sustained fractures to his left rib, left third belonging to Skeleton 1232, an 18–25 year old female, metacarpal and right fibula shaft. Fractures invol- suggested that she may have been decapitated prior ving the left third and fourthmetacarpals were to her burial.However, there was no evidence on her recorded for Skeleton 1118, an adult male. It is not skeletal remains that thishad been the case. Decap- possible to say whether these skeletons sustained itation may be identified on dry bone as cut or chop these breaks at the same time, or whether they were marks on the cervical vertebrae, clavicles, basioccipi- the victims of more than one traumatic event. tal of the skull andthe mandible. The remains of Cranial trauma: One skeleton, Skeleton 1112, had a Skeleton 1232 were in afragmentary condition, small ellipticaldepressed fracture that was superior although the cortical bone was well preserved and to the nuchal line of the left occipital bone(Plate 3.6). all of the aforementioned elements were available for The lesion was healed and was probably the result of examination. None of these displayedchanges that an impact from ablunt object. The individual was an were consistent with cut or chop marks. adolescent male. Very similar traumaticlesions were Peri-mortem cut marks: Formal analysis of the identified on four skulls from Cirencester (south)by London Road skeletons for peri-mortemmodification Wells(1982a). These were attributed to ‘ ....an was beyond the scope of the present report, but any accidental blow on the head ....or.....a fall against observations relating to thiswere notedinpassing. Of the cornerofatable... .[or]....a battle wound caused particular note were the remainsofSkeleton 1496, a25 by an accurately directed slingshot’ (ibid., 163–4). to 35 year old male, that had numerous parallel and Spondylolisis: Spondylolisis is astress fracture in sub-parallel linear incisions that cut into the cortical which the neuralarch separates from the body of the surfaces of several bones. These striations were fifth lumbar vertebra. It may be causedbyan located perpendicular to the long axisofthe right underlying congenital weakness in this partofthe and left femora, right ilium and left ulna. Most were spine (Aufderheide and Rodrı´ guez-Martı´ n1998). sharp,straight and had ‘v’ shaped profiles. Striations One skeleton, an adultmale (Skeleton 1118), showed that have this appearanceand occupy aperpendicular evidence for thiscondition which gives acrude orientation are indicative of deliberate anthropogenic

Plate 3.6Healed blunt trauma involving the occipital bone of the skull from Skeleton 1112, an adolescent male.

56 Chapter Three cut marks, such as may be observed as aresult of vertebraisseparated from the rest of the sacrum and dismemberment. However, in the presentexample, assumesthe form of the last lumbar vertebra the striations tended to be locatedaway from (Aufderheide and Rodrı´ guez-Martı´ n1998). The articulations, which is where the marksofdismember- assimilation may involve one sideonlyorboth ment are usually observed. Further, some of the marks sides, it may be complete or incomplete, unlateral or had distinctlyerodedmargins and were quite bilateral and symmetrical or asymmetrical. The shallow. These factors would combine to suggest that condition would not have causedany symptoms they were the result of excavation damagerather than and is more common among females (Aufderheide any deliberate modification that had occurred in and Rodrı´ guez-Martı´ n1998). Only the rightside of antiquity. They have been classified as pseudo-cut the sacrum was present and,therefore, it was not marks (Plate 3.7). possible to determine what the involvement was. Congenital and developmental conditions Joint disease This referstoabnormalities in growth or develop- Osteoarthritis: Osteoarthritis (OA) is the most comment. They may notbecome evidentuntil the period mon pathologicalcondition in both archaeological of growth or young adulthood, or theymay be and modern populations (Rogers and Waldron presentatthe fetal stage or at birth. The most 1995). It is therefore not surprising to findthat it common abnormalities are relatively minor and involvedthe most skeletons of all otherdisease cate- involve the spinal column (Barnes 1994). gories described in this report. Seventeen skeletons Spina bifida occulta and lumbarisation: Two skeletons (32.1% of all adults); 10 males (41.7%), four females exhibited bonychangesthat place theminthis (36.4%)and three of undetermined sex, presented category. Spina bifida occulta, in the form of changesthat are diagnostic of thiscondition(Table incomplete fusionofthe posterior neuralarches of 3.50). Thechanges included eburnation (polished the third, fourth and fifth sacral segments, was bone) or acombination of pitting, bony contour identified on Skeleton 1390, afemale of between 25 change and/or osteophytosis (new bone growth and 35 yearsofage.Skeleton 1393, an adult male, aroundthe marginofajoint or, less commonly, on a showedevidence forshifting of the lumbosacral joint surface) (Rogers and Waldron 1995). border in acaudal direction, aconditionalso known Two skeletons (1153 and 1206), who were young as lumbarisation. In this condition, the first sacral adults of between approximately 18 and 25 years of age, had OA affecting their costovertebral joints. OA is not usually observed among young adult skeletons because it is associated with old age. Thus, activity and trauma may have played asignificant role in the manifestation of the disease in these individuals. Comparedto124–130 LondonRoad, rates for OA were higher.For 124–130 LondonRoad the CPR was 17.4% for alladultmales and no females (out of a sampleof21) were affected(Clough 2003). London Road rateswere,however, similar to those calculated forthe Cirencester populationwhich had a CPR for males of 51.5% and for females,32.9% (Wells 1982a). These comparisons should be viewed with caution becausethey may be biased by the different methods employed to diagnoseOAbe- tweenobservers (Rogers and Waldron 1995). The most frequently affectedjoints were the cervical spine (TPR 17.1%), followedbythe hip (TPR 15%) (Table 3.51). It was notpossible to identify the number of vertebrae involved foreach spine owingto fragmentation. Skeleton 1281 had OA that involved the metacarpophalangeal joint (the knuckle) of the third finger.The proximal and middle phalanges belonging to thisfinger bone had ankylosed(fused). The changesmay be secondary to trauma. It is not unusual to find that the knee was rarely affected. This is typical of archaeological populations where both the hip and cervical spine are frequently involved(Rogers et al. 1981). Rates for the knee are, however, perhaps low for this period, whereas rates for the hip would seem to be high for this period. For Plate 3.7Pseudo-cutmarksobserved on the left femur example, Roberts (1989) reports aTPR of 4.2% for from Skeleton 1496, a25–35 year old male. both the hip and the knee for skeletons from

57 Life andDeathinaRoman City

Table 3.50 Age and sex distribution of skeletons with OA.

Males Females UnsexedTotal

Adult (any adult age category) 3126 Young adult 11 2 Middle adult 31 4 Mature adult 11 Old adult 314 Total 10/24 (41.7%) 4/11 (36.4%) 3/18 (16.7%) 17/53 (32.1%)

Kingsholm, and the calculatedTPR forPoundbury is responsible for this disease (Waldron and Cox 1989). 1.1% for the hip and 3.1% forthe knee (Farwell and The distribution of the disease in the presentsample Molleson1993). would notseem to be directly associatedwith a Three individuals(1089, 1216and 1262) had specific activity or occupation. generalised OA which means that the disease had Spondylosis deformans andSchmorl’s nodes: Spondy- affectedmultiple joints in their skeletons. Skeleton losis deformans and Schmorl’s nodes are two condi- 1089, amale aged over 45 years, showed involve- tions that involve the spine and are extremely common ment of the right and left hips and the cervical in bothmodern and archaeological populations. vertebrae. Theright and left hips were also involved Schmorl’s nodes are caused by intervertebral disc on Skeleton 1216, amale aged over 45 years, who herniation into the vertebral body. They appear on dry also had costovertebral involvement. Finally,skele- bone as depressions, either on the superiororinferior ton 1262, a25–35 year old male, showed involve- surface of the body.Although associated with degene- ment in the rightand left hips, the rightelbow, left rative disease,Schmorl’s nodes have been linked to knee and cervical and lumbar spines. activityand trauma, especially in adolescence,or Great attentionhas been focusedinthe archae- metabolic disorders (Jurmain 1999). Four skeletons, ologicalliterature on the association between OA and three males and one female, showedevidence forthis activityand occupation (see Jurmain1999). However, condition. They ranged in age from 18–25 years(one it is veryunlikelythat occupationand activitywill male), to 25–35 years(one female and one male). One have played auniquerole in the manifestation of OA male couldnot be aged more specifically than adult. in this population (possibly with the exception of The changesinvolvedthe thoracic spine of three Skeletons 1153 and 1206 described above). Many skeletons and the lumbar spine of one skeleton. factors, including age, sex, ancestry andgenetic Spondylosis deformans is identified on dry bone predisposition,aswell as activity or occupation, play as increased porosityonthe surfaces of the vertebral apart in the manifestation and course of the disease. bodies. The conditionismainly causedbydegenera- Except in rare instances, when apattern of OA occurs tion of the intervertebral disc and is associated with that is uniquetoanactivity or occupation, it is usually increasing age. There were four skeletons with this impossible to determine which of these factors was condition, including one female adult, a25–35 year

Table 3.51 Frequency of joints affected with OA.

Skeletal Total %Males Females Unsexed adults element Right %Left %Right %Left %Right %Left %

Costovert. 11/274 4.0 3/62 4.8 3/87 3.4 2/60 3.3 2/47* 4.3 0/10 0/8 Cervical 6/35 17.1 4/18 22.2 1/10 1.0 1/7 14.3 spine Thoracic 1/30 3.3 1/17 5.9 0/8 0/5 spine Lumbar 1/28 3.6 1/16 6.3 0/9 0/3 spine Shoulder 0/37 0/8 0/12 0/7 0/6 0/3 0/1 Elbow 1/56 1.8 1/14 7.1 0/15 0/8 0/7 0/7 0/5 Hand 1/61 1.6 0/15 0/17 0/9 0/9 1/6 16.7 0/5 Hip 9/60 15.0 4/15 26.7 4/16 25.0 0/9 0/9 1/11** 9.1 Knee 1/58 1.5 1/16 6.25 0/15 0/9 0/10 0/4 0/4

Key: costovert.=costovertebral joints *excludes one unsided costovertebral joint (this has been included in the total); **hip bone is unsided; Hand =any carpal or metacarpal joints; or more joints; All spines =number of skeletons with one or more vertebrae present

58 Chapter Three old individual of undetermined sex and two males of Skeleton 1510,anunsexed adult(Plate3.9). The 18 to 25 yearsand over 45 yearsold.Inall, the lesionshad penetrated thecorticalboneand were changesinvolved the cervical spine. Spondylosis is large(approximatelysix millimetresby13milli- uncommon on young individuals and is probably metres). This skeleton waslessthan25% complete the result of trauma. andwas representedbythe righttibia andfibula and Possible Legg-Calve´ -Perthes disease: The osteoarthri- an incomplete setoffootbones.Nochanges were tis that was observedonthe left hip bone(Plate 3.8) observed on theother bonesthatwerepresent.Inthe belonging to Skeleton 1286,amale aged between 25 absenceofany inflammation, thechanges areunlikely and 35 years, may have been secondary to Legg- to be theresultofinfection involvingthisjoint. Calve´ -Perthesdisease (Perthes). Perthes affects chil- Diagnosisisdifficultinthe absenceofother hand and dren between three and 12 yearsofage,ismore foot bones, butanerosive jointdisease is apossibility. common among males and is associatedwith growth This includes gout,which commonly involves thefeet retardation. Repeated trauma, infection and endo- andpresentspunched-out oval or roundlesions. crine disease have all beenimplicated in the aetiology Peri-articular scalloped erosions were present on of Perthes, but it is generallyregarded as an idopathic the heads of the rightand left first metatarsals condition (ofunknown cause). It arises when im- belonging to Skeleton 1453, an adultmale (Plate paired bloodsupply during growth of the femoral 3.10). Thesame changesalso involved the phalanx of head causes necrosis and deformity leadingto the left greattoe. Periostitiswas present on the shaft premature joint disease in this partofthe hip. In the of the first left metatarsal, in additiontothe left tibia presentexample, the changesincluded amushroom and fibula(the latter may have beenfractured). The shaped,flattened femoral headand osteophytosis skeleton was represented by lower limb bones and (new bone growth aroundthe marginofthe joint). forearm bones and had incompletesets of hand and However, notenough of the bone had survivedto determine whether it displayed coxavara(an abnormal femoral neckangle), or whether the femoral neck was shortened and widened, and there were no metaphyseal cysts. These are other changesthat are seen in this condition. Right andleft acetabulae were shallowbut both were too incomplete to determine whether they were lengthened (also diagnostic of Perthes). The left femur was missing from the top of the diaphyseal shaft upwards. Differential diagnoses includebad osteoarthritis and hip dislocation. Otherjoint disease: Therewerefive skeletons that presentedchanges involvingtheir joints that do notfit into theabove categories.Two smooth-walledante- mortem oval,lytic lesions wereobservedinthe region 0 50 mm of thefibula articularsurface of therighttalus bone of 1:1

Plate 3.9Oval, lytic lesions involving the lateral side of the right talus bone of Skeleton 1510, an unsexed adult.

Plate 3.8Possible Legg-Calve´ -Perthes disease involving the left femur belonging to Skeleton 1286, a25–35 year Plate 3.10 Scallopederosions involving the heads of the old male. right and left great toes from Skeleton 1453,anadult male.

59 Life andDeathinaRoman City foot bones. Erosive changes affecting foot bones may peoplebetween the ages of 15 and 35 yearsand be caused by relatively minor conditions such as involvesmore males than females (Rogers and hallux valgus (bunions) or may be the result of Waldron 1995). Inheritance has astrong influence something more systemic such as gout. In the over the manifestation of thisdisease which is of presentcase, the erosionswere not typical of those unknown cause.Itisimpossible to say whether the seen in gout (Rogers and Waldron 1995). However, changesthat were observedinSkeleton1367 were a bones of the feet are typically involved in this result of this disease because the lower spine had not disease.Gout, therefore,cannotberuledout. survived. The sacrum and innominate bones were Asmall circular ante-mortem lytic lesion was also also very incomplete. Other conditions that may observedonthe joint surface of each of two left wrist cause ankylosis includediffuse idiopathic skeletal bones (scaphoid and lunate) belongingtoSkeleton hyperostosis (DISH), another condition that involves 1328, a25–35 year old male. Therewas no indication fusionofthe spine in acharacteristic way, and that these changeswerepresent on other joints from trauma. DISH tends to affect males over the age of 45 this skeleton which was mostly complete but had a yearsand is associated with obesity and diabetes. very limited number of hand and foot bones Circulatory disorders surviving. This was arelatively mild lesionthat Reduction or lossofthe blood supply to bone may was probably the result of abenign cyst. result in necrosis (bone death) in the affected area Left temporomandibular joint dysfunction of the and subsequent joint dysfunction. Examples of this jaw bone was identified on Skeleton 1112, an in the LondonRoad inhumations were presentinthe adolescent male (Plate 3.11). Here, extension of the form of osteochondritis dissecansand possible Legg- joint surface area, apossiblepseudo-joint, and alytic Calve´ -Perthesdisease. defect may have been the result of temporomandib- Osteochondritis dissecans: Osteochondritis dissecans ular joint dislocation with secondary cystic forma- involved1.6 %(CPR)ofthe populationand was tion. Alternatively the lytic defect may have been observedonthe rightcapitulum of the humerus (the causedbyneoplastic disease. elbow) of apossible female aged between 25 and35 Early ankylosis (fusion) involving the left hip and years(Skeleton 1337). In thiscondition, necrosis sacrum belonging to Skeleton 1367, amale adult, occurs in asmall focal area on the convexsurface of was indicated by the presence of aspur of new bone diarthrodial joints and resultsinpartial or complete on the left iliac joint surface of the pelvis. This detachment of asegment of the subchondral bone skeleton also had osteoarthritis of the left hip. Anky- and articular cartilage. The aetiology of this condi- losis involving the sacroiliac joint is observedin tion is notfullyunderstood but it may be caused by ankylosing spondylitis, ajoint disease in which the low grade chronic trauma or micro-trauma. In the spine fuses in acharacteristicway so that is resembles Romanperiod, this condition is seeninincreasing bamboo. The disease begins in the sacroiliac joints frequency in the knee joint, possibly the result of and then progresses up the spine. It usually affects occupational trauma (Roberts and Cox 2003, 151). In modern populations, the elbow is the third most common site to be involvedand the condition is more common among males andindividuals aged between 10 and 25 yearsold. The predominant involvement of the rightside is possibly as aresult of right-sided dominance in strong activities involving the arms (Aufderheide and Rodriguez 1998). In the presentexample,the lesion would nothave greatly impactedonthe daily life of the individual, who may have experienced aclickinginthe joint and some pain with movement.For the Romano-British period, the crude prevalence rate, based on seven populations, rangesfrom between 0.2% for the Eastern Cemetery, Londonand 5.2% for Baldock 2 (Robertsand Cox 2003, 152), with an overall prevalence of 0.4%. Theprevalence for London Road, therefore, is among the highest. Neoplastic disease The only form of neoplastic disease that was identified in the sample were two ivory osteomas (or button osteomas), one on each of the cranial vaults of Skeleton 1114, a25–35 year old of unknown sex, and Skeleton1537, an adult female. Ivory osteomas are benign tumours and are common Plate 3.11 Lytic lesion involving the head of the left among archaeological and modern populations. mandibular condyle. Skeleton 1112, an adolescentmale. They may be described as smooth dense expansions

60 Chapter Three of cortical bonethat are usually small and tend to presentlesions,their relatively benign appearance occupythe outertable of the skull or the sinuses. andsmoothmargins,suggest that non-pathological Their cause is unknown but some have suggested skeletal variationisamore likely explanation. that they are aresponse to trauma (Aufderheide and Rodrı´ guez-Martı´ n1998). The mass grave Miscellaneous conditions This analysis includes aminimum number of 76 Hyperostosis frontalis interna: Hyperostosis frontalis adults and 15 subadults. Sexed individuals include a interna (HFI)was present on the endocrania of four minimumnumber of 15 females (includingone skeletons (TPR 8.3%, 4/48;CPR 6.3%, 4/63) includ- adolescent) and 38 males (including three adoles- ing ayoung adultfemale, amiddleadultfemale and cents). Crudeprevalence rates have been calculated ayoung adult and amature adultofunknown sex. based on the minimum number of individuals by HFI is identified on dry bone as thickening and employing the methods described above. noduleformationonthe endocranialsurface of the frontalbone. This condition has associationswith Infection virilism and obesity and it is commonamong post- Non-specific inflammation involving the skull: This menopausal women. Its cause is unknown but it has was observedonthe outersurface of skull vaults been known to occur in pregnancy and accompany only and affected two skeletons (CPR 2.2%; TPR 1.6%) changesassociated with acromegaly, therebyimpli- including two males aged 35–45years(Skeleton 1518) cating some sort of pituitary gland disorder in and 18–25 years(Skeleton1541). On both, the changes its aetiology (Aufderheide and Rodrı´ guez-Martı´ n included porousnew boneformation that was located 1998). on the frontaland parietal bones. The changes involving Skeleton 15 were associated with trauma Unclassified pathological changes involving the frontal bone of the cranial vault. Changes that do not fit into any of the above Periostitis: At least16(CPR 17.6%)individuals had categories in this section are discussedhere. They periostitis involving one or more elements. Periostitis concernskeletons that would require radiography to was more frequent among males (18.4%)than it was explorediagnoses. Thecranial fragments belonging among females (13.3%)and it involved few indivi- to Skeleton1238, amale aged over 45 years, were duals who were below the age of 18 years very light and had awoolly appearance, both (Table 3.52). The most frequently affectedelement suggestive of poor boneturnover. One condition was the tibia (TPR 11.3%), followedbythe fibula that resultsinpoor bone turnover is Paget’sdisease, (TPR 6.5%) (Table 3.53). In one caseperiostitiswas adisease that results in the progressive enlargement associated with healed trauma involving the fifth and thickening of bones as aresult of disruption to metatarsal (1516). cellular activity. Comparedtothe discrete skeletons, asimilar Skeleton 1337,a25to35yearold female,had range of skeletal elements were affected with shallow, oval lyticlesions on theanteriorsurfacesof periositis (Fig. 3.6). However, unlike the discrete rightand left femoralnecks adjacent to themarginof skeletons,there was involvement of the forearms but thefemoral heads(Plate3.12).These lesionswerevery not the innominate bones, the metacarpals, or the subtle,werenot associated with anyinflammation calcaneus—patterns that are probably not signifi- andhad regularmargins.Lytic lesionsinthe region of cant owingtothe small numbers involved. Based on thefemoral neckscan occurinpathologicalconditions the crude prevalence, the rate for periostitis is high such as tuberculosis.However,the subtlety of the for aRomano-British population (Roberts and Cox

Table 3.52 Age and sex distribution of skeletons with periostitis.

Male Female Unsexed Total (%) (%) (%) (%)

Infant 0/0 0/0 0/1 0/1 Young child 0/0 0/0 0/3 0/3 Older child 0/0 0/0 1/4 1/4 Adolescent 1/3 0/1 0/3 1/7 Young adult 1/19 0/7 1/10 2/36 Middle adult 3/7 1/4 0/2 4/13 Mature adult 1/2 1/2 0/3 2/7 Older adult 0/3 0/1 1/1 1/5 Adult 1/4 0/0 4/11 5/15 Total 7/38 2/15 7/38 16/91 Plate 3.12 Oval lytic lesions involving the necks of right (18.4) (13.3) (18.4) (17.6) and left femora. Skeleton 1337 a25–35 year old female.

61 Life andDeathinaRoman City

Table 3.53 Periostitis, absolute prevalence.

Element Total Unsexed adult Male Female Subadults involved (%) Right (%) Left (%) Right (%) Left (%) Right (%) Left (%) Right (%) Left (%)

Ribs 1/793 (0.1) 0/77 1*/77 (0.6) 0/195 0/172 0/116 0/101 0/28 0/27 Ulna 1/123 (0.8) 0/20 0/18 1/25 (4.0) 0/29 0/10 0/12 0/5 0/4 Radius 1/111 (0.9) 0/18 0/14 1/24 (4.2) 0/27 0/9 0/11 0/4 0/4 Femur 3/199 (1.5) 1/30 (3.3) 0/40 1/38 (2.6) 0/34 0/22 0/21 0/7 1/7 (14.3) Tibia 16/142 (11.3) 0/26 3/31 (9.7) 4/26 (15.4) 3/20 (15.0) 2/14 (14.3) 2/15 (13.3) 1/2 (50.0) 1/8 (12.5) Fibula 8/124 (6.5) 1/18 (5.6) 4/28 (14.3) 1/21 (4.8) 0/20 1/12 (8.3) 1/16 (6.3) 0/4 0/5 Metatarsals 3/319 (0.9) 0/34 3/68 (4.4) 0/69 0/57 0/36 0/42 0/8 0/5

*rib fragment is unsided; APR calculated out of atotal of 154

2003, 124). However, true prevalence rates forthe with infection, or they may have beenthe result of tibia are lower than Kempston and Kingsholm but trauma. Therewas no evidence for osteomyelitis. similar to Cirencester (see above). It is impossible to exploreevidencefor multiple Metabolic disorders element involvement in this group owingtothe Cribra orbitalia: Cribra orbitalia affectedaminimum fact that the sample comprises articulated and dis- number of six skeletons, four males (CPR 10.5%) and articulated elements. However, consideration of only two females (CPR 13.3%). The femaleswere aged the articulated skeletons with periostitisand/ or between 18 and 25 yearsand 25 and 35 years and the inﬂammationonthe cranial vault (twelveintotal), males included oneadult, one adolescent, one aged suggests that, like the cemetery population, multiple between 18 and 25 yearsand one aged between 25 element involvement (and hence systemicdisease) and 35 years. No sub-adults were affected by the was also commontothis group (but less frequent) condition. The condition was also not observedon (Table 3.54). any older adults.The lack of sub-adult involvement Other infection: There was asmall fragment of is unusual giventhat cribraorbitalia developsin innominate bone(possibly ilium) that displayed childhood and has usually healed by the time signiﬁcant remodelled lamellar bone. Smallprolif- individuals reach adulthood. erativeerosionswere also present. Thefragment was More left orbits were involved than rightorbits so small that it was notpossibletotell exactly what (Table 3.55). Therewas little difference between the the full extent of the changeswere (ie whether they number of affectedmale orbits (10.1%)compared to involvedajoint and/or were in an area of muscle the number of affected female orbits (10.7%). True attachment). Thechanges may have been associated prevalence rates for males, females and subadults %

Figure 3.6 Periostitis distribution: discreteburials and mass grave (TPR).

62 Chapter Three

Table 3.54 Distribution of elements with non-specific inflammation by skeleton (articulated skeletons only). mber nu us ial erus phals. la it ur ius ls a a phals. ton Ts Cs Ts Cs rbit um cranial rb lna umer ln um adius ad ibula ibia emur lium ibu ibi em li Rf Lf RM LM Foot Tarsa Rt Lt En.- Ec.-cran RO LO Ribs CV TV LV Ri Li Rh RM LM Hand Rf Lf Age Sex Skele Lh Rr Lr Ru Lu

Adult u/k 1516 X Adult u/k 1517 XX 35–45 M1518 XX X 25–35 F1539 XX 25–35 M1540 XX X 18–25 M1541 XXX Adult u/k 1543 XX 25–35 M1544 XX 46+ u/k 1546 X 12–18 M1547 XX 35–45 F1552 XXXX 25–35 M1553 XX Adult u/k 15162 X 18–25 u/k 15652 XX Adult M15762 X

Table 3.55 Cribra orbitalia, absolute prevalence. limbs,one from the torso and five from the lower limbs (Table 3.56). All fractures had probablybeen Orbit Male %Female %Unsexed %Sub- % sustained along time before death owing to the fact adult adult that they displayed long termbone turnover. All were well alignedexcept for the fractured humerus belong- Right 4/36 11.1 1/13 7.7 0/7 00/10 0 ing to Skeleton1524 which displayed markedangular Left 3/33 9.1 2/15 13.3 0/8 00/7 0 deformity and shortening (Plate 3.13). This skeleton Total 7/69 10.1 3/28 10.7 0/15 00/17 0 may also have had adislocated right shoulder, but fragmentation has prevented aconclusive diagnosis. No fractured bones displayedevidence for OA in were low compared to Poundbury, Kempston and associated joints, although many joint surfaces were Cirencester (see above). missing. Three fractured elements, afifthmetatarsal Porotic hyperostosis: Increasedporositywas fre- (Skeleton1516) and aleft tibia and left fibula (Skeleton quent on outer skull vaults throughout the assem- 1543), were associatedwith periostitis. blage, achange that may have been caused by The distribution of fractures was similar to that porotichyperostosis (this also applies to the discrete described forthe cemetery population: no females skeletons). However, none of these displayedthe were affected,rib fractures were infrequent, there thinnedouter tables and expanded diploic space that were no radial or ulnar fractures and the overall is consistent with the disease. This porositywas fracture rate forupper limb bones was lower than that probablynot of any greatconsequencetothe overall for lowerlimb bones (0.3% compared to 0.7%). health of the individuals and was probablyrelatedto With the exception of the clavicle, frequencies were conditions such as minor scalp infection. consistently lower forall elements among the mass Trauma grave population compared to the cemetery popula- Ante-mortem fractures: Healed fractures were identi- tion (Figure 3.7). fied on at leastseven skeletons (CPR7.8%) including The most frequently fractured elements were the one 18 to 25 year old male (1524),two 25 to 35 year old clavicle(TPR1.8%) and the fibula(TPR 0.9%) males (1544 and 1648) and four unsexed adults (427, followed by the tibia (TPR 0.8%) (Table 3.57). 609, 1516 and 1543). Thus, the CPR for males was 7.8% Compared to other populations,the crude preva- (3/38) which is far lower than rates observedat124– lence rate for claviclefractures is higher: 3.2% 130 London Road (CPR 30%) (Clough 2003) and compared to 1.8% for Ancaster, 4.2% for Kingsholm Cirencester (south)(CPR (26.7%) (Wells 1982a, 167). (Robertsand Cox 2003), 0.8% for Cirencester (south) Therewas no macroscopic evidence that any of the (Wells 1982a), and 2.3% for Poundbury(Farwell and femaleshad sustainedfractures. Molleson1993). Fracturesinvolvedatotal of nine elements Generally speaking, clavicles may fracture in three (excludes sternal end of clavicle), three from the upper places; at the sternal end of the shaft, along the mid

63 Life andDeathinaRoman City

Table 3.56 Distribution of fractures, absolute prevalence.

Skeletal element Total % Males Females Unsexed adults

Right %Left %Right %Left %Right %Left %

Clavicle 2/109 1.8 1/24 4.2 1/27 3.7 0/10 0/12 0/19 0/17 Ribs 1/738 0.1 0/195 0/172 0/116 0/101 1*/154 Humerus 1/135 0.7 0/29 1/32 3.1 0/11 0/11 0/24 0/28 Radius 0/103 0/24 0/27 0/9 0/11 0/18 0/14 Ulna 0/114 0/25 0/29 0/10 0/12 0/20 0/18 Hand phalanges 1/585 0/313 0/148 1/124 Femur 2/185 1.1 0/38 2/34 5.9 0/22 0/21 0/30 0/40 Tibia 1/132 0.8 0/26 0/20 0/14 0/15 0/26 1/31 3.2 Fibula 1/115 0.9 0/21 0/20 0/12 0/16 0/18 1/28 3.6 MTs 1/321 0.3 0/69 0/57 0/57 0/36 0/34 1/68 1.5 Talus 1/76 1.3 0/18 0/14 0/9 0/10 0/9 1/16 6.3

*rib is unsided. Prevalencecalculated out of all ribs.

may also have been fractured (and has been counted here),but requires radiological investigation.Frac- tures that involve the acromial end may result from being displaced following aconsiderable pull on associated muscles, most often as aresult of ablow to the shoulder (Galloway 1999, 115). Fractures that involve the sternal end tendtoarise as aresult of direct violence (ibid.). Plate 3.13 Fractured left humerus belonging to Skeleton Cranial trauma: Therewas littleevidence for cranial 1524, an 18–25 year old male. trauma. Skeleton1518 had healed trauma on the superioraspect of the left orbital margin. The lesion, shaft, or at the acromial end of the shaft (the end aremodelledlinear defect, was consistent with sharp nearest the scapula). Most commonare those that force trauma, such as that which mightbedelivered involve the mid shaft and these are often associated by someone yielding ablade.The skeleton was a35– with ablow as aresult of afall or being struck with 45 year old male. Healed sharp force trauma was an object (Galloway 1999, 115). In the present also identified by the presence of ahealed linear sample, one fracture involved the acromial end of lesion on aparietal bonefragment of asmall find the left claviclethat belonged to Skeleton 1544, a25– (number 2032), and identified as a25–35 year old 35 year old male. Thesternal end of the right clavicle male. Taken together, the lesions indicate an overall

Bone Figure 3.7 Ante-mortem fracture distribution: cemetery and mass grave (TPR).

64 Chapter Three

Table 3.57 OA frequency of sitesaffected.

Skeletal element Total % Males Females Unsexed adults

Right %Left %Right %Left %Right %Left %

Costovert. 0/738 0/195 0/172 0/116 0/101 0/154 Cervical spine 2/80 2.5 0/47 0/19 2/14 14.3 Thoracic spine 1/72 1.4 0/37 0/19 1/16 6.3 Lumbar spine 2/68 2.9 0/34 0/20 2/14 14.3 Shoulder 1/90 1.1 0/19 0/22 0/12 0/10 0/12 1/15 7.7 Elbow 0/99 0/24 0/26 0/9 0/10 0/13 0/17 Hand 1/104 1.0 0/25 0/25 0/10 0/13 1/14 7.1 0/17 Hip 0/158 0/37 0/35 0/26 0/24 0/15 0/21 Knee 1/114 0.9 0/21 0/23 0/12 0/14 1/20 0/24 crude prevalence rate of 2.2% (2/91) and an absolute otherRomano-British populations. For example, prevalence rate of 1.6% (2/122). Wells(1982a) calculated aCPR of 44.8% for adults Dislocations: A25to35year old male (1553) had a from Cirencester (south), although inter-observer dislocated rightfinger that involved the joint variability in diagnosing OA should be borne in between the first and second phalanges (metacarpal mind here. unidentified). The dislocation had resulted in abnor- Unlike the skeletons from the cemetery, the mal articulation between these two bones and skeletons from the mass grave displayed adifferent secondary changesincluded osteoarthritis and peri- patternofjoint involvement (Table 3.57). Thespine ostitis.One otherskeleton (1554) had apossible was the most frequently affected,but instead of the dislocated rightshoulder andhas been mentioned cervical spine, the lumbar spine showedthe highest above. prevalence (2.9% compared to 2.5%). There was no Peri-mortem fractures: No peri-mortem fractures evidence forhip OA and, unlike the cemetery were identified. Breakage patterns were not formally population, there was evidence for shoulder OA. studied (this was beyond the scope of the present Schmorl’s nodes and degenerative disc disease: Five report) but in general, there were numerous trans- skeletons (CPR5.5%) had changesintheir spines verse breaks among the long bones, afeature that is that were consistent with Schmorl’s nodes (Table typical of post-mortem fracturing (Galloway 1999) 3.58). They involved one unsexed adult(Skeleton and is usually the result of soil compaction (Trin- 1517), two adult males (Skeletons 1542 and 1551) and khaus1985). one male (Skeleton 1518) and one female (Skeleton 1552) who were both35–45 yearsofage. Three Congenital and developmental anomalies skeletons 1517, 1518 and 1542) had thoracicand Skeleton1547, an adolescent male, had one bifid lumbar spine involvement andtwo (1551 and 1552) unidentified left rib. had thoracic spine involvement only. Spines were Joint disease too incompletetoallow calculation of the number of Osteoarthritis:Owingtofragmentation, the prevalence affectedvertebrae from each spine. for spinal OA was calculated based on the number of Degenerative disc disease affected just one skele- cervical,thoracicand lumbar spines that were avail- ton (CPR 1.2%),anindividual of over 45 years and of able for analysis among the articulated skeletons that unknown sex. The changesinvolved the cervical were excavated from the pit only. Only one small find spine (TPR 1.3%) and the thoracic spine (TPR 1.4%). had spinal OA (701). This was an adultmale who had Circulatory disorders: The distal joint surface of the changesinvolving the cervical spine (this has not been left tibia belonging to Skeleton 1540, a25–35 year old included in Table 3.57). All other spinal OA was adult, had asingle delineatedpit that resembled the observedonskeletons that had been articulated (ie were notsmall finds). Theprevalence for all extra- spinal OA has been calculatedbased on the total Table 3.58 frequencyofspines affected with Schmorl’s number of all joints that were present among the nodes. articulated skeletons and the small finds. Total Males Females Unsexed At least three individuals had OA involving one or (%) (%) (%) (%) more of their joints (articulated Skeletons 1516, 1546 and 1555). These were two adults and one old adult Cervical 0/80 0/47 0/19 0/14 ( 5 45 years), all of undetermined sex (three out of 76 spine adults,or3.9%).The old adult had OA affecting Thoracic 5/72 (6.9) 3/37 (8.1) 1/19 (5.3) 1/16 (6.3) multiplejoints including the righthand, the left spine shoulder, the cervical spine and thoracic spine. This Lumbar 3/68 (4.4) 2/34 (5.9) 0/20 1/14 (7.1) is perhaps notsurprising given the age of the spine individual. The CPR (3.9%)islow compared to

65 Life andDeathinaRoman City changesseeninosteochondritis dissecans. However, Afurther adult mandible (small ﬁnd 716) had a in thisdisease the changes affect the concave surface large lytic lesion that involvedthe alveolar bone of of diarthrodial joints, not the convexsurface (Auf- the left mandible (at the site of the second premolar derheide and Rodrı´ guez-Martı´ n1998). Thetalus and the left molars) and is consistent with an bone belonging to thisskeleton did notshow any ameloblastoma (Plate 3.15) (Dias and Tayles 1997). changes. Ameloblastomas are rare,benign neoplasmsthat are Neoplastic disease: Oneskeleton (1542) had an ivory slow growing and cause severe abnormalities to the osteomaonthe cranial vault. This was an adultmale. face and jaw. The mandible was incompleteand One upper thoracicvertebrabelonging to this lacked the rightmandibular ramus and bothascend- individual had destructive lesions that involvedthe ing rami.All the incisors, canines and left premolars superiorsurface of the body (the laminae were had been lostpost-mortem.The amenoblastoma was missing) (Plate 3.14). The height of the vertebral 28 mm in diameter. Aperiapical cyst (approximately body was also uniformily reduced. There was no 7.83 mm wide) was located on the site of the left involvement of other vertebrae. Infection and neo- canine and has been counted among the dental plasmsare among the diseases that may cause these conditions recorded in the relevant sectionabove. changes. The latter is the favoured diagnosishere becauseofthe solitary involvement of the vertebra THE DISARTICULATED MATERIAL and the absence of inﬂammatory changes. by Nicholas Ma´ rquez-Grant Disarticulated human remains(bones and teeth) were recovered from 45 contexts throughout the site (except forthe massgrave)and are listed in Appendix 2. Allcontexts but four provided a minimumofone individual.Contexts (1182),(1454) and (1457) had aminimum of two individuals represented,while (1409) had aminimum of three adult individuals. Thus, adding all the MNI calculations for each context, the total MNI is 50. However, it is unlikely that these bones representdiscrete skeletons that are additional to the number of skeletons from the cemetery overall.Theyprobablybelongtothe discrete skeletons, or may form partofaskeleton that has remainsinvarious contexts due to post- mortemdisturbance. Plate 3.14 Destructive lesions involving athoracic All individualswere adult except in contexts vertebra from Skeleton 1542, an adult male. (1285) and(1418) where they were subadult

Plate 3.15 Ameloblastomainvolving the left side of the mandible from an adult male(small ﬁnd 716).

66 Chapter Three

( 4 18 years) individuals. Apossible subadult was have similar high values are Cornwall, Ireland and also represented by the remains in context (1182). the east coast of Italy. The bones from five contexts (1247, 1341, 1361, 1387, These analyses do not point with confidence to 1397) did not provide any ageestimates. In addition, the geographic locationofthe individuals from bone recovered from contexts (1282) and(1365) either the cemetery or the massgrave. Thereisno provided anarrower adult agerange,suggesting patterninthe distribution of isotope valuesof that the remainsrepresent young adultindividuals different skeletons in terms of age, sex, burialtype (18–25 years). Sexdetermination was possible for and associated artefacts. However, sample size was contexts 1333 (sf154), 1182 and 1457 which suggest small. male, male and apossible female respectively. Overall,the results would suggest that individuals The only pathological changesthat were observed from the mass grave and cemetery were of diverse on these remainswas periostitis. This involvedthe origin and possiblyinclude individuals from regions left tibia and left fibulabones that were recovered with climatesthat are warmer than the UK.The fact from the fill of grave 1402. The changes were consi- that values did not show any correlation with burial derableand are in the sameleague as those identi- locationwould indicate that individuals from the fied on the remainsofSkeleton1286 and illustrated mass grave and the cemetery were partofthe same in Plate 3.4 above. No changeswere identified population. Thus, the selection of individuals for macroscopically that would confirm that the lesions burial in the mass grave was random rather than one had beencaused by osteomyelitis, although thisis determined by geographic origin. likely. DISCUSSION OF THE UNBURNT HUMAN Isotopeanalysis REMAINS by Louise Loe, Nicholas Ma´ rquez-Grant and Jonny Geber Asinglemolar tooth was taken from21skeletons, including 10 discrete inhumations and 11 mass grave Asummary of the main findings described in this skeletons,for analysis of oxygen and strontium report is presented in Table 3.59. Overall, the mass isotopesatthe NERC Isotope Geoscience Labora- grave skeletons and the discrete inhumations did not tory. Adetailed report is provided in Appendix 3. differ significantlyinterms of their physical attri- The main findings are summarised here. butes, demography and disease. They shared a The analyses were undertaken to explore the similar range of non-metric traits,bothgroups geographic origin of the skeletons. In particular, comprised males and females, the young and the they were undertaken to identify any ‘non-local’ old, and, although frequencies differed,they shared individuals and, if present, establish their origin. the samerange of pathological conditions. These Originwas explored in relationtocultural (for osteological findings therefore suggest that there is example, burial locationand associated artefacts) no reason to believe that the discrete inhumations and biological (for example, age and sex) variables. and mass grave skeletons were from different The findings of the analyses indicated two distinct populations. groups, as reflected by their combined strontium and oxygen values(seeFigure A3.1, Appendix 3). The Some caveats first group comprised eight discrete skeletons and six mass grave skeletons. Among these individuals Of the discrete skeletons that could be assignedtoan 87Sr/86Sr values were wide ranging but most refer- age category, few were assigned to the older adult age red to geology that lies within a20kmradius of range.Inaddition, infants were under-represented. Gloucester, in particular the MalvernHills(Chenery, These findings were also observed among the mass Appendix 3). Oxygen valuesrelatedtodrinking grave assemblage. Wheninterpreting these trends water zones throughout the UK, therefore suggest- there areseveral caveats that should be considered. ing wider geographic origin. Overall, the results Perhaps the most important is the fact that estimating would suggest that the individualsprobablyorigi- biological ageisfraught with problems, notleast nated from the UK. becausepresent standards have been derivedusing The second group comprised two discrete skele- populations that are distanced in timeand space from tons and three massgrave skeletons.Strontium archaeological populations (Bocquet-Appel and Mas- valuestended to cluster, suggesting acommon set 1982). Studies using adult skeletons of documen- geographic originfor these individualsthat has a ted age have demonstrated that presentmethodstend different geology to Gloucester. This is confirmed by to over-age the young and under-age the old (Cox Oxygen values which exceeded the expectedrange 2000), and thismust also be borne in mindinrelation for the UK and were reflective of drinking water that to the present results. mightbefoundonthe extreme westernorsouthern In the present assemblage, most age estimations coast of Europe, North Africa, or extremely arid were based on observations relating to dental locations. However, veryfew values have been attrition. This method is widely recognisedfor its published for these locations to say with confidence tendency to under-age skeletons (Cox 2000). More whether or notthe LondonRoad values relate to relevantly, it has been notedfor its tendency to drinking water zones in these areas. Other areas that under-age Romanpopulations in particular (Boston

67 Life andDeathinaRoman City

Table 3.59 Summary of the main osteological findings.

Criteria Summary

Discrete inhumationsMass grave

Minimum number 64 (63 analysed) 91 (422 small finds and 201 skeletons) of individuals Preservation Most very good or fair and almost half less Most good or very good than 25% complete Sex 11 females, 24 males 14 females, 35 males Age Deaths highest among young adults and middle Deaths highest among young adults. Few adults. Few children below the age of five years. deaths among children under five years More older adult deaths than the mass grave. and among older adults. More females than males in all adult age Young adult and older adult age categories categories, except older adults (more males). have more males than females. Many more un-aged male adults than Middle and mature adult age categories female adults. have more females than males. Stature Average female height: 160 cm (5’3’’) One unsexed skeleton: Average male height: 169 cm (5’5’’) 160 cm (5’2’’) For164 cm (5’3’’) M Non-metric traits Most frequent cranial traits: Parietal foramen Most frequentcranial traits: Lambdoid and supraorbital notches. ossicles, parietal foramen and Access Most frequent post-cranial traits: tibial infraorbitalforamen. squatting facets. Most frequentpost-cranial traits: Hypotrochanteric fossa, tibial squatting facets and vastus notch. Dental health and disease More enamel hypoplasia among subadults More carious cavities among subadults than mass grave. than discrete inhumations. More carious cavities, ante-mortemtooth loss, More periapicalcavities among adults calculus and periodontitis among adults than than discrete inhumations. mass grave. Pathology Trauma Ante mortem fractures: Ante mortem fractures: 14.1% of population (CPR) 7.8% of population (CPR) 29.2% of males (CPR) 7.9 %ofmales (CPR) No female fractures No female fractures Tibia most common (TPR 3.9%). Clavicle most common (TPR 1.8%). More lower limb than upper limb. More lower limb than upper limb. Non- specific Periostitis: Periostitis: Infection 44.4% of population (CPR) 17.6% of population (CPR) 63.6% females (CPR) 13.3% females (CPR) 50% males (CPR) 18.4% males (CPR) Tibia most frequently involved element Tibia most frequently involved element (TPR 28.9%). Frequent multiple element (TPR 11.3%). Frequentmultiple element involvement. involvement. MetabolicCribra: Cribra: conditions21.6% of all adults (CPR) 7.9% of all adults (CPR) 22.2% of all subadults (CPR) No subadults affected males 27.8% (TPR) males 10.5% (CPR) females 25% (TPR) females 13.3% (CPR) subadults 50% (TPR) No porotic hyperostosis No porotic hyperostosis Joint disease OA: OA: 32.1% of adults (CPR) 3.9% of adults (CPR) 41.7% of males (CPR) All unsexed skeletons 36.4% of females (CPR) LV most common (TPR 2.9%) CV most common (TPR 17.1%) followed No hip OA by hip (TPR 15.0%) Shoulder OA (TPR 1.1%) pers com.). Furthermore, many skeletons, both reflectedinthe large number of adult skeletons that discrete and commingled, were incomplete. This could notbeassigned to an age category. If it had has meantthat age had to be estimated by employ- been possibletoestimate more preciseages forthese ing fewer indicators than the desired range.This is skeletons,and employ more accurate methods, the

68 Chapter Three mortality profiles forthe discrete inhumationsand Obstetric casualties and in- and out-migration are the massgrave may appear quite different.The fact factors that may explain the higher numbersof that some of the youngeradults had pathological deathsamong the 25–35 year oldsand the 35–45 year conditions (for example OA) that are more typical of olds (Grauer 1991; Chamberlain2000). In- and out- older agegroups, may suggest that the individuals migration are very plausible influences given the from LondonRoad were notdying at such young socio-economic and cultural activities taking place in ages as the mortality profiles would perhaps other- Gloucester at this time. wise suggest. The mortality profilefor the mass grave does not conform to the usual curve forRoman populations, but reflects the profileofaliving population. Unlike The under-representation of the youngest the mortalityprofilefor the discrete burialsthe The under-representation of the youngest indivi- majority of individuals, both males and females, duals (from two years and below) is not uniqueto died before theyhad reached 35 years of age. The the London Road assemblage and has been observed kolmogorov-Smirnov statistical test (at the signifi- among otherRomancemetery populations (Clough cance level of p 5 0.05 the maximum difference 2003; Mays2000). Several factors may have caused between the two samples (0.183) was less than D m,n this including excavationbias, cultural practices and (0.263))indicatedthat there was no significant preservation. For example, it was common practice difference between the mortality profiles. However, to bury infants in settlements rather than afford them this does not mean that the profiles are the same, burial in acemetery (Philpott 1991). It has been rather there is insufficient evidence to suggest that suggested that the practice of burying babies beneath they are different (Shennan 1997, 61). doorsteps was adopted during the Roman period Apeak in deaths in young adulthoodiscompa- becauseofthe belief that theywould bring good luck tible with the mortality profiles of other populations to the inhabitants (Scott 1991; Scheuer and Black dying as aresult of acatastrophic event (Margerison 2000, 15). Infanticide may have resulted in the and Knu¨ sel 2002). Catastrophic events that might disposal of infants in sewers and ditches, or their cause sudden death in apopulation include war, abandonment outside the settlement (Mays 1993; massacre and plague. War may be identified in Smith and Kahila 1992; see Chapter6). Drivenbya skeletal remains by ahigh mortality among young belief in ‘changelings’ mothers may have also males,high frequencies of peri-mortem trauma, temporarily abandoned their infants and thiswould embedded weaponsand partially articulated skele- have resulted in deaths and,byproxy, infanticide tons (Boylston et al. 2000; Hurlburt 2000; Ostendorf (Mays 2000, 185). Philpott (1991) considers that, Smith 1997). These trends were reflected among during the Romanperiod, infants below the age of skeletons from amass grave in Towton, Yorkshire about 18 months were given burialsthat were often (Fiorato et al. 2000). Sharp, blunt and projectile casual and separate fromcemeteries becausethey traumas,largely involving the skull, were frequent were not regarded as full members of society.There among this group, who were victims of the Battle of may perhaps have been separate child cemeteries, Towton(1461).Similar patterns have been observed not yet discovered (Dennehy 2001; Pearce 2001). for aNapoleonic period mass grave in Lithuania The full extent of the LondonRoad cemetery was (Signoli et al.2004) and an early/middle Bronze Age not identified during excavation and it is therefore multipleburial from Wassenaar, the Netherlands entirely possible that young infants were buriedin (Kooijmans 2005). No such patterns were observed areas that were not excavated.Further, infants may among the remainsfrom LondonRoad. not have survived the burial environment because Epidemics such as famine or plague result in acute their small, fragile bones are more susceptible to death,thus pathological changes associatedwith disintegration in the ground(Chamberlain 2000). these conditions do not show on skeletal remains. Although these factors have probablycontributed to Faminetends to affect the most vulnerable members the rate of recovery, the fact that the small, fragile of acommunity, these being the oldest and the bones of older childrenwererecoveredsuggests that youngest (Margerison 1997). This is not reflectedin these factors did not significantly affect this. the mortality profileofthe London Road assemblage, although problems with age estimationand the under-representation of infants meansthat this The mortality profiles suggestion cannot be fully explored with the present For un-biased, pre-industrial populations the mor- data. Nevertheless, the overall impression is of a tality profileshould approximate the U-shaped mixed population that comprised both sexes span- curve that is characteristicofundeveloped countries ning all ages. An epidemic, such as plague, that did where the highest deathsoccur in the youngest and not target either sex or any particular age group, oldest agecategories (Waldron 1994). The mortality therefore provides amore powerful explanation for profilefor the discrete burialsisbroadly compatible the deathsamong thisrandom cross section of the with this. As discussed above the youngest are population. Asimilar cross section of the East under-represented forvarious reasons (in particular Smithfield populationdiedasaresult of the Black cultural), andadults are probablyunder-aged Death during the 14th century,and supportsthis becauseofthe methodsthat have beenemployed. interpretation (Kausmally 2007).

69 Life andDeathinaRoman City

The dateofthe mass grave is contemporary with ing exercises were limited to large bones only (for the Antonine Plague, which may have been an example, skulls and major limb bones), whereas outbreak of measles or smallpox (see Chapter6for Duday (2006, 52) considers that small bones, namely discussion). Measles causesnounique or recognisa- the patellae and those of the hands and feet, provide ble changes in the skeleton, although smallpox may the most matches in assemblages of commingled be identified skeletally. Changes include osteomye- remains. The fragmentary nature of the assemblage litis and arthritis which tend to involve the limb has also severely hindered element matching which bones, in particular thosefrom the arm (Aufderheide would benefit from adedicated programme of and Rodrı´ guez-Martı´ n1998). However, such lesions reconstruction.There was limited opportunity to are notseeninadult skeletons and they tend to fully integrate biological and archaeological data disappear up to ayear following the virus (ibid.). with the three-dimensional representation of the Secondary changesincludebone ankylosis (or mass grave, Crossbones.Again, this more specialised fusion), degenerative joint disease and retarded analysis would, no doubt, yield positive results. growth.Inthe presentsample, there was no The high level of incompleteness observed among confirmed evidence for osteomyelitis and changes the mass grave skeletons is probablythe result of a that involvedjoints were not consistent with those combination of in situ decomposition and recording. described for smallpox (Resnick and Niwayama Post-excavation analysis has determinedthat, in 1989). Evidence for secondary changeswere present many cases, different parts of the sameskeletons in the form of growth arrest(manifested as enamel were assignedadifferent skeletonnumber during hypoplasia), osteoarthritis and bonyankylosis. excavation. This is aresult of the sheer complexity of However, these changesmay arise in association the deposit which could notbefully understood with many other conditions. Without any primary until it had beenexcavated and digitally analysed. evidence, the presence of smallpox cannot be proved This has therefore artificially increased the apparent or disproved from osteological evidence alone. high level of incompleteness. In situ decomposition relates to environmental factors combined with the physical and chemical Taphonomy properties of the skeleton. The sameskeleton may It was suggested during the excavation of the mass have had its elements displaced to different areas of grave that some bodies were in the early stages of the grave as aresult of piled cadaversdecomposing putrefaction priortobeing buried. Basic analysis of (Duday2006, 50). Other factors that have probably the remains to explore whether theyhad been resulted in the pattern of disarticulation of bones in deposited as fleshedcorpsesorskeletonised remains the mass grave relate to variation in the resilience of was undertaken. In particular, the remains were ligaments between some articulations in the skeleton analysed to explore why the disarticulation of compared to others, the effects of gravity and the skeletons had occurred. The results revealed that creation of empty spacesfor bones to fall into as soft the condition of all bones was notmarkedly different tissues decayed (Galloway 1999; Duday 2006). Rising between the different burial types(articulated and falling water tables and body fluids also skeletons,small finds and disarticulated bone) probablycontributed to the displacementofele- implying uniformity in their treatment. There was ments,particularlyoflighter spongy bones and no evidence that human remainshad been exposed skulls which float more easily than otherelements to scavengers or sunlightand there was also ahigh (Boazand Behrensmeyer1976; Duday 2006). recovery rate of elements. Environmental evidence These conclusionsare consistent with the fact indicatedthat the pit had containedadistinct lack of that displaced elements, recorded as small finds, snails,insect remains, or otherfactors that would were found in the part of the pit where there would point to amassofdecomposing bodies. However, have been the highest water levels (the western this does notnecessarily mean that skeletons rather side),being the deepestpart. In addition, spatial than corpses had been deposited in the pit, particu- analysis showedthat upper limbshad been dis- larly if the pit was filled in quickly (Nicholson pers persed more than lowerlimbswhich are denserand com.). have more resilient connecting soft tissues. Skulls Overall,these observations suggest that fleshed were also among the most frequent type of element corpseshad been buriedinthe pit, probably around amongst disassociated bones that comprised the the sametime, and that all were complete when small finds. deposited. Thus, the pit and remains therein probably An interesting observation was the fact that many represent aprimary, rather than secondary, burial. broken bones displayedfracture margins that were The results of the matching and conjoining indicative of having been broken when there was exercises unfortunatelydonot lend support to this still some organic matrix in them (ie they were ‘dry’ conclusion. These did notidentify any matches rather than ‘green’). This implies that some of the between elements or fragments and this therefore breakage did not occur immediately aroundthe time may suggest that incomplete skeletons or corpses of death,but neither during recent times. These are were buried in the pit. However, it is more likely that most likely to be natural fractures resultingfrom these results are areflection of the preliminary environmental processes (for example,soilcompac- nature of the present analysis.For example,match- tion) rather than the result of deliberate human

70 Chapter Three activity. Galloway (1999, 16) points out that bone inadequate in many urban Romano-British environ- may not necessarily always completely dryout in its ments,despite the ‘apparent ‘superficial’ sanitary burial environment, if it is lying below the water organisation forthis period’ (2003, 125). The evi- table or is in aconstantly wet environment. Further, dence from London Road is in keeping with this. in the context of amassgrave, body fluids and Two skeletons, one from adiscrete burial (Skeleton moisture can be trapped for many yearsand thereby 1277) and one from the mass grave (Skeleton1672) reduce the degreetowhich bones dryout (Galloway had severe enamel defects which are possibly 1999). Thus, in these cases, dry fracture properties of indicative of congenital syphilis.Syphilis is diag- bone may be retained and may result in some of the nosed on dry bones and teeth based on the presence changesdescribed here. of proliferative and destructive lesions. Involvement of the entireskeletonand ‘cariessicca’ (distinctive lesions on the outside surface of the skull) are Pathology and physical attributes among the common in the veneralform of the disease and discreteskeletons and the mass grave skeletons dental defects (‘mulberry molars’ and ‘Hutchinson’s Compared to the discrete inhumations, prevalence teeth’), perforations in the nasal area and palate,joint rates for pathological conditions among the skele- involvement and changesinvolving the tibia (‘sabre tons from the mass grave were lower. However, a shin’) are common in the congenital form of the similar range of conditions was experienced by both disease (Aufderheide and Rodrı´ guez-Martı´ n1998). groups. For example,periostitiswas more common Apart from the enamel defects,there were no skeletal among the discrete skeletons than the mass grave changesidentified in the assemblage that match skeletons,but several skeletons from both groups those observedineither congenital of venereal forms showedapatternofskeletal involvement that impli- of the disease.Ifthese cases are confirmed they will cated systemicdisease,rather than mild trauma, in join an increasing number of importantpalaeopatho- its aetiology. Thus, similar disease processes were logical examples which, if the diagnosesare correct, being experienced by both groups.Itispossible that suggest that congenital syphilis was presentinthe the mass graveskeletons had fewerpathological Old World before 1493 (see Erdal 2006, 30). lesions because they represented those individuals of Overall,the prevalence of ante-mortem fractures the population who had weaker constitutions and among the male skeletons from the discrete burials were more susceptible to acute disease.Interestingly, (CPR 29.2%) was similar to otherRomano-British enamel hypoplasia was more frequent among the assemblages including 124-130 London Road (30%) discrete inhumations than the mass grave skeletons. (Clough 2003) and Cirencester (26.7%) (Wells 1982a, Some have suggested that this evidence for growth 167). This was not reflected in the prevalence that arrestduring childhood shows in the skeletons of was calculated for the massgrave skeletons (CPR individuals who had robust constitutions,orwere 7.8%),although fragmentation has probably reduced ‘survivors’, not those who were more susceptible to the number of fractures that couldbeidentified here. ill health and acute death (Duray 1996; Lewis and Fracture patterns suggest limited interpersonal Roberts 1997). violence in the population (just three skull wounds) Alow prevalence of pathology among the mass and involved only males.Thismay suggest that grave skeletons is not an uncommonfinding for this males engaged in particular activities that predis- type of burial.Thishas also been observedamong posed themtotrauma, unlike the females. Patterns othermassgravesresulting from natural disasters of trauma (in terms of elements involved) were such as an epidemic. For example, alow prevalence similar to thosereported by Wells (1982a;b) for was also observed among the medieval skeletons Cirencester (south and north), namely the blunt force from the East Smithfield Black Death cemetery skull trauma, the under-representation of forearm (1348–50) (Kausmally2007). fractures and the frequency of lower limb fractures. ComparedtootherRomano-British populations, While the CPRfor trauma was low among the the London Road population is notablefor its high individuals from the mass grave,fragmentation has levels of non-specific infection, with acrude pre- probablyreducedthe number of fractures that could valence rate of 44.4% for the discrete skeletons and be identified here. 17.6% for the massgrave skeletons. These rates are Evidence for nutritional status indicated apopula- higher than that (CPR6.7%) reported by Roberts and tion that may have had high levelsofiron deficiency Cox (2003, 124) for 5,716 skeletons from 52 different anaemia compared to other Roman groups. The true Romano-British sites.Patterns of elementinvolve- prevalence forthe cemetery population was 27.8% ment suggest that systemicdisease, including infec- for males, 25% for females and 50% forsubadults. tion, was prevalent. That few cases seem to have These figures are higher than thosereportedfor been associated with infection resulting fromtrauma Kempston, Bedfordshire(18.3% for males,10.5% for supportsthissuggestion. High levelsofinfection are femalesand 26.7% for subadults) (Boylston and typical of close-contact societies whereby over- Roberts 1996) and Cirencester (19.9% for males, crowding and poorsanitation facilitate the transmis- 13.3% for females and 35.1% for subadults) (Wells sion of disease (Larsen 1997). Areview of the 1982a, 186); and similar to Poundbury (24.4% for evidence by Roberts and Cox (2003) concludesthat males,27.7% for females and 36.4% for subadults) sanitation (for example, sewerage management) was (Stuart-Macadam 1991). However, rates forthe mass

71 Life andDeathinaRoman City grave are lower (10.1% for males, 10.7% for females tions would benefit from radiography to explore and 0% for subadults). differential diagnoses. In particular, this concerns Iron deficiencyanaemia may be caused by a joint disease and infection to explore the presence or deficient diet. However, there was no convincing absence of specific infections (for example syphilis evidence in the assemblage formalnourishment.In and tuberculosis) and gout. The possible examplesof fact, the evidence (as it stands) indicated that the congenital syphilis would benefit from histological population enjoyedanadequate diet. The high levels analysis. The study of DNA from the massgrave has of iron deficiencyanaemiamay therefore be aresult the potential to make asignificant contribution to the of increased pathogen loads. For example, there question of why the individuals were afforded this were no confirmedcases of scurvy or rickets, as has type of burial,once problems such as contamination been observed among otherRoman populations are resolved (Thomas et al.2005;but see Drancourt et including, Kingsholm, Gloucester (Robertsand Cox al.2004). Palaeopathology is arapidlyexpanding 2003, 142–3) and Kempston, Bedfordshire(Boylston field in which knowledge of skeletal involvement in and Roberts 1996). Compared to otherpopulations, certain diseases is constantly being developed (for there were high ratesofcaries and this maysuggest example Brickley et al. 2007). Those pathological that the London Road skeletons consumed higher skeletons in the present assemblage, in particular levels of carbohydrates, or employed different diet- those which presently remain of uncertain diagnosis ary regimesand food preparation techniques com- (for example,with endocranial lesions), will surely pared to their contemporaries.Evidence for the benefitfrom future advances in understanding. qualityand quantityofdiet is also indicated,ifthe erosive joint changes that have been described turn out to have been caused by gout. Gout is adisease THE BURNT HUMAN BONE that is associatedwith excessive alcohol intake, by Nicholas Ma´ rquez-Grant obesity, high bloodpressureand kidney problems Introduction (Robertsand Cox 2003, 137). Interestingly, gout is common in the Cotswolds (both today and in the In accordance with recommended practice (McKinley past) and the oldest examplesdate from the Roman and Roberts 1993), cinerary urns (1061 and 1251) period (Robertsand Cox 2003, 137–8). were lifted largelyintact and the contents of Fragmentation provided little opportunity to ex- cremated bonewereexcavated undercontrolled plore the physical attributes of the population. laboratory conditions.These deposits were exca- However, limited data for stature indicate that the vated in 20 mm spits, each of which was planned at a population may nothave experienced significant scale of 1:10 and atotal soil sample taken. The growth disruption during childhoodbecause male samples were thenwet sieved and sorted into and female statures tended to fall above the average 4 10 mm, 4 5mmand 4 2mmsize categories.The that has been reportedfor males (1.69 m) and females sorted bone andthe residues were then assessed. (1.59 m) from this period (Roberts and Cox 2003). The fragmentation and/or truncation of the other Growth disruption can be caused by prolonged four cinerary urns (1068, 1223, 1200, 1191), meant periods of malnutrition, disease and/or heavy man- that it was not possible to excavate and processthe ual stress (Steckel1995). Therange of non-metric traits remainsinthe sameway. In some cases, the deposits in the population suggest that perhaps inheritance were excavated in spits. In others,the boneand was amore significant component in their manifesta- surrounding soil was bulk sampledasasingle unit. tion in thispopulation than mechanical stress. These samples were wet-sieved and residuessorted into boneand other material such as charcoal and artefacts. All bone was sorted into 4 10 mm, Conclusions and future potential 4 5mmand 4 2mmsize categories. The human skeletons from LondonRoad representa The colour of bonefragments, the weightofeach highly significant group that includes discrete deposit, fracture patterns, fragment sizes and the articulated skeletons and commingled skeletons skeletal and dental elements present were recorded to from amass grave that are probablyfrom the same explorethe cremation process and pyre technology. population. This has affordedauniqueopportunity Colourisareflection of the temperatures achieved to compare patterns of demography and disease during cremation (see below). All deposits were between attritional and catastrophic groups from a examined to determine the MNI and estimate ages single population. Themassgrave comprises in- and sexes,where possible. Bones were also examined dividualswho probably died over arelatively short for any evidence of bonyabnormality or pathological period of timeand therefore affords the opportunity lesions. to explore aspects of demography and disease among individuals who lived at the same time. Methods The assemblage has considerable potential for furtherresearch. Amore dedicated programme of Where possible, each fragment was identified and element matching and reconstruction to explorethe classifiedbyanatomical region (eg neurocranium, burial taphonomy of the mass grave has already splachnocranium,upperlimb and lowerlimb). The been discussed above. Several pathological condi- presence of any specific anatomical landmark (eg

72 Chapter Three orbit, lineaaspera )was noted. The representationof bag or awooden box). If they have beencollected in skeletal elements may indicate whether certain an organic container, this will not survive the burial parts of the skeleton were favoured over others for process, so it will appear that the crematedboneis burial (see below), although it must be taken into ‘loose’inthe soil. Collection of the entire cremated accountthat the prevalence of unidentifiable bone is skeletal remainsfor burial rarely occurs. This may largelydependent on the degreeoffragmentation, have beenbecause the relatives(or undertakers) whereby largerfragments are easier to identify than were not very efficient at collecting all the cremated smaller ones. It must also be taken into consideration bone from the pyre (the pyre would comprise aheap that some skeletal elements (for example cranial of charcoal, partially burnt wood, and possibly fragments) are more diagnostic and more easily iden- stones or burntclay from the lining of the fire pit). tified than others and therefore more often recorded. Alternatively it may have been unimportant to This may create bias in calculations of the relative collect all of the burnt bone, only atoken handful quantities of skeletal elements collected for burial. or two. Perhaps only certain body parts were delibe- Bone fragmentswere weighed to the nearest 0.5 g. rately selectedfor burial.Cremated human bone Skeletal remains were classified as ‘unidentified’ if may be recoveredfrom anumber of different con- the specific element theyrepresented was unclear. texts and these are summarised in Table 3.60. It is not This largelyrelates to fragments of trabecular bone alwayspossible to determine the type of deposit structure. The average thickness of the cranial vault (particularly if it is not containedwithin an urn) bones was also measuredusing asliding calliper to during excavation. Only by considering, during ana- the nearest 0.01 mm.The presence of pyre goods, lysis, factors such as the degree of fragmentation, the grave goods and faunal remains within the deposits skeletal elements that are present, and the mini- was also noted. mum number of individuals buriedisitpossible to Material smaller than 5mmwas not sorted. establish the most likely type. Analysis at thislevel focusedonthe general colour and boneelements only. Anote was also made of Results any identifiable fragments, such as dental crowns, hand and foot bones and otherfragmentsthat may Therewere six urned and three un-urnedcremation provideadditional information regardingthe MNI burials(Table 3.61). Adetailed record for each is count and estimation of biological parameters. The provided in Appendix4.Inaddition, contexts methods described for the unburnt human remains throughout the site included scatters of human also apply to the burnt remains in terms of esti- cremated bone. These are discussedseparately. mating the MNI,biological parameters and diagnosing pathological conditions. Onemethod,based on Condition and completeness cranial vault thickness (Gejvall 1963), which is specifically aimed at crematedbonedeposits, was not All but two (1266, 1767) burialswere incomplete, employed because the preservation of the remains having undergone considerable truncation. The was not sufficient. ceramic containers in cremation burials 1766, 1196 and 1209 were fragmented in situ,and it is unclear whether the deposits had been dispersed. This Background:the cremation process meansthat it is impossible to distinguish between The information that may be derivedfromcremated the effects of taphonomy and the weight of bone bone is less extensive than that derived from unburnt originally selected forburial. Overall, bonepreserva- bone. However, it is possible to makeinferences about tion was good, there was no markedweathering on pyre technology and whathappened to the bone the cortical surface of the bones and large quantities during and after the cremation process. When abody of trabecular bone were present in most deposits. is cremated,the skeleton is not destroyed but changes occur to the colour and composition of the bones. Weight and skeletal partrepresentation During the cremation, all the moisture evaporates out of the bone and the organiccomponent (chiefly The deposits rangedinweightfrom 1.5 gto1255.5 g collagen) is combusted, leaving only the mineral (Table 3.62). Theweightofbone of an adult portion. The bones also fragment and can become cremation from amodern crematorium rangesfrom distortedinshape with some shrinkageoccurring. about 1000 to 3600 g(McKinley 2000b,404). There- It is mostly the body fat that fuelsthe combustion. fore, only context 1767 can be considered to repre- Observations in modern crematoriasuggest that sent one complete cremated adulthuman skeleton. once the temperature has reached about 800– Cthe The highest represented bone fragmentswerethose fat will ignite,and the fuel jets can actually be turned from the upper and lowerlimb bones (919.5 gor off (McKinley 2000a). Whenthe body has been 38.7% of the entireassemblage). The skull vault cremated, and the pyre has cooled down, the bone was represented mainly by the frontal, parietal and fragments are collected. These may be buried occipital bones (335 gor14.1% of the entireassem- directlyinthe ground, in asmall pit, but more blage). Temporal bones were rarely represented by usuallythey are collected together and placed inside the petrouspyramid (cremation burial 1766), which is an urn or an organic container(for example, acloth usuallyacharacteristic anatomical landmark that is

73 Life andDeathinaRoman City

Table 3.60 Definitions of deposit types.

Deposit type Deﬁnition

Bustum Pyre site which also functioned as agrave. The pyre burnt down into the under-pyre pit and the human remains are buried in situ.Where no secondary manipulation has occurred, the cremated remains are expected to lie in the correct anatomicalposition on abed of charcoal.The effect of the burning has been observed to penetratethe soil by 2–5 cm. The average weight of bone retrievedfrom crematedadult is between 1600–2000 gbut may be as little as 1000 g. Urned burial Depositofcremated bone within acontainer.May be surroundedby, on top of, or overlain by a deposit of redeposited pyre debris. Unurned burial Concentrated deposit of bone, which may have been in an organic container,which may also include asecondary deposit of pyre debris within the backfill. Unurned burial or redeposited An apparentlymixed deposit of cremated human bone and charcoal which may represent the pyre debris remains of one or more cremated individuals. Pyre site Large quantity of charcoal with relatively small amount of burnt bone fragmentssituated on the ground surface or in under-pyre pits. The pits may also be TorL-shaped to aid draught and are shallow (0.10–0.20 mdeep). The soil beneath the pyre should show evidence of burning that may penetratethe soil by 2–5 cm. Redeposited pyre debris Amixture of fuel ash, fragmentsofcremated bone and pyre goods, and possibly burnt flint, burnt stone, burnt clay, fuel, ash and slag depending on the local environment. May contain a relativelylarge quantity of bone since asmall deposit of bone may have been collected for burial. The deposit may be present in the backfill of the burial, over the cremation burial, within pre-existing features, uncontained in spreads and in deliberately excavatedfeatures. Redeposited cremated remains Small amounts of cremated bone situated or recovered from features,such as pits and ditches, and in the backfill of intercuttingcremation burials. Cremation-related deposit Unknowndeposit type including cremated human bone.

Table 3.61 Summary of cremated human bone.

Cremationburial Context(s) Weight of MNI and identiﬁable Predominant colour/ human bone fragments Observations

1196 (late 1st-early 1189 (fill of pot 1191), 1190 269.5g MNI =1Skull, vertebrae, White; Fauna present 2nd AD) (backfill), 1193 and 1194 ribs, pelvis, shoulder girdle, (1g) (residues of fill of pot) long bones 1209 (late 1st-early 1197 (disturbed backfill), 281.5g MNI =1Skull, vertebrae, White; Fauna present 2nd AD) 1201 (fill of ancillary pot), ribs, pelvis, long bones (7.5g) 1202 (fill of pot sf 71) 1227 (late 1st-early 1225 (disturbedback fill) 1.5g MNI =1Skull White 2nd AD) 1266 (late 1st-early 1251 (fill of pot 1252), 1258 143g MNI =2Skull, vertebrae, ribs, White 2nd AD) (backfill) shouldergirdle, long bones 1766 (Pre-Flavian 1049 (fill of pot 1068) 132.5g MNI =1Skull, vertebrae, White; Fauna present AD 49–68) ribs, pelvis, shoulder girdle, (78.5g) long bones 1767 (Pre-Flavian 1062 (fill of pot), 1377 1255.5g MNI =3Skull, vertebrae, White; Fauna present AD 49–68) (backfill) ribs, pelvis, shoulder girdle, (11g) long bones 1768 1095 Unurned cremation 17.5g MNI =1Skull, ribs, White; Fauna present deposit long bones (1g) 1769 (AD 60–220) 1187 Unurned cremation 261g MNI =1Skull, vertebrae, White; Fauna present deposit ribs, pelvis, long bones (130.5g) 1770 (50 BC–AD 70) 1324 Unurned cremation 8.5g MNI =1Skull, ribs, White deposit long bones

presentinhuman cremated bonedeposits. There had the coronal and sagittalsutureswerepresent. were very few identiﬁable skeletal landmarks,con- Therewas one occipital protuberance amongstthe dylesand some portions of orbit (cremationburial occipital fragments (cremation burial 1196) and 1767) being amongst these. Some skull fragments that fragile,facial bones, such as the sphenoid and maxilla

74 Chapter Three

Table 3.62 Summary of weights (g) for each of the human cremated bone deposits.

Anatomicalarea Burial number

1196 1209 1227 1266 1766 1767 1768 1769 1770 Total (g)

Cranium 42.5 51 0.5 32 18.5 156.5 6271 335 Mandible 0.5 10 0.5 0.5 3.5 0006 Teeth 01032.5 50.5 70 19 Rib 4608.5 5550.5 3.5 0.5 83 Vertebra 2101223.5 01030.5 Scapula/pelvis 4.5 2.5 00.5 6.5 60 00.5 074.5 Long bone 125.5 92.5 03740527 8.5 83 6919.5 Patella 0200000002 Hand/foot0.5 6.5 010.5 23.5 02.5 034.5 Unidentiﬁed 90 118 159.5 57 401.5 2136.5 1866.5 Total (g) 269.5 281.5 1.5 143 132.5 1255.5 17.5 261 8.5 2370.5

(palatefragments in cremation burials 1766 and were also found in cremation burial 1266. Amongst 1767), were rarely present or could notbeidentified. the foot bones, metatarsals, tarsals(talus, calcaneous) Aright zygomatic (cremationburial 1766) was and phalanges were presentincremationburial 1209. present. Metatarsals, metacarpals and phalanges were also Few mandibular fragmentswere identified (6 gor identified in cremation burials 1266, 1767 and 1769. 0.2% of the total weight of the assemblage), mainly Many fragments (36.5%)could not be identified. fragments of the body, the ascending ramus,gonion These came largelyfrom the unsorted 4–2 mm sieve and the condyles (cremation burials1766 and 1767). size bulk,which were mostly long bone fragments. Some dental rootswere recoveredwhich provided Other commonunidentified fragmentswere portions information on biological age. of trabecular bone. Vertebrae and ribs were very fragmented and not complete. Most remains belonged to spinous pro- Minimum number of individuals, sex and age cesses or pedicles, bodies of vertebrae and shafts of determination ribs. Two odontoid processes from two axisvertebrae were identified in cremationburial 2. Atlas The material represented aminimum of 12 indivi- fragments were also identified in cremation burials duals (Table 3.63). Only two deposits contained 5and 7. material that indicated the presence of more than one The shoulder and pelvic girdle accounted for74.5 individual. These were contexts 1266 and 1767 and gor3.1%ofthe total weight. Some body portions of they contained aminimum of two and three indivi- the scapula and some clavicle shaft fragments duals respectively. represented the shoulder girdle. With regard to the The assemblage comprised nine adults,two pelvis, most fragments were from the body of the subadults and one individual of unknown age. The ilium and the acetabulum. There was no sciaticnotch adults were assumed to be over the age of approxi- identified. mately 15 years because of the dimensions of the All the major long bones were present, although long bones and dental formation. Unfortunately, no many of the epiphyses were missing post-mortem. epiphyses were preserved to assess their degreeof Therewas aportionofthe proximal epiphysis of the fusiontothe metaphyses or shafts of the long bones, radius (cremation burials1766 and 1767), and several and also no third molars were present to confirm the specific landmarks in cremationburial 1767: aportion adult age of the individuals. Thesubadult age was of the fovea capitis of afemur, distal radius, the radial established from dental formation. Tentatively, the tuberosity and nutrient foramenofthe radius, presence of first and second molars in cremation olecranonofthe ulna, proximal tibia, proximal and burial 1769, the lack of exposed dentine in aportion distal humerus, and distal femur. Humerus, radius, of amolar,the non-edentulous state of the mandible ulna, femur, tibia and fibula shaft fragments were and the non-obliterationofthe cranial sutures, clearly identified in most of the cremations. The suggest ayoung adult individual (18–25 years). skeletal catalogue in Appendix 4states what bones Considerable toothloss may be regarded as an were presentfor each cremationburial.Sothat, for indicatorofold age (over 45 years). Despite several example, in cremationburials1768 and 1769 only factors, for examplecultural practices, which may upper limb bones could be identified. One portionof influence tooth loss in younger adults, mandibles or patella (2 gor0.08% of the total assemblage weight) maxillae with over 50% of teeth lostante-mortem was identified in cremationburial 1209. have been assignedbysome to the older age Hand metacarpalsand phalanges were present in categories (Miles 2001;Mays 2002; also see Wols cremation burials1209 and 1767. Hand phalanges and Baker 2004). If considerable ante-mortem tooth

75 Life andDeathinaRoman City

Table 3.63 Summary of cremated human bone assemblage.

Cremation MNI Age Sex Observations burial (ageing criteria)

1196 1?Adult ?Bone dimensions 1209 1?Adult ?male >12 (dental formation) 1227 1? ?– 1266 22–6 years and >6–12 years ?Dental formation 1766 1?Adult ?>15 years (dental formation) 1767 3?Adult 2x?male; 1x?>12 (dental formation) 1768 1?Adult ?>12 (dental formation) 1769 1?Adult ?>15 (dental formation) 1770 1?Adult ?Bone dimensions loss can be considered to be an indicator of old age longest shaft fragment measured59mmand was ( 5 45 years), the skeleton from thislatter cremation from atibia from burial 1209. This was followed by burial did not show any signs (regeneration) of a fragments from burial 1767 which included afibula considerable ( 5 50%) number of teeth having been fragment measuring 58 mm and others from the lost during life. This is based on three tooth sockets femur and the tibia which were 55 mm in length. (alveoli) observed for the anterior dentitionand 29 dental fragments, which probably correspond to a Colour minimumofnine teeth.Hence,thisadultmay not be regarded as an old adult. Most of the cremated bonewas white or predomi- The sex of the adults couldnot be determined with nantly white with hues of blue/grey (93.2% of all any certainty owing to the absence of diagnostic classifiedfragments) (Table 3.65). Fragments that bones. Based on robusticity and the dimensions of were adarker colour,ranging from deep blue long bones and mandible fragments, it is tentatively through to grey, brown and black, accounted for suggested that the remainsfrom context 1767 repre- 6.7% of all classifiedfragments (143 g/2109 g).Thus sent the remainsofmale individuals. In addition, for most of the assemblage complete or intense certain cranial features (for example, the supraorbital combustion of all the organic component of the bone margin, the posterior zygomaticarch and the nuchal had taken place.Thisgenerally occurs at tempera- crest)that were identified in context 1209 suggested tures above 700 – C(Holden et al. 1995). apossible male, although all of these reflected a degreeofambiguity. Palaeopathology

Non-metric traits Most of the recovered fragmentscould be observed for pathological changes. Conditions that were Only one non-metric trait, the multi-foramina in the presentinclude periostitis and joint disease. zygomatic, couldbescored for presence or absence. The right zygomatic foundincremation burial 1766 Periostitis presented multi-foramina. No ossicles were found in any of the portions of cranial suturesthat were Periostitis was observedonthe bones from three observed. deposits (context numbers 1767, 1769 and 1770). Their distribution is summarised in Table 3.66. In Fragmentation addition to the usual periosteal changes, some bone fragments, in particular one metatarsal from burial Fragments between 10 and 5mmwere the most 1769, showeda‘coral reef’ appearance. This is remi- frequent, with the exception of cremation burials niscent of pulmonary osteoarthropathy. However, 1767 and 1770 which were mainly represented by this diagnosis, which relies on being able to establish fragments larger than 10 mm (Table 3.64). The the patternofskeletal involvement, cannotbe

Table 3.64 Percentage (in brackets) of bone fragments larger than 10 mm, between10–4 mm andsmallerthan 4mm.

Context number

1196 (%) 1209 (%) 1227 (%) 1266 (%) 1766 (%) 1767 (%) 1768 (%) 1769 (%) 1770 (%)

>10 mm 98g (36.3) 103g (36.5) 0g (0) 24g (16.7) 36g (27.1) 730.5 (58.1) 8g (45.7) 74g (28.3) 5g (58.8) 10–4 mm 153.5g (56.9) 145.5g (51.6) 1.5g (100) 86g (60.1) 87.5g (66.0) 468 (37.2) 9g (51.4) 145g (55.5) 3g (35.2) 5 4mm18g (6.6) 33g (11.7) 0g (0) 33g (23.0) 9g (6.7) 57 (4.5) 0.5g (2.8) 42g (16.0) 0.5g (5.8) Total (g) 269.5 281.5 1.5 143 132.5 1255.5 17.5 261 8.5

76 Chapter Three

Table 3.65 Percentage of fragments accordingtoweight that presented aparticular colour.

Context number

1196 1209 1227 1266 1766 1767 1768 1769 1770

Black 0g 0.5g 0g 0g 0g 0g 0g 0g 0g Blue 10g 0g 0g 10g* 7g 42g 0g 6g 0.5g* Brown 0g 1g 0g 0g 0g 18g 0g 1g 0g Grey 1g 18.5g 0g 0g 1g 22g 0g 15g 0g White 177g 94g 1.5g 86.5g 111.5g 251g 17.5g 22g 8g White +blue/grey 76g 91g 0g 36g 12g 830g 0g 152g 0g

*Dark blue with shades of light grey

Table 3.66 Periostitis distribution in the human cre- fragment. In cremation burial 1196 there was one mated bone assemblage. parietal fragment (2 g) with ectocranial pitting. These examples of ectocranialpitting were probably Context Bones affected weight %ofweights in each not caused by porotic hyperostosis because they did deposit (long bones) not display all of the features that are characteristic of this condition. 1767 Tibia Fibula 2.75g 0.5% 1769 Long bones 16g 19.2% (of long bones) Hand and foot 1.5g 60% (of hand/foot Discussion of the cremation burials bones bones) 1770 Long bone 5g 8.5g Aminimum of nine adultindividuals, two subadults (humerus?) and oneofunknown agecomposed the human cremated bone assemblage. confirmed with the present remains. The periosteal Skeletal biologyand palaeopathology reactionsthat were identified for burials 1767 and 1770 were typical of this condition. As previously stated, amongst the 12 individuals identified from the ninecremated human bone deposits, two were subadults youngerthan 15 years Joint disease of age, the ageofone skeletonwas unknown and the Joint disease,inthe form of osteoarthritis (OA), was remainder were likelytobeadults based on bone observedonthe bones from context 1767. The dimensions and dental formation. Three adults were changes, osteophytosis and an altered bonycontour, possible males whilethe remainder were of un- involvedthe first metatarsophalangeal joint (the joint known sex. Thus the assemblage includes the at the base of the greattoe). No other joints from this remainsofindividuals of different ages, subadult deposit1767 had OA.This includes two vertebral and adult, and at leastofmale individuals. facets,two rib facets and aleft mandibular fossa of Health indicators were presentinthe form of the temporo-mandibular joint. periostitis and OA.Interms of periostitis, the involvement of the tibia may suggest that relatively inconsequential factors,such as mild trauma, were Dental conditions causing this bony reaction.However, multiple Dental elements were obtained from six cremation element involvement and the suggestion of more burials(1196, 1209, 1266, 1766, 1767 and1769) and thorough-goingperiostealreactions(in the form of these are included in the skeletal catalogue in ‘coral reef’ bone), indicatesthat systemic, chronic Appendix 4. No pathological conditions were obvious disease was experienced by some of these indivi- on the dental fragments. No ante-mortem tooth loss duals. This finding is in keeping with that observed was observedin15sockets from adultindividuals for the unburnthuman remains from LondonRoad. correspondingtosix mandibular (contexts 1196 and Gross dental caries has akey role in the develop- 1209) and nine maxillary alveoli (contexts 1766, 1767 ment of ante-mortem toothloss. Thus, although no and 1769). Periapical cavities couldnot be scored complete teeth survive it is possible to say that none becauseofthe incompleteness of the remains. (0/15 sockets) had beenlost during life as aresult of dental caries or abscesses. Cariesand abscesses are heavilyinfluenced by adiethigh in carbohydrate Further observations and poororalhygiene. In the absence of information Increasedporositywas present in the ectocranial pertaining to more precise age estimatesand without surface of several fragmentsofskull from cremation the complete dental crowns and the sockets for burial 1769. This was presentonthree fragments posterior ratherthan anterior teeth,moredetailed (2 g) from the skull,including one occipital bone information on diet cannot be obtained. None of the

77 Life andDeathinaRoman City seven alveoli from the adultindividualspresented cremated bone, different temperatures arecorrelated any regeneration indicative of ante-mortem tooth with different colours. These range from orange loss. (200– C), throughtoblack (300– C), grey (600– C) and Overall,the limited pathological conditions re- finally white at atemperature usually between 700– C ported to be present may be reflective of living con- and 800– C(Shipman et al. 1984; Holden et al. 1995). In ditions perhaps associated with inadequate nutrition modern crematoria where temperature, fuel avail- and an environment which would have facilitated abilityand air circulationare optimised,full crema- the spread of infection. However, alargersample size tion of an adultcorpse generallytakes between 1and would be required to confirm thisstatement together 1.5 hourstocomplete when the temperature within with afurther exploration of the living conditions at the crematorismaintained between 700–1000– C the time as inferred by other sourcessuch as material (McKinley 2000b, 404). The colour of the boneinthe culture and environmental data for the period (see London Road assemblage was predominantlywhite Chapter5). reflecting successful cremationatatemperature over 600–700– C. Limited incompleteburning was indicated by the presence of some brown/black/blue Efficiency of cremation and the funerary rite fragments. This would suggest that although the Overall, there was no apparent pattern with regard mourners and/or funerary attendantstook consider- to the presence or absence of skeletal elements. The able pains to ensure thorough burning of the corpse, relatively high proportion of cranial fragmentsis some body parts escaped complete cremation, prob- most probablyaresult of the relative ease with ably depending on the depth of overlyingsoft tissue which theyare identified compared to other bones, and/or their position on the pyre. Other factors that and the small quantityoffragments fromthe axial may influence colour variation are the temperature of skeleton is probably the result of preservation and the firing, oxygen supply,the size of the pyre, the identification ratherthan deliberate exclusion. Col- duration of exposureand proximity of the body to lection of the bones seems to have been undertaken the flames, the amountofsoft tissue, position of the with care because there has clearlybeenanattempt corpse,and movementsofbody parts during the to select as many bones as possible, including small cremation (McKinley 2000a and 2000b). fragments such as dental crowns and roots. Fracture patterns and surface texture may indicate Cremated bone may range in colour from brown- whether abody was burntwith its flesh or whether ish-black (slightlycharred), through hues of blue and only the dry skeleton was selected for the pyre after grey, to white, or fullycalcined bone (McKinley 2000b, decomposition or defleshing of the soft tissue. As 405). Colourmay indicate the temperature at which opposedtodry bones, burning askeleton covered the body was burnt, although this also depends on with flesh produces curved transverse fracture lines, how much soft tissue is around the bone. On human irregular longitudinal splitting, twistingand marked

Table 3.67 Human bone from non-cremation burial contexts.

Context Sample/ Sieve size Bone Weight Colour Observations SF (mm) (g)

1063 310–4, 4–2 1rib shaft 14 unidentified 1White +grey Dimensions=>10 years fragments 1106 –>10, 10–4 2upper limb, 1long bone. 4.5 White Adult dimensions. 4unidentifiedfragments Longest fragment =38mm 1132 11 10–4 1?radius shaft 0.5 White Adult dimensions 1166 28 >10, 10–4 3skull vault fragments 1.5 White Dimensions=>10 years 1166 29 >10, 10–4 1skull vault, 1rib shaft, 2White +hues of blue – 5unidentifiedfragments 1166 30 10–4, 4–2 3skull vault fragments, 1White – 1unidentified 1201 53 >10 mm 1long bone shaft, 1vertebral 3White-light grey Spit 2 body, 2unidentified fragments 1212 112 10–4 1ulna or fibula shaft fragment <0.5 White – 1246 95 10–4 1rib shaft, 2unidentified 0.5 Light grey – 1333 145 10–4 2long bone shaft fragments <0.5 White – 1515 242 10–4 2unidentifiedand 2long 0.5 Light brown, grey – bone fragments 1707 –>10, 10–4 1ulna, 1tibia, 1long bone, 4White, white +grey, Adult dimensions. 1metatarsal? shaft fragments hues of blue (MT shaft) Longest fragment =28mm. 2fauna fragments (5 g)

78 Chapter Three warping (Ubelaker 1989; Buikstra and Ubelaker of the aperture of most vessels was around c 100 mm 1994). This latter was the case at London Road. and the size argument heremay be irrelevant (Booth, Investigations in modern crematoria have found pers. comm.). In addition, McKinley(1997, 251) that the averageboneweight of acremated adult points out that fragment sizes recorded during post- female individual weighs on average 1615.7 gand a excavation analysis will unlikely represent the size of male body an average of 2283.5 g(McKinley 1993). the fragments during the timeofdeposition due to Predictably, individualsofsmaller and more gracile anumber of factors including site disturbances and build, such as children, or those older individuals the excavation process. with osteoporosis will have alower bone weight and possibly poorerbonesurvival (McKinley 2000b, 404). Crematedhuman bone from non-cremation Overall, bone weights were low in the LondonRoad, burial contexts Gloucester, assemblage with the exception of one undisturbed, urned adult cremation burial (burial Anumber of non-cremation burial related deposits 1767, context 1062), which weighed approximately with human crematedbone were present and are 1255.5 g. This may suggest that, at least forthose merelylisted here (Table 3.67). The estimatedbiologi- undisturbed burials, collection and deposition of the cal ages are based on bone dimensions only. In no case entirecremated remains of these individualswas not was sex determinedand no further information could deemed important, and that only ‘token’ deposits be obtained other than that included in Table 3.67. had been buried. In addition, threebagswere analysed containing Fragment size represents the measurements taken humancremated bone from the mass grave fill by the human osteologist during analysis. Factors that (1545).From this context, sample 244 containedone affect fragmentation include the cremationprocess, trabecular bone fragment (10–4 mm, 4 0.5 g). The collection and burial of the humanremains, deliberate remaining fragments from the other two bags inclu- fragmentation by the mourners or grave diggers, ded apparently adult bones. These comprised long taphonomic factors (forexample,soil characteristics bones smaller than 10 mm and two unidentified and ploughing), and the muchlater processof fragments, in total weighing 6g.The colour varied archaeological excavation and post-excavation pro- according to the fragment from black to light blue, cessing(McKinley 1994). It is possiblethat cremated grey and white.Thereisalso 5gof skull vault (two bone was fragmented deliberately in order to fit it into fragments possiblyoffrontalbone) which are white the urn but there is no evidence to suggest this was the in colour with hues of blue and grey. Due to the case. In fact, while most fragmentswere larger contextual informationand the small quantity of ( 4 10 mm) than would perhaps be expectedifthey these remainsitislikely that the inclusions in the had been deliberatelytreatedinthismanner, the size mass graveare accidentalrather than deliberate.

79 Chapter 4The Grave Catalogue

INTRODUCTION Sex: Male Age: 13–17 years Each grave is referred to in the catalogue by its grave Completeness: 50–75% group number, the number usedthroughout the text. Condition: Good Each entry gives details of the grave,comprising Pathological conditions: healed blunt cranial trauma, temporo- mandibularjoint anomaly, enamel hypoplasia orientation; shape and dimensions; the nature of the fill; stratigraphic relationships with other features; Finds: Five sherds of 1st–early 2nd century pottery were recovered from the back-fill. evidence for grave furniture; adescription of any grave goods, pyre goodsand other finds recovered from the grave;datewhere thiscould be established Inhumationgrave 1144 from artefactual evidence or radiocarbon dating. The Grave cut 1122 locationofeach grave is indicated on Figure 4.1 and Orientation:SSW-NNE plans (Figs 4.2–8) are only provided for those graves Fill: Greyish-brown silty clay back-fill (1123=1269) containing grave goods. Relationships: Cuts grave 1143 The following information is given for each skeleton, when applicable: adescription of the posture of Skeleton 1114=1270 the body, including the position of the arms and legs; Posture: Supine, legs extended and parallel sex; age; an indication of the completeness of the Arm position: Left arm extended beside body, right arm absent skeleton as recovered; an assessment of the condition Age: 26–35 years of the bone; and details of any pathological abnor- Completeness: 25%–50% Condition: Very good malities. The completeness of discrete skeletons was Pathological conditions: cribra orbitalia, neoplastic disease categorised as >75%, 75–50%,50–25% or <25%. Bone (button osteoma), enamel hypoplasia, dental caries, dental cal- condition was assessed according to criteria pro- culus, periodontitis posed by McKinley(2004). Musculo-skeletal markers: enthesophytes on calcanei For cremations the following informationisgiven: Charnel 1115 the weight of the calcined fragments; maximum dimension of the largest fragment;abrief note of Shafts of left and right tibia, which had been placed directly on identifiable fragments; colour of the bone; the the skull of skeleton 1114=1270 minimumnumber of individuals (MNI)represented Finds: An assemblageof11hobnails and afragment of colourless by the remains; the age and sex of the individual(s). glass that may be amodern intrusion were recovered from the back-fill. Age at death and biological sex were estimated as Date: 2nd centuryADorlater described in Chapter3. Comments: The middle part of the grave was truncated by a Every object is described in detail with the excep- modern feature, leaving only the head and left upper arm and the tion of material derivedfrom the backfill of the lower parts of the legs. grave.The bracketed numbers following each entry are the context number and small find number Inhumationgrave 1145 (Fig. 4.2) ascribedtothe object (the latter prefixedbysf), and Grave cut 1126 areference to any illustration of the item located within the relevant specialist report. Orientation:SW-NE Shape: Sub-rectangular. Vertical sides, flat base Dimensions: 1.52 x0.54 m, 0.12 mdeep CATALOGUE OF GRAVES Fill: Orange-brown silty clay back-fill (1129) Inhumationgrave 1143 Skeleton 1127 Grave cut 1111 Posture: Prone, extended. Lower parts of legs removed by later truncation Orientation:NE-SW Arm position: Left arm extended beside body, right flexed Shape: Rectangular. Straight vertical sides, flat base beneath torso with hand lying beneath left shoulder. Dimensions: 1.97 x0.69 m, 0.25 mdeep Sex: ?Male Fill: Greyish-brown silty clay back-fill (1113) Age: 18–25 years Relationships: Cuts grave 1759, cut by grave 1144. Relationship Completeness: 25–50% with grave 1167 uncertain Condition: Fair Pathological conditions: dental calculus Skeleton 1112 Musculo-skeletal markers: enthesophytes on ulnae and Posture: Prone, extended. Legs extended and parallel, trun- femora cated at mid-tibia Grave goods: Part of aring-necked flagon in oxidised TF 11A (1128, Arm position: Hands and forearmsabsent but the angle of the Fig. 5.2, no. 10) lay on its side against the side of the skull. The rim upper arms suggests that they were behind the back has an old chip out of the circumference.

81 Life andDeathinaRoman City

Figure 4.1 Plan of all burials.

82 Chapter Four

Figure 4.2 Plan of inhumation burial 1145.

Figure 4.4 Plan of inhumation burial 1334.

Figure 4.3 Plan of inhumation burial 1243.

Other ﬁnds: Asingle iron nail (sf 24) was recovered from the backﬁll. Date: 1st–early 2nd century

Inhumationgrave 1146 Figure 4.5 Plan of inhumation burial 1352. Grave cut 1130 Orientation:N-S Age: 26–35 years Shape: Sub-rectangular, concave base Completeness: <25% Dimensions: 1.90 x0.40 m, 0.18 mdeep Condition: Fair Fill: Greyish-brown clay back-ﬁll (1132) Pathological conditions: active periostitis on left femur and left calcaneous, healed periostitis on right femur, dental calculus. Skeleton 1131 Finds: Three sherds of 1st century pottery were recovered from the back-ﬁll, as well as eight hobnailsand asingle structural nail Posture: Supine, legs extended with feet together (sf 23). Arm position: Left arm extended beside body, right arm Comments: This individual has been buried in avery narrow grave. absent The legs and feet are very close together, suggesting that they may Sex: Female have been bound within ashroud.

83 Life andDeathinaRoman City

Dimensions: 1.80 x0.50 m, 0.10 mdeep Fill: Orange-brown clay back-ﬁll (1136)

Skeleton 1135 Posture: Supine Age: Adult (>18 years) Completeness: <25% Condition: Good Pathological conditions: enamel hypoplasia, dental calculus, periodontitis Finds: One sherd of 1st century pottery and two structural nails (sfs 25, 26) were recovered from the back-ﬁll.

Inhumationgrave 1148 Grave cut 1137 Orientation:E-W Shape: Indeterminate due to severe truncation Dimensions: Indeterminate due to severe truncation Fill: Orange-brown silty clay back-ﬁll (1139)

Skeleton 1138 Posture: Supine, extended Arm position: Left arm extendedbeside body, right side of body absent Age: 18–25 years Completeness: <25% Condition: Good Pathological conditions: healed periostitis on tibiae, enamel hypoplasia, dental caries, dental calculus Finds: Asingle nail was recovered from the back-fill. Comments: Grave severely truncated by modern disturbance, only the left side of the body surviving. Figure 4.6 Plan of inhumation burial 1374. Inhumationgrave 1150 Grave cut 1104 Orientation:NE-SW Shape: Rectangular. Vertical sides, flat base Dimensions: >1.5 x0.6 m, 0.1 mdeep Fill: Orange-brown silty clay back-fill (1102)

Skeleton 1103 Posture: Prone, legs extended and parallel Arm position: Left arm extended beside body, right arm slightly ﬂexed with hand beneath pelvis Sex: Female Age: 36–45 years Completeness: 50–75% Condition: Very good Pathological conditions: osteoarthritisonribs and vertebrae, osteophytosis on ribs, vertebrae and long bones, active and healed periostitis on right tibia, enamel hypoplasia, dental caries, dental calculus, periodontitis, periapicalcavity Finds: Asingle sherd of 1st–2nd century pottery was recovered from the back-ﬁll. Comments: The north-eastern end of the grave was truncated by the foundation of amodern wall and the south-western end by machining,the latter having removed the feet.

Figure 4.7 Plan of inhumation burial 1505. Inhumationgrave 1151 Grave cut 1108 Inhumationgrave 1147 Orientation:E-W Grave cut 1134 Shape: Sub-rectangular. Steep sides, ﬂat base Dimensions: 2.16 x0.76 m, 0.21 mdeep Orientation:NE-SW Fill: Yellowish brown silty clay back-ﬁll (1110) Shape: Sub-rectangular Relationships: Cuts grave 1759

84 Chapter Four

Figure 4.8 Plan of cremation burials 1196, 1209, 1266 and 1767.

Skeleton 1109 Comments: Asmall patch of green staining indicating the former location of adegraded copper alloy object was observed in the Posture: Supine, legs extended with feet together chest region. Skull absent due to truncation. Arm position: Arms slightly flexed, hands crossed over pelvis Sex: ?Male Age: 26–35 years Inhumationgrave 1167 Completeness: >75% Grave cut 1164 Condition: Very good Pathological conditions: healed trauma on right rib, healed Orientation:NNE-SSW periostitis on tibiae and fibulae, dental caries, dental calculus Shape: Sub-rectangular. Vertical sides, flat base Musculo-skeletal markers: enthesophytes on tibiae and left Dimensions: 1.76 x0.42 m, 0.22 mdeep patella Fill: Orange-brown clay back-fill (1166) Relationships: Relationships with graves 1143 and 1759 are Grave furniture: The remains of acoffin were represented by four uncertain iron nails (sfs 16–19) located in situ around the left side and foot of the burial. sf 19 was located above the skeleton, indicating that this Skeleton 1165 was acoffin rather than abier. Other finds: Adamaged and incompletecopper alloy mount (sf Posture: Supine, legs extended and parallel 20) was recovered from the back-fill. Arm position: Arms flexed with hands crossed over pelvis

85 Life andDeathinaRoman City

Sex: ?Female Dimensions: 1.47 x0.4 m, 0.18 mdeep Age: 18–25 years Fill: Orange-brown clay gravel back-fill (1185) Completeness: >75% Relationships: Cut by graves 1229 and 1230 Condition: Poor Pathological conditions: healed periostitis on left calcaneous, Skeleton 1184 enamel hypoplasia, dental calculus Grave furniture: The remains of acoffin were represented by five Posture: Supine, legs extended iron nails (sfs 35–39) lying in situ around the lower part of the Completeness: <25% skeleton. Condition: Poor Grave goods: Atotal of 20 hobnails were recovered. Agroup of 17 Date: AD 70–240 (NZA 27006) (radiocarbon assay) hobnails (sf 34) were located at the foot of the grave and may re- Comments: The middle and south-western parts of this grave were present an item or items of footwear placed as grave good(s). The removed by the digging of subsequent graves 1230 and 1229 location of the remaining three hobnails (sf 28) was not recorded. respectively. Other finds: Arim sherd of 1st century Dorset black burnished ware was recovered from the back-fill, which joined with asherd recovered from grave 1230. Inhumationgrave 1229 Grave cut 1154 Inhumationgrave 1218 Orientation:S-N Grave cut 1215 Shape: Sub-rectangular Dimensions: 1.18 x0.58 m, 0.07 mdeep Orientation:SE-NW Fill: Orange-brown clay back-fill (1155) Shape: Rectangular. Vertical sides, flat base Relationships: Cuts grave 1228, cut by grave 1230 Dimensions: 1.93 x0.55 m, 0.28 mdeep Fill: Orange-brown silty clay back-fill (1217) Skeleton 1153

Skeleton 1216 Posture: Supine. Lower part of body absent from mid-pelvis Arm position: Left arm flexed across pelvis, right arm absent Posture: Supine, legs extended with feet together Sex: ?Male Arm position: Left arm flexed with hand near right shoulder, Age: 18–25 years right arm extended beside body with hand beneath pelvis Completeness: <25% Sex: Male Condition: Very good Age: 45+ years Pathological conditions: rib osteoarthritis, Schmorl’s nodes, Completeness: >75% enamel hypoplasia, dental caries, dental calculus, periodontitis, Condition: Fair periapical cavity Pathological conditions: cortical defect on left clavicle, osteoar- Grave furniture: The remains of acoffin were represented by five thritis on ribs, vertebrae and hip joints (femora and os coxa affec- nails (sfs 29–33) located in situ around the edge of the grave. ted), healed periostitis on ribs and metatarsals, cribra orbitalia, enamel hypoplasia, dental caries, dental calculus, periodontitis Other finds: Asingle small sherd of 1st century pottery and two Musculo-skeletal markers: enthesophytes on patellae nails were recovered from the back-fill. Comments: The northern and eastern parts of the grave have been Grave furniture: Three iron nails (sfs 78–80) were recovered from truncated by the digging of subsequent grave 1230, removing the the back-fill. right side and lower part of the body. Other finds: Four sherds of 1st–2nd century pottery were recovered from the back-fill. Inhumationgrave 1230

Inhumationgrave 1219 Grave cut 1180 Grave cut 1208 Orientation:SSW-NNE Shape: Sub-rectangular. Vertical sides, concave base Shape: Sub-circular Dimensions: 2.06 x0.41 m Dimensions: 1.40 x1.22 m, 0.15 mdeep Fill: Brownish-orange clay gravel back-ﬁll (1182) Fill: Reddish-brown silty clay back-ﬁll (1207) Relationships: Cuts grave 1229

Skeleton 1206 Skeleton 1181 Posture: Crouched.Lying on right side with head to NE and Posture: Prone, legs extended with feet together legs flexed Arm position: Left arm extended beside body, right arm flexed Arm position: Left arm flexed, right arm extended with hand beneath pelvis Sex: ?Female Sex: Female Age: 18–25 years Age: 26–35 years Completeness: 50–75% Height: 1.62 m Condition: Fair Completeness: >75% Pathological conditions: rib osteoarthritis, possible healed Condition: Very good trauma on right humerus,possible abnormalhealed inflamma- Pathological conditions: healed rib periostitis, periostitis on tion on sphenoid, cribra orbitalia, enamel hypoplasia, dental occipital bone, vertebral osteophytosis, healed maxillary caries, dental calculus, periodontitis sinusitis, enamel hypoplasia, dental caries, dental calculus, Finds: Asingle unclassified nail was recovered from the back-fill. periodontitis Musculo-skeletal markers: enthesophytes on left femur Finds: An iron ring (sf 141), acoin dated c AD 270–295, four Inhumationgrave 1228 fragments of tile, an iron nail, ahobnail head and seven sherds of Grave cut 1183 2nd century pottery were recovered from the back-fill, the latter including arim sherd from aDorset black burnished ware vessel Orientation:S-N that joined with asherd from the back-fill of grave 1167. Shape: Rectangular. Vertical sides, flat base Date: Late 3rd–4th century

86 Chapter Four

Inhumationgrave 1234 Date: AD 1–130 (NZA 27005) (radiocarbon assay) Comments: The western half of grave was truncated by amodern Grave cut 1231 pit, which had removed the upper part of body with the exception of the right forearm and hand. Orientation:NNW-SSE Shape: Rectangular. Vertical sides, flat base Dimensions: >0.92 x0.38 m, 0.12 mdeep Inhumationgrave 1246 Fill: Orange-brown silty clay back-fill (1233) Grave cut 1245 Skeleton 1232 Orientation:E-W Posture: Prone, legs extended and parallel. Severely truncated, Fill: Greenish-orange clay (1246) only the legs surviving. Skull placed on backs of knees Sex: ?Female Skeleton 1235 Age: 18–25 years Completeness: 25–50% Posture: Crouched,lying on left side with head to east Condition: Good Arm position: Tightly flexed in front of torso Pathological conditions: possible active periostitis on mandible, Age: 5–12 years cribra orbitalia, hyperostosis frontalis interna, enamel hypo- Completeness: <25% plasia, dental caries, dental calculus Condition: Good Grave goods: Two groups of hobnailswere found, representing Finds: An iron finger ring (sf 88), lackingits intaglio, was re- items of footwear worn at the time of burial—a group of 32 (sf 87) covered from the grave and is likely to have been placed as agrave associated with the right foot and agroup of 37 (sf 91) associated good. with the left foot. Comments: Southern side of grave truncated by modern concrete Other finds: Atotal of 27 sherds of 1st century pottery and three pad, removing legs. iron nails were recovered from the back-fill. Comments: The northern half of this grave has been truncated by a modern pit, leaving only the southern end. Inhumationgrave 1264 Grave cut 1261 Inhumationgrave 1240 Orientation:N-S Shape: Rectangular. Vertical sides, flat base Grave cut 1237 Dimensions: 1.82 x0.50 m, 0.22 mdeep Shape: Sub-circular. Steep sides, flat base Fill: Orange-brown silty clay back-fill (1263) Dimensions: 0.90 mdiameter, 0.22m deep. Relationships: Cuts graves 1243 and 1288 Fill: Brownish-yellow sandy clay back-fill (1239) Relationships: Overlain by Roman soil layer 1106 Skeleton 1262 Posture: Supine, legs extended and parallel Skeleton 1238 Arm position: Left arm extended beside body, right arm flexed with hand over pelvis Posture: Crouched, lying on left side with the head to the east Sex: ?Male Sex: ?Male Age: 26–35 years Age: 45+ years Completeness: <25% Completeness: <25% Condition: Fair Condition: Fair Pathological conditions: osteoarthritis (including both eburna- Pathological conditions: cortical defects on left femur, possible tion and osteophytosis) on vertebrae, left os coxa, right Paget’s disease, enamel hypoplasia, dental caries, dental humerus, right radius and right femur, healed trauma on right calculus clavicle, active and healed periostitis on right tibia and fibula, Finds: Three sherds of 1st century pottery were recovered from the healed periostitis on right metatarsal, enamel hypoplasia, back-fill. dental caries, dental calculus, periodontitis Comments: Truncated by amodern pit. Musculo-skeletal markers: enthesophytes on femora and patellae Grave furniture: The possible remains of acoffin were represented Inhumationgrave 1243 (Fig. 4.3) by four nails (sfs 106, 110, 111, 116) located around the foot of the Grave cut 1321 grave and afifth (sf 112) at the head. Other finds: Asingle small sherd of 1st century pottery was Orientation:WNW-ESE recovered from the back-fill. Shape: Sub-rectangular Date: Assigned a3rd–4th century date on the basis of its apparent Dimensions: 2.0 x0.60 m, 0.30 mdeep spatial associationwith grave 1370. Fill: Orange sandy clay back-fill (1322) Relationships: Cut by grave 1264 Inhumationgrave 1275 Skeleton 1211 Grave cut 1276 Posture: Supine, legs extended Orientation:SE-NW Arm position:Right arm extended beside body. Left arm Shape: Sub-rectangular. Steep sides, flat base absent Dimensions: 0.95 x0.36 m, 0.17 mdeep Age: Adult (>18 years) Fill: Yellow clay back-fill (1278) Completeness: <25% Condition: Poor Skeleton 1277 Grave goods: The upper part of aring-necked flagon in oxidised fabric TF 11A (sf 155, Fig. 5.2, no. 11) lay on the right foot. Posture: Supine. Left leg absent, right leg extended Other finds: Two sherds of 1st century pottery were recovered from Arm position: Left arm absent, right arm extended beside body the back-fill. Age: 2–5 years

87 Life andDeathinaRoman City

Completeness: <25% Skeleton 1286 Condition: Fair Pathological conditions: cribra orbitalia, enamel hypoplasia Posture: Extended on left side but slumped slightly onto front. Legs slightly flexed, feet together Date: AD 50–220 (NZA 29423) (Radiocarbon assay) Arm position:Left arm beneath torso, right arm extended beside body Sex: ?Male Inhumationgrave 1283 Age: 26–35 years Grave cut 1210 Completeness: 50–75% Condition: Good Orientation:N-S Pathological conditions: rib osteoarthritis, osteophytosis on Shape: Sub-rectangular. Steep sides, flat base vertebrae and femora, trauma, periostitis on right tibia, osteo- Dimensions: 1.70 x0.73 m, 0.16 mdeep myelitis on left tibia and fibula, cribra orbitalia, enamel Fill: Brownish-grey silty clay back-fill 1212 hypoplasia, dental calculus, periodontitis, circulatory disease Relationships: Cut by grave 1284 (possible Perthes’ disease). Musculo-skeletal markers: enthesophytes on femora. Skeleton 1279 Grave furniture: The remains of acoffin were represented by an assemblage of 15 coffin nails (sfs 119–125,128, 129, 131, 132, 136–8, Posture: Supine, extended 140) located in situ around the edges of the grave. Arm position: Right arm extended beside body, left arm absent Grave goods: The partial remains of adomestic fowl were placed Age: Adult (>18 years) behind legs. Completeness: <25% Other finds: Ten sherds of 1st century pottery were recovered from Condition: Fair the back-fill. Pathological conditions: healed periostitis on left metatarsal Date: AD 60–240 (NZA 27007) (radiocarbon assay) Grave goods: An assemblageof44hobnailswas recovered from a soil sample taken from around the feet. Inhumationgrave 1315 Other finds: Four nails and ten sherds of 1st century pottery including asingle fragment from an amphora were recovered Grave cut 1314 from the back-fill. Comments: The northern end of the grave is truncated by amodern Orientation:NW-SE feature and the eastern side by grave 1284, leaving only the feet, Shape: Sub-rectangular. Steep sides, flat base right leg and right arm. Dimensions: 2.10 x0.50 m, 0.37 mdeep Fill: Greyish yellow clay back-fill (1312) Relationships: Cuts grave 1330 Inhumationgrave 1284 Skeleton 1313 Grave cut 1280 Posture: Supine, right leg extended,left leg flexed with foot Orientation:N-S beneath right ankle Shape: Sub-rectangular. Steep sides, flat base Arm position: Left arm extended at slight angle to body, right Dimensions: 1.80 x0.75 m, 0.44 mdeep arm flexed with forearm across pelvis Fill: Brownishgrey silty clay back-fill (1282) Sex: Male Relationships: Cuts grave 1283 Age: 18–25 years Completeness: 50–75% Skeleton 1281 Condition: Good Pathological conditions: vertebral osteophytosis,degenerative Posture: Supine, legs extended and parallel disc disease, enamel hypoplasia, dental caries, dental calculus, Arm position: Left arm extended beside body, right arm periodontitis slightly flexed with hand over pelvis Musculo-skeletal markers: enthesophytes on ulnae and femora Age: Adult (>18 years) Completeness: <25% Grave goods: An assemblageof37hobnailswas recovered from a Condition: Fair soil sample taken from around the feet. Pathological conditions: osteoarthritis on metacarpal and Other finds: Two iron nails (sfs 139 and 143) were recovered from hand phalanx, osteophytosisonmatacarpals, phalanges, right the back-fill. femora and left os coxa, possible trauma (followed by anky- Comments: The irregular posture may result from the body shift- losis) on hand phalanges ing inside acoffin, although only two possible coffin nails were Musculo-skeletal markers: enthesophytes on left ulna and recovered. femora Grave goods: Three hobnails were recovered from the area of the Inhumationgrave 1319 feet (sf 130, 133). Other finds: Eleven sherds of 1st centurypottery and three frag- Grave cut 1316 mentary iron nails (sfs 94, 113 and 127) were recovered from the Orientation:SE-NW back-fill. Shape: Sub-rectangular. Steep sides, flat base Comments: The northern end of the grave is truncated by amodern Dimensions: 1.48 x0.52 m, 0.14 mdeep feature. Fill: Orange-brown gritty clay back-fill (1318)

Inhumationgrave 1288 Skeleton 1317 Grave cut 1287 Posture: Supine, legs extended and truncated at approximately the knees Orientation:SE-NW Arm position: Left arm flexed at an angle to body, right arm Shape: Sub-rectangular. Near vertical sides, flat base flexed across stomach Dimensions: 2.00 x0.60 m, 0.29 mdeep Age: 13–17 years Fill: Brown clay back-fill (1285) Completeness: 25–50% Relationships: Cut by grave 1264 Condition: Good

88 Chapter Four

Pathological conditions: healed periostitis in femora, endocra- Skeleton 1340 nial lesions, enamel hypoplasia, dental caries, dental calculus Posture: Supine, legs extended and parallel Grave furniture: The remains of acoffin were represented by eight Arm position: Arms extended beside body nails located around the edges of the south-eastern half of the Age: 36–45 years grave (sfs 141, 144–150). Completeness: <25% Other finds: Asingle sherd of 1st century pottery was recovered Condition: Poor from the back-fill. Pathological conditions: cribra orbitalia, hyperostosis frontalis Comments: The NW end of the grave has been truncated by a interna, neoplastic disease (button osteomata) modern feature, removing the lower parts of the legs. Musculo-skeletal markers: enthesophytes on femur Grave furniture: The remains of acoffin were represented by five Inhumationgrave 1330 iron nails (sfs 163, 168–71, 182, 203) located around the edges of the north-western end of the grave. Grave cut 1329 Grave goods: An assemblageof178 hobnails (sf 170) was recovered from the area of the feet. Orientation:NE-SW Shape: Sub-rectangular. Steep sides, flat base Other finds: Two sherds of 1st century pottery were recovered from Dimensions: 2.00 x0.65 m, 0.18 mdeep the back-fill. Fill: Greenish yellow clay back-fill (1327) Relationships: Cut by grave 1315 Inhumationgrave 1352 (Fig. 4.5) Skeleton 1328 Grave cut 1345 Posture: Supine, legs extended and parallel Orientation:NE-SW Arm position: Both arms flexed, crossing over pelvis Shape: Sub-rectangular. Steep sides, flat base Sex: Male Dimensions: 1.10 x0.60 m, 0.18 mdeep Age: 26–35 years Fill: Yellowish-brown gravely clay back-fill (13470) Completeness: 50–75% Skeleton: No skeletal material was present Condition: Fair Grave furniture: The remains of abox or coffin were represented by Pathological conditions: osteoarthritisinvertebrae, benign an assemblage of seven nails (sfs 176–81, 247) located around the cysts in carpal bones, healed infection in frontal sinus, cribra edges of the grave. orbitalia, dental caries, dental calculus, periodontitis,periapical Grave goods: cavity Grave goods: Two groups of hobnailswere found, representing 1. Ring-necked flagon in oxidised TF 11A, represented by the footwear worn at the time of burial—a group of 22 (sf 156) upper part only (1348, sf 174, Fig. 5.1, no. 8) associated with the right foot and agroup of 13 (sf 157) associated 2. Upper portion of aflagon in oxidised TF 11A (1349, sf 175, with the left foot. Fig. 5.2, no. 9) Other finds: Three iron nails (sfs 158–160) were recovered from the 3. Tubular unguent bottle in blue/green glass. sf 173 back-fill. 4. Tall conical unguent bottle in blue/green glass. sf 172 5. Some mammal bone (sf 183) that was lost prior to identification

Inhumationgrave 1334 (Fig. 4.4) Other finds: Three sherds of 1st century pottery were recovered from the back-fill. Grave cut 1331 Date: 1st–early 2nd century AD Orientation:NNE-SSW Shape: Oval Dimensions: 1.52 x0.83 m, depth not recorded Inhumationgrave 1353 Fill: Reddish-brown gravelly clay back-fill (1333) Grave cut 1336 Relationships: Cut by grave 1344 Orientation:S-N Skeleton 1332 Shape: Rectangular. Near vertical sides, flat base Dimensions: 1.76 x0.44 m, 0.10 mdeep Posture: Crouched,lying on left side with legs flexed Fill: Orange-brown gravelly clay back-fill (1338) Arm position: Left arm absent, right arm flexed Relationships: Relationship with grave 1374 could not be Age: Adult (>18 years) established Completeness: <25% Condition: Poor Skeleton 1337 Musculo-skeletal markers: enthesophytes on femora Grave goods: The base of awell-fragmented grey ware jar with a Posture: Supine, legs extended and parallel triangular-shaped rim in fabric TF11A was positioned at the base Arm position: Left hand extended with hand on pelvis, right of the grave near the feet. arm flexed across pelvis Other finds: Asingle sherd of 1st century pottery was recovered Sex: ?Female from the back-fill. Age: 26–35 years Date: 1st–early 2nd century AD (Radiocarbon date AD 50–230 Completeness: >75% (NZA 27004)) Condition: Very good Pathological conditions: Schmorl’snodes, osteochondritis dissecans, active periostitis in tibiae, erosive lesions in femoral Inhumationgrave 1344 neck, cribra orbitalia, hyperostosis frontalis interna, enamel hypoplasia, dental caries, dental calculus, periodontitis Grave cut 1339 Grave furniture: The remains of acoffin were represented by seven Orientation:SE-NW iron nails (sfs 206–12) located around the western side and ends of Shape: Sub-rectangular. Steep sides, flat base the grave. Dimensions: 1.96 x0.38 m, 0.23 mdeep Grave goods: Agroup of 17 hobnails(sf 204) was found associated Fill: Greyish-yellow clay gravel back-fill (1341) with the right foot and afurther two (sf 205) associated with the Relationships: Cuts grave 1334 and probably grave 1756 left foot, as well as 19 recovered from asoil sample.

89 Life andDeathinaRoman City

Other finds: Four sherds of 1st century pottery were recovered Skeleton 1364 from the back-fill. Posture: Supine, legs extended Arm position:Both arms flexed at right angles across stomach Inhumationgrave 1362 Sex: ?Female Age: 18–25 years Grave cut 1359 Completeness: 50–75% Condition: Very good Orientation:SW-NE Pathological conditions: dental caries, dental calculus Shape: Sub-rectangular. Moderate-steep sides, concave base Dimensions: 2.08 x0.83 m, 0.20 mdeep Grave furniture: The remains of acoffin were represented by nine Fill: Brownish-grey silty clay back-fill (1361) iron nails (sfs 220–6, 245, 246) located around the edges and northern end of the grave. Other finds: Three sherds of 1st century pottery were recovered Skeleton 1360 from the back-fill. Posture: Supine, legs extended and parallel, left tibia displaced Comments: It is difficult to establish the original position of the legs toward right leg as the tibia and fibula of both legs have become displaced after Arm position: Left arm extended beside body, right arm deposition. slightly flexed across pelvis Age: 18–25 years Completeness: 25–50% Inhumationgrave 1374 (Fig. 4.6) Condition: Poor Grave cut 1371 Pathological conditions: cribra orbitalia, hyperostosis frontalis interna, dental caries, dental calculus. Orientation:SSE-NNW Grave furniture: The remains of acoffin were represented by six Shape: Rectangular. Vertical sides, flat base iron nails (sfs 213–5, 217, 218 and 241) located in situ around the Dimensions: 1.72 x0.77 m, 0.37 mdeep edges of the grave. Fill: Brownish-yellow gravelly clay back-fill (1373) Grave goods: Relationships: Cut by graves 1369 and 1370

1. Acopper alloy bracelet (sf 216) worn on the left arm Skeleton 1372 2. Acopper alloy finger ring (?) sf 219 Posture: Supine, legs extended and parallel Other finds: Eleven sherds of 1st–2nd century pottery were reco- Arm position:Left arm flexed at aright angle across the vered from the back-fill. stomach, right arm extended beside body Date: 3rd–4th century AD Sex: Male Age: Adult (>18 years) Completeness: 50–75% Inhumationgrave 1369 Condition: Good Pathological conditions: possible ligament trauma on os coxae, Grave cut 1366 healed periostitis on metacarpals, femora, tibiae and metatar- Orientation:N-S sals Shape: Sub-rectangular. Steep sides, flat base Grave furniture: The remains of acoffin were represented by nine Dimensions: 1.92 x0.48 m, 0.21 mdeep iron nails (sfs 289–97) located around the east side and foot of the Fill: Reddish-brown gritty clay back-fill (1368) grave. Relationships: Cuts grave 1374 Grave goods: The substantial part of aflared rim jar in micaceous greyware fabric TF 5(1375, sf 239, Fig. 5.2, no. 12) was placed Skeleton 1367 beside the feet, possibly within the coffin. Other finds: Four iron nails were recovered from environmental Posture: Supine, legs extended and parallel samples. Arm position: Left arm extended beside body, right arm flexed Date: 3rd–4th century AD behind back Sex: Male Age: Adult (>18 years) Inhumationgrave 1388 Completeness: >75% Condition: Very good Grave cut 1385 Pathological conditions: osteoarthritis on acetabula, cortial Orientation:NE-SW defects on right humerus,dental caries, periodontitis Shape: Sub-rectangular with aroundedend Musculo-skeletal marker: enthesophytes on femora Dimensions: 0.8 x0.63 m, 0.27 mdeep Grave furniture: The remains of acoffin were represented by 14 Fill: Yellowish brown gravely clay back-fill (1387) iron nails (sfs 184, 227–37, 242, 243) located in situ around the edges of the grave and two nails recovered from environmental Skeleton 1386 samples. Grave goods: An assemblage of 104 hobnails was found in the area Posture: Supine of the feet. Arm position:Left arm flexed at aright angle across the Date: 3rd–4th century AD stomach, right arm absent Sex: Male Age: 18–25 years Inhumationgrave 1370 Completeness: <25% Condition: Poor Grave cut 1363 Pathological conditions: endocranial lesions, enamel hypoplasia Orientation:N-S Other finds: Two small sherds of 2nd century pottery were Shape: Rectangular. Near vertical sides, flat base recovered from the back-fill. Dimensions: 1.82 x0.54 m, 0.35 mdeep Date: 2nd centuryADorlater. Fill: Orange-brown gravely clay back-fill (1365) Comments: Amodern foundation trench had removed the right Relationships: Cuts grave 1374 side and lower part of the body.

90 Chapter Four

Inhumationgrave 1398 Comments: Severely truncated grave, only the feet and the lower part of the right leg surviving. Grave cut 1395

Orientation:S-N Inhumationgrave 1422 Shape: Sub-rectangular Dimensions: 1.70 x0.10 m, 0.10 mdeep Grave cut 1419 Fill:Brownish-orange siltyclayback-fill(1397) Relationships: Cut by pit 1408 Orientation:SW-NE Shape: Sub-rectangular. Vertical sides, flat base Dimensions: 0.70 x0.60 m, 0.07 mdeep Skeleton 1396 Fill: Orange-brown silty clay back-fill (1421) Relationships: Cuts gully 1358 Posture: Supine, legs extended and parallel Arm position: Arms slightly flexed, hands crossing over pelvis Skeleton 1420 Age: 13–17 years Posture: Supine Completeness: 25–50% Arm position: Left arm flexed across stomach. Right arm Condition: Poor extended beside body Pathological conditions: possible active periostitis on femora Completeness: <25% Grave furniture: The possible remains of acoffin were represented Condition: Poor by four iron nails (sfs 302–5) located in situ around the western Grave furniture: The possible remains of acoffin were represented edge of the grave. by five iron nails (sfs 310–3, 319) located in situ at the south- Comments: The grave was truncated by the foundation of a western end of the grave and afurther three recovered from modern wall defining the eastern boundary of the site, which has environmental samples. removed the head and the upper right part of the chest.

Inhumationgrave 1427 Inhumationgrave 1403 Grave cut 1424 Grave cut 1400 Orientation:E-W Orientation:NW-SE Shape: Sub-rectangular, tapering to west. Vertical sides, flat Shape: Sub-rectangular base Dimensions: 1.30 x0.45 m, 0.33 mdeep Dimensions: 2.63 x0.92 m, 0.19 mdeep Fill: Brownish-orange clay back-fill (1402) Fill: Orange-brown silty clay back-fill (1426) Relationships: Cut by pit 1408. Cuts grave 1407 Relationships: Relationship with pit 1301 uncertain

Skeleton 1401 Skeleton 1425 Posture: Supine, legs parallel Posture: Supine Age: Adult (>18 years) Completeness: <25% Completeness: <25% Condition: Fair Condition: Good Musculo-skeletal markers: enthesophytes on left femur Pathological conditions: active periostitis on right femur, Comments: Preservation was very poor and only the legs survive. healed periostitis on tibiae and fibulae Musculo-skeletal marker: enthesophytes on femora, tibia and left patella Inhumationgrave 1501 Finds: Two sherds of 2nd century pottery and an intrusive medi- Grave cut 1452 eval sherd were recovered from the back-fill. Date: 2nd centuryADorlater. Orientation:W-E Shape: Cut not seen, inferred from the presenceofskeleton 1453 Fill: Brown silty clay back-fill (1454) Inhumationgrave 1407 Relationships: Cuts grave 1764. Cut by ditch 1055 Grave cut 1404 Skeleton 1453 Orientation:N-S Shape: Grave cut not seen, but inferred from the presence of Posture: Supine, legs extended with feet together skeleton 1405 Arm position: Arms extended beside body Dimensions: Grave cut not seen, but inferred from the presence Sex: Male of skeleton 1405 Age: Adult (>18 years) Fill: Brownish-orange clay back-fill (1406) Completeness: 25–50% Relationships: Cut by grave 1403 Condition: Good Pathological conditions: osteoarthritis on left os coxa, osteo- Skeleton 1405 phytosis on os coxa, erosive lesions on left and right metatarsals and phalanges, possible trauma on left foot phalanx, Posture: Supine, the left foot crossed over the right healed periostitis on left tibia and left metatarsal, active and Age: Adult (>18 years) healed periostitis on left fibula Completeness: <25% Musculo-skeletal markers: enthesophytes on femora, tibiae and Condition: Fair patellae Pathological conditions: healed periostitis on right tibia and Finds: Atotal of 43 sherds of 3rd–4th century pottery and asingle fibula, possible healed trama/haematoma on right tibia iron nail were recovered from the back-fill. Musculo-skeletal markers: enthesophytes on left calcaneous Date: 3rd–4th century AD Finds: Asingle small sherd of 1st–2nd centurypottery was reco- Comments: The western half of the grave was truncated by ditch vered from the back-fill. 1055, removing the upper part of the body.

91 Life andDeathinaRoman City

Inhumationgrave 1503 Age: Adult (>18 years) Completeness: 25–50% Grave cut 1458 Condition: Fair Orientation:N-S Pathological conditions: osteoarthritisonunsided os coxa and Shape: Sub-rectangular. Steep sides, flat base femur, osteophytosis on unsided os coxa and femur Dimensions: 1.59 x0.50 m, 0.10 mdeep Musculo-skeletal markers: enthesophytes on left ulna Fill: Brown silty clay back-fill (1460) Grave goods: Relationships: Cuts graves 1505 and 1764. Cut by grave 1. 1489 (sf 323). Several sherds from aDOR BB1 jar decorated with 1504 an oblique burnished line lattice. Most of the rim is missing. Placed between the legs. The vessel contained apartial domestic Skeleton 1459 fowl skeleton (28 fragments including foot phalanges, acarpal a tarsometatarsus and indeterminate fragment) Posture: Supine, legs extended with feet together 2. Acopper alloy finger ring (sf 319) worn on the third or fourth Sex: Male finger of the left hand Age: 45+ years Completeness: <25% Other finds: Eight sherds of pottery dating from the 2nd century or Condition: Good later and three iron nails (sf 321–3) and ahobnail recovered from Pathological conditions: osteoarthritis on vertebra, osteophy- the back-fill. tosis on vertebrae, healed periostitis on right tibia, enamel Date: 3rd–4th century AD hypoplasia, dental caries, dental calculus Comments: The north end of the grave had been truncated by Musculo-skeletal markers: enthesophytes on patellae graves 1503 and 1504 and the south end by ditch 1729, removing the head and most of the torso and the left leg. Finds: Three sherds of 1st century pottery and three sherds of 3rd– 4th centurypottery were recovered from the back-fill. Date: 3rd–4th century AD Inhumationgrave 1507 Comments: The central part of this grave has been truncated by the digging of grave 1504. Grave cut 1495 Orientation:S-N Inhumationgrave 1504 Shape: Rectangular. Flat base Dimensions: 1.96 x0.56 m, 0.12 mdeep Grave cut 1466 Fill: Orange-brown silty clay back-fill (1497) Relationships: Cuts mass grave pit 1483 Orientation:W-E Shape: Rectangular. Flat base Skeleton 1496 Dimensions: >1.60 x0.60 m, 0.10 mdeep Fill: Brown silty clay back-fill (1468) Posture: Supine, legs extended and parallel Relationships: Cuts grave 1503 Arm position: Left arm flexed with hand over sternum, right arm extended beside body Skeleton 1467 Sex: ?Male Age: 26–35 years Posture: Supine, legs extended and parallel Completeness: 50–75% Arm position: Arms flexed with hands crossed over pelvis Condition: Very good Age: 45+ years Pathological conditions: Schmorl’s nodes, osteophytosis on Completeness: <25% right femur, healed periostitis on tibiae, enamel hypoplasia, Condition: Fair dental caries, dental calculus Pathological conditions: osteoarthritis on vertebra, verteb- Musculo-skeletal markers: enthesophytes on right femur ral osteophytosis, enamel hypoplasia, dental caries, dental calculus Grave furniture: The remains of acoffin were represented by five iron nails (sfs 327, 329–31 and 335) located in ahalo around the Grave furniture: The possible remains of acoffin were represented north end of the grave and asingle nail (sf 328) at the south end. by seven iron nails (sfs 312–8) located along both sides of the Other finds: Asingle iron nail was recovered from an environ- skeleton. mental sample. Other finds: Nine sherds of pottery dating from the 2nd century or Date: 3rd–4th century AD later, two iron nails and asingle hobnail were recovered from the back-fill. Date: 3rd–4th century AD Inhumationgrave 1508 Comments: Ditch 1055 truncates the western end of the grave, Grave cut 1498 removing the top of the skull of skeleton1467. Orientation:E-W Shape: Grave cut not seen, but inferred from the presence of Inhumationgrave 1505 (Fig. 4.7) skeleton 1499 Grave cut 1479=1486 Fill: Orange-brown silty clay back-fill (1500) Relationships: Cuts mass grave pit 1483 Orientation:SW-NE Shape: Sub-rectangular. Near vertical sides, flat base Skeleton 1499 Dimensions: 1.70 mx0.50 m. 0.12 mdeep Fill: Brown sandy clay back-fill (1481=1488) Posture: Supine, legs extended with feet together Relationships: Cuts mass grave pit 1483. Cut by grave 1503 Arm position: Left arm flexed with hand across pelvis, right arm tightly flexed with forearm over chest Skeleton 1480=1487 Sex: Male Age: Adult (>18 years) Posture: Supine Completeness: 50–75% Arm position: Left arm flexed with forearm over left side of Condition: Fair pelvis, right arm flexed at aright angle across stomach and the Pathological conditions: active periostitis on left metatarsal hand resting on the left forearm. Left hand displaced, located Musculo-skeletal markers: enthesophytes on right ulna, right beneath the right side of the pelvis patella and tibiae

92 Chapter Four

Grave goods: An assemblage of 15 hobnails (sf 332, 333), was Skeleton 1712 recovered but their location within the grave were not recorded. Other finds: Asingle sherd of Dorset black burnished ware was Posture: Supine recovered from the back-fill. Arm position:Left arm flexed with hand resting on pelvis, Date: 3rd–4th century AD right arm extended beside body with the forearm displaced Comments: The upper part of the skeleton has been removed by somewhat truncation. Age: Adult (>18 years) Completeness: <25% Condition: Poor Inhumationgrave 1720 Grave furniture: The remains of acoffin were represented by ten Grave cut 1705 iron nails (sfs 1018–25, 1028, 1029) located in situ along the sides of the grave. Orientation:W-E Comments: The west end of the grave has been truncated by the Shape: Sub-rectangular. Steep sides, flat base digging of grave 1724, removing the head of the skeleton. Dimensions: >1.76 x0.50 m. Depth not recorded Fill: Brownish-yellow sandy clay back-fill (1707) Relationships: Cut by ditch 1729 Inhumationgrave 1723

Skeleton 1706 Grave cut 1714 Posture: Supine, legs extended and parallel Orientation:NNE-SSW Arm position:Left arm flexed at aright angle across the Shape: Sub-rectangular stomach. Right arm absent Dimensions: >0.75 x0.60 m, 0.10 mdeep Sex: ?Male Fill: brownish-orange silty clay back-fill (1716) Age: 18–25 years Relationships: Cut by ditch 1729. Relationship with grave 1724 Completeness: 50–75% not established Condition: Poor Pathological conditions: enamel hypoplasia, dental calculus Skeleton 1715 Grave furniture: The remains of acoffin were represented by an assemblage of 18 iron nails (sfs 1000–17) located along the north Posture: Supine side and both ends of the grave. Arm position: Arms beside body, truncated at elbows Other finds: Atotal of 26 sherds of pottery were recovered from the Sex: ?Male back-fill, including afragment from an amphora of 2nd century Age: 18–25 years date. Completeness: 25–50% Date: 2nd centuryADorlater. Condition: Fair Comments: The south side of the grave has been truncated by Pathological conditions: enamel hypoplasia, dental calculus ditch 1729 and the east end by amodern feature, removing Finds: Two sherds of pottery and three iron nails were recovered much of the right side of the skeleton, the feet and the ends of from the back-fill. the legs. Comments: The grave is truncated by ditch 1729 to the south and modern disturbance 1700 to the west, leaving only the upper part of the body. Inhumationgrave 1721 Grave cut 1708 Inhumationgrave 1724 Orientation:E-W Shape: Rectangular. Near vertical sides, flat base Grave cut 1717 Dimensions: >1.77 x0.62 m, 0.14 mdeep Fill: Orange-brown sandy clay back-fill (1710) Orientation:SE-NW Relationships: Cut by grave 1722 Shape: Sub-rectangular. Steep sides, flat base Dimensions: >1.40 x1.38 m Fill: back-fill (1719) Skeleton 1709 Relationships: Cuts grave 1722, relationship with grave 1723 Posture: Supine, legs extended and parallel not established Arm position: Arms extended beside body Age: Adult (>18 years) Skeleton 1718 Completeness: <25% Condition: Good Posture: Supine Pathological conditions: active periostitis on tibiae Arm position: Left arm extended beside body, right arm Other finds: Three iron nails (sfs 1026, 1027, 1030) were recovered extended beyond the edge of the trench at mid-humerus from this grave. Sex: Male Comments: The grave has been severely affected by truncation. The Age: 45+ years feet have been removed by grave 1724, the central part of the body Completeness: <25% from the knees to the abdomen by grave 1722 and the head during Condition: Poor machine stripping of the site. Pathological conditions: enamel hypoplasia, dental caries, dental calculus Grave furniture: The remains of acoffin were represented by 10 Inhumationgrave 1722 iron nails (sfs 1035–44)located in ahalo around the surviving Grave cut 1711 south east end of the grave. Other finds: Two iron nails and asingle hobnail were recovered Orientation:W-E from the back-fill. Shape: Rectangular. Near vertical sides, flat base Comments: The north-west part of the grave, containingthe lower Dimensions: >1.41 x0.60 m, 0.08 mdeep part of the body, lay beyond the edge of the trench and so this part Fill: Orange-brown sandy clay back-fill (1713) of the skeletonwas not recovered. The end of the left arm had Relationships: Cut by grave 1724. Cuts grave 1721 been removed by modern truncation 1700.

93 Life andDeathinaRoman City

Inhumationgrave 1755 Inhumationgrave 1758 Grave cut 1065 Grave cut 1117 Orientation:N-S Orientation:E-W Shape: Oval. Steep sides, flat base Shape: Sub-rectangular Dimensions: 0.90 x0.32 m, 0.23 mdeep Dimensions: >1.10 x0.73 m. Depth not recorded Fill: Orange-brown silty clay back-fill (1067) Fill: Reddish-brown silty clay back-fill (1119) Relationships: Cuts soil layer 1020=1106.Cut by pit 1003 Relationships: Cut by ditch 1055

Skeleton 1066 Skeleton 1118 Posture: Crouched,lying on left side Posture: Supine Arm position: Both arms flexed in front of face Arm position: Left arm absent, right arm extended beside Age: 2–5 years body Completeness: 25–50% Sex: Male Condition: Very good Age: Adult (>18 years) Pathological conditions: enamel hypoplasia Completeness: <25% Condition: Very good Comments: The grave had been truncated by pit 1003, causing Pathological conditions: spondylolysis, possible healed trauma some damagetothe skull. on left metacarpals Musculo-skeletal markers: enthesophytes on right ulna Inhumationgrave 1756 Finds: Two iron nails (sfs 21 and 22) were recovered from the back- fill. Grave cut 1071 Comments: The western half of the grave had been truncated by ditch 1055, removing the lower part of the body, and the left arm Orientation:NE-SW and head had been removedbytruncation. Shape: Rectangular, tapering from NE to SW. Vertical sides, flat base Dimensions: 0.65 x0.40 m, 0.25 mdeep Inhumationgrave 1759 Fill: Yellowish-brown sandy silt back-fill (1078) Relationships: Cut soil layer 1020=1106. ?Cut by grave 1344 Grave cut 1170 Orientation:N-S Skeleton 1073 Shape: Sub-rectangular. Near vertical sides, flat base Dimensions: 1.89 x0.84 m, 0.27 mdeep Posture: Insufficient remains were present to determine the Fill: Brownish-grey silty clay back-fill (1171) posture Relationships: Cut by graves 1143 and 1151. Relationship with Age: 0years (perinate) grave 1167 not established Completeness: <25% Condition: Poor Grave goods: Apartial domestic fowl skeleton was recovered. Other finds: Asingle small sherd of 1st century pottery was Grave furniture: Cist formed from five pieces of limestone recovered from the back-fill. measuring between 0.25 x0.17 x0.05 mand 0.42 x0.21 x0.05 Comments: No skeletal remains were recovered from this grave. m. Five of these stones were set on edge to form atrapezoidal cist measuring c 0.5 x0.3 mand 0.2 mhigh, and the sixth placed flat to form abase for the south-westend of the grave. Inhumationgrave 1760 Other finds: sf 12 (Iron object) recovered from the back-fill between the cist and the grave cut (1080). Grave cut 1382 Orientation:W-E Shape: Sub-rectangular Inhumationgrave 1757 Dimensions: >0.60 x0.45 m, 0.15 mdeep Grave cut 1088 Fill: Yellowish-brown sandy clay back-fill (1380) Orientation:NW-SE Skeleton 1381 Shape: Sub-rectangular. Sloping ides, concavebase Dimensions: 0.87 x0.53 m, 0.13 mdeep Posture: Supine Fill: Orange-brown silty clay back-fill (1090) Age: 26–35 years Completeness: <25% Skeleton 1089 Condition: Very good Pathological conditions: degenerative disc disease, remodelled Posture: Skeleton 1089 was atightly packed cluster of new bone on zygomatic bones, cribra orbitalia, enamel disarticulated bones hypoplasia, dental caries, dental calculus Sex: Male Comments: The eastern part of the grave was truncated by a Age: 45+ years modern foundation, leaving only the upper part of the body. Completeness: 50–75% Condition: Good Pathological conditions: osteoarthritis on vertebrae and on os Inhumationgrave 1761 coxae, osteophytosis on left femur, right humerus and left auricular surface, degenerative disc disease, healed trauma on left Grave cut 1391 rib, on right fibula and on left metacarpal, healed periostitis on Orientation:NW-SE right fibula Shape: Sub-rectangular with roundedends Musculo-skeletal markers: enthesophytes on right ulna, femora Dimensions: 1.80 x0.60 m, 0.15 mdeep and tibiae Fill: Yellowish-brown silty clay back-fill (1389) Other finds: Animal bone (sf 14) recovered from the back-fill. Relationships: Cut by grave 1762

94 Chapter Four

Skeleton 1390 Condition: Fair Pathological conditions: healed lytic lesions on right talus, Posture: Supine, legs extended and parallel possible osteochondritis dissecans on foot phalanges, healed Arm position: Both arms extendedbeside body periostitis on right fibula Sex: ?Female Age: 26–35 years Grave goods: Ahobnail shoe comprising 42 nails (sf 334) was worn Completeness: >75% on the surviving foot. Condition: Very good Date: 3rd–4th century AD Pathological conditions: rib osteoarthritis, rib osteophytosis, Comments: The grave has been severely truncated by subsequent pseudo-jointsurface on left os coxa, healed periostitis on tibiae, burials, leaving only the lower part of the right leg. spina biffida occulta on sacrum, enamel hypoplasia, dental caries, dental calculus, periodontitis Inhumationgrave 1765 Comments: The south-westcorner of the grave was truncated by grave 1762, removingthe right foot. Grave cut 1535 Orientation:E-W Shape: Oval. Vertical sides, flat base Inhumationgrave 1762 Dimensions: 1.88 x0.60 m, 0.30 mdeep Grave cut 1394 Fill: Mottled brown and orange silty sand back-fill (1536) Relationships: Cut by grave 1764 Orientation:W-E Shape: Sub-rectangular Skeleton 1537 Dimensions: >1.30 x0.60 m, 0.12 mdeep Fill: Yellowish-brown silty clay back-fill (1392) Posture: Prone, legs parallel Relationships: Cuts grave 1761 Arm position: Left arm extended beside body, right arm flexed behind back with the hand resting on the back of the Skeleton 1393 pelvis Sex: Female Posture: Supine, legs extended and parallel Age: Adult (>18 years) Arm position: Left arm flexed across stomach, right arm Completeness: >75% extended beside body Condition: Good Sex: ?Male Pathological conditions: osteoarthritis on vertebrae,vertebral Age: Adult (>18 years) osteophytosis, degenerativedisc disease, cribra orbitalia, neo- Completeness: 50–75% plastic disease (button osteoma), enamel hypoplasia, dental Condition: Very good calculus, periodontitis Pathological conditions: rib osteoarthritis, vertebral osteophy- Musculo-skeletal markers: enthesophytes on left patella tosis, Schmorl’s nodes, possible haematoma on left femur, Grave furniture: The probable remains of acoffin were represented healed periostitis on tibiae, congenital conditions(lumbarisa- by 12 iron nails (sfs 354–63, 368 and 369). tion). Grave goods: Agroup of 42 hobnails (sf 367) was located in the area Musculo-skeletal markers: enthesophytes on tibiae and left of the left foot. femur Date: 3rd–4th century AD Comments: The west end of the grave had been truncated by a modern feature, removing the head. Inhumationgrave 1772

Inhumationgrave 1763 Grave cut 1077 Grave cut 1410 Orientation:SE-NW Shape: Oval Orientation:NE-SW Dimensions: 0.83 x0.56 m, 0.23 mdeep Shape: Sub-rectangular with roundedends. Fill: Orange-brown silty clay back-fill (1087) Dimensions: 0.65 x0.35 m. Depth not recorded. Relationships: Cuts soil layer 1020=1106 Fill: Brownishorange clay back-fill (1411). Skeleton: No skeletal material was present. Skeleton 1086 Finds: Fourteen sherds of 1st century pottery were recovered from Posture: Crouched, lying on left side with legs drawn up tight the back-fill. to the chest Arm position: Arms wrapped around legs Inhumationgrave 1764 Age: 5–12 years Completeness: 50–75% Grave cut 1509 Condition: Very good Pathological conditions: active rib periostitis, cribra orbitalia, Orientation:SSW-NNE enamel hypoplasia. Shape: Grave cut not seen, but inferred from the presence of skeleton Comments: The upper part of the grave was filled by adeposit of Fill: Orange-brown silty clay back-fill (1511) limestone rubble that was visible in the surface of soil layer Relationships: Cuts grave 1765. Cut by graves 1501, 1503 and 1020=1106 and may have acted as agrave marker. 1508 Inhumationgrave 2027 Skeleton 1510 Grave cut 2000 Posture: Posture not established as only the lower part of the right leg survived Orientation:SW-NE Age: Adult (>18 years) Shape: Grave cut not seen, but inferred from the presence of Completeness: <25% skeleton 2001

95 Life andDeathinaRoman City

Fill: Yellowish-brown silty clay back-fill (2002) large mammal long bone and 14 fragments of unburnt large Relationships: Cut by ditch 1725 mammal long bone. Date: 1st–early 2nd century AD Skeleton 2001 Cremation burial 1209 (Fig. 4.8) Posture: Supine Age: Subadult (<18 years) Grave cut (1197) Completeness: <25% Condition: Fair Shape: Sub-rectangular, oriented NE-SW Dimensions: 0.55 x0.40 m, 0.17 mdeep Comments: The grave has been severely truncated by modern Fill: Orange-brown clay back-fill (1198) disturbance, which has left only the right femur, right side of the pelvis and part of the left femur. Cinerary urn (1200, sf 71):The lower part only of agreyware jar in fabric TF 25. The cremation urn also contained one small sherd of In addition to the skeletal remainsrecoveredfrom grey sandy ware (TF 39). graves, three partialskeletons were recorded which Cremation deposit (1202) Weight: 281.5 g had been disturbed and for which no grave could be Largest fragment: 59 mm identified. These are listed below. Identified bones and teeth: Skull, vertebrae, ribs, pelvis, long bones Colour: White Skeleton 1057 MNI: 1 Sex: ?Male Age: Adult (> 18 years) Age: 36–45 years Sex: Male? Completeness: <25% Pyre goods: Condition: Fair 1. The lower neck and reservoir of atubular unguent bottle in Pathological conditions: enamel hypoplasia, dental caries, blue/green glass, melted and collapsed dental calculus, periodontitis 2. Amelted drop of blue/greenglass 3. Abase fragment from atubular unguent bottle in blue/green Skeleton 1116 glass (sf 74) 4. Two fragments of blue/green glass Sex: ?Female 5. Two fragments from alarge mammal skull fragment Age: 45+ years (possibly human) and apig phalanx from ajuvenile animal Completeness: <25% were recovered from the cremation deposit, which may Condition: Very good represent the remains of pyre goods Pathological conditions: None Grave goods: Skeleton 1294 1. An oxidised flagon in fabric TF 11A with no survivingrim (1199, sf 70). An indeterminatemedium mammal sized Completeness: <25% fragment,asingle hobnail and afragment of nail shank were Condition: Fair inside the vessel 2. Body fragments of glass unguent bottle

CATALOGUE OF CREMATIONS Other finds:The backfill produced 27 sherds of 1st century pottery including ahandle from an oxidised flagon and two rim sherds Cremation burial 1196 (Fig. 4.8) from aSevern Valley ware lid, six iron nails, four hobnails and some burnt animal bone fragments. Grave cut (1188) Date:1st–early 2nd century AD Shape: Sub-rectangular, oriented N-S Dimensions: 0.70 x0.50 m, 0.08 mdeep Cremation burial 1227 Fill: Light yellowish-red silty clay back-fill (1190) Grave cut (1222) Cinerary urn (1191, sf 61):Agrey, necked jar in the local Gloucester kiln ware TF 11A. The vessel is fragmented and most of the rim is Shape: Sub-rectangular, oriented N-S missing. Fig. 5.1, no. 5. Dimensions: 0.65 x0.44 m, 0.05 mdeep Cremation deposit (1189) Fill: Mid yellowish-brown silty clay back-fill (1225) Weight: 269.5 g Cinerary urn (1223, sf 79):The lower part of an oxidised flagon in Largest fragment: 43 mm the local Gloucester kiln ware TF 11A. Identified bones and teeth: Skull, vertebrae, ribs, pelvis, Cremation deposit (1224) shoulder girdle, long bones Weight: 1.5 g Colour: White Largest fragment: <10 mm MNI: 1 Identified bones and teeth: Skull Age: Adult (>18 years) Colour: White Pyre goods:Five fragmentsofcalcined probable pig bone including MNI: 1 cranial and long bone fragments were recovered from the cremation deposit. Finds: The cremation urn fill produced four sherds of TF 11A with Grave goods: afurther small sherd from the cremation backfill. Date: 1st–early 2nd century AD 1. An oxidised flagon in fabric TF 11A (1192, sf 62), well fragmented and missing the rim Cremation burial 1266 (Fig. 4.8) 2. An oxidised flagon in fabric TF 11A (1193, sf 63), well fragmented and missing the rim Grave cut (1253) Other finds: The cremation backfill contained 24 sherds of 1st Shape: Sub-circular century pottery, one small intrusive sherd of medieval date and a Dimensions: 0.60 in diameter, 0.08 mdeep single iron nail, as well as eleven fragmentsofburnt medium and Fill: Brownish-orange silty clay back-fill (1258)

96 Chapter Four

Cinerary urn (1251, sf 97): Areduced everted rim jar in fabric TF Cremation deposit (1062) 11A (Fig. 5.1, no. 6) with asquared rim.. Most of vessel is present Weight: 1255.5 including the entire rim, but broken. Two small nails were inside Largest fragment: 58 mm the vessel. Identified bones and teeth: Skull, vertebrae, ribs, pelvis, Cremation deposit (1252) shoulder girdle, long bones Weight: 143 g Colour: White Largest fragment: 38 mm MNI: 3 Identified bones and teeth: Skull, part of the mandible, verte- Age: Adult (>18 years) brae, ribs, shoulder girdle, long bones Sex: Mandibularfragmentsfrom two individualshad male Colour: White traits MNI: 2 Pyre goods: Atotal of ten fragments(48.6 g) of cremated animal Age: 5–10 years and 10–15 years bone were recovered from the cremation deposit, comprising the Grave goods: remains of apossible cattle rib and medium mammal-sized long 1. Acomplete ring-necked flagon in oxidised TF 11A (1256, sf96). bones and cranial fragments, as well as an unburnt fish rib/ The rim circumference is complete apart from an old nick in one process and an unburnt indeterminate fragment. part. Fig. 5.1, no. 7 Grave goods: Collared rim (Hofheim) flagon in oxidised sandy ware 2. Agamingset comprising 43 bone and glass counters and two TF 25, almost complete althoughpart of the base is missing (1063, dice together with afrit melon bead were found in the Fig. 5.1, no. 3). The fill of the vessel contained amedium mammal cremation urn long bone fragment and three indeterminatebone fragments weighing atotal of 8g. Other finds: Amedium mammal long bone fragment, one medium Date: Pre-Flavian mammal or bird long bone fragment and five indeterminate fragments were recovered from the backfill. Date:1st–early 2nd century AD Cremation burial 1768 Grave cut (1094) Cremation burial 1766 (Plate 2.6) Shape: Circular Grave cut (1048) Dimensions: 0.25 mindiameter,0.03 mdeep Fill: Cremation deposit 1095, ayellowish-brownsilty clay with Shape: Circular pieces of calcined bone Dimensions: At least 0.44 mindiameter,0.13 mdeep Cinerary urn: Un-urned Fill: Orange-brown sandy clay back-fill (1072) Cremation deposit (1095) Relationships: Cuts pit 1074 Weight: 17.5 g Cinerary urn (1068): Grey ware jar in fabric TF 25, asubstantial part Largest fragment: 28 mm of which survives, with an everted thickened rim. Fig. 5.1, no. 1. Identified bones and teeth: Skull, ribs, long bones Cremation deposit (1049) Colour: White Weight: 132.5 g MNI: 1 Largest fragment: 47 mm Age: Adult (>18 years) Identified bones and teeth: Skull, vertebrae, ribs, pelvis, shoul- Pyre goods: Atotal of five fragments(1.8 g) of burnt medium der girdle, long bones mammal sized long bone and indeterminate bone fragments were Colour: White recovered from the cremation deposit. MNI: 1 Age: Adult (>18 years) Non-metric traits: The only trait available for observation was Cremation burial 1769 the multi-foramina in the zygomatic which was present. Grave cut (1186) Pyre goods: 1. Melted and vesicular fragment of blue/green glass, possibly the Shape: Rectangular, oriented E-W collapsed reservoir of atubular unguent bottle Dimensions: 0.43 x0.28 m, 0.05 mdeep 2. Aburnt sherd of everted rim jar in oxidised TF 11A was Fill: Cremation deposit 1187, aloose black charcoal-rich clay recovered from the fill of the cineraryurn with calcined bone Cinerary urn: Un-urned 3. Animal bone was recovered from the cremation deposit Cremation deposit (1187) representingatleast two pigs. The pigs were represented by Weight: 261 g cranial and tooth fragments, aleft tibia, aleft astragalus,a Largest fragment: 44 mm calcaneum,aleft humerus,and two left ulnae. Further Identified bones and teeth: Skull, vertebrae, ribs, pelvis, long medium or large mammal sized scapula, long bone, tooth and bones indeterminatefragmentsmay also be pig. Colour: White Other finds: The fill of the cinerary urn contained four sherds of 1st MNI: 1 century pottery, two very small fragments of fired clay and four Age: Adult (>18 years) iron nails. Pyre goods: The cremation deposit contained 325 fragments(119 g) Date: Pre-Flavian of animal bone comprising pig skull, right radius and left tibia Comments: The northern half of the feature was truncated by a fragments, and further medium mammal-sized cranial, vertebral, modern pit. rib, long bone and indeterminatefragmentswhich may also be pig. Aburnt possible bird long bone fragment was also recovered. Cremation burial 1767 (Fig. 4.8) Grave goods:Anassemblageof27hobnails was scattered throughout the cremation deposit. Grave cut (1059) Other finds: Eight iron nails were recovered. Shape: Square Date: AD 60–220 (OxA-16811)(radiocarbon assay) Dimensions: 0.48 x0.45 m, 0.15 mdeep Fill: Back-fill(1377) Cremation burial 1770 Cinerary urn (1061): Necked jar with athickened rim in grey sandy Grave cut (1323) ware TF 25. Most of the vessel is present. Traces of athin white calcareous deposit are visible on the interior of the basesherds. Shape: Sub-circular Fig. 5.1, no. 2. Dimensions: 0.22 mindiameter,0.07 mdeep

97 Life andDeathinaRoman City

Fill: Loose yellowish-brownsilt clay with some charcoal and Identiﬁed bones and teeth: Skull, ribs, long bones calcined bone (1324) Colour: White Cinerary urn: Un-urned MNI: 1 Cremation deposit (1324) Age: Adult (>18 years) Weight: 8.5 g Date: 50 BC –AD70(OxA-16792)(radiocarbon assay) Largest fragment: 39 mm

98 Chapter 5The Finds and Environmental Evidence

THE POTTERY appear from the Neronian period through to the by Jane Timby early 2nd century. Small amounts of Dorsetblack burnishedware (DOR BB1) were reaching Kingholm Introduction in the 1st century but the bulkofsuch material The excavation resulted in the recovery of 2156 comingtoGloucester dates from the Hadrianic sherds of pottery weighing 18.1 kg. Most of the period through to the late Romanperiod. This pottery dates to the Romanperiod but also present together with the south-west variant (SOW BB1) are 61 sherds of medieval and 80 sherds of post- accountfor 21.6%ofthe London Road assemblage medieval date. Thepost-Roman material is not and largelycomes from the later Romanfeatures. considered in detail otherthan where it has provided SevernValley ware is quite poorlyrepresented at datingevidence. 6.2%. This is surprising as once the Gloucester kilns The Romanpottery rangedindate from the had ceased production Severn Valleywares tended Claudio-Neronian period through to the late 3rd– to flood the market, but can probablybeexplained 4th century. Approximately 77% of the assemblage by by the chronology of the site and ageneral absence sherd count came from burial contexts, the remainder of mid-later2nd and 3rd century wares. All other deriving frompits and post-Roman features. The wares are presentinvery small amounts. The only condition of the material is thus quite variable. Nearly otherfabric, occurringasacomplete (broken)vessel all the vessels deliberatelydepositedasgrave goods in aflared rim jar in TF 5, is amicaceous grey ware, or as containers forthe crematedremainswerein which may comefrom south Gloucestershire. semi-complete state and only one vessel, asmall flagon (1256, sf 96, Fig. 5.1, no.7), had survived intact. Pottery from burials Pottery vessels were recovered from six cremation Fabrics and forms burials(1196, 1209, 1227, 1266, 1766, 1767). None of The pottery assemblage as awhole is fairly typical for the vessels survived intact and in many cases only a Gloucester forthe periods covered in terms of the few sherds were present so it was unclear whether fabricsand forms present, although slightlyatypical these represented fragmented and dispersed burial in the balance of wares present. Locallymade urns or accessory vessels, were deliberately depos- Gloucester kiln wares,notablyTF11A, afine textured ited isolated sherds connected with the burial ritual, ware madeinbothoxidised and reduced versions or were simply stray finds. At leastsix inhumation (Timby 1991), dominatethe early Romanassemblage gravesmay have also had deliberately deposited (Table 5.1). Withthe sandier counterpart TF 25, the vessels (grave groups 1145, 1243, 1334, 1352, 1374 local Gloucester kiln wares accountfor 55.4%by and 1505), whileaninverted pot was placed in a count of the total recovered assemblage, with 52.9% possible ‘memorial’ pit (1149; see Chapter 6). In being the finer TF 11 variant. This is slightly unusual addition, potsherds were recovered from the fills of away from the kiln sites and demonstrates the 31 inhumation burials. preference shown for these vessels as burial urns and accessory vessels. Also local in source and Pottery from cremation burials potentially slightly earlier are fabrics 24, 36, 39 and 213 which are all associated with the legionary Group1196.Cremation urn: grey, necked jar in the fortressatKingsholmand are military wares, again local Gloucester kiln ware TF 11A (1191, sf 61, Fig. produced locally in the Neronian period (Hurst1985; 5.1, no. 5). The vessel is fragmented and most of the Timby 1999). Other contemporarywares include rim is missing. Accompanying the urn were two very small quantities of imported fineware (South oxidised flagonsinfabricTF11A (1192, sf 62 and Gaulishsamian and Lyons ware) and asmall amount 1193 sf 63), bothwell fragmented and missing the of amphora. Afine white ware flask from the fill of rims. Thefill of the cinerary urn also produced four gully 1358 couldalso possibly be an import. Four sherds of oxidised TF 11A. The cremation backfill sherds of grog-tempered handmade native ware produced onesmall sherd of early Seven Valley (TF 2A) arealso typical for 1st centuryGloucester, ware, 24 very fragmented sherds of oxidised TF 11A and it is possible that the nearby river crossing was and one small intrusive sherd of medieval date. the site of anative settlement prior to the establish- Date: Flavian-Trajanic. ment of the Kingsholm fortress(cf Timby 1990). Group 1209.Cremation urn:The lower partonly of Other local wares at this time include ablack agreyware jar TF 25 (1200, sf 71). The urn was sandyburnishedware (TF 201) which tends to accompanied by an oxidised flagon, TF 11A with no

99 Life andDeathinaRoman City

Figure 5.1 Roman pottery (1–8). surviving rim (1199, sf 70). The cremationurn complete ring-necked flagon in oxidised TF 11A, backfill produced onesmall sherd of grey sandy (1256, sf 96,Fig. 5.1, no.7). The rim circumference is ware TF 39. The cremation backfill produced further complete apart from an old nickinone part. Date: sherds of oxidised flagon,including ahandle in TF Flavian-Trajanic. 11A, one sherd of grey TF 11A, five sherds of TF 39, Group 1766.Cremation urn: Substantial part of a one of TF 25 and six sherds fromanearly Severn grey ware jar with an everted thickened rim,TF25 Valley ware jar as well as two rimsherds from a (1068, Fig. 5.1, no. 1). Asherd of evertedrim jar in SevernValley ware lid. Date: Flavian-Trajanic. oxidised TF 11A was mixed in with these sherds. The Group 1227. Cremation urn: the lower partofan cremation backfill produced four oxidised sandy oxidised flagon in the local Gloucester kilnware TF sherds TF 25. Date: Neronian. 11A (1223, sf 79). The cremation urn fill produced Group 1767.Cremation urn: Necked jar with a four further sherds of TF 11A with afurthersmall thickened rim in grey sandyware, TF 25 (1061, Fig. sherd from the cremationbackfill. Date: Flavian- 5.1, no. 2). Most of the vessel is present. Traces of a Trajanic. thin white calcareous deposit are visible on the Group 1266.Cremation urn: everted rim jar with a interior of the base sherds. Theancillary vessel is a squaredrim in reducedfabricTF11A (1251, sf 97, collaredrim (Hofheim) flagon in oxidised sandy Fig. 5.1, no. 6). Most of vessel is presentincluding the ware TF 25, almost complete although partofthe entirerim, but broken. Accompanying the urn is a base is missing (1063, Fig. 5.1, no. 3). Date: Neronian.

100 Chapter Five

Table 5.1 Quantification of the pottery assemblage by fabric.

Roman Fabric Description No %Wt(g) %Eve %

Imports SGSAM South Gaulish samian 40.2 10 0.1 30.2 CGSAM Central Gauilsh samian 40.2 18 0.1 00.0 LYO CC Lyons ware 10.0 90.1 00.0 GAL AM Gallic amphora 10.0 68 0.4 00.0 10 miscellaneous amphorae 50.2 263 1.6 00.0 Regional DOR BB1 Dorset black burnished ware 257 12.8 1592 9.6 148 8.9 SOW BB1 South-west black burnished ware 178 8.8 1333 8.0 101 6.1 OXF RS Oxon red slipped ware 90.4 43 0.3 00.0 5grey micaceousware 43 2.1 654 3.9 90 5.4 19 wheelmade Malvernian ware 20.1 30 0.2 00.0 Native ware 2A hm grog-temperedware 40.2 36 0.2 13 0.8 Glos kilns 7white slipped oxidised ware 20.1 10 0.1 00.0 11A (ox) fine oxidised ware 897 44.5 5835 35.2 576 34.8 11A (gy) fine grey ware 170 8.4 1126 6.8 190 11.5 25 (ox) sandy oxidised ware 42 2.1 884 5.3 213 12.9 25 (gy) sandy grey ware 79 3.9 1528 9.2 133 8.0 24 Kingsholm flagon fabric 53 2.6 340 2.1 00.0 36 Kingholm oxidised ware 40.2 15 0.1 00.0 39 Kingsholm grey sandy 80 4.0 1118 6.7 29 1.8 213 Kingsholm grey ware 30.1 14 0.1 00.0 Local SVW OX Severn Valley ware 114 5.7 1305 7.9 63 3.8 11D early Severn Valley ware 13 0.6 153 0.9 11 0.7 11E limestone-tempered SVW 10.0 20.0 00.0 23 hm Severn Valley ware 10.0 27 0.2 00.0 201 wm black burnished sandy 29 1.4 70 0.4 30.2 12R local colour-coated ware 30.1 90.1 12 0.7 Unknown WW miscellaneous whiteware 90.4 27 0.2 50 3.0 WWF fine whiteware10.0 10 0.1 00.0 GREY unattributedgrey sandy wares 60.3 54 0.3 19 1.1 Roman total 2015 100.0 16583 100.0 1654 100.0 Med 61 0.0 793 0.0 00.0 Pmed 80 0.0 710 0.0 00.0 Total 2156 0.0 18086 0.0 1654 0.0

Pottery from inhumations and ‘memorial’ Grave group 1374.The substantial part of aflared rim jar (1375) in micaceous greyware TF 5came from Grave group 1145. The proneinhumation burial was this burial.(1375, sf 239, Fig. 5.2, no. 12). Date: 3rd or accompanied by several sherds from aring-necked 4th century. flagon in oxidised TF 11A (1128, Fig. 5.2, no. 10). The Grave group 1505.Several sherds from aDOR BB1 rim has an old chipout of the circumference. Date: jar (1489, sf 323). Most of the rim is missing. Flavian-Trajanic. Decorated with an obliqueburnished line lattice. ‘Memorial’ 1149.Ring-necked flagon represented by Date: 3rd or 4th century. the upper part of the vessel in oxidised TF 11A (1141, sf 80, Fig. 5.1, no. 4). Date: Neronian-earlyFlavian. Grave group 1243. The upper partofaring-necked Pottery from gravebackfills flagon in oxidised TF 25 (1322, sf 155, Fig. 5.2, no. Pottery was recoveredfrom the backfillsof31ofthe 11). Date: Flavian-Trajanic. inhumation graves. In most cases thisappears to be Grave group1334.Awell-fragmented grey ware jar isolated sherds rather than deliberate grave goods. in fabricTF11A with atriangular-shaped rim.Only The quantity rangedfrom singlesherds to amaxi- the base survived in situ,positioned at the base of the mum of 31 in grave1501. Asherd join was observed grave nearthe feet. Date: Flavian-Trajanic. between grave groups 1167 and 1230, which lie Grave group1352.Ring-necked flagon in oxidised adjacent to oneanother but do notintercut. TF 11A (1348, sf 174, Fig. 5.1, no. 8). Accompanying the flagon is the upperportion of asecond flagon in Pottery from the mass grave the same fabric (1351, sf 175, Fig. 5.2, no. 9). The back- fill produced three sherds in the same fabricprobably The mass grave produced alarge assemblage of from the samevessels.Date: Flavian-Trajanic. pottery amounting to some 267 sherds amongst

101 Life andDeathinaRoman City

Figure 5.2 Roman pottery (9–15). which were the substantial parts of at least six Medieval soillayer 1025/1105 produced 99 sherds vessels (sf 64, sf 531–2, sf 619–20),all South-West (1131 g) including Malvernian ware cooking pot black burnished ware jars (SOW BB1) (Fig. 5.2, (GlosTF40), Gloucester limestone-tempered ware no.13). Some of SOW BB1 jar sherds have right- (TF 41b) and Minety ware (TF44), collectively angled to slightly oblique burnished lattice. Many suggesting adate in the 13th century. Residual sherds are heavily sootedand most of the rims are Romanmaterial from thislayer spans the 1st to 4th missing. The latticing would suggest alater 2nd–3rd centuries. centurydate. Other material from this deposit was confined to sherds of TF 11Aand aDOR BB1 plain- Discussion rimmed dish. Pottery from the upper part of the backfill comprisedamixture of earlier and later Chronologically the assemblage falls into two main Romanpottery, in particular TF 11A, SVWOX, DOR periods,the 1st–early 2nd century and the 3rd–4th BB1 including aflanged conical bowl, SOW BB1, century, with something of ahiatus fromaroundthe OXF RS,the latter mainly from aflask, and single end of the Trajanic period through to the 3rd sherds of samian and grey ware.The terminus post century. Theearliest burialsare dated to the pre- quem forthis assemblage lies in the later 3rd–4th cen- Flavian period on the basis of the fabrics present and tury, but its mixed nature suggests that it has been the typology of the vessels.These includecremation subject to some degree of contamination, probably groups1766 and 1767, ‘memorial’ 1149 and pit 1306. associated with the later burialscut into the backfill. FabricsTF24, 36, 213 and 39, represented in these features, all have direct linkswith the early military occupation at the Kingsholm fortress, occupied from Pottery from buried soils c AD 49 through to AD 66/7 (Hurst 1985, 122; see Substantial amounts of Romanpottery were recov- Chapter1), and arelikely to have been produced by ered from the various buriedsoils defined on the the army, probably locally. Theproducts of these site. The soillayer associated with the use of the military kilns includeflagons, jars, bowls, platters, cemetery (1106/1020/1440) produced some 180 mortariaand open lamps (ibid., 78 ff) but the ex- sherds (1310 g) of 1st–2nd century pottery including amples found at 120–122 London Road are restricted asignificant number of DOR BB1 sherds. The frag- to flagons, jars and one beaker. mentation rate is quite high with an overall aver- Pottery recoveredfrom the cremation burials age sherd weightof7gsuggesting well-disturbed entirely comprised products from local kilns, with material. the fine oxidised and reducedgrey ware fabric TF

102 Chapter Five

11A and its sandier counterpart TF 25 most colour had some relevance. Biddulph (2005, 37) has prevalent. These wares have been identified from considered the possibility of vessels being provided at least two kiln sites in Gloucester: the College of throughburial societies through which the costs Art site (Rawes1972) and Berkeley Street(Timby and equipment of burial would be provided on a 1991). Both these siteswere producing wares from member’s death,assuming regular payment of dues. the Flavian-Trajanic periods but an earlier phase of kilns at Berkeley Street would suggest apotentially Catalogue of illustrated sherds(Figs 5.1 and 5.2) earlier phase of production (ibid.). Athird unpublished kiln from the Kingsholm rugby groundwas 1. Cremationurn 1068: jar with an everted thickened rim in grey ware TF 39. Cremation burial 1766. also producing fine grey wares identical to TF 11A in 2. Cremationurn 1061: necked jar with athickened rim in grey the Flavian-Trajanic period. sandy ware TF 25. Traces of athin white calcareous deposit There appears to be apattern of using grey ware are visible on the interior of the base sherds. Cremationburial jars as cremation urns,accompanied by one vessel, 1767. or in the case of burial 1196 two vessels, of a 3. Ancillary vessel 1063: collared rim (Hofheim) flagon in oxidised sandy ware TF 25. Cremationburial 1767. contrasting colour, possibly hinting at some form of 4. Ancillary vessel 1141, sf 80: ring-necked flagon in oxidised TF symbolism represented in the choice of form and 11A. ‘Memorial’ pit 1149. colour.This pattern was also found in six of the 17 5. Cremationurn 1191, sf 61: everted rim jar with asquared rim cremation burials excavatedat124–130 London in grey ware TF 11A. Cremation burial 1196. 6. Cremationurn 1251, sf 97: jar in grey ware TF 11A. Road, but the rest of the burialsatthat site displayed Cremationburial 1266. agreater variety of vessels,including at least four 7. Ancillary vessel 1256, sf 96: ring-necked flagon in oxidised TF burialswithDorsetblack burnished ware jars, and 11A. Cremationburial 1266. one with two South Gaulishsamian dishes. The 8. Grave good 1348, sf 174: ring-necked flagon in oxidised TF number of vesselsper grave also varied, ranging 11A. Grave group 1352. 9. Grave good 1351, sf 175: flagon in oxidised TF 11A. Grave from one to amaximumoffour, the latter including group 1352. aflagon,jar, bowland lamp (Timby forthcoming). 10. Grave good 1128. upper part of aring-necked flagon in Thereissome evidence that the use of no more than oxidised TF 11A. Grave group 1145. one or two vessels was fairly standard at this time: at 11. Grave good 1322: upper part of ring-necked flagon in oxidised TF 25. Grave group 1243. the King HarryLane cemetery at Verulamium, one 12. Grave good 1375, sf 239: flared rim jar in micaceousgreyware of the largestlate Iron Age-early Romancremation TF 5. Grave group 1374. cemeteries to be published, some 66% of the 455 13. South-Westblack burnishedware jar (SOW BB1), sf 323. cremation burials were urned and 64% of these Context 1489, mass grave deposit, pit 1483. comprised asingle vessel, with afurther 21% 14. Dorset black burnishedware beaded rim jar decorated with diagonal burnished lines. Joining sherds from contexts 1182 containing two vessels (Stead and Rigby1989, table (grave group 1230) and 1166 (grave group 1167). 3). Abroadly contemporary cremationcemetery ex- 15. Whiteware flask of unknown provenance in ahard, fine cavatedatCaerleon contained121 cremation burials white ware with aslightly sandy texture. Context 1265, datingtothe later 1st–2nd centuryAD, of which curving gully 1358. over 75% of the urned cremation burialscomprised a single vessel (Evans and Maynard 1997). THE COINS None of the vessels found with the burials can be by Paul Booth considered exclusively made for the purpose of burial as they are moderately common forms and all Four Romancoins were recovered during the occur in other ‘domestic’ contexts acrossGloucester. excavation, all of late 3rd–4th century date. All were It has beensuggested that poorerqualityvessels in poor condition and required cleaning and stabili- were often usedfor burial(Philpott 1991, 36) but sation before identification was possible. The coins none of the pots usedatthe London Road site appear are listed in approximate chronological order: to have been wasters or seconds,although their 1. Irregularradiate, 13–16 mm. Obv: DIVO CLAUDIO, legend fragmented state may be masking slight deformities. mostly off the flan. Rev: CONSECRATIO, altar 1b. c AD 270– The use of jars with relatively wide mouths for con- 295. Context 1182, grave 1230, sf 28. tainingthe crematedremains makes practical sense. 2. Antoninianus, 22–25 mm. Obv: IMP CARAUSIUS PF AUG. The flagonsaccompanyingthem probably contained Rev: PAX AUG, Pax standing lholding olive branch and liquid,whether wine for the deceased or some other vertical sceptre. Mint mark lost, but no letters in field. AD 286–293.Context 1254, pit 1255, sf 95. form of libation.Due to the poor preservation of most 3. AE2, 20 mm. Obv: ?]CONSTANTINUS[, head l. Rev: BEATA of the vessels it is impossible to determine whether TR[ANQUILLITAS, altar. Mint mark lost. AD 321–324. they were deposited intact or whether the vessels Context 1300, pit 1301, sf 152. were deliberatelybroken as partofthe burial rites. 4. Minim, 5mm. No details visible. ? c 350–364.Context 1300, The limited range of vesselsused in the sample pit 1301, sf 134. excavated would strongly indicate aprescribed None of the coins is intrinsically remarkable, and range of pots availablefor the burials and that these the only item recovered from aburial feature was the may have been specifically made available as apack- irregular radiate from context 1182, the back-fill of age on demand, perhaps with apottery workshop inhumation grave 1230. For the type and date of the having acontractwith the overseers of the cemetery Consecratio issues of Claudius II see Bland and and of the burialprocedures. It is possible that vessel Burnett (1988, 144).

103 Life andDeathinaRoman City

THE SMALL FINDS was not universal, but does seem to have been a by Hilary Cool common part of the funeral riteinover aquarter of the cremation burials excavated.Given how com- Introduction mon these little bottles must have beeninthe cemetery, the presence of asmall unburntfragment Finds other than pottery, hobnails and coffin fittings in cremation burial 1209 may be fortuitous rather were recoveredfrom seveninhumation graves, three than indicating the use of the contents bothbefore cremation burials and from the massgrave. It was and after burning. only amongst the glass vessels that the sametype The only other fragment of aglass vessel was was foundwithmorethan one burial. These can be foundinthe fill of inhumation grave 1144. If it is more usefullyconsidered as awhole. The otheritems Roman, and the possibility of it being amodern are considered according to the burial theywere intrusion can not entirely be ruledout, it would be foundin. most likely to be of 2nd or 3rd century dategiven that it is made of colourless glass. The glass vessels Small finds from cremation burials The most commonly occurring vessel type was the tubular unguent bottle (Price and Cottam 1998, 169– Cremation burial1209 71). This is an extremely common form in the mid1st century, going out of use during the 80s. They For discussion see the section on the vessel glass. occurred as pyre goods in cremationburials 1209 1. Tubular unguent bottle; lower neck and reservoir. Blue/green and 1766, as acomplete unburntvessel in the melted and collapsed. Length 39 mm, weight 5g.1198 somewhat enigmatic feature 1352 (sf 173; see Sample 68 spit 1. (Plate 5.1) Chapters 2and 6), and as small unburntfragments 2. Melted drop of blue/green glass. Strain-cracked. Weight 1g. in cremation burial 1209 (sf74). This type has been 1202 sample 67. foundatKingsholm, where at least eighteenexam- 3. Tubular unguent bottle (?); base fragment. Blue/green. Convex-curved thick fragment. Dimensions10x7mm, ples were recovered(Price and Cool 1985, 44)and at thickness 3mm. 1198 sf 74. 124–130 London Road (Foundations Archaeology 4. Body fragments (2); blue/green. Convex-curved. Dimensions 2003) where exampleswerefound in three of the 15 x4.5, 11 x5.5 mm, wall thickness 2mm. 1198 Sample 70 burials. Theexamplesfound at 120–122 London spit 2. Road were madeinblue/green glass which is in contrast to thosefound during the earlier excavations, where oneofthe cremation burialshad examples in light green, light yellow/green, and blue as well as blue/green. The other two burials only had blue/green examples as here. Another type of unguent bottle also came from feature1352 (sf 172). This had atall conical reservoir but like the tubular form had asheared rim.Conical unguent bottles were contemporary with the tubular form (Price and Cottam 1998, 172–4) with some examples continuing to be foundintothe 2nd century. The later examples tend to have rolled rims, and the sheared rim of this example would suggest that it is a1st century AD example. The combination of the two in the grave-like feature would suggest that it dates before c AD 90. These sorts of unguent bottles were used for holding oil for bathing prior to the development of the two-handled bath flask (Price and Cottam 1998, 188), and presumably the contents of thosefoundon the pyreshad beenused to cleanse the body before burning. Atotal of 28 cremation burialshave come from the excavations at 120–122and 124–130 London Road, of which five definitely had tubular unguent bottles as pyre goods, another had an unguent bottle of uncertain type, and two others had glass vesselswhich were so completely melted that identification was impossible,but which are likely to have been unguent bottles given the pattern amongst the glass that has been identified. It would appear Plate 5.1Melted unguent bottle from cremation burial that the inclusion of unguent bottles as pyre goods 1209.

104 Chapter Five

Cremation burial1266 Fig. 5.3). With the exception of nos. 15 and 16 which are oval, all the plano-convex counters are approxi- Agaming set comprising 43 boneand glass counters mately circular in outline. There are also ten flat bone and two dice was found in the cremationurn of this counters with acentral dot on one face (nos 1–10). grave together with afrit melon bead and asmall One of these (no.3) possibly has agraffito in the form scrap of unidentified iron (Plate 5.2). Neither the of asingle groove. Thebonecounters belong to gamingset or the melon bead showed any signs of Greep’s Type 1which he has identified as the early burningand so theycan be regarded as grave rather form in use during the 1st and 2nd centuries,being than pyre goods. supplanted by his type 2inthe 2nd century (Greep All of the glass counters (nos. 12–43) have the 1998, 272). The number of glass counterswould typical plano-convex form with the exception of no. agreewithadate in the 1st or earlier 2nd centuryfor 42 which is aflat oval. In total there are 19 white this set, as it can be shown that their use had counters, 11 of verydarkglass which appears black, declined markedly by the later 2nd century (Cool two of amid blue shade appearing opaque, and one et al. 1995, table 125). It is likely that the set had been which is averypale opaque blue/green (Plate 5.3; used forsome time priortodeposition as one of the black counters (no. 11) showedadistinct areaof wear on one edge,and some of the othercounters also showeddulling or wear on their bases. Both of the dice belong to the common form where the markings add up to sevenonopposite faces. This type is not closely dateable. They are not apair as no. 44 is cubic with the markings indicatedbyadouble ring and dot, whilst no. 45 is more rectangular with single ring and dot markings. Asmall flake of no. 45 has become detached which enables it to be seenthat it is madeofivory. It is unclear whether no. 44 was made of bone or ivory. Themelon bead no. 46 is contemporary with the counters as such beads were in use during the mid 1st to mid 2nd centuries AD. The range of sizes of the circular counters(ie excluding nos. 15, 16 and 42) is showninTable 5.2. As can be seenthe bonecounters are, on the whole, Plate 5.2Cineraryurn 1252, burial 1266, before larger than the glass ones. Though the black counters excavation of the contents in the laboratory. tend to be slightly smaller than the white ones there is no major difference in size between the different coloursofglass used. In the gaming set found laid out on aboardinthe doctor’s grave at Stanway one of the players had aspecial piece that was clearly muchsmaller than the rest of the counters(Crummy et al. 2007, 217), but in this set there is no obvious candidatefor such apiece unless the flat oval counter no. 42 functioned in such away. Asmall number of the counterswererecovered on site as some of the fill had spilled in antiquity. The rest were recoveredduring the excavationofthe urn in the laboratoryand so their position can be plotted (see Table 5.3). As can be seen the different types of counters were foundatall depths of the urn. If the various spits are amalgamatedintoupper and lower fills, with thosefound on site considered as coming from the upper fill, it is possible to test formally whether there is any association between the type of counter and its position in the urn. Such tests show there is no association and so it is reasonable to assume theywereadded randomlyasthe calcined bones were poured into the urn. Both of the dice were addedwhenthe urn was almostfull as was the melon bead no. 46. The set shows some unusual features given the number of counters that are present. This can best be Plate 5.3Gaming counters, dice andmelon bead from appreciated from Table 5.4 which summarises the cremation burial 1266. gamingsets from burialsofthe 1st to early 3rd

105 Life andDeathinaRoman City

Figure 5.3 Gaming counters from cremation burial 1266. centurywhere we can be reasonably sure that also excludes aset from aClaudio-Neronian grave at everything depositedinthe grave was recovered. It Alton which was recovered from part of the deposit thus excludes the set discovered by metal detector- excavated in 1860 (Millett 1986, 51–8). No record of ists as partofwhatappears to have been arich later the ﬁnd was madeatthe time. An early museum 2nd century grave at Elsenham, Essex(Hayes 1991; registerrecords 19 counters and an unspeciﬁed Frere 1992, 289) and that from Grave 66 at the St number of dice ( ibid 43) thoughonly ninecounters Pancras cemetery in Chichester as the gravewas were catalogued when the cemetery was published described as disturbed (Down and Rule 1971, 97). It in 1986. The set from Grange Road, Winchester has been included though it shouldbenoted that the bag that contained the counterswas found closetothe Table 5.2 The size of the counters in the gaming set point where the grave was cut by apipe-trench, and within burial 1266. so it is possible that other elements of the set could have been lost. Type Number Mean Median Range As can be seen from this table, where the set was (mm) (mm) (mm) designed for agame of chance using dice as here, the Bone 10 18.8 19.0 21–17 normal number of counterswas 24. Threeofthe four White glass 18 16.5 16.5 19–14 dice gravesinTable 5.4 have that number,and given ‘Black’ glass 915.8 15.5 18–13 that the bone countersinthe grave at Chichester Blue glass 215.8 –15.5–16 were described as very friable so that ‘the exact Pale glass 115.0 – number could notbedetermined’, there must be the possibility that thisgrave toocouldhave had 24

Table 5.3 Distribution of the material in the urn (burial 1266). ‘Site’ refers to those recoveredonsite.

Location White Black Blue Pale green Bone Die Melon Total counter counter counter counter counter bead

Site 14–– 4––9 Spit 1––– –1–1 2 Spit 23–– ––1– 4 Spit 3–11 ––1– 3 Spit 4221 –2–– 7 Spit 541– –––– 5 Spit 66–– 1––– 7 Spit 711– –2–– 4 Spit 822– –1–– 5 Total 19 11 2110 2146

106 Chapter Five

Table 5.4 Gaming sets from selection of 1st to 3rd century burials.

Place Date Counternumbers Dice numbers Reference

Stanway Claudian 26 None Crummy et al. 2007, 217–20 London 40–80 24 4Barber and Bowsher 2000, B435 Verulamium 80–90 22 None Frere 1991b, 259 Anon 1990, 120 Winchester 85–95 18 None Biddle 1967, 230–45 ColchesterMid or late 1 st century 24 4May 1930, 275 Pins Knoll Late 1 st century? 20 None Bailey 1967, 157–8 Chichester Antonine 25 or 26 2Down and Rule 1971, 117 no. 250 Ospringe Later 2 nd or earlier 3 rd century242Whiting 1925, 39 Group XXXVIII counters. Greep (1998, 272) has noted that despite units found. That the different types founddonot the literary references to games played with dice neatly divide intomultiplesofsix would not such as xii scripta needing 15 pieces, the evidence of necessarily have mattered as whilst some sets seem the sets themselves shows that theyoften consisted to divide neatly into two different sorts, others do of multiplesofsix with 18 and 24 being the not. Of the sets with dice in Table 5.4,onlythose commonest. The set from thisgrave does not consist from London and Ospringe are published in of amultiple of six, but it may be notedthat it would sufficient detail for the types to be divided. The do if only the ‘normal’ plano-convex glass counter London burial has 11 white and 13 black counters and bone counters were considered, and the very whilst that from Ospringehad four each in yellow unusual flat oval piece no.42was ignored. That no. and black glass, one each in blue and green glass, 42 shouldberegarded as partofthe set, however, is two bonecounters with the rest being of white glass. suggested by its position within the urn. It came Table 5.4 suggests that agame requiring dice from spit5and so was not added at alate stage as needed24counters. The set at Winchester did not the dice and melon bead were. have dice and consistedof18counters. Aset found The number of counters from this burial is in abuilding occupied in the second quarter of the exceptional andwithin Britain is only exceeded by 2nd century at Castleford also consisted of 18 the 46 bone countersfound amongst the pyre debris counters and lacked dice (Cooland Philo 1998, in alate 2nd century cremation burial at Trentholme 362). The Winchester and Castleford sets do not Drive York (Wenham 1968, 32 no. 41, 97 no. 46). The divide neatly into two sets of nine counters. In the excavator suggested that these had been pyre goods, case of the Winchester counters (all of glass), there though the description of themmakes no mention of were twelve white,fourblack and two blue counters. whether they were burntornot. If he was correct, the The Castleford set consistedofsevenbone counters number in the set could have beenmuch larger, given decorated by acentral dot (Greep type 1), nine bone that not all of the pyre debris is likely to have been counters with faces decorated by concentric rings collected. This brings to mind the 126counters that (GreepType 3) and two opaquewhite glass coun- had been stored in abag in abarrack block in the fort ters. Both of these sets thus divide intothree at Ravenglass when it burnt downinafire during the recognisablegroups but the numbersineach group late 2nd or early 3rd century (Potter 1979, 75–6). are different. Based on the types present, it was suggested that the If we assume the game needingdice took the Ravenglass group included one set of 42 pieces and a London figuresitwould be easy to suggest that second more miscellaneous group. Equally though the Gloucester group had one set consisting of 13 the numbersrecovered could have been the equiva- white and 11 black, which would leave the second lent of three sets of 42. The figure is interesting in the set consisting of eight glasscounters(five white,one light of the observation that this set consists of 42 very pale blue/green, two blue) and ten boneones. standard counters and one unusual piece. Given the verypale blue/green counter couldeasily It is always possible that the large number of pass as awhite counter, the postulated second set counters from this burial relatestothe fact that it was would have three groupsofclearly different coun- the grave of two individuals although,given that all ters justlike the two groups from Winchester and gamingsets assume that at least two people will be Castleford described above. This is of course spec- using them, the increased number of counters would ulationwhich is not susceptible to proof, but given not necessarily be requiredbecause of that. As Table the fact that it is the burial of two individuals, it is a 5.4 shows, two dice appear to have been regarded as possibility. The age rangesgiventothe remains the minimum needed for aset, so it seems unlikely meansthat we could be dealing with two children in that we are looking at two sets forthe same type of the age range of nine to eleven. Equally thoughthere game. Acase could be advancedfor the counters could have been adifference of almost ten years representing two different games, one needing dice between them, and the board games afive or six year and one which did not. Forty-twocounters could old plays tends to be more restricted than those represent oneset of 18 and one set of 24 which, as availabletoafourteen or fifteen year old. Two Greep pointed out (see above), are the commonest different sets might have been thought appropriate.

107 Life andDeathinaRoman City

An interesting question is why these individuals grave at Gloucester would suggest that they were were thought to be in need of one or two gamingsets. either the children of someone attached to the legion In general,though countersare averycommon site or part of afamily of an early colonist. find, they are rare in burials. Philpott’s review of The difference in the deposition patterns between grave furnishings noted ten sets accompanying the nativeelite and the incomers suggests that the cremation burials and three accompanying inhuma- sets were perceived differently in the two commu- tions (Philpott 1991, 185,279). Since then afew others nities.For native elite they could be viewed as being have been foundincluding the ones from the London part of the ‘Roman’ cultural package alongside the East Cemetery and Stanway (see Table5.4). (See writing equipment, items associated with visiting the Crummy et al.2007, table 5.5,which appearedwhilst baths, items for making sacrifice and tablewares for this was in press. Full consideration of the discussion polite dining that are variously seeninthese graves. has not beenpossible.) Recent excavations by Oxford For the incomers there was no need to flaunt symbols Archaeology on the A2 near Springhead in Kent of Romanitas as they,self-evidently, were ‘Roman’. revealed another set (with 23 glasscountersand two In the Gloucester burial the gamingset(s) are not dice) from arich early Roman cremation grave indicating that the deceased were part of an elite (Allen, pers com.). Despite these additional sets, this family, but perhaps that these individuals were seen Gloucester find still only brings the total known as in need of special care in their burial rite. It is from Britain to 26, so clearlytheycontinue to be rare noticeable that young people are often singled out for as aburial find. In cremation burialsthey occur special treatment (see forexample Martin-Kilcher primarily as grave goods ratherthan pyre goods,as 2000; Gowland 2001;Coolforthcoming),and this is the case here. In the 1st century two different may be the reason for the deposition here. Themelon patterns of deposition can be identified. In the caseof bead also included mightbeafurther indication of the Stanway, Alton, Verulamium, Springhead and this. Melon beads occur in large quantities on Winchester gravestheyare occurring as partofthe domestic sites,occasionally in circumstances that burial furnishings of elitemembers of native society. suggest they may have been regarded as talismans. Ihave argued elsewhere (Cool 2006, 193) that these At both Scole and Colchester for example they have types of graves can be interpretedinterms of the been found deposited with abell in amanner that native aristocracy in southern Englandexploring the suggests they mayhave formed part of afoundation parameters of whatitmeant to be ‘Roman’ and what deposit(Seeley 1995; Crummy 1992, 187no. 1663). was necessary forthe good life. The Elsenham and They are also sometimes found singly or as parts of Chichester finds continuethispattern into the small assemblages of amulets in graves(Philpott Antonine period. The young man of between 15 1991, 130). It is noteworthy that the individual buried and 17 years buriedatPins Knoll, Dorsetmay be at Grange Road, Winchester also had eight melon exhibiting the sametendency in adifferent cultural beads.These had been placedinabag with the milieu.Hewas buried in the Durotrigian tradition as counters as well as various othertrinkets including a acrouched inhumation burial,possiblyinthe second finger ring,aseal box lid decorated with an animal, a half of the 1st century. In the crookofhis left arm bell and abanded flint. This group of material is very there were 20 counters that had been in abag, and an reminiscent of the crepundia that Martin-Kilcher iron stylus and two studs.The counters were made of (2000) has identified as being afeature of the burials stone apart from two fashioned from potsherds and of girls and young women in the Romanworldwho one from an oyster shell. Thecounters were thus not died before marriage. The inclusion of abellinthe of the typical ‘Roman’ types but the inclusion of a group is also significant as when bells occur in graves stylus is paralleledinseveral of the contemporary they tend to be in those of children and an amuletic elite gravesfurthertothe east where the inclusion of significance can be assumed(Cool2004, 401). writing equipment seems to be anotherfacet of what There tendstobeanassumption that gaming sets it was to be ‘Roman’. were appropriate for males.The (male) authors of The othertradition of depositing gamingsets is the reports on boththe Winchester and the Alton set seen at Colchester (Crummy et al. 1993, table 8.2), at make thisexplicit assumption (Biddle 1967, 248; London and here at Gloucester. This is where they Millett 1986, 51), though in the case of the Alton form virtually the only offering.AtColchester the set burial the bone specialist could only identify the had been on the pyre. At London the grave was remainsasadult (unsexed) (Green in Millett 1986, 77 suggested to be an inhumation but it was disturbed and table 4) and at Winchester the expert view was and no human bone was recovered.The only other that the remains suggested ‘immaturity (perhaps a find was apottery vessel. The Colchester burial is teenager) or afairly slender female’ (Brothwell in not closely dated but the inclusion of burnt Biddle 1967, 231). As already discussed the bag of fragments of glassunguent bottles would strongly trinkets which included the countersinthe Winche- suggest it was of mid 1st century date or possibly a ster grave suggeststhe deceased may well have been little later in the century.Itwill thushave been the an adolescent female. Of the graves with gaming burial of acolonist at the timewhen most of the sets where the sex of the deceased is known, few inhabitants of Colchester canbeassumed to be non- can be shown to be undoubtedly that of amale. At native.Asimilar origin might be proposed for the presentthe admittedly sparse evidence suggests that London individual, and the date proposed for this gamingsets were not seen as agender specific type

108 Chapter Five

Figure 5.4 Other objects from cremation burial 1266. of grave good but that they were seenasmore 13. Counter, plano-convex. Very dark glass appearing black; base appropriate for young people (here, Winchester, pitted. Diameter 17.5 mm, thickness 7mm. sf 269. Level 5. (not illustrated) Pins’ Knoll) than for adults (Alton). 14. Counter, plano-convex. Very dark glass appearing black; base pitted and dulled. Diameter 17.5 x16.5 mm, thickness 7mm. Catalogue of small finds from grave 1266 (Figs 5.3–4) sf 101. Site find. (not illustrated) 15. Counter, oval plano-convex. Very dark glass appearing black; 1. Counter; worked bone. Circular with flat face and bevelled base pitted. Diameter 17 x13mm, thickness 6.5 mm. sf 286. edges and central dot on one face. Diameter 21 mm, thickness Level 8. 4mm. sf 258. Level 4. 16. Counter, oval plano-convex. Very dark glass appearing black; 2. Counter; worked bone. Circular with flat face and bevelled base smoothed. Diameter 16 x13mm, thickness 6.5 mm. sf edges and central dot on one face. Diameter 20 mm, thickness 288. Site find. (not illustrated) 3.5 mm. sf 280. Level 7. (not illustrated) 17. Counter, plano-convex.Very dark glass appearing black; 3. Counter; worked bone. Circular with flat face and bevelled base pitted. Diameter 15.5 mm, thickness 7mm. sf 105. Site edges and central dot on one face. Possible adeliberate find. groove on other face. Diameter 20 mm, thickness 4mm. sf 18. Counter, plano-convex. Very dark glass appearing black; base 281 Level 7. (not illustrated) pitted and dulled. Diameter 15 x14mm, thickness 7mm. sf 4. Counter; worked bone. Circular with flat face and bevelled 104. Site find. (not illustrated) edges and central dot on one face. Diameter 20 mm, thickness 19. Counter, plano-convex. Very dark glass appearing black; base 4mm. sf 103. Site find. (not illustrated) pitted. Diameter 14 mm, thickness 6mm. Sample 203. Spit 8. 5. Counter; worked bone. Circular with flat face and bevelled 20. Counter, plano-convex. Very dark glass appearing black; base edges and central dot on one face. Diameter 19 mm, thickness pitted. Diameter 13.5 mm, thickness 6mm. sf 262. Level 4. 3mm. sf 102. Site find. (not illustrated) (not illustrated) 6. Counter; worked bone. Circular with flat face and bevelled 21. Counter, plano-convex; Very dark glass appearing black; base edges and central dot on one face. Diameter 19 mm, thickness pitted. Half extant. Diameter 13 mm, thickness 5.5 mm. sf 4mm. sf 279. Level 7. 282. Level 7. (not illustrated) 7. Counter; worked bone. Circular with flat face and bevelled 22. Counter, plano-convex. Cloudy mid blue appearing opaque, edges and central dot on one face. Diameter 18 mm, thickness bubble voids on surfaces; base smooth. Diameter 16 x15mm, 6mm. sf 259. Level 4. (not illustrated) thickness 7mm. sf 257. Level 3. (not illustrated) 8. Counter; worked bone. Circular with flat face and bevelled 23. Counter, plano-convex oval. Opaque mid blue, base pitted. edges and central dot on one face. Diameter 17 mm, thickness Diameter 15.5 mm, thickness 13.5 mm. sf 260. Level 4. (not 3.5 mm. sf 98. Site find. (not illustrated) illustrated) 9. Counter; worked bone. Circular with flat face and bevelled 24. Counter, plano-convex. Opaque white glass; base pitted. edges and central dot on one face. Diameter 17 mm, thickness Diameter 22 x20mm, thickness 7mm. sf 284. Level 8. (not 4mm. sf 100. Site find. (not illustrated) illustrated) 10. Counter; worked bone. Circular with flat face and bevelled 25. Counter, plano-convex. Opaque white glass; base smooth. edges and central dot on one face. Diameter 17 mm, thickness Diameter 19 x17mm, thickness 7mm. sf 271. Level 6. (not 3mm. sf 249. Level 1. illustrated) 11. Counter, plano-convex. Very dark glass appearing black; base 26. Counter, oval plano-convex.Opaque white glass; base pitted. pitted, one edge worn smooth. Diameter 18 mm, thickness One edge chipped. Diameter 19 x16mm, thickness 6.5 mm. 7mm. sf 256. Level 3. sf 283. Level 7. (not illustrated) 12. Counter, plano-convex. Very dark glass appearing black; base 27. Counter, plano-convex. Opaque white glass; base smooth; pitted, some wear. Diameter 18 mm, thickness 7mm. sf 264. bubble voids on surface. Diameter 18.5 x16.5 mm, thickness Level 4. (not illustrated) 7.5 mm. sf 276. Level 6. (not illustrated)

109 Life andDeathinaRoman City

28. Counter, plano-convex.Opaque white; base smooth. Diameter 18 mm, thickness 6mm. sf 267. Level 5. (not illustrated) 29. Counter, plano-convex. Opaque white glass; base pitted. Diameter 17 mm, thickness 6mm. sf 275. Level 6. (not illustrated) 30. Counter, plano convex. Opaque white glass; slightly pitted base. Diameter 17 x15mm, wall thickness 6mm. sf 99. Site find. (not illustrated) 31. Counter, plano-convex. Opaque white glass; base pitted. Diameter 17 x15mm, wall thickness 6.5 mm. sf 274. Level 6. (not illustrated) 32. Counter, plano-convex.Opaque white, bubble void on surface; base smooth. Diameter 16.5 mm, thickness 6mm. sf 252. Level 2. 33. Counter, plano-convex.Opaque white glass appearing black; base pitted. Diameter 16.5 mm, thickness 6mm. sf 268. Level 5. (not illustrated) 34. Counter, plano-convex. Opaque white glass; base smooth. Diameter 16 mm, thickness 7mm. sf 272. Level 6. (not illustrated) 35. Counter, plano-convex. Opaque white glass; base smooth. Diameter 16 mm, thickness 6mm. sf 261. Level 4. (not illustrated) 36. Counter, plano-convex.Opaque white; base smooth. Diameter 15.5 x15mm, thickness 6mm. sf 285. Level 8. (not illustrated) 37. Counter, plano-convex. Opaque white glass; bubble void on surface; base smooth. Diameter 15.5 mm, thickness 6.5 mm. sf 251. Level 2. (not illustrated) 38. Counter, plano-convex.Opaque white glass; base smooth. In two joining pieces. Diameter 15 mm, thickness 7mm. sf 263. Level 4. (not illustrated) 39. Counter, plano-convex. Opaque white glass, bubble void on surface; base smooth. Diameter 14.5 x13.5 mm, thickness 6mm. sf 250. Level 2. 40. Counter, plano-convex.Opaque white; nase smooth. Dia- meter 14 x13mm, thickness 6mm. sf 266. Level 5. (not illustrated) 41. Counter, plano-convex. Opaque white glass; base pitted. Diameter 14 x12mm, thickness 6mm. sf 273. Level 6. Plate 5.4Melted unguent bottle from cremation burial 42. Counter; opaque white glass. Oval flat. Dimensions17x 1766. 12 mm, thickness 4mm. sf 265. Level 5. 43. Counter, plano-convex.Opaque glass appearing very pale blue/green; base smooth. Diameter 15 mm, thickness 6mm. Inhumation grave1151 sf 277. Level 6. (not illustrated) 44. Die, workedbone or ivory. Cube with markings formed by The only item in thisgrave was adamaged and double ring and dot; opposite faces adding to seven. Dimensions16x16 x16.5 mm. sf 253. Level 2. incomplete mount, fastened to whatever it decora- 45. Die, ivory. Rectangular-sided cube marked with ring and dot. ted by asplit pin. The configuration suggests it could Opposite faces add to seven. Dimensions16.5 x13x12mm. have actedassome form of hinge. Whilst the depth sf 254. Level 3. of the split pin suggests it would have been fastened 46. Melon bead, turquoise frit with glaze surviving. Complete. to something substantial such as awooden board, it Diameter 18 mm, length 14 mm, perforation diameter 8mm. sf 248. Level 1. seems unlikely that this is abox mountassheet 47. Unidentified. Iron strip (2 fragments) retaining fragment of mountings on those are normally fastened by studs copper alloy sheet. Lengths 17 and 14 mm, width 2mm. sf (see for example Crummy 1983, 85–9). This item was 255. Level 3. (not illustrated) recoveredfrom the back-fill, and givenits damaged and incompletestate, it seems most likely to have been achance inclusion, rather than adeliberate Cremation burial1766 grave good. For discussion see the section on the vessel glass. 1. Mount. Copper alloy. Rectangularsheet with one short end 1. Melted and vesicular fragment of blue/green glass. Possibly broken; centrally perforated at other short end which is the collapsed reservoir of atubular unguent bottle. Present concave; perforation retains split pin through perforation; one length 37 mm, weight 2g.1049. (Plate 5.4) leg of pin bent back for c 6mm. Sheet has central vertical rib possibly formed from having been bent. Sheet: dimensions 43 x23mm. Split pin: length 27 mm, width 4mm. 1110 sf 20. Small finds from inhumation burials Inhumation grave1144 Inhumation Grave 1230 For discussion see the section on the vessel glass. The fragment from an iron ring was foundinthe abdomenarea of the skeleton in the grave.Had it 1. Body fragment; colourless body; strain-cracked. Possibly been agrave good more could have been expectedto modern. Dimensions21.5 x17mm, wall thickness 3mm. survive as it retains asolid metal core. Given the

110 Chapter Five range of items found in the ﬁll, it seems best to Table 5.5 The occurrence of intaglios in complete and regard this item as an accidentalinclusion rather incomplete simple expandedfingerrings (Sourceofdata: than adeliberately placedgrave good. Cool 1983, Finger ring group IVA excluding those with enamelled bezels). 1. Ring; approximately one-third extant. Iron. Square-sectioned. Diameter c 25 mm. 1181 sf 141. State Intaglio Present Intaglio missing Total

Complete 54 862 Inhumation grave1246 (Fig. 5.5) Incomplete 32 21 53 The type of finger ring foundinthis grave (sf 88) was Total 86 29 115 in use primarily during the 1st and 2nd centuries AD (Henig1974, 46 types II and III). It is currently incomplete, lacking asmall partofthe hoop but the part that is missingiswhere the hoop would have be foundwith empty bezels when complete. As been narrowest and it may have disappeared Henig has frequentlyobserved,itisexceptional for throughcorrosion. The intaglio it would have held intaglios to be foundinRomano-British burials is also missing. It does not appear to be directly probablybecausetheyneededtobehandedonto associated with the skeleton, but there are some the heirs (eg Henig in Millett 1986, 57). The grounds forthinking it was deliberately placedinthe deliberate removal of the intaglio from this ring grave.Iron rings such as this were amark of rank, as would fit this patternand would explain why it is so iron rings were appropriate forthe ordinarycitizen muchatvariance with the pattern seeninTable5.5. whereas gold ones were technically reservedfor the These factors, and the likely importance of it being senatorial and equestrian classes(Pliny Natural made of iron, all arguefor the ring having been History 33.iv–viii). In acommunitysuch as the deliberately included in the grave,rather than just fortressand colonia at Gloucester it mightbe being achance inclusion. expected that status indicators such as the correct It is notuncommonfor childrentobesingled out metal for afinger ring mightberespected. There is for special treatment in their burial rite, and it is also some evidence that iron may have been possibly the age of the deceased in this grave (5–12 perceived as particularlyappropriate for placing years) that triggered the deposition of thisring. It with the dead.Apreliminarysurveycomparing the mightbethought that the child was too young to be materials used for personal ornaments in domestic in aposition to have heirs to whom the intaglio and in funerary contexts at Baldock shows that iron passed, but there is epigraphic evidence in Britain for items are muchmore likely to be present in graves childrenasyoung as 13 having formal heirs than would be expected in the domestic assemblages (Collingwood and Wright 1965, no. 696). (Rosten 2007, 117–9). The absence of an intaglio is also significant. As 1. Finger ring. Iron. Slender hoop expanding mainly at oval already noted, the incomplete state of the ringmay bezel, empty setting now infilled with corrosion products. In two fragments with part of back of hoop missing. Diameter c have come about through corrosion. Normally 27 mm, bezel dimensions 14 x10mm. 1244. sf 88. complete rings such as this can be expectedtoretain their intaglios when found. This can be demon- stratedinTable 5.5. This is derived fromasurveyof Inhumation grave1352 (Fig. 5.6) metal personal ornaments acrosssouthern Britain. For discussion see the section on the vessel glass. As can be seen it is unusual for rings such as these to 1. Tubularunguent bottle;blue/green.Outbent rim, edge sheared; cylindricalneck; tubular body with slight flattening at base. Complete apart from asmall fragment missing in the rim. Neck/body junction tooled. Rim diameter 15 x14mm, maximumbody diameter 14 mm, height 74 mm. 1347. sf 173. 2. Tall conical unguent bottle; complete. Blue/green. Outbent rim, edge sheared;cylindrical neck;conical body with flattened base. Rim diameter 23 x22mm, maximum body diameter 29 mm, height 102 mm. 1347. sf 172.

Inhumation grave1362 (Fig. 5.7) Braceletsf216 belongstothe family of late 3rd and 4th century bracelets decorated with multiple units. These appear to be an insular fashion. Thenormal patternisfor there to be azoneofidentical decoration behind each terminal, and forthe central third of the circumference either to be decorated with anumber of different motifs symmetrically Figure 5.5 Iron ring from inhumation burial 1246. arranged aroundthe centre point or forittobe

111 Life andDeathinaRoman City

Figure 5.6 Unguent bottles from inhumation grave 1352.

occupied by just asingle motif (Cool 1983, Bracelet Groups XXXI and XXXII; see also Swift 2000, 145). This example belongs to the latter group. The decoration is very faintbut, insofar as can be judged, is unparalleled. Neither of the motifs occurred in the c 190 examples cataloguedaspart of my survey. The bracelet was worn on the left arm of an adultof18–25 years. It was not possible to sex the individual from osteological data but elsewhere where the skeleton can be sexed,the association is normally with afemale. This was certainly the patterninthe three graveswithbracelets fromthe earlier excavations in thiscemetery where two of the skeletons could be sexed as female. One was of the same ageasthe individual in 1362 (20–25 years) and the other awoman of c 50 years. The third individual was described as amature adult and appears to have worn at least one of her bracelets, Figure 5.7 Bracelet from inhumation grave 1362. also on the left arm. Thepresence of abracelet with an adultwoman in this grave continues the trend seen in the earlier excavations that suggests that amongst this communitybracelets were seenas and punched decoration in three zones, now very faint with appropriate foradults. At some sitessuch as unit of vertical grooves between them and behind terminals; central unit of three horizontal grooves with punching on two Lankhills in Winchester bracelet wearing seems zones between them producing amilled appearance; zones particularly associated with younger girls (Gowland betweencentral unitand terminals probably consist of 2001, 162), but this does notappear to be the case at grooves parallel to each edge with aline of punched dots Gloucester. between; Diameter c 55 mm, section 7x1.5 mm. 1361. sf 216. 2. Finger ring (?), copper alloy. Rectangular-sectioned with The function of ring sf 219 cannot be ascertained rounded edges, vertical groovesacross broken bezel. Central with certainty. The piece is broken acrossthe area part of bezel missing. Diameter 18 mm, section 2.5 x2mm. where the bezelwould have been if it were afinger 1361. sf 219. (not illustrated) ring, but it is possible that these breaks represent damaged terminals of apenannular ring. If that is correct, then it couldwell have been an earring Inhumation grave1505 (Fig. 5.8) (Allason-Jones1989a, Type 2b). Equally, however, it The ringfoundonthe third or fourth finger of the mighthave been an example of the light trinket rings left hand of the skeleton belongstoagroup of trinket popular in the 3rd and 4thcenturies. finger rings decorated with patterns similar to those 1. Bracelet (4 fragments).Copper alloy. Rectangular-sectioned, on the light bangles in use during the later 3rd and widest to wrist; hook and eye terminal, eye broken. Grooved 4th centuries (Cool 1983,Finger ring Group XX) and

112 Chapter Five

intagliowhich was the seal of its owner, these trinket rings were purely decorative.

1. Finger ring. Copper alloy. D-sectioned band; oval outline. Wide vertical grooves forming abeaded outer face. Diameter 19 x16mm, section 2.5 x1.5 mm. 1480. sf 319.

Small finds from the mass grave (Fig. 5.9) The brooches from thisfeature suggest that burial took place in the later 2nd century.Wroxeter Figure 5.8 Finger ring from inhumation grave 1505. brooches such as Figure 5.9 no. 1are 2nd century in date. The examplesfrom dated contexts that Mackreth cites belong to the middlepartofthe century(Mackreth 1995, 963 no. 27) but they seemto is probably contemporary with them. Apair of rings have continued in use until at least the end of the decorated with verticalgrooves were found appar- century. An example came from ahoard of items ently worn on the fingers of the left hand of an dated to AD 200 at Chepstow (Bayley and Butcher individual of about 20 to 25 yearsburied in the third 2004, 169) and another camefrom aditch fill at quarter of the 4th century at Lankhills (Clarke 1979, Somerford Keynes with mid to late 2nd century 68 grave 326, fig. 87 nos. 401–2), and asimilar ring pottery (Miles et al. 2007, 251 table 9.6 see also from the Fordington cemetery at Dorchester is digitalsection5.3). The knee brooch (Fig. 5.9, no. 2) preserved still in situ around afinger bone(Cool belongs to the variety that appears to be aBritish 1983, 1108 no. 2). Unlike the intaglio ringssuch as sf development of the form (Hull form 171) though the 88 in grave1246, whose function was to hold the bow is more angular than normal. Adate in the later

Figure 5.9 Small ﬁnds from mass grave 1483.

113 Life andDeathinaRoman City

2nd to early 3rd century is appropriate (Bayley and as that in grave 1246, so a1st to early 3rd century Butcher 2004, 179–80).The third brooch is an iron date is appropriate. penannular brooch which is notclosely dateable The presence of three brooches in this deposit (Fig. 5.9, no. 3). The proposed date would be in stresseshow unusual it is. Despite brooch wearing agreement with the hairpin found in the upper fill being very common in the 1st to mid 2nd century (Fig. 5.9, no. 6) as this is an example of aCrummy judged by site finds, brooches seem almosttobe Type 2(Crummy1983, 21) which is the commonest activelyshunned as grave or pyre goods at that time. 2nd century form. 2nd or 3rd century cemeteries with brooch gravesare None of the brooches was directly associated with not common as can be seen from Philpott’s distribu- any of the skeletons in the gravebut all had their tion maps (Philpott 1991, figs. 9and 26). Though pins in the position that would be expected if they these were based on research carried out in the had been fastening items of dress. It seems likely that late 1980s,discoveries since then have not materi- they entered the deposit as part of worn items of ally changedthe picture.The habitofincluding clothing. brooches was commonest in the 1st centuryand in An adult female aged 26–35 (1596) wore the the 4th. To findthree brooches in this burial pit bracelet (Fig. 5.9, no.5)onher right forearm (Plate would strongly suggest that the dead were not being 5.5). It is averysimple form with hooked terminals. prepared forburial in the normal way and there Most similar bracelets have beenrecovered from 4th was apressing need to dispose of them, perhaps in centurycontexts (Cool1983, 195 BraceletGroup the garments theywere wearing when they died. XXXIII), but acomplete example was recovered from The dates of the brooches would suggest that if these an occupation level at Fishbourne dated to the early unfortunate people were victims of ahistorically 2nd to later 3rdcenturies(Cunliffe 1971, 107no. 47, attested plague, thenitisthe Antonine one that is fig. 41). Figure 5.9, no. 5provides verywelcome most likely (see discussion in Chapter 6). The later evidence that the form was being worn in the later one in the mid 3rd century would be toolate for 2nd century. In general bracelet wearing was not these styles of brooches as by that time one would be fashionable for the majorityofwomenbefore the end expecting thosepeople who chose to wear brooches of the 3rdcentury but there is agrowing body of to be favouring P-shaped brooches and light cross- evidence that in the Gloucestershire area it was a bows. fashion that developed in the 2nd century (Cool in Miles et al. 2007, 346). 1. Wroxeter brooch. Copper alloy. Cast headloop with flat D- An adultmale (skeleton 1542) had several items shaped head plate. Two perforated lugs behind head plate holding spring of two turns, chord forms loop below spring; associated with him including the copper alloyring pin complete and fastened inside triangular catch plate. D- (Fig. 5.9, no. 7) and two fragments of iron bar(nos. sectioned upper bow with three vertical grooves joins 8–9). Unfortunately quite what the iron fragments headplate with lentoid moulding on either side; junction of were part of cannot be identified, but as one of upper and lower bow marked by three transverse horizontally ribbed mouldings; Lower bow has one vertical groove. them was over the left hip, it might perhaps be sug- Cylindricalfootknob with two mouldings at top. Length gested theywerepart of some form of belt or purse 55 mm, head width 15 mm. 1545. sf 406. fitting. 2. Knee brooch.Copper alloy. D-shaped headplate with two The finger-ring (Fig. 5.9, no. 4) was not associated perforated lugs behind holding spring of four turns, cord with any particular skeleton so it is notknown forming aloop below spring; pin complete and fastened in trapezoidal catch plate. Slightly faceted, sharply angular bow whether it was worn at the time of deposition. It expanding out to flat foot plate. Length 41 mm, head width belongs to the samefamily of simple expanded rings 13 mm. 1483. sf 437. 3. Penannular or annular brooch, two joining fragments; iron. Square sectioned hoop with broken ends. Pin wrapped around hoop and tapering to point resting on hoop on opposite side. Diameter c 40 mm, hoop section 5mm. 1485. sf 345. 4. Finger ring in four joining fragments. Copper alloy. Circular- sectioned hoop expanding to flattened rectangular bezel. Diameter 24 x20mm, bezel section 5.5 x2mm, hoop section 2mm. 1526 sf 383. 5. Bracelet, copper alloy. D-sectioned hoop, double hook terminal, part of one hook broken. Diameter 66 x64mm, section 3.5 x2mm. 1596. sf 464. 6. Hair pin, worked bone (three joining fragments).Circular- sectioned shank; conical head with three grooves; broken tip. Present length 68 mm, shank section 3mm. 1485. sf 341. 7. Ring, copper alloy. Rectangular-sectionedhoop. Diameter 19 mm, section 3.5 x2.5 mm. 1545 sf 378 8. Bar. Iron. Rectangular-sectioned;three fragments, one curved. Largest fragment 21 mm, section 8x4.5 mm. 1545. sf 399. (not illustrated) 9. Bar. Iron. Rectangular sectioned,tapering to point Length 46 mm, section 8x4mm. 1545. sf 408. (not illustrated) Plate 5.5Copper alloybracelet (sf 464) worn on the 10. Angle bracket? Iron. Length c 55 mm. 1544. sf 386. (not right forearm of skeleton 1596. illustrated)

114 Chapter Five

Small finds from othercontexts Nailsfrom inhumation burials The only item from the non-funerarycontexts that Structural nails warrants discussion is the fragment which appears The majority of structural nails came frominhuma- to be from asmall barrel padlock from the cemetery soil. If this identification is correct, thenitwas lostor tion burials, aminimum of 224 definite and six possible nails being recovered from 40 graves. The discarded when it was locked. Barrel padlock. Iron. Part of case with curved top and size of groups from individual gravesvaried from squared base retaining part of the haft, stop ridge and central one nail to 18. The nails in 19 graves were interpreted internal rod. Present length 46 mm, casing 23 x20mm. 1440. on the basis of their locations withinthe grave as (not illustrated) being in situ,representing the remains of coffins. Mineralised wood was identified adhering to nails in seven of these graves(1330, 1353, 1369, 1374, 1422, THE NAILS 1505, 1507). The lengthsofcomplete or near complete by Kelly Powell nails generallyfell within the range of 50–80 mm, Introduction and indicatedahigh level of consistency in the construction of coffins. The overall lack of smaller An assemblage of 1440 nails, including hobnails, was nails suggests smaller wooden grave goods were recoveredfrom the excavation, of which 999 came unlikely to be have beeninterred with the burials. from burial features. Thecondition of the nails was generally poor, the majoritybeing heavily corroded and fragmentary. However, the lengths of most of Hobnails the structural nails fell withinthe range 50–80 mm, Aminimum number of 730 hobnails were recovered typical of many Roman nail assemblages. Atotal of from atotal of 20 inhumation graves. The groups of 145 were classified as Manning type 1structural hobnails from six graveswere recorded as represent- nails (47 of which were type 1b), with an additional ing items of footwear that were worn at the timeof five being tentatively identified as type 1(Manning burial (see Footware in Chapter 2). The footwear 1985). Atotal of 1146 nails could be positively worn by burial 1232 in grave 1234 were represented identified as hobnails (Manningtype10) with five by groupsof32hobnails associatedwith the left foot additional tentative identifications. The assemblage and 37 associatedwith the rightfoot,while13 contained some complete hobnails, generallyupto hobnails were associatedwith the left foot of burial 18 mm long with headdiameters of 9–10 mm. The 1328 in grave 1330 and 22 with the right foot,aswell remainder of the nails were unclassifiable due to as afurther 28 hobnails recovered from asoil sample. their fragmentary nature and none of the less The only surviving foot of burial 1537 in grave 1765 common nail types were positively identified within was shod in an item of footwear comprising 42 the assemblage. hobnails. The burialsingraves1344, 1353 and 1369 wore footwear comprising 178, 38 and 104 hobnails respectively,and an unidentified iron object from Nailsfrom cremation burials grave 1353 mayhave been further corroded hob- An assemblage of 13 structural nails and 32 hobnails nails. Seventeen of the 20 hobnails recoveredfrom were recovered from five cremation burials. The grave 1167 were located beyond the feet and may largestgroup came from cremation burial 1769 and come from an item of footwear placed at the foot of comprised 24 definite and three probablehobnails, the grave ratherthan worn. Graves 1283, 1315 and and eight structural nails. Four structuralnails were 1765 produced assemblages of 44,37and 42 hobnails recoveredfromthe cinerary urn in grave 1766 and respectively that are also likely to represent foot- two from the urn in grave 1766, and asingle wear, whether worn or placed as grave goods. The possibly complete hobnail and afragment of nail hobnails from grave 1315 were mostly corroded shank were found withinaccessory vessel 1199 in together, some with possible mineralised organic burial 1209. The remainder of the assemblage came material holding them in their original formation, from grave backfills. but it is unclear whether this item of footwear was The majority were relatively small compared to worn or simply placedwiththe burial. Theremain- those from inhumation graves, measuring less than ing groups were smaller in number, and four graves 37 mm long, and couldoriginate from anumber of contained only singlehobnails, which are likely to be funerary fittingssuch as biers and boxes, as foundat chanceinclusions. otherRomano-British cemeteries such as the M6 Toll (Powellforthcoming), Pepper Hill,Kent (Biddulph Nailsfrom the mass grave 2006) andBrougham, Cumbria (Cool2004). No clear examples of the use of boxes or caskets as containers Five skeletons in the massgrave appear to have been for cremated bone or pyre goodswere identified, and shod at the timeofburial.Agroup of 69 hobnails (sf it is likely that the nails recovered from cremation 2004) associatedwith skeleton2005 was found in its burialswerethe remains of items included on the original formation, held together by corrosion and pyre. missingonlythe toe end. Hobnails were observed

115 Life andDeathinaRoman City aroundthe edge of the sole with threelongitudinal 2005). Photomicrographs of the tombstone geology rows at the toe end andafigure of eight at the heel. can be found in Appendix 5. The patterning of hobnails on the soles of shoes has been discussedbyvan DrielMurray (1999), who Tombstone 1(Plate 5.6) argues that they were not merely fashionable decoration, but carried avariety of symbolic mean- Tombstone 1was recovered from pit 1003, which ing, possibly associated with protection. was cut into the soil layer associatedwith the use of Skeleton 1666 was also shod, with two groups (sf the cemetery (see Fig. 2.2 above). It is carved from a 677, 676) of aminimum of 80 and 79 hobnails found block of local oolitic limestone (from Painswick Hill, associated with the left and rightfoot respectively. In Glos, see Appendix5), broken but mostly surviving addition to this 15 hobnails were recorded as being in three main joining pieces with anumber of smaller foundinassociation with skeleton1638, eight with fragments. It measures 0.59 minbreadth, and the skeleton 1552 and five with skeleton 2009. Afurther surviving height is 1.03 mthough with apossibly two groups of 41 hobnails (sf 2018) and 21 hobnails gabledtop and abase it may have reached some (sf 517)are likely to represent items of footwear that 1.30 m. Thestele is 0.14 mthick. had become separated from their original wearers The upper part of the stele displays within aniche, during or after burial, and atotal of 59 stray hobnails perhaps originally pedimented, aman reclining on a (sf 249, 516, 517) were also found. The mass grave couch to the rightand facing front. He wears atoga, contained no obviouslystructuralnails. pulleduptocover the back of his head, and holds the scroll containing his will in his left hand and a cup with flutedsides in his right hand. He is clean- Discussion shavenand his coiffure is characteristic of the Groups of in situ structural nails indicated that at Claudio-Neronian period thoughitshould be noted least 19 burialswere buried in coffins.Smaller num- that the banquet tombstonesfrom Germany dis- bers of nails recovered fromafurther21gravesmay cussed below whose male subjects are shown with be evidence forfurther coffined burials, the small similar coiffure and physiognomyare all assigned size of these groups being the result of truncation or Flavian or even Trajanic dates. The couch has a differential preservation, although it is also possible raised back with curved sides but the foot on the left that some may represent residual material. of the stone is obscured by the figure of aslave-boy, Hobnails indicate that shoes appearcommonly as grave goodsinthe Romanperiod, though the relatively small numbers involvedinmany cases are unlikely to constitute whole pairs of shoes. In the case of cremation burials differential collection of material fromthe pyre for burial may result in the under-representation of hobnails, and deliberate selectionmay also play apart in the formation of the buried assemblage (McKinley 2000c, 41; Cool 2004, 391). In inhumation burialsthe presenceor absence of hobnails mayresult from differing burial traditions relatedtothe inclusion of grave goods.A number of the burials from LondonRoad are entirely devoid of hobnails, suggesting that the individuals were not interred with shoes, or were buriedwith footwear whose construction did not involve the use of hobnails. The small numbersofhobnails in several gravesmay be the result of differential preservation, as also appears to have affected some of the skeletal material. In contrast some of the burials appear to have enough hobnails for more than one pair of shoes,acircumstance that has previously been noted in Romano-British cemeteries (Philpott 1991, 168).

THE SCULPTURAL STONE by MartinHenig and Roger Tomlin The two tombstones recovered during the course of the excavation have previously been publishedinthe journal Britannia,largelyfrom the standpoint of their inscriptions though incorporatingafew comments on context andiconography (Tomlin and Hassall Plate 5.6Tombstone1after cleaning and conservation.

116 Chapter Five whilethat on the right was on amissingportionof the military community when required. It is surely the stone. The head of the slave-boy is preserved, as significant that the closest parallels in terms of the is his left foot but the front of his body, which seems shortened figure, the form of the table and the table to have beenwrapped in acloak, has sheered off. In top seen in bird’s eye view are from Chester, front of the man is atripod table with curving although these do not datebefore the later 2nd widely-splayed legs, only two of which are shown, centuryoreven the 3rd century, and differinpoints the position of the middleone being taken by ajug of detail, such as the position of the servantwho is which stands on the ground line. The circular top of often behind the couch,orinfront and notatthe foot the table is depicted in flattened bird’seye view with (Henig2004, 14–16). two vessels, another cup, of scyphosform, and a It is reasonable to see such stylistic links as there bowl. are between tombstones which are carved in The stone was probablyoriginally painted and different materials (those fromCheshire being of there seems to be areddish wash on the background sandstone not limestone) as well as belonging to to the head, though any othertrace of colour appears different periods as being due to acontinuing to have been lost. tradition in LegioXX’s lapidaryworkshop. The The inscription reads: legion was stationed at Gloucester during the mid- M[.]RTIALIS 1st century, but it should be noted that connections continued with legionaries retiring to the colonia as C.[.]LONI D SERV[.] exemplified by atombstone of aveteran, L. Valerius [...]O D XIIII Aurelius foundatKingsholm some yearsago. [...] E M[a]rtialis |G(ai) [Ci]loni serv[i] |[ann]o(rum) XIIII |[h(ic) s(itus)] e(st) Tombstone 2(Plate 5.7) ‘(To the shades of the dead and) of Martialis, the Tombstone 2islikewise carved from local oolitic slave of Gaius Cilonius, aged 14. He lies here.’ limestone, and is broken intoanumber of pieces.It The reading is fully discussed in RomanInscrip- measures 0.51 minwidth by at least1.00 min tions of Britain, III (forthcoming), but it may be noted height and is some 0.11 mthick. The recessed panel that the absence of an introductory ‘Dis Manibus’, on the front face of the stele is divided into two the genitive caseofthe name of the deceased, and Gaius Cilonius’ lack of cognomen all point to a Claudio-Neronian date. His Italian nomen is uncommon, but probablycognate with the more common cognomen Cilo.

Discussion The fact that the main ﬁgure is depicted wearing a toga and with his will suggests the tombstone was intended for aRomancitizen, though in the event it was usedfor a14-year old slave. Thecarving must therefore have beenexecuted prior to the inscription being added,and the tombstone was presumably bought ‘off the peg’. The character of the carving is of considerable interest. It is crisply cut,asisespecially apparent in the folds of the toga, though the execution of the ﬁgure is much more schematised than,for example, the near-contemporaryLegioXXtombstone of M.Favonius Facilis from Colchester (Huskinson 1994, 23). 1st century military grave stele from Germany include banquet scenes that are iconogra- phically similar to the Gloucester stone, with a reclining clean-shavenman facing the spectator and aslave standing near the foot of the couch (Bauchhenss 1988, 43–9; Boppert 1992, 157–163), although stylistic contrasts between these thorough- ly ‘Roman’ productionsand what was produced at the sameperiod in Gloucester are striking. It must be concluded that the latter workshop was a‘local’ one with links to the army or indeed staffed by soldiers whose task was to produce tombstones, and perhaps also religiousand commemorative monuments,for Plate 5.7Tombstone2after cleaning and conservation.

117 Life andDeathinaRoman City unequal parts. The upper edge of the top part is sculptor, or else the lapidarius himselfhad ashot at arched and displays afacing bust. Despite the carving the bust. Maybe something similar hap- brushed-forward Julio-Claudian hair, the style of pened in the caseatAlchester where acrescent moon presentation is native,asexemplified by the subject’s was inscribedremarkablylike aletter ‘C’. jug-ears, ovoid eyes,small wide-open mouth and wedge-shaped nose. Immediately below the head he OTHERWORKED OR UTILISEDSTONE sports alarge circular amulet. On the left edge a by RuthShaffrey lobate leaf is inscribed. The inscription reads: The only item of workedstone apart from the tombstones was afragment of roof tile recovered L D OCTAVI from back-fill of acremation burial of 1st–early 2nd L D POL centurydate (1227).The tile was composed of Oolitic MARTIALIS limestone weighing 401 gand had two surviving EPOREDIA M D LEG XX edgesindicating adiamond shape. Among the unworkedstonerecorded were three L(uci) Octavi|L(uci fili) Pol(lia tribu) |Martialis fragments of micaceouslower Old Red Sandstone,a |Eporedia |m(ilitis) leg(ionis) XX stone known to have been used locally as roofing ‘(To the shades of the dead and) of Lucius Octa- material, recovered from medieval and post-medie- vius Martialis, son of Lucius, of the Pollian voting- val contexts. They areall small fragmentswithout tribe, from Eporedia, soldier of the Twentieth clear evidence of working, although shallow grooves Legion.’ on one side of athickish piece from the medieval soil This reading again is fullydiscussed in Roman layer (1025) indicate that it had been usedasa Inscriptions of Britain, III (forthcoming), and the whetstone. absence of the introductory‘Dis Manibus’, and the genitive case of the name of the deceased indicate a THE CERAMIC BUILDING MATERIAL Claudio-Neronian date. The lack of the legion’scog- by Cynthia Poole nominaV(aleria) V(ictrix) also suggestsanearly date, but the omission of the age of the deceased and Atotal of 154 fragments(7418 g) of ceramic building his yearsofservice,and perhaps of aconcluding material of Romantomodern date was recovered formula like h(ic) s(itus) e(st), is unusual rather than from the excavation, as well as two fragments of significant of date. Martialis’ home town, Eporedia fired clay (4 g). Romanmaterial accounted for in Italia Transpadana, modern Ivrea north of Turin, nearly half the assemblage by weight(3348 g), but is alreadyknown to have contributed soldiers to four only 38% of the Romanmaterial occurred in Roman otherlegions. contexts, the remainder coming fromfeatures of later date. Discussion No complete tiles were recovered,the assemblage comprising fragmentary and heavily abraded mate- The inscription shows that thiswas the tombstone of rial, some pieces with score marks that may have asoldierofthe XXth legion, but it is far inferior in derivedfrom ploughing. Over halfthe material style to the otherGloucester tombstone,and perhaps identified as Romancould not be positively identi- the work of an amateurornovice mason. For the fied to form, and couldonly be described as flat tile, purpose of commemorating the dead, the portrait of brick or unidentified (Table 5.6). The remainder of the deceased was important, either as afree-standing the assemblage comprisedpieces of tegulaand bust or arelief, but accurate delineationofphysiog- inbrex,withasingle ceramic tessera. No markings nomy was clearly not important as the subject of the (signatures, stamps, tally marks,combing, imprints) ‘portrait’ would have beenidentified by the inscrip- were notedonany pieces. tion which would be readout by thosewho came to No evidence was foundfor the use of tiles as grave pay their respects or by the traveller passing the covers or markers,which has beenattested at Roman grave.The bearded head of Q. Cornelius, probably also of LegioXX, from 3rd-century Chester provides aparallel of sorts (Henig2004,17). Table 5.6 Quantification of Roman ceramicbuilding The leaf like form is reminiscent of the decoration material by form. of an early Legio II tombstone from Alchester, Oxfordshire, ornamented with triangular motifs in Form Nos. %Nos Wt (g) %Wt the spandrels which look likecarrot amphorae but are probablyleaves (Sauer2005, 103 and fig. 1). The Tegula 78.75 775 23.15 contrast between the low level of artistic competence Imbrex 13 16.25 703 21.00 here and the relatively high level of the epigraphy is Flat /brick 32 40 1418 42.35 intriguing. While in the case of Tombstone1surely Brick 56.25 223 6.66 two men, asculptor and alapidarius were involved, Tessera 11.25 14 0.42 it is possible that in the caseofTombstone 2either Unid 22 27.5 215 6.42 there was acompetent lapidarius and an amateur Total 80 100 3348 100

118 Chapter Five cemeteries elsewhere (Philpott 1991, 10–11 and Table 5.8 The animal bone assemblage from sieved 66–7). If any had been employed in such amanner residues quantified by NISP. *denotesthat NISP count here more complete examples would be expected, includes at leastone partial skeleton. This table includes whereas the pieces that were found in grave or cre- some burnt bones. mationfills were small and already heavily abraded when deposited. The character of boththe Roman Species Roman Post-medieval Total and later material is more typical of arural context subject to agriculture,with nearly two-thirds of the Cattle 33 assemblage being moderately to heavily abraded, Pig? 11 probablyasthe result of cultivation. This suggests Mouse 11 that muchofthe material may have arrivedthrough Domestic fowl 90* 64 manuring or dumping of refuse from the town. Bird 11 618 Megafauna 11 Large mammal 24 832 THE ANIMAL BONE Medium mammal 11 by FayWorley Small mammal 11 Introduction Micro mammal 11 Unidentified 290 114 458 The assemblage comprised 1464 fragments of un- Total 418 134 581 burnt animal bone, and afurther 540 fragments (2.7 kg) of cremated animal bone were recovered from cremationburial deposits. Theanimal bone 5.9). The modest size of the post-Roman assemblages was identified as preciselyaspossible to taxon and limits their potential for interpretation beyond element using faunal reference material and identi- presence or absence of taxa. The majority of data fication manuals(Cohenand Serjeantson 1996; related to the Romancemetery, although muchof Hillson 1996a; Schmid 1972). Afull methodology this material derivedfrom grave backfills and may for analysis can be foundinthe site archive. be accidental inclusions. The conditionofthe assemblage was variable, Results ranging from verygood to poor. The identified faunal assemblage was dominated by domestic The assemblage comprised 883 hand collected bone mammalsincluding cattle, sheep,sheep/goat and fragments, dated from the Romantothe modern pig. No bones were identified as goat. Equid periods (Table 5.7), plus 581 fragments recovered (probably horse) was also identified in the Roman from sieved residues (Table 5.8). In addition to these, and post-medieval assemblages, and dog in the 540 fragments (2.7 kg) of crematedbonewere Romanand medieval assemblages.Domestic fowl recoveredfrom cremationburialdeposits (Table was prevalent in the Romanassemblage, but also

Table 5.7 The hand collected animal bone assemblage. *denotesthat NISP count includes at leastone partial skeleton. This table includes some burnt bones.

Species Roman Medieval Post-medieval Modern Total

Cattle 28 11 13 557 Horse 20 525 Dog 11 241* 45 Sheep 112 Sheep/goat813 10 738 Pig 6* 57523 Hare 11 Rabbit 22 Hedgehog 11 Rat 11 Vole 11 Domestic fowl 29 1131 Goose 11 Pigeon 11 Bird 15 116 Bird? 11 Large mammal 78 30 61 17 186 Medium mammal 40 25 30 32 127 Unknown 261 11 49 3324 Total 488 98 183 114 883

119 Life andDeathinaRoman City

Table 5.9 The cremated animal bone assemblage quantified by NISP, MNI, MNT and weight.

Data Cremationburials Inhumation Total

Burial 1196 1209 1766 1767 1768 1769 1720

Total number of fragments (NISP) 16 16 166 10 5325 2540

Total weight (g) 20.8 11.9 86.8 48.6 1.8 118.7 5.8 274.7

pp Pig -

pp p ?pig -

p ?cattle -

pppppp p Medium mammal - ent p ?medium mammal - pres

xa pp p Medium/large mammal - Ta pp ?large mammal -

p ?bird -

p Indeterminate -

Minimum Number of Taxa (MNT) 112212 1-

Minimum Number of Individuals (MNI) 113212 1-

presentinthe medieval and modern assemblages. (see Chapter 6), as no human bonewas recovered. One goosebone was present in the medieval assem- The feature containedaflagon and eight unburnt blage and one pigeon bone in the post-medieval animalbones including aright forelimb pork joint assemblage. Other wild taxa identified include (humerus, radius and ulna) from apig aged between rabbit,rat, mouse and vole in Romandeposits. 12 and 42 months old at death.The context also included acattle calcanuem and three medium Animal bone from cremation burials mammal sizedbone fragments. Cremated animal bone was recovered from four of Material from grave back-fills the five urned cremation burialsand two unurned cremation burials(Table 5.9). Full details of these Animal bone was recovered from the back-fills of 21 assemblages can be found within the cremation graves. The majorityofthis assemblage is unlikely to records in Appendix 4. comprise grave goodsbut rather residual material incorporated into grave back-fills accidentally. This material is not considered further here. No burnt Animal bone from inhumation burials bones were recovered and only one bone exhibited Faunal remains, all comprisingdomestic fowl butchery marks (see inhumation 1288 above). skeletons,were deposited as grave goods with three burials(Table 5.7). In burial 1505 the domestic fowl Animal bone from the mass grave skeleton was recoveredfrom within avessel in the grave.The fowl in 1288 exhibited asmall cut mark The massgrave contained two groups of small- on its humerus. The third domesticfowl skeleton foundanimal bone. Small-find sf 340 comprised a was recoveredfrom the backfill of inhumation 1759, fragmented sheep/goat left femur and tibia, acattle which may have been acenotaph rather than a carpaland some indeterminate bonefragments. The normal burial.Perhaps unsurprisinglywithsuch a sheep or goat bones may have been depositedwhile small dataset, there is no association of the fowl articulated and may represent ameat joint or grave goodswith aparticular demographic profile. butchery/food debris. The bones were in fair to very poor condition and no butchery marks were observed. Small-find sf 405comprised 24 domestic Possible inhumation burial or ‘memorial’ 1149 fowl bones from an adult cockskeleton. The It is uncertain whether thisfeaturewas agrave or identified skeletal elements comprised pelves, fe- some other form of deliberate ‘memorial’ deposit mora, tibiotarsi and tarsometatarsi (with spurs), a

120 Chapter Five right humerus and radius, aleft carpometacarpus, fied as pig, possible pig, possible cattle and possible the sternum and synsacrum and several vertebrae bird. Further fragmentswere recorded as medium, and foot phalanges. Soil samples from this deposit medium/large or possible large mammal sized. The contained additional bird bones which may have burnt bonegrave assemblages each containthe originatedfrom this skeleton. The burial pit also remainsofone or two taxa. Most contain aminimum contained adog-gnawedmedium mammal long of oneindividual of each taxon with the exception of bone and three indeterminate bonefragments and 1766 which contains the remains of at least two pigs. two cattle carpals, alarge mammal patella and No butcherymarks or pathological lesions were eighteen further domestic fowl bones. Several large notedinthe burntbone assemblages. Given the mammal,medium mammal and indeterminate bone taphonomic pathways of cremationburialsand fragments were recovered from sieved soil samples. problemsofidentification of the highly distorted and fragmented burntbones, the pyre goods can be interpreted as complete pig carcasses in burials 1209, Animalbone from other features 1766 (togetherwith remainsofasecond pig) and An assemblage of 132 fragmentsofanimal bone was 1769, aminimum of apig skull and limbinburial collected fromthe soillayer associatedwith the use 1196, aminimum of amedium mammal skull and of the cemetery (1020/1106/1440) including cattle, limb, with alarge mammal rib in 1767 and amedium sheep/goat,horse and dog. This group included an mammal limb in 1768. As no othermedium articulating horse rightradiusand ulna, left and mammalswere specifically identified in the assem- right femora, aright humerus, left and right inno- blage, the medium mammal pyre goodsmay also minates, skull fragment and sevenmaxillary teeth. be pig. While epiphysealfusiondata could indicate that the Animal pyre goods have been recovered from long bones originate fromasingleanimalaged Romancremation burials elsewhere in Britain, the approximately 36–42 months old at death,tooth largestassemblage being that fromthe Eastern attritiondatasuggests the presence of aconsiderably Cemetery of RomanLondon(Barber and Bowsher older horse’s skull. The small size of this assemblage 2000). The frequency of cremationburials containing together with its provenance negates further inter- recognised faunal pyre goodsatRoman cemeteries pretation of animal husbandryatGloucester. tendstobelowerthan that observed at London Road (typically5–47% of burials in cemeteries Discussion where seven or more burialswere analysed for faunal pyre goods:Worley forthcoming a), although Burntanimal bonewas recoveredfrom seven of the 80% of urned burialsatRyknield Street, Wall, nine cremationburials. The amountofburntanimal Staffordshire, contained faunal pyre goods (McKin- bone in each burial varied from 1.8g(5 fragments) ley forthcoming). Many other sitesinclude pig to 118.7 g(325 fragments). Themajority of the pyre goods, although they are generally less com- animalbone fragments were off-white in colour indi- mon than domestic fowl pyre goodsinRoman cating complete calcination. Afew fragments were Britain (Worley forthcoming a). The Severn region less well burntbut this may be of no significance as includes published Romanfaunal pyre good assem- variation in pyrolysis between fragmentsneed not blages from Cowbridge (Parkhouse and Evans 1996) indicate that theyoriginate fromdifferent pyre and Lodge Hill Cemetery, Caerleon (Evans and goods or different stages in the cremationprocess Maynard1997) in south Wales, the Bathgate (Worley forthcoming a). Theanimal remains were all (McWhirr et al. 1982) and OakleyCottage (Reece highly fragmented and exhibited fine fracture lines 1962) cemeteries in Cirencester, and Abonae (Bennett and fissures characteristic of burntbone. No con- 1985) in Bristol.Noneofthese Severn region centricfractures were identified and no conclusions cemeteries exhibits adominance of pig pyre goods can be drawn as to the fleshing levels of the pyre as recorded for the London Road, Gloucester burials. goods when cremated. However, pigs were the most commonmammalian The vastmajority of burntbonefragments were pyre goodsatRomancemeteries in Eastern London unidentifiable but some fragments could be identi- (Barberand Bowsher 2000), Derby Racecourse

Table 5.10 Fish bones recovered from soil samples taken from burials.

Burial Context Sample/spit Context type Quantity Taxon Element

1767 1062 <213>spit 5Cremationdeposit 1pike vertebra 1767 1062 <212>spit 4Cremationdeposit 1?clupeid vertebra 1196 1190 <59> Cremationbackﬁll 2clupeid vertebra 1209 1198 <68> Cremationbackﬁll 2clupeid vertebra 1114 1123 <22> Inhumation: chest area 1Eel vertebra 1131 1132 <12> Inhumation: abdomen 1clupeid vertebra 1360 1361 <172> Inhumation: abdomen 1clupeid vertebra

121 Life andDeathinaRoman City

(Wheeler 1985), St StephensinHertfordshire (author’s THE FISHBONE observation) and Ryknield Street, Wall (Worley by Rebecca Nicholson forthcoming b). Animal grave goodsare also often recovered from Asmall number of fish bones were recovered from inhumationsinRomanBritain. Birds,particularly several soil residues afterprocessing soil samples domestic fowl,are often identified (Lauwerier 1993; from the cremations and inhumations (Table 5.10). In Philpot 1991). Both domesticfowl and pigs provided all cases the bones were unburnt. The bones from asource of food in RomanBritain, consumptionof cremation burial 1767 were recovered from soilsam- pork and chicken may even have been relatively ples taken from the contents of the cinerary urn and high statusand an expression of Romanised identity. may have been placed deliberately, but the material However, both species were also associatedwith from cremation burials1196 and 1209 comes from deities and funerary sacrifice in Roman literature the back-fill and may be accidental inclusions.The (although the literary sources refer to practice in bones from inhumation burialswere recovered from Rome rather than Britain).Male domestic fowl were samples taken from the chest and abdomen areas of associated with chthonic deities and are often the skeletons, but whether they were associated illustrated on tombstones (for example at Abonae; directlywith the body or are accidental inclusions Bennett 1985, 62). While the pig and fowl pyre and is uncertain. Speciesrepresented included small grave goods at London Road mayrepresent food clupeids (herring Clupea harengus or sprat Sprattus offerings, their potential symbolic significance sprattus), eel ( Anguilla anguilla)and juvenile pike shouldnot be forgotten. We cannotdetermine the ( Esox lucius). While pike and eel are bothfoundin intention of those arranging the funerals regarding rivers and ponds and could have been fished locally, the animal remainsbut the growing body of evi- the clupeids were most likely to have been salted or dence suggests that the use of pig pyre goodsand pickled fish. Small oily fish such as clupeids were a domestic fowl grave goodsinRomanburials in common ingredient in garum, afermented fish sauce Britain may have conformed to broad traditions. muchfavoured by the Romans.

122 Chapter 6Discussion of the Roman Cemetery

INTRODUCTION truncation mayhave occurred acrossthe remainder of the site due to later cultivation, and may have resulted The site at 120–122 LondonRoad is the first substantial in the loss of an unknown number of shallower area of the Wotton cemetery to be excavatedand burials, with the result that the burialsexcavated may published to modern standards.The excavation be the survivors of an originally much greater comprised an area of 0.1 hectares and resulted in the concentration. Such destructionbylater development recovery of the remains of at least 165 individuals is afeatureofthe Wotton cemetery, and indeed of from graves, cremation burials and the massgrave. Romanurban cemeteries in general.The customary Although the existence of the cemetery has been siting of Romancemeteries aroundthe peripheryof known since the 19th century, discoveries have the urban areahas made them particularlyvulnerable mainly beenpiecemeal, informationderiving from to depredation from the effectsofthe expansion of the observations made during the 19th century develop- towns over the past two centuries. The development ment of the area and from more modern records of 19th century suburbs may have exposed the during small-scaleground disturbances. This has existence and scale of these cemeteries, but it also resulted in aratherunsatisfactory and disjointed inflictedpermanent and widespread damage on hotchpotch of observations, with no overall coher- many of them. ence. The only previous excavation of asubstantial area of the cemetery was carried out at the adjacent property, 124–130 LondonRoad, but the results of this THE DEVELOPMENT OF THE WOTTON investigation have not yet beenpublished (Founda- CEMETERY tions Archaeology 2003). The siting of the cemetery As one of only three coloniae established during the early partofthe Romanperiod in Britain, Roman The locationofthe cemetery at the junction of the Gloucester had an unusualstatus and history, being a main approach roads leadingtoboth the fortress and community of Romancitizens established de novo in a the river crossing is unlikely to be accidental. Roman recently conquered territory. The excavation of part of law dictated that cemeteries shouldbeplaced out- one of its main burial grounds affordsanopportunity side the city limits, and it was customarytolocate to study the effects of the imposition of thisalien them alongside the main roads approaching the community, and the resultant meeting of native settlement. Funerals in the Romanworldwere very and immigrant culturesasexpressed through their public affairs, intended as adisplay of the social funerary practices. The early date forthe first burials, statusofthe deceased and his or her heirs, and it was contemporary with the legionary fortressatnearby also considered important for the well-being of the Kingsholm that preceded the establishment of the spirit of the deceased that the grave should be colonia,has provided arare opportunity to examine located where it would be visible to the most the burial practices of the military community and the passers-by, who could then pay their respects while changesfollowing the transition to civilian occupa- passing. The establishment of aburial ground at a tion. Furthermore, the discoveryofarare mass grave, very visible location in the landscape,onthe crest of possibly representing the victims of an epidemic, has ahillock in the otherwiseflat base of the Severn affordedanopportunity to examineacross-section of Valley, may also have been intended as aphysical the populationselected in all likelihood by chance expression of the permanence of the Romanpresence rather than by rules governing access to the cemetery and their dominance of the surrounding landscape (See Chapter 3and discussion below). and its native population. Over the following three Certain caveats must be borne in mind when centuries the cemetery developed intoanecropolis examining the evidencefromthe cemetery. Obvi- extending for more than 500 malong the road to the ously, the area investigated by the excavation is only a colonia and would have formedadramatic sight for small sample of alarge cemetery, and conclusions visitors approaching the city. based on the analysis of this sample may not hold true for the cemetery as awhole. In particular, it is The growth of the cemetery considered likely that this area of the cemetery was used by arelatively low status section of the The cemetery appears to have been established on population (see below). Even within the areaexca- land previously usedfor farming, as indicated by the vated, survival of gravesmay be only partial: the ploughmarks recorded in the north-eastern part of the northern and western ends of the site have been site. Theearliest burials recorded during the excava- truncated in modern times resultinginthe loss of any tion are two cremation burials containing pottery burialsinthese areas. Some degree of horizontal datable on stylistic grounds to the pre-Flavian period

123 Life andDeathinaRoman City

(AD 49–69) and athird that yielded aradiocarbon from within the Wotton cemetery and it is possible determination of 50 BC–AD70(OxA-16792 cal 2 that at thisstage the cemetery was limited to ahandful sigma). These earliest burials, and six cremation of burialsinasmall areatothe south of the junction of burialsofsimilar date recorded on the adjacent plot Ermin Street with the roadlinking it to the site of a at 124–130 London Road (Foundations Archaeology presumed crossing of the RiverSevernnear the 2003) are contemporary with the use of the fortress at modern city centre (Fig. 6.1; andsee Chapter1, Kingsholm, which is believedtohave beenoccupied Fig. 1.3). from AD 49 until the mid 60s.Their presence here There appears to have beenasubstantial increase suggests that the Wotton cemetery was one of the in both the number and the areaofburialsduring the main burial grounds associatedwith the fortress, late 1st to early 2nd centuryAD. This expansion is along with the eastern partofthe cemetery which later broadly contemporarywiththe occupation of the occupied the Kingsholm area, where asmall number legionary fortressatthe city centre and the establish- of similarly early burials have been found(Rawes ment of the colonia, and it is tempting to attribute it 1991, 231). Alone inhumation burial from 124–130 to adeliberate act of policy, extending the burial London Road was attributed to this period by the groundofthe formerfortresstoserve the needs of excavator on the basis of astratigraphic relationship the new settlement. However, the dating evidence that indicated that it was earlier than alate 1st–early for both the expansion of burials and the founding 2nd century cremation(Foundations Archaeology of the settlement is not sufﬁciently clear-cut to 2003), but its 1st century dateisnot certain,particu- demonstrate this beyond doubt. Burials of this date larly as it contained sherds of 2nd centurypottery. No have beenfound over an area measuring c 375 m otherburials of this early date have been reported in length,extending from 83–89 LondonRoad to

Figure 6.1 The growth of the Wotton cemetery.

124 Chapter Six

St Catherine’s Church at the junction with Ermin cemetery peteredout gradually or came to an abrupt Street (Fig. 6.1). At 124–130 London Road the late 1st halt, but the evidence indicatesthat it persisteduntil to early 2nd century burialsextended further south, at leastthe mid-300s, and in all probability continued suggesting that the cemetery may also have ex- until organised urbanlife withinthe colonia had panded further back from the road at this time. It is largelyceasedaround the end of the 4thcentury or also from thisperiod that the earliest evidence has early 5th centuryAD. been found forburialslocated to the east of Ermin Street, at St Catherine’s Church (Wills 2000, 226) and The layout and organisation of the cemetery 3Denmark Road (Atkins and Garrod 1988, 216; Rawes1992, 221), and on the north side of the road Little evidence survivestoindicate how the ceme- linking the colonia to Ermin Street, at 93 London teries of RomanBritain were managed, but clearly a Road (Heighway 1980, 63). At St Margaret’s Hospi- major urbancemetery such as that at Wotton, tal, agroup of five cremation burials of this date was serving the population of the colonia and possibly arranged around the edge of an areaofmetalling, of the surrounding territorium,would require some and this may be evidence that the formal layout of degreeofcentralised administration. Such manage- the cemetery was established at this time. The four ment would be necessary to ensure that asufficient crouched burialsfrom the present site were also area was available forburials, maintainits upkeep, interred during this timeand are likely to be the administer the allocation of individual plots, possi- remainsofmembers of the native population who bly to providefor the proper carrying out of burial had settledatthe colonia or who had been integrated rites, and perhaps to ensure that only burials of in some way intoits populationbut continued to eligible individuals were allowed withinthe pre- practice their traditional burial customs. scribedarea. Aprofessional class of individuals During the 2nd century cremationwas gradually entrusted with the task of burial is known to have superseded by inhumation as the main funerary rite existedinRome (Bodel2000, 135–44) but it is not throughout the Romanempire (Jones1987, 816), and known whether similar individualsexisted in the majority of the otherwiseundated gravesare Britain. likely to belong broadly within the 2nd–4th centuries The clearestsurviving evidence for the manage- AD. Theareaused for burial appears to have ment of the cemetery lies in its spatial organisation. expanded again during thisperiod, with inhumation The boundaries of the cemetery were not located graveshaving been recorded furtheralong London within the excavation, and indeed no boundary Road toward the colonia,at50–52 London Road and features have beenidentified in other investigations 67 London Road, as well as being foundthroughout within the Wotton cemetery. The provision of clearly the areaalready in use (Fig. 3.1). It is unclear definedphysical boundaries to enclose the area whether the growth of the area usedfor burial given over to burial and separate it from land represents expansion into an area that had been occupied by the livingwas an importantconsidera- reservedfor burialsbut had hitherto been unused, or tion in the Romanworldascontact with the dead the extension of the cemetery into an areaformerly was considered to be rituallypolluting (Toynbee given over to some other use. Hurst (1999b, 121) has 1971, 50–1). Urban cemeteries elsewhere in Britain suggested that an areaofopen ground existed as a have been foundtolie within boundaries defined by buffer between the colonia’s north-westernsuburb ditches, presumably accompanied by banks (eg and the cemetery, but the discovery of the beam slots Poundbury, Farwell and Molleson 1993, fig. 33; Butt of abuilding during an evaluation at 59–65 London Road, Colchester, Crummy et al. 1993, 13). At Road (Atkins and Garrod 1987, 233) may indicate Winchester the boundaries of the cemetery at Lank- that the suburb extended further than he allowed, hills School were defined by both aditch and a and that the expansion of the cemetery in its final hedgeline (Clarke 1979, 109–10). It is likely that phase broughtitalmost intocontact with the extra- similar boundaries existed at Wotton and either mural settlement. await discovery or have been destroyedbythe more It is not possible to establish an accurate datefor recent development of the area. The excavation at the end of the cemetery, although there is evidence 124–130 London Road demonstratedthat burialsdid that it continued in use well into the 4th century. not extendmuch furthersouth than the southern This much is clear from the two sequences of graves edge of the current site (Foundations Archaeology in the south-eastern and western parts of the 2003) and it is likely that the southern boundaryof excavation, bothofwhich started with burials the cemetery lies only ashort distanceaway. accompanied by vessels dating from the 3rd to 4th No evidence was found for the sub-division of the century. In addition to this, the ring and bracelet part of the cemetery within the area of the excavation, worn by the individual buriedingrave 1362 are of but urbancemeteries elsewhere have been found to be 4th century date. Threegravesat124–130 London sub-divided internally by boundaryditches to form Road contained4th centurycoins,with two of these discrete plots or enclosures (eg Barber and Bowsher datingtonoearlier than AD 348, indicating that 2000; Farwelland Molleson1993). However, three burial continued until at leastthe middle of the sides of arectilinear ditched enclosure recorded at century. Due to the small number of datable burials 124–130 London Road may be evidence for such it is not possible to assess whether the use of the enclosures at Wotton. The enclosure appears to have

125 Life andDeathinaRoman City originatedduring the 2nd century as asingle group (Foundations Archaeology 2003). Theapparent boundaryditch,orientedNNW-SSE at rightangles preference for cardinal orientations displayed by to the line of Ermine Street. This was subsequently most of the graves to which a3rd–4th century date superseded by the north-westernside of the enclosure, could be assignedmay suggest an increase in the the ditch of which silted up during the 3rd century. level of organisation of the layout of the cemetery at The enclosure did notmark the edge of the cemetery, this time, although the apparent clusteringofthe as furthergraveswere identified beyond it to the gravesdated to this period in two distinct areas east, but it appears to have beenkept clear of graves, as suggests that thisstill comprised discrete clusters of the only burialsidentified within it were apre-existing burialsdistributed sporadically acrossthe cemetery, cremation burial,a3rd or 4th century grave that perhaps in individual plots,withunusedareas or may post-date the use of the enclosure, and asingle areas of older gravesinbetween. Asimilar shift from undated inhumation. The function of this apparently scattered, haphazardburialstoregular rows of east- empty enclosurewithin the cemetery is uncertain, west orientedgraves also occurred at around this although the identification of gravelsurfaces else- time at the cemetery at Kingsholm (Atkins 1987). where within the cemetery suggests that open areas The burialswere not distributed evenly acrossthe were not uncommon. The excavators suggested that area but were locatedinclusters of varying size and the enclosure may have been aprivateburial plot density, separated by ‘blank’ areas. While the effects containing ahigh status burial, although no evidence of later truncation may have played some part in of such aburial survivedtosupport thisinterpretation creatingthis pattern, it is likely that this is atrue (Foundations Archaeology 2003). Although it would reflection of the layout of the cemetery as similar have beenasignificant feature of thispart of the variations in the density of burials have been cemetery, the enclosure does not seem to have had any observedincemeteries elsewhere in Roman Britain influence on the orientation of nearby graves. (eg Barber and Bowsher 2000, 300–1). This clustering The distribution of burialsisostensibly hapha- may derive from the desire of members of family or zard, although some evidence for more deliberate othergroups to bury their dead in close proximity, organisation is apparent when examined in more and may represent discrete plots within the cemetery detail. The earliest burials are too sparsely scattered that were maintained over along period of time. The to enable anycoherent pattern to be identified, and group of 3rd–4th century burialsinthe south-east asimilar circumstance was recorded at 124–130 part of the excavationhas the appearance of being London Road, where the pre-Flavian cremations buriedinsuch aplot, and the clusters of burials were quite dispersed, although they were generally formedbygraves 1143, 1144,1151, 1167 and 1759, located in the northern partofthe excavation, and graves1228, 1229, 1230, 1761 and 1762 may be nearest to the road linking Ermine Street with the similar groups.Inthe latter group, dating evidence site of the second fortress and colonia.Burialsatthis from 1st to 4th century suggests that these plots may time appear to have beendotted around the area of have persistedover aconsiderable period of time. the cemetery in arelatively disperseddistribution. The areas between graves may have been occupied The layout of the cemetery from the 2nd century by surface features that did not survive to be recorded onwardismoreeasily understood, due to the greater during the excavation, such as trees or hedges, or may number of burials and the replacement of cremation have resulted from the need to maintain openspaces by inhumation as the dominantrite. Graves lie on a in which to conduct funerary rites, to provideaccess to varietyoforientations,but two main trendscan be gravesand to separate burial plots belonging to identified (see also discussion on grave orientation different groups.There is evidence from anumber below). The first of these is anumber of graves of locations within the Wotton cemetery forgravel orientedNE-SW or NW-SE, parallel and at right surfaces, whether these were paths or more substan- angles respectively to the line of the road linking the tial open areas. Awatchingbrief undertaken during colonia to Ermine Street. The aligning of graves the diggingofservice trenches at St Margaret’s according to pre-existing boundaries is acommon Hospital in 1976, for example, recorded an extensive occurrence in Romano-British cemeteries, and this metalled surface with cremations arranged around its road would clearlyhave beenadominant feature in fringes(Heighway 1980, 64) and parts of gravel the layout of the cemetery, forming the route along surfaces associated with the cemetery have also been which funeral processions would have proceeded recorded at 50–52 LondonRoad (Rawes 1983, 192) and from the colonia to the burial site. The eight inhuma- St Catherine’s Church (Wills 2000, 226). The largest tion burials excavatedat3,DenmarkStreetwere open area thusfar identified is the enclosure at similarly alignedparallel and at right angles to the 124–130 London Road, although no evidence was line of the adjacent partofErmin Street. recoveredtoindicate that thisarea was metalled. Afurther group of graves was orientedN-S or The overall pattern of burials at 120–122 London E-W. These included the group of 3rd–4th century Road then is of adiscontinuous and irregular burialstoward the western edge of the distribution distribution of burials, with little evidence for of burials (1264, 1353, 1369, 1370) and most of the any attempttoimposeacontrolled layout. This con- similarly dated graves in the south-eastern partofthe trasts with the evidence from some urbanceme- excavation, as well as anumber of gravesat124–130 teries,such as Lankhills School, Winchester (Clarke London Road that may be an extensionofthe latter 1979, fig. 105) and Poundbury, Dorset(Farwell and

126 Chapter Six

Molleson1993, fig. 10), where the graveswere of society and are geographically biasedtoward the placedinneatly regimented rows. Althoughsome core provincesofItaly and the Mediterranean. low level of organisation can be demonstrated at Neverthelessthey provide an indication of the sort London Road, in the form of small groups of graves of funerary ritual and practices that we mightexpect with common orientations, this seems more likely to to encounter at an urbancemetery such as that at result from the decisions or preferences of the groups Wotton. Obviously this is not intended to suggest conducting individual funerals than from any form that all the individuals buriedwithin the cemetery of central control.This pattern may have arisenfrom necessarily sharedthe same beliefs or were buried sporadic use of thispart of the cemetery over an according to precisely the same rites, as it is clear extended period of time, although it is also possible from the physical evidence that arange of different that the site lay in aperipheral area of the cemetery customs was currentwithinthe cemetery. As a where control was more lax. fortress, colonia and port Gloucester is likely to have The areaexamined is set back somewhat from the had adiverse population with acorresponding frontage of the road,particularly since the Roman varietyoffunerary and burial practices, but at its road was located to the north of the line of the heart it was acommunity of individuals who are presentLondon Road (Fig. 6.1, Spry 1971, 4). Burial likely to have aspired to the values and customs of plots on the road frontage are likely to have been the the classical world. most prestigious, due to their greater visibility and The funeral processbegan whenthe deceased the consequently increased opportunityfor display breathed his or her last breath, and could last fora that they afforded, anditispossible that the area number of days until the remains were ultimately investigated, being more marginal,was of lower interred. During this period, the body was prepared status. This would explain the apparently rather and dressed in clothes appropriate to the status of modesttreatment of the dead,withfew grave goods the individual and was laid out in his or her house and little in the way of exoticmaterial, and in parti- for mourners to pay their respects. On the day of the cular the absence of practices associated with higher funeral the corpse was carried in aprocession to the statusburials elsewhere during the late Roman cemetery, where valedictory speeches may have period, such as lead or stone coffins, mausolea or been addressed to it before the final disposal, plasterburials. Although the presenceoftombstones although it is uncertain howcommon the latter to mark at leasttwo graveshints at more provision practice was outside aristocraticcircles. During the for these burials, these date from the military phase early Romanperiod the body would then be and may not be relevant to the status of this part of cremated and the cremated bone buried, sometimes the cemetery subsequently. Thelow statusattached in aceramic or glass vessel, but during the 2nd to this area may also explain why the massgrave centurycremationdeclined in popularity throughout was dug here. the Romanworldand was superseded by inhuma- Mausolea have been identified within the Wotton tion as the main burial rite(Jones 1987, 816). cemetery at 50–52 London Road (Rawes 1983), Much of the funerary process will have left no St. Margaret’sHospital (Heighway 1980, 63–4; archaeologically identifiable evidence. Clearly the Rawes1993, 224) and 3DenmarkRoad (Rawes lying-in-state and the funeral procession took place 1992, 221) and may representburials of amore high away from the place of burial,and even if artefacts statuspartofthe population. In additiontothis, buriedwiththe body or otherwisedepositedatthe burialshave also beendiscovered accompanied by site of the grave had been used during these parts of multipleancillary vessels at 3Denmark Road (ibid., the funerary processitwould not be possible to 221) andStCatherine’s Church (Wills 2000, 226), and establish this. This is particularlyunfortunate as accompanied by ceramic lamps at St Margaret’s these may have been the most importantparts of the Hospital (Rawes1993, 224) suggesting that awider funeral, with the act of burial itself being arelatively range of funerary rites was practised by the popu- minor element. Certainlythe lying-in-stateand lation of the colonia than was evident within the area procession would have provided an opportunity excavated. The burials in thisarea may therefore not for displays of grief and demonstrations of the be characteristic of the wholeofthe population of wealthand status of the deceased to afar larger the colonia,orrepresentative of the entirerange of audience than couldbeaccommodated at the funerary customs practised within the cemetery, but graveside. It is importanttherefore to bear in mind perhaps represent the burials of only alower stratum that the evidence recovered from the excavation of of society. the graveislikely to relate only to the final act of the funerary process,and to avoid assumptions that this was the focal part of the funeral. FUNERARY RITES

The funerary process Cremation burials Romantraditionsregarding the treatment and The pyre disposal of the dead are preserved in numerous written, epigraphic and iconographic sources, most Cremation burials during the Romanperiod took of which refer to the practices of the upper echelons one of two forms: in situ cremation, in which the

127 Life andDeathinaRoman City body was burnt at the place of burial with apit dug are probablyalso the remainsofpigs (See Worley, beneath the pyre to receive the remains (a bustum Chapter5). The significance of these offerings is burial), or cremation at alocation reserved for this difficult to ascertain. In particularitisuncertain activity(an ustrinum), with the ashes subsequently whether theywere intended as food offerings forthe buriedelsewhere. All eight cremation burialsdis- deceased, or as offerings to the chthonic deities covered were of the latter type, representing burials intended to ease the individual’s passage into the of burntremains collected fromapyre. No physical underworld. In eithercase, the selection of aspecies evidence for ustrina or dumpsofpyre debris were whose consumption may have been associated with excavated, andnone has been identified elsewhere high status would have been appropriate. The within the Wotton cemetery, although clearly such possible presence of the entire carcass clearly indi- installations can be assumedtohave existedsome- cates that the offerings were not limited to joints of where in the vicinity.Possible evidence for the use of meat, and may represent the remainsofanimals such sites was recovered from cremation burial 1767. sacrificedaspart of the funerary ritual. The deposit of crematedbone in this grave contained Apart from the animal remains very few pyre three mandibular condyles from the right sideofthe goods were identified from the cremation burials, mandible (see Chapter3), and although thismay although this maypartly be an accident of the represent the deliberate inclusion of three indivi- processofcollecting cremation remainsfor burial as duals in asingle burial, it is equally possiblethat it there is no reason to believe that the inclusion of such resultsfrom the accidental incorporation of remains items in the burial deposit was done intentionally. from individuals cremated previouslyonthe same The remains of glass unguent bottles were found in ustrinum.The absence of direct evidence for pyre three cremations, and there were two such vessels in sites is particularlyunfortunate as items used in feature1352 which dated from the late 1st-early2nd rituals conducted at the pyre may have been left on centuryand would have been contemporary with the pyre site itself (Polfer 2000). However, evidence the use of cremationinthe cemetery (See Cool, for funerary rites conducted at the pyre can be Chapter5). Similar vessels have been recorded extractedfrom material from the pyre incorporated, elsewhere within the Wotton cemetery at 124–130 whether deliberatelyoraccidentally, into the sub- London Road (Foundations Archaeology 2003) and 3 sequent burial deposit (Pearce1998). DenmarkRoad (Rawes 1992, 221). These sorts of Thequantity of fuel required to cremate ahuman bottles were commonly usedfor holding oil for body represented asignificant investment of re- bathing,and presumably thosefound with these sources, and the burning of the corpse would have cremation burials had beenusedtocleanse the body provided aconsiderable spectaclewithin an impor- before cremation. It is possible that the deposition of tant and dramaticpartofthe funeral. Thecolour of these, and otheritems, involveddeliberately throw- the bone recovered from the cremationburialswas ing them onto the pyre in order to smash them, predominantly white, reflecting successful cremation whether to symbolically kill them and thus trans- at atemperature over 600–700– C, although limited form them into the same state as the deceased, or to incomplete burningwas indicated by the presence of commit them to the deceased by placing them some brown/black/blue fragments (See Chapter3 beyond the use of the living. In relation to this it is for discussion on efficiency of cremation). This interesting to note that Wigg (1993) has suggested would suggest that although some considerable that material used in funerary feasts was deliberately pains were taken to ensure thorough burning of the destroyed, and has identified anumber of deposits at corpse,some body parts escaped complete crema- cemeteries in Britain and elsewhere in the northern tion, probablydepending on the depth of overlying provincesasresultingfrom this practice. The only soft tissueand/or their position on the pyre. evidence for ceramic offerings at the pyre that was Evidence for pyre goods incorporated into preserved in the buriedremainswas asingleburnt cremation burials is generallymorecommon than rimsherd from alocally made jar recoveredfrom the occurrence of grave goodsininhumations. This cremation burial 1766. Although so workaday a reflects the central role that the pyre played in vessel may appear out of place in the context of cremation rites, as the main focus for the dedication funerary ritual, it is likely to be the contents rather and destruction of objects (Barberand Bowsher 2000, than the container that were of significance. 308). The most common pyre goods recorded at The pyre would have had to be left for some time 120–122 LondonRoad were animal remains, which to cool before the cremated bones couldbecollected were recoveredfrom six of the nine cremation for burial—perhaps it was left overnight and the burials. Thequantity of the animal remains placed funeral party, or some partofit, returned the on the pyre is difficult to establish due to the partial following day for the final act of the funeral. nature of the collection of debris from the pyre for burial,but it is likely that these remains represent The grave complete pig carcasses in three burials and at least part of apig carcassinburial 1196.The evidence The weights of the deposits of cremated bone demonstrates aclear preference for pigasapyre recoveredfrom the cremation burialswere consis- offering, particularly as the bones from medium tentlyless than the average weightof c 1000–3600 g sized mammals recovered from two further burials that would be expectedfrom the cremationofa

128 Chapter Six complete human body (see Chapter 3for details). Ancillaryvessels were placedwith four of the six This may partly be explained by the truncation urned cremationburials(1196, 1209, 1266, 1767), and which had clearly affected most of the features, exclusively comprised flagonsinoxidised fabrics. removing the upper parts of the cinerary urns in Threeofthese burials contained asingleancillary burials1196, 1209 and 1227 and reducing the three vessel and one (1196) was buriedwith two such unurned cremationburialstodepths of no more than vessels.Both cinerary urns and ancillary vessels 70 mm. However, the two deposits of crematedbone were sourced locally, and there is some evidence that that had not been truncated were similarly under- they may have been re-used rather than made weight. Deposit 1062 was more substantial but it specifically for deposition with burials. All are forms may represent the combined remains of three adults that are commonindomestic assemblages in (see above). Thelow weights forthese untruncated Gloucester, and cinerary urn 1061, in burial 1767, bone deposits suggest that the apparent under- had tracesofathin white calcareous deposit on the representation of boneinall nine cremationburials interior of the base sherds that may be the result of may be aproduct of funerary practice ratherthan of prior use. These vessels were clearly notselectedfor taphonomy. McKinley (1993b, 41–3) has suggested deposition with burialsdue to any intrinsic value, that all cremation deposits are essentiallytoken, and and were presumably viewed as being as utilitarian the contemporary sourcesindicate that the burial of in this context as in more domestic circumstances, only aportion of the corpse was arequirementof serving simply as containers for offeringstoaccom- cremation ritual (Toynbee 1971, 49). There is, then, pany the dead. The preference for flagonssuggests evidence that the deposits of crematedboneburied that these offerings were most commonly liquid, at 120–122 LondonRoad result from the collection whether intended as adrink or libation. Strangely, from the pyre of only aportion of the crematedbone the cinerary urns and ancillary vessels in burials as atoken to symbolically representthe entireindi- 1266 and 1767 lay on their sides, having presumably vidual. Evidence has beenrecorded at othersites been knocked over during the funeral or the back- for the disposal of the remainder of the debrisfrom filling of the grave pit and suggesting that little care the pyre in avariety of locations including grave was taken to avoid or correct such accidents. back-fills, pre-existing features, deliberately exca- The practice of using greyware jars as cinerary vated features and as surface spreads (McKinley urns,accompanied by one, or in one case two vessels 2000a, 41–2), but no such remains were identified in acontrasting colour was also found in six of the 17 here. cremation burials excavatedat124–130 London Six of the ninecremation burialsrecorded in the Road (Timby forthcoming).The coloursofthese course of the excavation were buried in ceramic vessels may to some extent reflect the general trend containers (1196, 1209, 1227, 1266, 1766, 1767). This in RomanBritain forusing dark fabrics for cooking appears to have beenthe standard form of cremation vessels,suitable for containing crematedremains, burial at Wotton, although cremationburials in and oxidised fabrics for vessels used in serving and glass vessels have been recorded at St. Margaret’s consumption, such as the flagons selected here to Hospital. At both this site and at St. Catherine’s accompany the burials. However, the consistency of Church,cremation burialsplacedwithinurns were the pattern at Wotton raises the possibility that a furthercontainedinsidestone cists (Heighway 1980, standardised package of vessels was being usedfor 63–4). Apart from the three examples excavated here, deposition in burials, perhaps resulting from the only two examples of cremation deposits without employment of professionalundertakers or repre- containers have been recorded within the Wotton senting astandard burial rite provided by aburial cemetery, at 124–130 London Road (Foundations club (Biddulph2005, 37). Membership of such a Archaeology 2003) and St. Margaret’s Hospital club was ameansbywhich an individual could (Heighway 1980, 64). However, it is possiblethat guarantee an appropriate funeral, in returnfor the similar features have simply not been noticed else- payment of aregular subscription during life. where due to the nature of the investigations. Cremation burialselsewhere within the Wotton Five of the six cinerary urns at 120–122 London cemetery display greater numbers and types of Road are locally made jars in agrey fabric, and there vessels,anexceptional example including aburial is some doubt as to whether the only exception, at St. Catherine’s Church that had beenburied in a burial 1227, is in fact agrave (see below). It is large pear-shaped storage vessel accompanied by possible that this consistency in the type of vessel three pear-shapes jars and two fineware cups, as used represents adeliberate choice.Similar vessels well as asamian dish placedwithin the urn (Wills were usedfor cremation burialsat124–130 London 2000, 226) Aburial at 3DenmarkRoad in alarge Road (Timby forthcoming),and although urns double-handledwine flagon was surroundedbyan from agreater range of sources have been recorded assemblage of ten samian vessels including two elsewhere within the Wotton cemetery, the prefer- small and two large form 27 cups, three small form ence for vessels in reduced fabrics is still common, 35 dishes, and two large and one small form 18/31 with Black Burnished ware usedat124–130 bowls, with an additional form 27 cup insidethe urn London Road (ibid.), 83–89 LondonRoad (Sermon (Rawes1992, 221). 2003, 55) and 93–95London Road (Heighway The only non-ceramic grave goods recovered from 1980, 63). cremation burialswere an assemblage of 27 hobnails

129 Life andDeathinaRoman City scattered throughout burial 1769 and four hobnails burial during the late 1st century or early 3rd century from burial 1209, including one that was inside the (see Chapter 2for radiocarbon dates). The crouched cinerary urn.Due to the corroded conditionofthe burialsare best seen as members of the native hobnails it is not possible to determine whether community, whose poorlyunderstood funerary they are burnt, and so it is uncertain whether they traditions seem to have comprised mainly excarna- accompanied the body on the pyre, perhaps worn, or tion, along with some evidence for crouched burial were deliberatelyaddedtothe deposit of burntbone (Moore 2006, 111). Two examples of this practice are prior to burial. In eithercase, the symbolism of the known elsewhere in the Wotton cemetery, from provision of footwear is presumably associatedwith 124–130 London Road (Foundations Archaeology the belief that the deceased required footwear in 2003, table 2), and St. Margaret’s Hospital (Heigh- order to undertake ajourney to the afterlife (Philpott way 1980, 64). Slightly further afieldseven crouched 1991, 173). Thepossibility that organicmaterials burialswere foundamongst an assemblage of 125 were placed in the grave which have not survivedis burialsatKingsholm (Garrod and Heighway 1984, raised by the positions of the cinerary urn and 68). Asimilar ‘native’ interpretation mightalso be ancillary vessel in burial 1266, which were placedto possible for the otherearly burials, as extended one sideofthe grave pit, leaving alarge space in the inhumationshave been found in a2nd century rural south-westernhalf where such items could have cemetery associated with asmall native farmstead at been placed. Other grave goods recorded from Hucclecote, 5kmeast of the colonia,where they cremation burials within the Wotton cemetery surely representmembers of the local population includeceramic lamps at 124–130 London Road (Thomas et al. 2003, 65). These apparently early (Foundations Archaeology 2003) and St. Margaret’s extended burials may be part of alocal or regional Hospital (Heighway 1980, 64), coins at St. Margaret’s native trend for such aburialritepre-dating the Hospital (Heighway 1980, 64), 3Denmark Road widespread adoption of inhumation, although if so (Rawes1992, 221) and St. Catherine’s Church (Wills this tradition was clearly very circumscribed, as 2000, 226), and ametal pin and part of acopper Booth’s (2001) survey of the Oxfordshire Roman spoonat124–130 London Road (Foundations cemetery evidence identified no inhumations dated Archaeology 2003). to before the 4th century. Additional sherds of pottery were recoveredfrom The precise date of the widespread adoptionof the back-fills of four cremationburials (1196, 1209, inhumation at LondonRoad is uncertain due to the 1227, 1766). These assemblages were more varied small number of burialsaccompanied by datable than the pottery placedinthe burials, including grave goods. No grave goods were founddating products of the SevernValley industry in addition to from the 2nd century,either at 120–122 London the more local wares. The inclusion of such sherds Road or in previous excavations within the cemetery, within cremation burials has also been notedat but it is unclear whether this is because few burials several sites within the Wotton cemetery (eg Sermon were made during this period or if the practice of 2003, 55) and the likelihood is that at least some of depositing such items was not in vogue at this time. this material was incorporated into the graves Inhumation was certainly the only burial practice accidentally during back-filling. However, the pre- identified during the 3rd and 4th centuries, when a sence of the smashed remains of two wholeflagons number of burials were accompanied by grave in aburial at 3Denmark Road (Rawes 1992, 221) goods. suggests that some element of deliberate deposition may also be involved. This may be part of some Placingthe body ritual associated with the closing of the grave, possibly representing aseparate form of offering The inhumations were placedingrave pits that were from thoseaccompanying the burial, or involving generally rectangular or sub-rectangular in plan and the deliberate destruction of vessels rendered of asize to accommodate the body of the deceased. ‘unclean’ by their use in the funeral rites. The choice of orientation for burial provides an opportunity for the expression of cosmological beliefs,asisfound in the more recent Christian Inhumation burials tradition of orienting burials west-east. Unlike some Inhumation replaced cremation as the principal urbancemeteries from the Romanperiod (eg Clarke meansofdisposing of the dead throughout the 1979; Farwelland Molleson 1993) the burials RomanEmpire between the late 2nd centuryand the recorded during the excavationdid not conform to late 3rd century (Jones 1987, 816), but there is some asingleconsistent orientation, although this may evidence that inhumation was practised at London partly be due to the gravesrepresenting apalimpsest Road at an earlier date. Five inhumationswere of burials madeover aconsiderable length of time. attributed to the late 1st–early 2nd century, includ- Neverthelesssome patterns may be detectable. ing aproneburial and four crouched burials. In Just over half the burials(50.8%)lay on cardinal addition, three extended inhumation burials pro- orientations, divided almost equally between north- duced radiocarbon determinations centred on the south andeast-west. Howeverthere was little 2nd century, although the date rangesfor the latter discernible patterning in the orientation of the bodies are admittedly rather large and do not discount themselves in these groups, with at least five

130 Chapter Six

The majorityofthe burials were placedinasupine posture, possiblydue to arequirementfor the body to be facing the mourners for the speeches which mourners addressed to the corpse (Barberand Bowsher 2000, 306). Within the broadly uniform adoption of this posture, the position of the arms was somewhat more variable. The arrangement of the arms of those individuals that were sufficiently well preservedfor this information to be established did not appear to be random,but conformedtoarestricted range of positions (see Chapter2for details).These arrangements were presumably deliberately selected when the corpse was preparedfor burial,but there appeared to be little consistency. Of the 20 adult burialsinwhich botharms were preserved eight of the 16 possible combinations of arm positions were recorded, the most frequently encountered being with bothhands resting on the pelvis (five instances), both beside the body (four instances) and with the Figure 6.2 Orientation of inhumation burials, according left arm beside the body and the rightflexed with the to the position of the head (eachsegment represents an hand lying on the pelvis (four instances). No other individual inhumation burial). combination was represented by more than two examples, and asymmetrical arrangementswere slightlymore common than symmetrical ones. There was no apparent correlation between arm position examples recorded of each of the four possible and other aspects of the burial.The three most variations, although the placing of the headatthe frequent combinations of arm position are each east and north ends of the graves was the most divided equally between the sexes, excluding uncommon with 11 and 10 examplesrespectively (Fig. sexed individuals, and do not correlate with specific 6.2). The current evidence is insufficient to establish age groups. Unfortunately toofew gravescould be whether these alignmentscontainedsome form of dated sufficiently closely to establish whether there symbolism or simply resulted from an attemptto was any chronological patterning. The posture in imposeregularity to the arrangement of the ceme- which the body was placed in the grave was amatter tery throughconsistency of orientation. Of the of deliberate choice on the part of thoseconducting remaining graves, 21 (34.4%) were orientedNW-SE the burial (at least for non-shrouded bodies), and or NE-SW, that is parallel or at aright angle to the affordedanopportunity for subtle display of the line of the adjacent road,and it is likely that the statusand identity of the individual, and the alignments of these burials were influenced primar- positions of the arms may have played arole in ily by the physical boundaries of the cemetery and this. Unfortunately the apparent lack of coherent had no symbolic significance. As the majorityof patterns in the arrangementsrecorded indicatesthat gravescouldnot be assigned an accurate date it was if this were so it is acode we are unable to read, not possible to analyse changesinorientation over although Barber and Bowsher (2000, 87) have poin- time. ted out that the asymmetric arm positions may mimic Beyond the selectionofthe locationand orienta- the ‘toga position’ of many Romanstatues, with one tion of the grave there is no evidence that the pit arm folded across the waist and the otherfree, and it required further preparation before receivingthe may be that the position of the arms is aconsequence burial.Nevertheless, elsewhere withinthe Wotton of the choice of clothing for the corpse. It may be cemetery athindeposit of carbonised wood and fire- relevant to note that the classical sources specifically reddened soil on the base of the grave of achild state that aRoman citizen would be buried wearing aged c 6years, associated with sherds of aBlack atoga (Toynbee 1971, 44). Burnished ware jar, has been interpreted as the residue of aritual preceding the placing of the body Discrepant burial rites: prone and decapitated (Wills 2000, 226). burials There is evidence that in some cases gravesmay have beendeliberatelyre-opened for the insertion of The burial of six individuals face down in aprone new burials, perhaps due to afamilial or other position represents apractice that was carried out on relationship between the individuals concerned (see aminority of burialsatmany cemeteries of the discussion on grave markers below). Asingle exam- Romanperiod (Philpott 1991, 71–6). The dates of ple of the deliberate insertion of alater burial into an these six burialsspan the period of use of the existing grave was notedat124–130 London Road cemetery, ranging from the late 1st–early 2nd (Foundations Archaeology 2003). century, when burial 1145 was buried accompanied

131 Life andDeathinaRoman City by aring-necked flagon,until the 3rd–4th century, abovethe elbows during machining and conse- when acoin dated to AD 270–95 was incorporated quentlyitisnot possibletobecertain whether they into the back-fill of proneburial 1230and burial 1537 were tied together at the timeofburial,while the was interred in grave 1765 in the group of late evidence from burial 1765 is inconclusive as the arms burialsinthe south-eastern partofthe excavation. of the individual are not sufficiently close together to The remaining threeexamples(1143, 1150, 1234) demonstrate whether they were bound. were undated and consequently it is not possibleto Anumber of possibleexplanations have been demonstrate whether the practice becamemore advancedfor the practice of prone burial, including popular here during the later partofthe Roman the possibility that they are accidental, resulting period, as has been found elsewhere (ibid., 78–9). from undertaker error when dealingwith sealed Although the riteappears to have beenmore coffins or shrouded bodies or asimple lack of care on common on rural sites, some indication of the the part of thoseconducting the burial (McWhirr et proportion of burialsinurban cemeteries positioned al. 1982, 78–81; Philpott 1991,72). Alternatively, the in thisway can be gatheredfrom the 3.3% of the rite may have been intended as an act of disrespect burialsfrom the eastern cemetery of RomanLondon towardthe deceased. Alack of care couldexplain the (Barberand Bowsher 2000, 87). In addition, 3.9% of slightlytwistedposture of skeleton 1127 in burial the 308 burials at Lankhills School, Winchester for 1145, and it is possible that the burial in grave 1230 which aposture could be confidently established was shrouded, but generally the proneburials were prone(Clarke 1979, table 10), and asingle appear to have been placed carefully, lying in fully example was excavated at Poundbury, Dorchester extended postures in graves no different to those (Farwell and Molleson 1993, 65). Afurther six prone dug for supine burials. The positions of the arms are burialswere found during the excavation at 124–130 varied, but no more so than those of the supine London Road,including two from the preceding burials, and there are no examples with the hapha- evaluation phase (Foundations Archaeology 2003), zard, sprawledposturestobeexpectedfrom bodies giving acombined figure of 12 proneburialsfor the carelessly thrownintothe grave,and demonstrated two areas. This is arather high figure,representing on this site by the bodies in the mass grave. The 11.2% of the 107 burials for which aposture couldbe apparently deliberate placing of the body, and the established, although 8.8% of the burialsexcavated provision of acoffin in grave 1765 and pottery grave at the Bath Gate cemetery, Cirencester, were prone, goods for this and burial 1145, suggest that the majo- and it is possible that the practice was more common rity of the prone burialswere interred with some in thisregion than elsewhere. Alternatively, the high degreeofcare, and that theyrepresent adeliberate proportion of proneburials recorded in the two burial rite used for selected individuals. excavations on London Road could be afunction of Afear of the potentially malign influence that the the locations of the two excavated areas towardthe dead could have on the world of the living was a southern marginofthe cemetery, as it has been concernthat shaped many aspects of funerary suggested that discrepant burialsare often distrib- practice in the Roman world (Toynbee 1971, 33–9). uted toward the edgesofthe areaofburial (Philpott Esmonde Cleary(2000, 137–8) has suggested that the 1991, 73). Although justover halfthe inhumation practice of enclosingcemeteries within physical burialslay on cardinal orientations none of the prone boundaries was intended to prevent the dead from burialswas on these alignments, but it remains wandering. Prone burial may have been one measure uncertain as to whether thisissignificant. taken to prevent this in the case of particular indi- The remainsofthe individuals buried in this viduals, intended to prevent the deceased from posture provide little evidence to explain why they escaping from the grave,whether bodily or in spirit were selected forthistreatment. They range in age form. Theminority of the buried population who from an adolescent male (1112, grave 1143) to a were selectedfor prone burial may thus be the mature female (1103, grave 1150), and although remainsofindividuals for whom this was thought to femalesoutnumber males by four to two, abias also be aparticular concern, perhaps due eithertothe seen at the eastern cemetery of London(Barber and circumstances, timing or nature of the particular Bowsher 2000, 87), the numbers of proneburials individual’s death or to their being outcastsorin from both sites are not large enoughtobeconsidered some way marginalised by societyduring life. statistically significant. At 124–130 London Road the Perhaps they were individuals who, though their sexes were represented equally and the age range appearance or behaviour, were feared or viewed as was even greater,the assemblage comprising two aberrant or unusual, but notsofar beyond the pale juveniles andfour adults aged 40 or over (Founda- as to be denied aplace in the community’s burial tions Archaeology 2003). ground. If this is correct, then the marginal status of The individuals buried proneinburials1143 and these individualsinlife may be mirroredbytheir 1765 were bothinterred with their hands behind burial in amarginal part of the cemetery away from their backs, and it is possible that they may have the more prestigious area on the road frontage. been bound, indicating some element of coercion, The individual in grave 1234 had been decapitated whether associated with sacrifice or judicial execu- in addition to being buriedface down. Due to the tion. (Philpott 1991, 72). Unfortunately the forearms loss of the upper partofthe body there was no of the individual in grave 1143 were truncated just evidence from this grave to indicate the means by

132 Chapter Six which the head was removed, but evidence from otherRomano-British cemeteries indicates that decapitation usually took place after death:the removal appears most often to have beenperformed from the front, and was accomplished with alevel of precision that would not be possible with asubject whose blood was still ﬂowing (Harman et al. 1981, 166; Philpott 1991, 78). Thereasons for decapitating the corpse are uncertain, although the occurrence of both rites in the same individual suggests that their purposes could overlap. Like proneburial,itmay have been intended to prevent the dead from walking, but it is equally possiblethat the rite was believedtobebeneﬁcial to the deceased, perhaps intended to release the soul and enable it to pass on to the afterlife. It is impossible to be sure whether those individuals who were buried proneordecapitated were affordedthe funerary rites that other individuals were entitled to, but these unusual practices would have had certain implications for some aspects of the funerary process.Atthe very least, placing the body in the grave face down or with the head removed would appear to be inconsistent with addressing valedictory speeches to it, and the opportunity for display of the corpse at the gravewould clearly have been diminished. It is also unknown at whatstagein the funerary processofdecapitatedburials the head was removed. It couldhave beendone at the death bed immediately after death, at the graveside, or even afterthe body was placedinthe grave, and would have introduced an additional elementtothe processofburial, possiblyone that formedthe Plate 6.1Burial 1313 in grave 1315. central act.

That some burials were made without the use of a Containing the body: coffins,cists and shrouds coffinisdemonstratedbybodies lying in positions Groups of nails indicative of the presence of coffins that preclude the presence of acontainer, such as were recorded in situ in atotal of 19 gravesand burial 1313 in grave 1315 which lies in asomewhat smaller numbersofnails that may also indicate the ‘sprawled’ posture and appears to have been formerpresence of coffins were recovered from a deposited rather casually (Plate 6.1). Due to the further21burials. Some of these assemblages may uncertainty as to whether some of the smaller nail comprise material incorporated into the back-fills of assemblages representcoffins it is not possibleto gravesaccidentally and it is thus not possible to establish the number of burialsplaced in nailed establish the exact number of graves that contained coffins and so comparison with othersites would be coffins. However, the presence of nails in 40 of the 66 unproductive. This uncertainty and the large num- inhumation graveswithin the area of the excavation ber of undated graves also hamper any attempt to indicatesthat asubstantial proportion of the graves assesschanges in the prevalence of coffined burials contained burialsincoffins secured by nails, and is over time. consistent with the results of the excavation at The groups of nails from most of the graves were 124–130 London Road, where only five of the small, and none contained definite evidence for an 37 gravesfailedtoproduce nails (Foundations upper layerused to secure the lid, raising the Archaeology 2003). It is also possible that graves possibility that some burials were interred on biers withoutnails may have containedcoffins whose or with open coffins,although coffins could have construction did not require nails, perhaps being been covered with un-nailed boards once the securedbyjointing or wooden pegs.The use of such appropriate ritualswerecomplete. The only coffined techniques in the construction of coffins has been burialswith grave goods were burial 1374, which demonstrated elsewhere by the identification of was accompanied by apottery vessel, and burial gravescontaining coffin stains without nails (Farwell 1360 in grave 1362, which was wearing abracelet and Molleson 1993, 44–5). No coffinstains were and ring. If the vessels were placedduring the identified at London Road, and in this case nothing funeral, or were at least intended to be displayed would survive to indicate the presence of the coffin. during it, the coffins must have beenopenfor at least

133 Life andDeathinaRoman City part of the ritual; perhaps they were open when date, perhaps indicating that the use of shrouds at loweredinto the grave,and the lid only securedonce Wottonwas aphenomenon of the later part of the the coffinwas in place,priortoback-filling. Barber Romanperiod, although the suggestion that any of and Bowsher (2000, 310)have suggested that coffins these bodies were wrapped in shrouds is admittedly may not have been closed until the last possible somewhat speculative.Burial 1151 was interred in a momentinorder to keepthe dead person’s face coffin, but as the gypsum burialsdemonstrate, the visible to the mourners whilethey recited the use of ashroud couldbecomplementary to the valedictory speeches. provision of acoffin ratherthan the two representing Asingleburial (1756) was provided with astone alternative modes of burial. cist formed from five slabs of limestone set on edge to form atrapezoidal box, apracticenot previously Adorning the body observedfor inhumation burialswithin the Wotton cemetery. No capstones were found, but these may Threecorpseswere buried wearing or accompanied have been removed by subsequent truncation or by items of jewellery. It is of course impossible to deliberate robbing. Although examples of cist burial know whether these items were placed on the body are known in the Gloucestershire area from the Iron specifically for burial or whether they were habi- Age the majority of datable instances recorded from tually worn during life, and perhaps at the time of Romancontexts belong to the 4th century(Philpott death.The individual in burial 1362 wearing a 1991, 61–2), and are thus too far separated in time copper alloy bracelet (sf 216) on the left arm could from the pre-Roman examples to argue convincingly not be assignedabiological sex from osteological for continuity. The placing of an inhumation withina evidence, but the wearing of bracelets is normally stone cist may have its origins in use of similar cists associated with female burials. At Lankhills School, to contain cremationburials, examplesofwhich Winchester, the wearing of bracelets appeared to be have been recorded within the Wotton cemetery at particularly associated with younger girls (Gowland St Margaret’sHospital (Heighway 1980, 63) and 2001, 162), but the individual at London Road was St Catherine’s church (Heighway 1980, 64). Burial aged 18–25 yearsand may be partofalocal tradition within astone cist may have been regarded as of adultwomen wearing such items, as the three equivalent to the provision of acoffin, albeit in a individuals discovered previously at Wotton asso- more robust material, but why this form of burial ciated with bracelets were all adults (see Cool, was chosen for aspecific individual is uncertain. It is Chapter5). This burial also containedaring or possible that the cist was intended to provide a earring(sf 219), although it is uncertain from the chamberfor the spirit of the deceased to inhabit,asit excavation records whether thisitem was worn at was popularlybelieved that the spirit remained at the time of burial or simply placed with the corpse. the place of burial (Toynbee 1971, 37–8). Alterna- In additiontothis burial, an adult of indeterminate tively the intention may have been to protect the sex buried in grave 1505 was buried wearing a body, as the late dateofmostburials of this sort copper alloy ring on the third or fourthfinger of the would be contemporary with the increasing influ- left hand and achild aged 5–12 yearsburiedingrave ence of Christianity andits doctrine of bodily 1246 was accompanied by an iron finger ring, minus resurrection, perhaps leadingtoagreater concern its intaglio.The latter item was not worn at the time with ensuring the physical preservation of the body. of burial but may have been deliberately placed as a Evidence forthe use of shrouds or winding-sheets grave good, as iron rings were worn as asign that is unlikelytosurvive in most circumstances due to the wearerwas aRoman citizen (see Cool Chapter the decomposition of the materials used. Some 5). In burial contexts they are commonly lacking the evidence has been preservedinthe form of textile intaglioasthiswas removed and handed on to the impressions in gypsum burialsatYork and Dorche- individual’s heir. ster (Crowfoot1981), but it uncertain howcommon the practice was. If the selectionofclothes in which Accompanyingthe body: food and drink the deceased was dressed for burial was an expression of his or her statuswithin societyashroud Six burialswere provided with ceramic vessels, would clearlynot be desirable, although such which presumably contained items of food or drink. indications of statusmay not have been universal. This total includes burial 1352 which contained no No materialevidence for the use of shrouds was humanremains and may have been acenotaph or foundinthe excavation, but it is possiblethat the memorial rather than agrave,and which is consi- posture of burials in which the arms are tight to the dered in more detail below. None of the burials body andthe legs close together indicates that they contained more than one vessel, and the vessels were wrapped in such away. Seven such burials themselves clearly played no part in displaying the were identified (1146, 1151, 1218, 1230, 1501, 1503, wealthorstatusofthe dead personastheyappear to 1508), including proneburial 1230. It is possible that have been standard domestictypes(see Timby, these examples of discrepantposturesresult from Chapter5), and achipmissingfrom the rim of the undertaker errorscaused by the shroudedbodies. flagon in burial 1145 suggeststhat they maynot All four potentially shrouded burialsfor which a have been new when deposited. It is probable that date could be established were of 3rd–4th century the vessels were purely utilitarian in their role in the

134 Chapter Six funerary rites, and that it was their contents that 1980, 103). Hobnailedfootwear is recorded more were significant. The Black Burnishedware jar commonly on rural sites than in urbancemeteries, placedingrave 1505 contained the partial skeleton albeit with some exceptions (Philpott 1991, 167), of adomestic fowl, and parts of similar birds had possibly because the more rigorous lifestyle of the been placedinburials1374 and 1759.Thiswould agricultural population required more robust foot- have beenaparticularly appropriate offeringto wear. The presence or absence of hobnails in any accompany the dead due to its association with individual grave may therefore be an indicationnot Mercury, who was believed to lead the souls of the of the provision or lack of footwear but an indication dead to the underworld (Philpott 1991, 206). No of the type of footwear worn by the individual evidence survivedfor the contents of the vessels in buriedtherein, and be dependent on his or her the otherburials. The flagonsinburials1145 and occupation or lifestyle ratherthan on liturgical 1243 would have been most appropriate to holding requirements. This suggestion would be consistent liquid,perhaps to quench the thirstduring the with the age range of the individuals provided with journeytothe underworld, although if drinking hobnailed footwear, who are all adults. Allsix vessels were provided to decant the contents into, individuals that could be assigned amoreaccurate then theymust have been of organic material as no age were aged between 18 and 45 years, with only evidence for them has survived. one aged over 35 years, and so would have been in The range of vessel types placed in the graves is the prime of their working lives. The biological sex of limited to locally made flagons, which only appear seven individuals could be established, of whom to have beendeposited with the earlier burials(1145, four were female and three male, but at 124–130 1243), and jars. Ajar produced at the Gloucester London Road males buried with hobnails outnum- kilns was placedwith crouched burial 1334 during bered females by 4to2,giving asmall overall the 1st–early 2nd century and jars from regional majority of males for the combined burial group. sourceswere usedduring the 3rd–4th centuries to The practice of placing items of footwear in graves accompany burials1374 and 1505,reflecting the away from the feet and thus clearlynot worn at the wider range of sourcesofpottery availableduring time of burial, as in burial 1167, indicates that they the later period. were of more than purely mundane significance in a Vesselswerenormally placed near the feet, the funerary context. Presumably the deceased was only exception being proneburial 1145, where the fla- expected to wear these items in the afterlife, or was gon lay beside the head, the inversion of the vessel’s required to undertake ajourneytoreach the afterlife. normal position perhaps mirroring the inversion of Either of these interpretations would imply that the the body. afterlife was envisaged as avery physical reality in which the dead existed in some corporeal form. Accompanyingthe body: footwear Commemorating thedead Items of footwear with hobnailed soles were provided in at least 12 graves, and in six of these The dead were not forgotten once buried, but were the individual was demonstrably wearing the shoes/ commemorated both by the erection of physical sandals at time of burial. Thenumbers of nails from memorials and by anumber of festivals and feasts individual graves varied, presumably representing that were features of the Romancalendar.The most different types of footwear (see Powell, Chapter 5). elaborate form of memorial was the construction of a The largest assemblages, comprising178 hobnails mausoleum to house the grave. No evidence for such from grave 1344and 104 from grave 1369, may structures was foundat120–122 LondonRoad, represent heavily shod boots,but couldalternatively although parts of possibleexamples have been be more than one pair of footwear. The groups of recorded elsewhere within the Wotton cemetery (eg hobnails associatedwith the right (37) and left (32) Heighway 1980, 63–4). It is possible that gully 1358 foot of the burial in grave 1234 and the hobnails formedthe eastern side of afeature enclosingagrave associated with the right (22) and left (13) foot of the or group of graves to the west (see Chapter2and burial in grave 1330 (the latter with afurther 28 below), as similarly shallow gullies of rectilinear plan hobnails unlocated), are consistent with Mills’ (1993, surrounded burials at Lankhills School,Winchester 99) observation that the number of nails per shoe/ and have been interpretedasbedding trenchesfor boot in gravesatPoundbury indicated the presence hedgelines (Clarke 1979, 109). Unfortunately this is of styles of footwear with approximately 10, 35 and not capable of proof as this partofthe site had been 50 hobnails. The smallest group that is accepted as subject to truncation associatedwithanadjacent representing footwear placed in the grave is the 18 track, and no features survive that could be confi- hobnails from grave 1508, which would be consistent dently associated with the gully. with Mills’ most lightly shod category,possiblya light sandal. Those burialslacking hobnails cannot Grave markers be assumed to have been without footwear, as shoes constructed without hobnails are likely to have been The two sculpted tombstones discovered during the common, and formed almosthalf of asample of course of the excavation (see Henig andTomlin, about 147shoes from Billingsgate, London (Rhodes Chapter5)bring the total now known from the

135 Life andDeathinaRoman City

Wottoncemetery to five. Sadly none of these can be Romanprohibition on disturbing burials. It is likely associated with aspecific grave.Itisuncertain how that either the earlier graves in such instances were common the use of stone grave markers was in the not marked, or the markerhad decayed by the cemetery, as many more may have been removed time the later grave was dug. The case of three and broken up during subsequent periods for use as intercutting burials towardthe western edge of the building stone, such as the example incorporated distribution of graves demonstrates that the digging into the chapel of the hospital of St Mary Magdalen of graves in such sequences couldbeseparated by a (Sermon 1995). It may also be significant that three considerable period of time. Theearliest grave in this of the five known tombstones are dedicated to group,grave 1243, contained aburial accompanied soldiers and one of the stones foundat120–122 by apot of Flavian-Trajanic date and yielding a London Road, although dedicated to aslave,has radiocarbon determination of AD 1–130 (NZA27005 been dated on stylistic and epigraphic grounds to cal 2sigma). This was partly truncated at its western the Claudio-Neronian period, contemporary with end by grave 1288, which lay on asomewhat the occupation of the Kingsholm fortress.This different alignment and contained aburial dated admittedly small sample may indicate that stone by radiocarbon to AD 60–240 (NZA 27007 cal 2 memorials were aphenomenon of the military phase sigma). Both graves were subsequently cut by grave of the cemetery that did notcatch on with the civilian 1264, which can tentatively be dated to the late population of the colonia,apattern that Mattingly 3rd–4th century by its shared orientation with the has observed throughout Britain more generally more securely dated graves1370 and1353. Over (Mattingly 2006, 202). such periods of time it is reasonable to presume that Presumably markers in less durablematerials wooden grave markers would decay—though these were more common, and have nowperished. The may nevertheless have been inscribedand paintedin absence of direct evidence for markers in materials the samemanner and to the same standard as their otherthan stone may simply mean that theydid not stone counterparts. penetratedeeply into the groundand that any evidence for their existence has been removed by Mourning and feasts subsequent truncation, although postholes found in proximity to two cremations at 124–130 London The completion of the funerary rites was not the end Road may have held wooden grave markers. That of the obligations of the livingtothe deceased, for graveswere marked by some meansmay be inferred the contemporary sourcesrecord that it was normal from the deliberate locating of gravesinrelationto practice in RomanItaly at least formourners to visit previous burials. Graves 1218, 1288 and 1344share a the grave on several occasions throughout the year common orientation and graves1264, 1370 and 1353 to commemorate the dead (Toynbee 1971, 50–4). may also have beendeliberately placed in an Indeed, it was not unusual for individuals to leave approximate line, bothofwhich would require the money in their wills to be spent on such commem- locations of the earlier gravesinthese groups to have orations (Hopkins 1983, 233). These occasions took been marked in some way. the form both of public festivals,the most important It is also possible that some instances of intercut- of which was the parentalia,afestival of the dead ting gravesmay represent the deliberate re-opening lasting from February 13th to 21st, and more private of the earlier grave,again requiring amarkerto observances such as the funerary feast eaten on the indicate its location. Grave 1722 was dug directly day of the funeral and celebrationsonthe birthday of into the back-fill of grave 1721 and may represent the the deceased. These celebrations usually took the deliberate opening of the earlier grave to receive a form of ameal eaten at the grave-side,aportionof subsequent burial, perhaps suggesting afamilial or which may have been set asidefor the deceased and otherrelationship between the two individuals. A libations poured. Scenes of dining depicted on similar situation may be represented by graves 1283 tombstones, such as Tombstone 1, may refer to such and 1284, whilethe diggingofgraves1228, 1229 and meals,and occasionally graveswereconstructed 1230 in approximately the samelocation, albeit on with holes or pipes through which food and drink slightlyvarying orientations, couldalso represent could be physically poured to the deceased, such as deliberate re-use of asingleburialplot. In the latter an example from Caerleon with alead pipe (Toynbee group,the burial in the earliest grave (1228) produ- 1971, 61 and plate 14). ced aradiocarbon determination of AD 70–240 (NZA At the Wotton cemetery these commemorative 27006 cal 2sigma) and the latest grave (1230) was meals may be the source of the pottery and animal dated to no earlier than AD 270 by acoin in the back- bone in the cemetery soiland incorporated inciden- fill, suggesting that the graveswere dug over some tally into gravesduring back-filling, as well as the fish considerable length of time, and that in order for remainswhich may derive fromthe consump- gravestobedeliberately placed at the locations of tion during these meals of garum sauce.The makeup existing burials the earlier gravesmust have been of these assemblages is more varied than that found marked. within the graves, featuringquantities of Severn The majority of occurrences of graves intercutting Valley ware and Dorset Black Burnished ware,small with earlier burialshowever appear to be accidental, amounts of imported fineware,and sherds of am- as onewould expect in view of the customary phora, and is more typical of a‘domestic’ assemblage

136 Chapter Six appropriate to the preparation and consumption of out withinthe cemetery. Furthermore,they form food than of afunerary assemblage. Thesetting aside arather heterogeneous group and may require of areas for such observances at the grave-side may different interpretations, although their locations be one explanation for the open areas and gravelled within the cemetery suggest that some association surfaces recorded at several locations within the with the dead is acommon link between them. cemetery (Heighway 1980, 64; Rawes1983, 192; Wills Feature 1759 has the physical appearance of agrave, 2000, 226). albeit with the body lacking, andmay be acenotaph. Another form of memorial may be represented by Acustom existedinthe Romanworld of creating an the vessel crushed beneath Tombstone 2, which empty grave for individuals whose body was not presumably lay on the ground surface when the availablefor burial,oroccasionally forthose buried tombstone fell and may have been placeddeliber- elsewhere (Toynbee 1971, 54), and this grave-shaped ately as an offering on or beside agrave. Clarke featuremay be such amonument. If the body was (1979, 185) noted otherpossible examplesofsuch not available, the placing withinitofadomestic ‘surface offerings’ at Lankhills, including an occur- fowl, the sacred animal of Mercury in his role as rence where two flagons and apileofthree coins guider of souls to the underworld would be parti- were placedonthe groundsurface aroundthe slight cularlyappropriate to such a‘lost’ soul. If feature moundleft by the back-filling of agrave. 1352 is also acenotaph, the objects in it represent a particularly interesting group. The feature dates from the late 1st–early 2nd century AD,when cre- Cenotaphs, memorials and ritual deposits mationwas still the dominantfunerary rite, and Four features originally recorded as burialscon- contained two flagons, two unguent bottles and a tained no skeletal material (1149, 1352, 1759, 1763). depositofmammal bone. Theinclusion of melted Bone preservation throughout the cemetery was fragments of unguent bottles in cremation burials quite poor, with almost half the skeletons less than 1209 and 1766 suggests that it was usual for these 25% complete, and it is possiblethat some of these items to be deposited on the pyre after use, but the features may have beengraves in which the bone two examples in thisfeature are intact and unburnt, had simply not survived—feature 1763 in particular as is the mammal bone. It is possible that this was the size and shape of achild’s grave and was assemblage represents the packageofitems neces- filled by asteriledeposit of re-depositedclay sary to carry out acremation ceremony, and that containing only sherds of pottery that are probably they have been buriedbecause the cremationcould residual, and it is possiblethat the fragile bones of its not takeplace forsome reason, perhaps becausethe occupant had simply decayed. However, the others body was notavailable. contained more complex deposits and it is possible Pits 1149 and 1306 are less obviously grave-like in that some or all of them were not intended as graves. form, although the vessels placed within themare An inverted flagon had beenplaced at the centre of similar to flagonsplaced as ancillary vessels with rectangular pit 1149, and two flagonsand some contemporary cremations 1145 and 1243. It is animalbonehad been placedatone end of feature possible that these deposits related in some way to 1352, the other half of which was occupied by a the funeral process,representing some form of wooden box,with two glassunguent bottles placed ritualised deposition of items rendered unclean by to one side of it. Whether feature1352 was aburial is use in funerary rites or debris from subsequent very muchopentodebate, as although the box may commemorativefeasts. Or else they may be memor- have contained the remains of achild it would still ials, the pot in pit 1149 perhaps having beeninverted be necessary to explain why the bones of the indi- to make it amoreappropriate offering forthe vidualdecayedwhile those of the animal did not. deceased, or to represent the deceased symbolically. Feature 1759 was the only one of this group of fea- Alternatively, they may nothave been associated tures with the dimensions of an adultgrave, but it is with burials, but representsome form of ritual unlikely that it formerly containedabody that has depositplacedinthe cemetery to communicate with since decayed completely as the partial skeleton of a or to draw on the power of the dead or the chthonic domestic fowl,which would be less robust and more deities.Indeed, anumber of classical authors refer susceptible to decay, was preserved within it. Also, to the popular belief that witches frequentedceme- the featurewas locatedamong acluster of other teries and madeuse of corpses in their practices graveswhich presumably experienced similar ground (Hope2000, 121–2). conditions and in which skeletons were preserved. In asimilar vein, McKinley (2000c, 42–3) has The items placed in these features have the appear- discussed agroup of cremationburials from a ance of deliberate deposits, representing an act of number of late Iron Age and Romanperiod some significance.Tothis group can probably be cemeteries in Britain and France that contain very added pit 1306, aseverely truncated feature within small quantities of crematedboneand has ques- which was the base of aflagon. tioned their identification as graves, suggesting that These anomalous deposits cannot be explained they may more accurately be viewed as memorials. satisfactorily by recourse to arguments based on the However, the difference between these two concepts decay of the boneortruncation of the feature, and is clearly blurred by the token nature of most seem to be the product of deliberate practices carried deposits of cremated remains in the Romanperiod.

137 Life andDeathinaRoman City

Cremation burials1227, 1768 and 1770, which Road were buriedprone, oneofwhich was also contained only 1.5–17.5 gofburntbone, may be decapitated. examples of this practice at the Wotton cemetery. BURIAL AND BELIEF Funerary rites for children The funerary rites of the individualsburiedinthe Children are commonly under-represented in excavated part of the Wotton cemetery were infor- Romano-British urban cemeteries (Philpott 1991, 98) med by the religious and superstitious beliefs of and the Wotton cemetery is no exception.Only seven the people conducting the funerals, and are our individuals under the age of 12 yearswererecorded best evidence for the beliefs of the populationof at 120–122 LondonRoad, with afurther three aged the colonia regarding the fate awaiting the dead. 13–17 years, and there is evidence that the younger However, the relationship between burial practice group were affordeddifferent funerary rites than and religious belief is not astraightforward one. those used for adults. Practice does not necessarily implybelief, and may Three of the four burialsofchildren below the age rather be associatedwithfashion, tradition or group of 12 years were placed in acrouched position rather identity, and it is axiomatic withinthe study of than the supine posture that was the norm for funerary archaeology that such practices may have adolescents and adults.This posture precluded the more to do with the deeds of the living than the dead use of cofﬁns or shrouds for these burials, and may (Parker-Pearson 1999). It can never be provedthat also have other implications regardingthe rites car- the individual within aparticular grave,orthose ried out at burial: if asupine posture was associated conducting the funeral, genuinely ascribed to the with funerary orations addressed to the deceased, beliefs expressed in the ceremony, but it is reason- the use of crouched postures for children may able to infer that these beliefs were currentamong suggest that such speeches were not apart of their the communityatthe time. It is intended here only to burial rites. As such speeches praisedthe life and sketch in outline some of the beliefs indicatedbythe achievements of the deceased, they may have been burial practices recorded in the excavation, although considered inappropriate to the burialsofchildren. If the evidence suggests that the beliefs of individuals the suggested association between the positions of varied considerably, albeit perhaps withinacom- the corpse’s arms and the clothing in which the mon overall cosmology. deceased was dressed for burial is correct, then the Although most of the burials appear to be use of acrouched posture may indicate that children ostensibly verysimilar, details such as the posture were not dressed for burial in the same way as of the body and the provision of grave goods were adults. very variable, indicating that awide range of belief Cremation burial 1266, which dated from the late and superstition was current among the community. 1st to early 2nd century, may provide evidence for The contemporary written and epigraphic sources the different treatment of children in the provision of are concerned largelywith the upper echelons of grave goods.Thisburial,which containedthe societyand are geographically biased towardthe remainsoftwo children aged 5–10 yearsand 10–15 core provinces of Italy and the Mediterranean,but yearsand was the only instance of the cremationof they record awide range of beliefs with no dominant non-adults, containedaset of gaming pieces and orthodoxy (Hopkins 1983, 226–35).The most com- dice, the only cremationburial provided with grave mon beliefs implied by the evidence from the goods otherthan ceramic vessels (See Cool, Chapter5 excavation appear to have been associated with the for discussion of gaming set). Taylor (2001, 125) has survival of the individual after death, and abelief notedthat grave goodsare more commonly pro- that the well-being of the dead couldbeaffected by vided to accompany the burialsofchildren, and it is the actions of the living: hence the need to providea possible that this was intended to placate the spirits place for them, accord themthe proper funerary of individualswho had died before their time: rites, and commemorate themthrough subsequent particularly if the gaming pieces buriedwith crema- feastsand physical memorials. Theprovision of tion burial 1266 were treasured possessions of oneor nourishment in the form of the food and drink held both of the individuals buriedtherein. The provision in the pots placedinthe grave,and of items of of astone cist for the burial of the perinatal footwear, indicate that the dead were believed to individual in grave 1756 maysimilarly be associated have some form of corporeal existence. Furthermore, with the need to quiet the spirit of the deceased, the placing in some graves of personal items such as whether intended to provide achamber for the spirit the gamingsets in cremation burial 1266 and the to inhabit or to contain apotentially troublesome jewellery worn by aminorityofindividuals may spirit.Proneburial is another practice that has been suggest that individual identity was thought to be interpreted as intended to contain an unquiet spirit, maintained in death,asuggestion strengthenedby perhaps as aresult of early death.Thereissome the provision of tombstones or other grave markers evidence that this practice was commoninburials commemoratingthe deceased. of children, as skeleton 1112 in grave 1143 was The provision of these things at the grave suggests an individual aged 13–17 yearsburiedinaprone that the dead were believed to remain at the place of posture, and two juvenile burialsat124–130 London burial,but it is possiblethat this sojourn at the grave

138 Chapter Six was only brief. Thefootwear placedinanumber of although they were described as measuring 9mlong gravesmay implythat the dead undertook ajourney by 4.5 to 6mwide with adepth of 3to3.5 m, and to reach their final destination, and the victuals contained adisordered jumbleofskeletons ‘thrown buriedwith themmay have been intended to nou- into them in large numbers withoutorder or respect, rish them only forthe period before this, or perhaps the feet often higher than the head’ (ibid.). This as supplies for the journey. There is also apossibility description would certainly be consistent with that some goods placedinthe grave or consumedon deposits comparable to that at LondonRoad,but the pyre were intended not for the dead person, otherwise the lack of parallels from other sites but as offeringstothe chthonic deities to ease the inevitably hampers attempts to interpret the circum- passageofthe deceased to the underworld. The stancesinwhich they were made. domestic fowl placedingraves 1374, 1505 and 1759, The aspect of the massgrave that most singles it for example, may have been offeringstoMercury, out from normative Romano-British burial practices who in Romanmythology guided the spirits of the is the overt lack of formalityinthe deposition of the dead to the underworld and whosesacred animal bodies. It was usually considered importantinthe this was. Perhaps the clearest example of agrave Romanworld that the dead should be treated with good placed as such an offeringisthe coin placedin respectand disposedofwith the proper ritual. the hand of an individual buriedat124–130 London Failuretoprovide asuitable resting place or to Road—presumably payment forCharon, the ferry- complete the burial rites adequately couldresult in man who conveyed the dead to the underworld repercussions for both the livingand the dead, and acrossthe RiverStyx. cautionary tales warned that spirits suffering this There are,however, indications that at leastsome fate might not pass to the afterlife and could of the dead were to be feared, and that measures consequentlyhaunt the living (Hope 2000, 106). were required to prevent them from exerting a The rituals surrounding burial were also thought malevolent influence on the affairs of the living. The necessary to remove the ritual pollution associated burial of six individualsface downinaprone with contact with the dead,whileemotional, familial position may have been intended to prevent them or social connectionswith the deceased would have from wandering, or perhaps even to imprison them provided amotivation to disposeofthemina in the grave and deny them access to the under- respectful and compassionate manner. At London world,and the motivebehind the burial of agaming Road, although the burials appear to be of indivi- set with the two young children in cremation burial duals of fairly low statusand are not well furnished 1266 (see Cool, Chapter5)may have been to placate with grave goods, there has nevertheless been some them by the provision of afamiliar and treasured effort madetoensure that the dead were buried possession. Indeed, even the placing of the dead in a appropriately. However, these obligationstoward discrete cemetery beyond the limitsofthe city and the dead appear to have been suspendedinthe case definedbyphysical boundaries indicates adesireto of those individualsinterred in the mass grave,and maintainaphysical separation from the world of the clearly any interpretation of thisdeposit must living (Esmonde Cleary 2000, 137–8). explain this. It has been suggested that burialswithgraves It is known that mass graves, referred to as orientedeast-west,with the head to the west, may puticuli,existedatRome for the disposal of the represent the remainsofChristians,particularly if remainsofthose who were too poor to afford a they dated from the later part of the Romanperiod properburial,orwho left no family or other and were interred in acoffinand without grave associates who could take responsibility forthe goods (Philpott 1991, 239). All five burials on this funeral, and it is possible that the pit at London orientation lackedgrave goods, and three were Road, and thoseatYork, represent asimilar practice. coffined, but as coffins were common and grave Although it couldbeargued that Rome,with its goods rare in the excavation as awhole they are in muchlarger impoverished under-class, was aspecial no way remarkableinthis, and so cannot reasonably case, and no examplesof puticuli are recorded be identified as forming adiscrete group with a outside Rome. It would nevertheless have been distinct burial rite. Indeed, the attributes considered necessary for urbancommunities such as that at to be characteristic of Christian burial rites are in fact Gloucester to have some mechanism for dealingwith typical of most burialsfrom Roman Britain in the 3rd corpsesfor which no-one else took responsibility. and 4th centuries,regardlessofany religious Such ascenario might also explain the demographic affiliations (Barberand Bowsher 2000, 321–2). profileofthe individuals buried in the pit, with its emphasisonyoung adults, perhaps representing individuals drawn to the town by its economicand THE MASS GRAVE commercial opportunities and who died without The discoveryofamass grave of Romandateis having developed the local social connections which almost unparalleledinBritisharchaeology. The only mightotherwise have given them the possibility of recordofpotentially similar features are two pits being afforded amore‘normal’ burial (J Pearce pers discovered in the 1870s during the construction of comm). It should be borne in mind, however, that York railway station (RCHME1962, 79). Unfortu- the skeletal remains provide no evidence that the nately no detailed recordofthese pits was made, individuals buriedinthe mass grave represent a

139 Life andDeathinaRoman City distinct section of the communitythat was poorer or that resulted in the deathsofthe individuals buried of lowerstatus than those buried in the discrete in the mass grave need notbeamong the recorded burials. They were no less well nourishedorless epidemics, it is nevertheless apossibility worthyof healthy, and appear to have derived from asingle consideration. population, exhibiting asimilar range of pathological As no pathologicalevidencewas detected on the conditions and non-metric traits.The resultsofthe skeletal remainsthat might help to establish the analysis of oxygen and strontium isotopesalso identity of the disease that may have caused these suggest that they had asimilar range of geographical deaths, its possible identification with ahistorically origins.Ifthe mass grave does indeed representthe documented epidemic must rely on the correlation of last resting place of the poor and indigent, this might its datewith thoseinthe ancientsources. The grave also explain why it was dug in amarginal part of the contained anumber of artefactsthat help in cemetery, away from the more prestigious areaon establishing adatefor the deposit. The pottery the road frontage. The pits at York were similarly assemblage included the substantial parts of at least described as being located‘on the outskirts’ of the six SouthWest Black Burnishedware jars, some of Railway Station cemetery, which may have been an which displayed right-angled to slightly oblique area of lower status burial relative to the more burnishedlattice indicative of adate during the later prestigious Mount Cemetery(RCHME 1962, 79). 2nd to 3rd century AD.Three brooches were also However, it could be questioned whether a present, probably worn by victims at the time of community the size of the colonia would have had burial,among which were aWroxeter brooch, alarge enough destitute populationtofill the pit, currentfrom the middle of the 2nd century until at unless it was left open for an extended period of time least the end of that century, and aknee brooch also during which the depositaccumulated through the datable to the later 2nd to early 3rd century. repeated deposition of further corpses.The latter Ahairpin found in the upper part of the back-fill scenario is notimpossible, as pits with asimilar was of Crummy Type 2, most common during the function in 18th century Paris are known to have 2nd century. In addition to the artefactual dating remained open for several months or even years evidence, samples from four skeletons were sub- (Arie` s1981, 56). There is, however, some evidence mitted for radiocarbon dating, but unfortunately that the deposit formed over amuchshorter length three of these yielded insufficient collagen for of time, and possiblyrepresents asingleevent analysis. The one successful sample,from skeleton (see Chapters 2and 3). The entangled nature of the 1630, produced adeterminationofAD70–240 remainssuggests that at least some of the corpses (NZA 27008 cal 2sigma). All the datingevidence were deposited together,perhaps dumpedfrom a would thus be compatiblewith adate during the cart or wheelbarrow, rather than accumulating more second halfofthe 2nd century or the early part of the gradually, and the absence of evidence for any of the 3rd century,although the predominantly2nd cen- remainshaving been exposed to sunlightorscaven- tury dates for the use of the Wroxeter brooch and gers would suggest that the pit was not openfor any hairpin maysuggest that adateinthe 3rd centuryis lengthyperiod. less likely. If the massgrave was buriedinasingleevent Of the epidemics recorded in the contemporary rather than accumulating over alongerperiod of sources, the only candidateconsistent with the date time, an alternative interpretation presents itself,in of the mass grave is the AntoninePlague, also which the individuals buried in the pit were the known as the Plague of Galen afterthe physician victims of acatastrophic occurrence that resulted in who reported it. This pestilence first appeared at the the need to dispose of alarge number of corpses with end of AD 165orthe beginning of AD 166in no opportunity for the normal formalities. As Loe et al. Mesopotamiaduring the emperor Verus’ Parthian have discussedelsewhere in this report (Chapter 3) campaign and quickly spread throughout the em- the most likely cause of this would be an epidemic, pire, reaching Rome within the year. The historian as the skeletons revealednoevidence for violent Ammianus Marcellinus records that the epidemic death and the mortality profile of the victims is spread to Gauland the Rhine (23.6.24;Gilliam most consistent with arandomcross-section of the 1961, 232), and there is no reason to suppose that population struck down indiscriminately by some Britain was spared. After the initial outbreak, it catastrophic event, which did not preferentially affect persistedfor many years, with its final and perhaps individuals accordingtoage or sex. most virulentoutbreak not occurring until AD 189. Symptoms described by Galen were ahigh fever, inflammationofthe mouth and throat,thirst, The mass grave and the Antonine Plague diarrhoea and skin eruption, fromwhich modern The ancientsources record anumber of epidemics scholars have been able to concludethat the plague that effected the Romanempire (Cartwright and was probablyanoutbreak of smallpox (Littman and Biddiss 2004, 10–17). Pestilence appears to have been Littman 1973, 252). This is aparticularly acute aperiodicoccurrence and occasionally had adirect disease,epidemics of which may infect 60–80% of impact on historical events, causing the deaths of the population and produce afatality rate of about several emperors, including Lucius Verusand 10% (ibid., 254). Estimates of the effects of the plague MarcusAurelius (ibid., 11–12). Although the disaster on the population of the Romanworldvary, some

140 Chapter Six scholars believing it to have had acatastrophic females, or aratio of 1.1:1. This couldbeinterpreted impact on manpowerthat precipitated the eventual as implying that different areas of the cemetery declineofthe Empire,whileothers have claimedthat were used by distinct groups within the community the resultant excess mortality was no more than 1% with differing balances between the sexes, perhaps (Greenburg 2003, 413). Dio Cassius reported that the representing families with differing traditionsof outbreak in 189 caused as many as 2000 deaths each burial based on an individual’s gender, or in the day in Rome alone,and although Gloucester had a case of the male dominated areas representing plots muchsmaller populationanoutbreak here would used by burial clubs or similar associations. Some certainly have causedanemergency severe enough support for this suggestion is provided by the to force the local authorities to resort to the sort of distribution of the burialstowhichasex could be mass burial of victims that has been identified by the attributed (Fig. 6.3). Agroup of four female burials at excavation. the eastern edge of the site is situatedadjacent to a The correlation of archaeologicalevidence with distinct cluster of female burialsinArea 1ofthe historically recorded events is alwaysproblematic, excavation at 124–130 London Road (Foundations and the identification of the emergency that resulted Archaeology 2003, fig. 6). Meanwhile, the concentra- in the mass grave with ahistorically recorded epi- tion of 3rd–4th century graves in the south-eastern demic can only ever be amatterofspeculation. part of the site,inwhich four of the five sexed burials Although there is some evidence that the deposit are male, is close to an area of similarly late burials may have been created in asingleevent this is not in Area 4ofthe Foundations site that are also certain and other interpretations may be equally predominantly male (ibid.). Nevertheless, the overall valid, particularly the possibility that the pit was imbalancestill persists,asthe combined figuresfrom used forthe disposal of the corpsesofindividuals the two excavations produce an overall male to too poor to afford aproperburial (see discussion female ratio of 1.5:1 from atotal of 87 sexed burials. above). The osteological evidence neither supports The under-representation of females is acom- nor refutes the case for these individualsbeing the monly observed featureofRomano-British ceme- victims of an epidemic, or for the identification of teries.The eastern cemetery of London (1.7:1) any possible epidemic as smallpox (see Chapter 3). It (Barberand Bowsher 2000, 311) and Lankhills is also entirely possible that the deposit represents School,Winchester (1.6:1) (Clarke 1979, 123) have the victims of an otherwise unknown and perhaps produced similar ratiostothat recorded at Wotton, purely local outbreak. However, in view of the un- and the excavation at TrentholmeDrive in the usual nature of the deposit it is aquestion worth Mountcemetery, York produced an even greater pursuing, and the weightofavailableevidence does disparity, with aratio of 3.6:1 (Warwick 1968, 147). It not contradict the suggestion, offered tentatively is unlikely that these figures are fortuitous, the result here, that the individuals buriedinthe massgrave of male and female burialsbeing located in different may have been struckdown as aresult of the parts of the cemetery,asifthis were the caseone Antonine Plague. would expect corresponding female-dominated areas to have been discovered. Theconclusion would appear to be either that some females received different funerary rites which cannot be detected THE PEOPLE OF ROMAN GLOUCESTER archaeologically or that the ratiosobserved in the excavated samplesare atrue reflectionofthe make- The structure of the population up of the population. The remainsfrom the mass The 64 inhumations, at leastten individuals from grave form auseful check on these figures, as these cremation burials, and 91 bodies from the mass individuals are likely to represent asample of the grave represent by far the largestsample of human population that has not been selectedaccording to remainsthat has been excavatedfromthe cemeteries any customs governing entitlement to burial in of RomanGloucester. As such they providean the cemetery, and may therefore provide amore excellent opportunity to investigate the structureand representative cross-section of the population. This health of the individuals buriedinthe part of the assemblage, perhaps surprisingly, exhibits agreater cemetery investigated by the excavation, although it imbalancethan that from the inhumations, with the must be borne in mindthat these individuals may 49 sexed individuals comprising 35 males and 14 have beenofrelatively low status and are not females, or amale to female ratio of 2.5:1. It is necessarily representativeofthe entirecommunity. possible that females lived more sheltered lives and Of the 37 inhumation and cremation burials from the were less exposed to infection, but the evidence from cemetery that could be assignedabiological sex 26 the massgravecertainly does not contradict the were male and 11 female, astriking imbalance suggestion that the populationofthe colonia con- between the sexes that represents amale to female tained considerably more men than women. Wells ratio of 2.4:1. Anote of caution in interpreting these (1982a, 135) has suggested that the imbalance figuresishowever suggested by the corresponding between the sexes at Cirencester may have been statistics from the adjacent excavation at 124–130 due to the prevalence of retired soldiers,and London Road, which yielded amoreequal division although there is no reason to believe that this was between the sexes consisting of 23 males and 21 the case at Cirencester it may nevertheless have been

141 Life andDeathinaRoman City

Figure 6.3 The distribution of inhumation burials to which abiological sex couldbeassigned.

afactor at Gloucester, since the colonia was originally also responsible, and the infanticide of female babies founded as asettlement for veterans and may have must also be considered apossibility (see discussion continued to attract veteransettlement afterits initial below). foundation (see below). However the general nature The age at death proﬁle of the cemetery indicated of the imbalance at numerous towns throughout that adult deaths were highest between the ages of 18 RomanBritain may indicate that otherfactors were and 35 years, representing 36.5% of the individuals

142 Chapter Six from inhumation burials, with sevenindividuals The most obvious expression of social statusfound (11.1%)livinginto the older age category, definedas within Romancemeteries are the mausolea and other over 45 years (see Chapter3formore detail). As with structures or enclosures that were usedtomark the female burials, those of subadults are typically under- burialsofthe wealthy or powerful.Nosuch represented in cemeteries of the Romanperiod structures were present withinthe area investigated (Barberand Bowsher 2000, 311), and this was the case at 120–122 Londonroad, which is likely to be an area at 120–122 London Road, where only nine individuals of the cemetery containing the gravesofindividuals (14.3%)wereaged less than 18 years, and clearlythis is of relatively low status. The only possible evidence not atrue reflection of the ageprofileofthe living for amorehumble form of funerary monument was population. Asimilar situation was recorded at part of agully (1358) that may have enclosed a 124–130 LondonRoad,where 12 burials(16.2%)from burial.However, even this possibility couldnot be the total of 74 excavatedweresubadults. Of the adults confirmed as the gully lay in an areaofthe site that at 120–122 LondonRoad, females appear to have been had been badly effected by truncation, and couldnot more likely to die before the ageof25years, be associatedwith asurviving grave. The only other presumably as aresult of the rigours of childbirth, burialsthat may have beenmarked out as being and at the opposite end of the agerange agreater special or different by virtue of their construction proportion of males (21.7%)than females (9%) lived were the graves of two children: the burial of a beyond 45 years. Theage profileofthe individuals in perinatal individual placed in astone cist (1756),and the mass grave has astriking peak in the 18–25 years the grave of achild aged 5–12 years, the upper part age group,suggesting that the emergency responsible of which was back-filled with limestone rubblethat for this deposit struck downmany individualsintheir may have formedamound defining the grave prime.Aswiththe assemblage from the cemetery, (1772).Ifthese different grave forms were intended males outnumber females in the older category. to express statusthen it surely refers to the status of the family ratherthan that of the children themselves, but it is equally possiblethat the particular Health forms of these gravesare associated with super- Overall, there was nothing remarkable about the stitions regarding the need to contain the spirits of population in general in terms of its health and the deceased rather than with displays of status. physical attributes. An exception is perhaps the Only asmall proportion of the burialswere levels of non-specific infection which seem to be provided with gravegoods, and thoseitems that higher than in other Romano-British populations are present are not of obvious material value. They (see Chapter 3). Patterns of element involvement are not made from valuablematerials, associated in suggest that among the London Road population non-burial contexts with high status, or of exotic systemicdisease, including infection, was prevalent. origin. This would suggest that grave goods and Evidence for injuries sustainedduring life was otherarchaeologically identifiable aspects of the foundinthe form of ante-mortem fractures observed burialswerenot generallyusedtoexpressthe status on the bones of 16 individuals from the cemetery and of the deceased in thispart of the cemetery, perhaps mass grave,all of them male, representing 16.9% of due to the low statusand relative poverty of those all male individuals. Therewere no traumatic buriedhere. It is interesting in this respectthat the injuries recorded on female skeletons. The majority child buried in grave 1246 was accompanied by an of these fractures are likely to have resulted from iron finger ring, which was traditionally worn as a accidents ratherthan inter-personal violence, atrend sign of citizen status, whileithas been suggested that is also seenamong otherassemblages of this (see above) that the asymmetrical positioning of the period (Robertsand Cox 2003, 157) and suggests that arm of some of the burialsmay be associatedwith the towns of RomanBritain were generally not the wearing of atoga, arightexclusive to citizens. unruly places. The presence of individuals with citizen status is not unexpected as the colonia was founded as asettlement of Romancitizens, and is not inconsistent with Status and society the suggestion that this was an areaoflow status The study of social complexity and individual status burialsascitizen status was of course no guarantee is amajor concern of funerary archaeology, but the against poverty. interpretation of cemetery evidence is fraught with Some objects, especially items of clothing or difficulties. It is possiblethat funerary customs may jewellery such as the bracelet and ring worn by the serve to create an idealised image of society rather individual in grave 1362 and the finger ring from than reflecting reality. Also, it is difficult to attribute burial 1505, mayhave been chosen for deposition in value to the grave goods or practices of past the grave becausethey were of personal significance societies, andmuchofthe symbolism usedin to the deceased, or because of afear that the the funerary rites may not nowbeunderstood deceased mightbeable to haunt items with which (Parker Pearson1999). Neverthelessitmay be possible he or shehad been closely associated in life. Other to draw some tentative conclusions regarding the items are likelytohave been depositedfor reasons population buried in the areaofthe cemetery associated with beliefs regarding the fate and investigated. subsequent needs of the dead (see above).

143 Life andDeathinaRoman City

There is some evidencethat the status of indivi- in the cemetery originatedfrom Britain and are likely duals may have been related to age and sex. The to derive from the native population. It is thus youngest members of the population are noticeably possible that we are seeing evidence for the loss of absent from the cemetery: Only asingle infant below the traditional role of womenwithinnative society the ageoftwo yearswas present among an that Allason-Jones (1989b) has argued may have excavated population of 64 individuals, and only arisen from the break-down of tribal structure two more examples were foundamong the 74 from associated with the migration to urbancentres. 124–130 London Road. The under-representation of this agegroup is common at Roman cemeteries The peopling of RomanGloucester (Philpott 1991, 98) and mayindicate that children of this young age were not usually entitled to burial in The arrival at Gloucester of the Romanmilitary in the community’s burial ground, perhaps because AD 49 occasionedthe introductionintothe region of they were not considered to be full members of anew population with alien practices. This is society(Philpott1991, 101). Pliny (Nat Hist VII, 15, particularly trueinthe case of funerary practice, cited in Philpott 1991, 101) claimed that childrendid where the establishment of the fortresses, first at not gain asoul until the age of teething, and it is Kingholmand later at the city centre on the future possible that until they were old enoughtowalk and site of the colonia,was accompanied by the establish- talk children were notconsidered to be full people. ment of the attendant cemetery and the introduction Children between this age and about 12 yearsappear of the rite of cremation. This rite had not been to have still been treated differently from adults as practised by the local native population, whose regards burial,with three out of four individuals funerary traditionsappear to have comprised mainly from this age group buriedinacrouched posture of excarnation, albeit with some evidence for rather than the more normal supine position, crouched burials(Moore 2006, 111). Theearliest implying that the people of the colonia recognised evidence for the burials of this new population are childhood as adistinct status. Childhood appears to two cremationburials, and afurther six excavated at have been considered to end at puberty, as the 124–130 London Road, which are associated with bodies of the threeindividualsinthe 13–17 age pottery of pre-Flavian date, contemporary with the group were buried according to the samepractices as occupation of the Kingsholm fortress. These burials the majority of the adults, including one individual are all of adults, comprising two males and afemale who was buriedprone. (wherethey couldbesexed), and it is tempting to Although the women in the cemetery received the interpret the males as being soldiers garrisoned at same burial rites as the men,their apparent under- the fortress. Thepresence of afemale, if correctly representation within the populationmay have identified, indicatesthat anon-militarypopulation implications regarding the status of women at the was also associated with the fortress.Settlementsof colonia, particularly regarding the possible practiceof civiliansoften grew up beside military bases, female infanticide (see above). If infants were not providing service industries forthe soldiery and regarded as full people, infanticide would have been even housing wives and families, although the latter an easy way of dealing with unwanted births, and it were notofficially recognised (Mattingly 2006, 170–6). is possible that among the poorer members of society The physical remains of such asettlement at so early social and economic pressures associated with a adate have not been identified at Gloucester but the requirementtoprovideadowry for daughters grave of this woman is evidence that civilians were resulted in the preferential selectionoffemale births not only present but were permitted to share the for thisfate.Awillingness to give up female children garrison’s cemetery, indicating that the two commu- in this way mightbeassociated with areduced nities were very closely connected. We cannot, statusfor females in general, at least in those groups however, know how substantial this civilian popula- within the population for whom such apractice was tion was, or whether it originated from camp economically necessary. No direct evidence for this followers who had arrivedwiththe army or from practice would be expected, as the unwanted female members of the local native population attracted by births would have joined the rest of the infant deaths the economicand social opportunities afforded by that were disposed of by some meansthat has left the presence of the military. little archaeologically detectabletrace. After the army had moved on,the site of the city The possible lower statusofwomen among the centre fortresswas reused for the founding of a part of the populationburiedhere may contrastwith civilian settlement, although the precise date of the situationamong the native population, and even neither of these events is known and there may with more high status sections of the colonia.The have beensome interlude before the town was group of apparently high status late Iron Age burials established. The full name of the settlement, Colonia at Birdlip are arranged aroundthe central grave of Nerva (or Nerviana) Glevensium (Wacher 1995, 150) an adult female, suggesting that women were able indicatesthat it was specifically founded as a colonia, to hold high statuswithin the native community asettlementfor retired soldiers.Research by Fulford (Staelens 1982). The programme of isotope analysis (1999, 178)has indicated that veterans typically carried out as partofthisproject has demonstrated settledinthe frontierregions in which they had that some, and probably most,ofthe females buried served rather than returning to their homelands, and

144 Chapter Six there is evidence that some of the burialsatWotton support from the programme of isotope analysis were of individuals with amilitary background. (see Appendix 3), which indicated that crouched The adult male (1542) associatedwith possible belt burial 1238 in grave 1240 was of local origin. During fittingsinthe mass grave may have been asoldier as the late 1st and 2nd centuries when these inhuma- the wearing of belts is not normally associated with tion burialswere made the custom of inhumation civilians. Two burials at 124–130 London Road had not yet been adopted as the main form of burial, associated with abrooch,buckle and belt plate, and so the partofthe population who buriedtheir and acoin and spearhead are also likely to be dead in this way would have clearlystood out from soldiers, the latter having been interpretedbyCool the rest of the community, who were practising (2003) as apossible auxiliary trooper. In addition, the cremation at this time. tombstone of aveteranofthe TwentiethLegion dated The resultsofthe isotope analysis, undertaken on to the 2nd centuryorlater (Hassall and Tomlin1984) asample of individuals from both the cemetery and has beenfound at the Coppice Corner cemetery the mass grave, has demonstrated that the popula- north of Kingsholm. It is therefore entirely likely that tion had arange of origins, deriving both from the soldiers retiring from the British garrison continued local area and also fromelsewhere in the province of to settle at Gloucester long after the initial settlement, Britannia, as well as asmall group probably fromthe perhaps due to acontinued link, formal or otherwise, Mediterranean area and two individuals from maintained between the colonia and the unit formerly another, unidentified non-UKsource (Appendix 3). based there. The individualsfrom the Mediterranean area were Wacher (1995, 165) has somewhat harshly char- clearly not restricted to military personnel as they acterisedthe military origins of the people of includeafemale aged 18–25 yearsburied wearing a Gloucester as producing‘atightly knit community copper alloybracelet and ring (1360, grave 1362). with rigid social strata accordingtoarmy rank and This group is presumably asmall sample of the yearsofservice’ with apopulation ‘introverted and many government officialsand traders and their ultra-conservative’, but thisunderestimates the con- families who settled, permanently or temporarily, at tributionofthe nativepopulation. The presence of the colonia.The local origins of the majority of the individuals fromthe native communitywithin the population is only to be expected, but the presence population of the colonia is suggested by the of individuals fromelsewhere in Britain, including occurrence of crouched burials, and perhaps also two who may have come from the Pennines, by the early extended inhumations(seeabove). It is suggests that Gloucester was sufficiently vibrant likely that these burials represent members of the economically to attract settlers from beyond the local population who were integrated into the immediate area. Theimplication of these results is population of the colonia,but had not adopted that the populationofthe colonia,ifnot exactly Romancustoms sufficiently to abandontheir tradi- cosmopolitan, was at least more varied than has tionalburial practices. This suggestion has received been previously appreciated.

145 Appendices

APPENDIX 1: DENTAL HEALTH STATUS death.Inthe case of discrete burialswhere there is a METHODOLOGY number of teeth and the alveoli have not preserved, the rate of AMTLwas scored according to the Enamel hypoplasia number of teeth present. Thus, if adiscrete skeleton Enamelhypoplasia was scored for the buccalcrown had 32 loose teeth present(the mandibular and surfaces of permanent teeth.All formed crownsfrom maxillaportions very fragmented and poorlypre- all permanent maxillary and mandibular teeth were served), the prevalence of AMTL was still recorded examined. Where present, defects were recorded by as none out of 32. employing the definitions and guidelines set out by the Fe´ de´ ration Dentaire Internationale (FDI, 1982). Dental calculus Calculus was scoredaspresent or absent.When Dental caries present, dental calculus was observed as slight, All developed and erupted teeth were examined for moderateorconsiderable following Brothwell caries. This was diagnosed whenthere was clear (1981),although post-mortem manipulation (eg evidence of lossoftoothsubstance (Hillson 2001). washing) mayhave affectedthe quantity that is The severity of the lesion was classified accordingto presentduring analysis in the laboratory. the categories drawn by Brothwell (1981) and Kelley et al. (1991). Periodontal disease The number of teeth affected was counted instead of the number of mandibles/individuals affected. In archaeologicalmaterial, examination of period- Many teeth were found loose andmany mandibles ontitis is based on the distance between the cemento- fragmented. Presenting caries as apercentage of teeth enamel junction (CEJ) of the tooth and the alveolar affectedwas advantageous because it included isola- crest for each tooth in the jaw. Recording of ted teeth and increased the sample size (Hillson 2001, periodontal disease followed the schemes devised 256). It also allowed obtaining frequencies according by Brothwell (1981). In general, lower degrees of to tooth category(anteriordentition, posterior denti- advancedperiodontitis tend to suggest agood tion). In addition, most mandibles or maxillae did standard of oral health. not preserve all the teeth, and making inferences about prevalence of DC in an individual when teeth Periapical cavities were missing would have been unreliable. Although information was obtained from each The identification of aperiapical cavity was diag- surface and tooth, dental caries frequencies were nosed by the presence of aperforating fistula in a pooled to overall rates according to toothclass (ant- specific alveolarlocation. These lesions were distin- erior, posterior) as recommended by Hillson (2001). guished frompost-mortem damage by observing Anterior teeth included incisors andcanines, while smooth,biological (healed) margins indicative of an posterior teeth included premolars and molars. ante-mortem condition (Roberts and Manchester 1995, 51). Ante-mortem tooth loss (AMTL) Dental anomalies Ante-mortem tooth loss was identified based on the degreeofalveolar boneresorption. Allthe alveoli Anomalies such as dental overcrowding, the pre- that could be observed were counted and any that sence of enamel pearls,congenital absence of teeth, had started to regenerate or that were completely and supernumerary dentitionwerealso recorded for regenerated were scored as having been lost before each of the skeletons when present.

147 Life andDeathinaRoman City

APPENDIX 2: SUMMARY OF DISARTICULATED HUMAN BONE

Ctx/Sf Context description Bones present MNI Age Sex sf154 context 1333, Occipitalbone 1Adult M backfill of 1334 1020 Roman soil 1fragment of cranial vault 1Adult 1024 Fill of modern pit 1023 Femur, humerus, fragments of long bones 1Adult 1025 Medieval soil 1fragment of cranial vault and 1lower left 1Adult second premolar 1090 backfill of grave 1757 1fragment of tibia 1Adult 1105 Post-medieval garden soil 15 fragments including 1fragment of ulna and 1Adult 14 fragmentsofskull 1106 Soil layer Fragments of skull, humerus, ulna, metacarpal, 1Adult femur, tibia, fibula, patella, tarsals (left and right calcaneous) and foot phalanx 1113 Backfill of grave 1143 33 fragments including 6cranial fragments 1Adult including supraorbitalmargin and 26 long bone fragmentsincluding the right distal humerus 1121 Fill of post-medieval ditch 1055 1fragment of radius and one of ulna 1Adult 1132 Backfill of grave 1146 3left maxillaryteeth (left canine and premolars), 2 1Adult fragmentsofcranial vault, 1fragment of humerus 1158 Fill of post-medieval ditch 1055 1fragment of cranial vault 1Adult 1175 Fill of post-medieval ditch 1725 1fragment of occipital bone, including nuchal 1Adult crest 1182 Grave 1230 Parietal, occipital, left femur and right femur, left 2Adult male and right tibiae, left fibula, right ulna, right (and sub?) scapula,right radius, right os coxa, right calcaneousand two left calcanei 1254 Fill of modern pit 1255 1fragment of distal humerus, distal ulna, rib, 1Adult scapula and 4unidentified fragments 1247 Natural 1fragment of tibia 1 1265 Fill of Roman gully 1358 3fragments of skull, 2ofscapula and 6long 1Adult bones fragments 1269 Grave 1144 Right scapula, left os coxa, ribs, right medial 1Adult cuneiform,left radius,left fibula, fibula, left MT3, frontal, parietal bones 1282 Backfill of grave 1284 7teeth including lower second molars, lower 1Young adult right third molar and lower left first molar, the latter with acarious lesion on the coronal (lingual) face. No other pathology apart from slight calculus on lower right premolars and slight periodontitis on lower molars, right MT3, right MT5, ribs, right ulna, right mandibleand thoracic vertebra 1285 Backfill of grave 1288 6teeth, left side: lower deciduous M2s, upper 1Subadult and lower M1 (roots open), unerupted 2nd molar and 1adult? and one uneruptd canine. All from individual c 7 years. In addition fragments of radius, ulna, mandible with 2unerupted incisors and one unerupted premolar, subadult skull fragments and some unidentifiedfragments. 2fragments of femoral condyles probably belonged to an adult skeleton 1312 Backfill of grave 1315 1fragment of right distal humerus 1Adult 1341 Fill of grave 1344 1thin skull fragment 1? 1361 Backfill of grave 1362 50 small (10–4 mm) fragments unidentified 1? 1365 Back fill of grave 1370 Skull bones: frontal, parietal, occipital, temporal 1Young Adult bones, and left zygomatic, mandible, right humerus, left tibia, teeth present (n=11) with no pathology

148 Appendices

Ctx/Sf Context descriptionBones present MNI Age Sex

1368 Grave 1369 Rib and right radius? 1Adult 1387 Backfill of grave 1388 1fragment of rib and c 20 small bone fragments 1? unidentified 1397 Backfill of grave 1398 3unidentified fragments1? 1402 Backfill grave 1403 1middle phalanx, vertebral body and 13 1Adult unidentified fragments, left femur, left tibia, left fibula, left ilium, sacrum, lumbar vertebra, hand phalanx 1406 Backfill grave 1407 14 fragments of foot: 6fragments of MTs, 8 1Adult phalanges (5 proximal, 1intermediate and two distal) 1409 Pit 1408 Metacarpal, rib, skull fragments, cremated tibia 3Adult fragment,left ilium, right femur, left ulna, right humerus, right humerus, humerus, right calcaneous, hand phalanx, right femur, left femur, left femur, left tibia, right tibia, left tibia, left tibia, right tibia, right tibia, right ilium, lumbar vertebra, left fibula, 2left petrous, right maxilla with M2, M1, P2 and P1 and on left LM1, LM2 and LM3, thoracic vertebra, left humerus,left humerus, radius, ulna, phalanx, left femoral head 1411 Backfill of grave 1763 1fragment distal tibia 1Adult 1413 Post-medieval ditch 1412 Left talus, right tibia, right ulna, sacrum, ilium, 1Adult left and right parietals, right frontal, metacarpal 1414 Post-medieval ditch 1412 Parietal, left humerus, right radius 1Adult 1416 Modern posthole 1415 Left ulna and one cervical vertebra 1Adult 1418 Modern posthole 1417 Skull fragments 1Subadult 1426 Grave 1427 Right radius, right ulna, hand proximal and 1Adult intermediate phalanx 1433 Fill of modern pit 1432 40 fragments labelled disarticulated, including 9 1Adult skull fragments, two teeth (upper right premolars), fragments of femur, tibia, left calcaneous and left talus, vertebra, ribs, fragments of long bones 1435 Post-medieval ditch 1fragment of cranial vault: an occipital with a 1Adult nuchal crest 1454 Grave 1501 Right femur, right femur, left femur, femur, 2Adult femur, femur, left patella, left tibia, left ilium, left ulna, right radius, ulna, radius, metacarpal, metatarsal, phalanges and skull fragments 1457 Backfill of grave 1502 13 skull fragments, 24 long bone fragments, 4 2Young Adult F? teeth (upper right premolars,upper left P2 and lower right M1), clavicle, ulna, patella, fibula, skull with right zygomatic, two left zygomatics, right temporal, axis, right navicular, right radius, right clavicle, right? Clavicle, right ulna, left scapula,femur, cervical vertebra, left mandible, left mandible, right mandible 1465 Post-medieval ditch 1729 Left ulna 1Adult 1468 Grave 1504 Left scaphoid, metacarpal, 1Adult cervicalvertebra and femur fragments 1480/sf319 Grave 1505 2phalanges inside afinger ring 1Adult 1481 Fill of grave 1505 19 fragmentsofskull and long bones 1Adult 1536 Grave 1765 1clavicle fragment 1Adult 1713 Grave 1723 Occipital1Adult

149 Life andDeathinaRoman City

APPENDIX 3: OXYGEN AND STRONTIUM Table A3.1 Details of individualsand teethsamples. ANALYSIS OF TOOTH ENAMEL Skeleton Sex Age Teeth Tooth (NERCIsotope Geoscience Laboratory present sampled ReportSeries No.221) by Carolyn Chenery 1103 FMiddle adult 26 M3 lower right Introduction 1127 ?M Young adult 2M3lower right 1131 FYoung adult 11 M3 lower right 87 86 18 This report presents Sr/ Sr and d Ooxygen 1181 FMiddle adult 24 M2 lower right isotope data for tooth enamel from permanent 1216 MMature adult 14 M3 lower left molarscollected from 21 individualsburiedina 1238 ?M Mature adult 14 M3 lower left Romancemetery at London Road, Gloucester, 1328 MMiddle adult 18 M3 lower left Gloucestershire. The samples were provided by 1340 IND Old adult 8M2lower left Oxford Archaeology and were analysed at the NERC 1360 IND Young adult 15 M2 lower left Isotope Geoscience Laboratory at the British Geolo- 1364 ?F Young adult 19 M3 lower left gical Survey (NIGL/BGS) as part of the AHRC- 1518 ?Male Middle adult 23 M2 upper left funded project ‘DiasporaCommunities in Roman 1520 Female Young adult 28 M3 upper left Britain’ based at The University of Reading. The 1539 Female Middle adult 29 M3 lower right purpose of the work is to further our understanding 1541 Male Young adult 31 M3 lower left of diversity in RomanBritain by assessing the 1544 Male Middle adult 29 M3 lower left proportion of individualswho: 1546 ND Mature adult 29 M3 lower right (a) mightbeoflocal origin andthosewho canbe 1553 Male Middle adult 30 M3 lower right defined as ‘non local’, 1560 Female Young adult 26 M3 lower right (b) to constrain the place of childhood origin for any 1561 Male Young adult 30 M3 upper left ‘non locals’ and 1565 Female Young adult 27 M2 lower left (c) to assess any variations found against archae- 1596 Female Middle adult 7M2upper right ologicalfactors such as sex, burial type and artefacts. Oxygenisotope analysis Methodology Biogenic phosphate was converted to silver phos- ToothSampling phate using the method of O’Neil et al. (1994) andis briefly summarisedhere. Thecore enamel samples At Oxford Archaeology, teeth were taken from 21 were crushed to afine powder and cleaned in individuals, 10 from individual inhumations and 11 hydrogen peroxide for 24 hours to remove organic from the massburial pit foroxygen andstrontium material. The peroxide was evaporated to dryness analysis. See Table A3.1 for details of individuals and and the sample dissolved in 2M HNO3 .The sample teeth sampledand Table A3.2 for grave and skeletal solutions were transferred to clean polypropylene descriptions. test tubes and each sample was treatedwith2M Photographs of each tooth (mesial, distal,buccal, KOH followed by 2M HF to removeCafrom the lingual and occlusal orientations and internal views of solution by precipitation. Thefollowing day, the sam- slices for most teeth)weremade prior to extraction of ples were centrifuged and the solution was added to samples from the crown and are contained in the site beakers containing silver amine solution and silver archive. phosphate precipitated, filtered, rinsed and dried. Analyticalmeasurement was by Continuous Flow Isotope Ratio Mass Spectrometry (CFIRMS) using the Isotope Analysis methodofVenneman et al. (2002).The instrumenta- Each tooth was cut in half using aflexible diamond tion is comprised of aTC/EA(thermo chemical edgedrotary dental saw. Thehalf selected for elemental analyser) coupled to aDelta Plus XL analysis was cleanedultrasonically for five minutes isotope ratio mass spectrometervia aConFlo III in high purity water and rinsed twice to remove interface, all by Thermo Finnigan.All reported looselyadhered material. Atungsten carbide dental isotope ratios are expressed using the delta ( d ) burr was usedtoabrade off the enamel surface to a notation in parts per thousand(permil: ‰ )relative depth of >100 microns. Secondary dentine was to astandard: removed and discardedand the enamel and primary d ð ‰ Þ¼ððR = R Þ1 Þ · 1000 dentine were separated. The dentine was reserved sample standard for future potential carbon, nitrogen and background The reference material NBS120C, calibrated strontium analyses and the enamel was preparedfor against certifiedreference material NBS127 (assum- oxygen and strontium analysis as follows. ing d 18OofNBS127 =+20.3‰ versus SMOW), has an

150 Appendices

Table A3.2 Grave and skeletal description.

Skeleton Grave Orientation Posture Grave furniture Grave goods

Single Burials 1103 1150 NE-SW Prone, legs extended None None and parallel 1127 1145 SW-NE Prone, extended.Lower None Ring-necked flagon parts of legs removed by later truncation 1131 1146 .N-S Supine, legs extended with None None feet together 1181 1230 S-N Prone, legs extendedwith None None feet together 1216 1218 SE-NW Supine, legs extended with Fe nails None feet together 1238 1240 n/a Crouched, lying on left side None None with the head to the east 1328 1330 NE-SW Supine, legs extended and Coffin, Fe nails None parallel 1340 1344 SE-NW Supine, legs extended and Coffin, Fe nails Two hobnail shoes parallel 1360 1362 SW-NE Supine, Legs extended and Coffin, Fe nails Fe bracelet worn on left arm, parallel, left tibia displaced Fe ring toward right leg 1364 1370 N-S Supine, legs extended Coffin None

Mass grave 1518 1485 Pit E-W Commingled human None Anumber of artefacts were found remains associated with the skeletons, including at least 3pairs of hobnailed shoes, two copper alloy trumpet brooches, one copper alloy bracelet, two belt buckles and two ﬁnger rings. At least eight near-complete, though fragmented,vessels had been deposited in mass grave 1483 probably during the later 2nd century, although pottery recovered from upper ﬁlls was of later 3rd or 4th century date. 1520 1485 Pit E-W Commingled human remains None 1539 1485 Pit E-W Commingled human remains None 1541 1485 Pit E-W Commingled human remains None 1544 1485 Pit E-W Commingled human remains None 1546 1485 Pit E-W Commingled human remains None 1553 1485 Pit E-W Commingled human remains None 1560 1485 Pit E-W Commingled human remains None 1561 1485 Pit E-W Commingled human remains None 1565 1485 Pit E-W Commingled human remains None 1596 1485 Pit E-W Commingled human remains None

expected value of 21.70‰ (Chenery 2005). Each (Chenery 2005) and the value for NBS120Busedby samplewas analysed in triplicate. Theinternal mass Levinson of d 18O=20.06 + 0.27‰ (2s ). spectrometry reproducibility for this set of analyses is + 0.18‰ (1s n=34). The batch reproducibility was Strontium isotope analysis 0.15‰ (n=6, 1 s ). Drinking water values are calculated using Levinson’s equation(Levinson et al., In aclean laboratory, the enamel sampleswere 18 18 1987), d O Drinking Water =(d O Phosphate Oxygen washedinacetone and cleaned twice, ultrasonically, 19.4)/0.46, aftercorrection for the difference in high purity water to remove dustand impurities. between the average published valuesfor NBS120C They were dried and weighed intopre-cleaned used at NIGL of d 18O=21.71 + 0.35‰ (1s )n=11 Teﬂonbeakers.Each sample was mixed with 84Sr

151 Life andDeathinaRoman City tracer solution and then dissolvedinTeflondistilled group,with higher valuesthan expected for the UK 16M HN03 .The sample was then converted to (see Fig. A3.1). Chloride and taken up in 2.5M HCl. Strontium was Statisticalcomparisons (F and t-tests) of oxygen collected using conventional, Dowex1 resin ion isotopesofindividuals from single graves with the exchange methods. mass burial pit show no significant difference in the The Sr isotope composition and concentrations two populations at the 95% level.Similarlycompar- were determinedbyThermal Ionisation Massspec- ison by sex also showed no significant difference at troscopy (TIMS) using aThermoFinningan Triton the 95% level. multi-collector mass spectrometer. Samples were run at c 5V using singleRefilaments loaded using TaF Strontium isotopes following the methodofBirck (1986).The interna- tionalstandard for 87Sr/86Sr, NBS987,gave avalue of The strontium isotope ratios(87Sr/86Sr) cover abroad 0.710275+ 0.000006 (1s ,n=12). All strontium ratios range of valuesfrom 0.70880 to 0.71347 (Fig. A3.1). have been correctedtoavalue for the standard of The bulk of the 87Sr/86Sr values plot withinthe range 0.710250.Strontium procedural blanks provided a of 0.7088 to 0.7106. Individuals in this group are negligible contribution. compatible with alimited set of values obtained from plants and bottled waters from Jurassic and Triassic sediments ranging from 0.7081 to 0.7100 (Montgom- Results ery et al., 2006 and Evans et al., in press). Two Oxygen and strontium results are presented in individuals GLR1131 and GLR1364 have similar Table A3.3 and Figure A3.1. values(0.71141 and 0.71145) and these along with GLR1541 (0.712200) are compatible with values foundinbottled water fromSilurian mudrocks and Oxygen isotopes Palaeozoic rocksinthe Pennines.The very high The enamel phosphate oxygen isotope ratios(d 18Op) valuesfor GLR1561 and GLR 1518 (0.71345 and from thissampleset presentabroad scatterofresults 0.71301) are compatible with Malvernand similar ranging from +17.05‰ to 19.18‰ ,with arange of Precambrian volcanic rocks. calculated drinking water oxygen isotope ratios Strontium concentrations vary from 51 to 177ppm ( d 18Odw) of 8.15‰ to 3.51‰ .These datasplit and fall within the typical UK range (Montgomery roughly intotwo groups; thosewho fall within the 2002). 18 expected UK Adult range ( d O enamel 17.00‰ to Statisticalcomparisons (F and tTests) of individuals 18 18.50‰ , d O drinking water -8.50‰ to -5.00‰ ), which from single graves with the massburial pit show no was calculated from the ‘Robust Mean’ + 2 s of d 18O significant difference in the two populations for ratiosof363 individuals fromthe UK, and asecond strontium isotope ratios and strontium concentration

Table A3.3 Oxygenand strontium isotope results.

18 18 Sample Burial Sr ppm 87Sr/86Sr Mean d O P Mean d O DW

GLR1103 Grave 114 0.709465 þ 18.44 + 0.12 5.12 + 0.27 GLR1127 Grave 121 0.709694 þ 17.19 + 0.13 7.82 + 0.29 GLR1131 Grave 133 0.711453 þ 17.09 + 0.14 8.05 + 0.32 GLR1181 Grave 76 0.710599 þ 17.60 + 0.15 6.93 + 0.32 GLR1216 Grave 93 0.709432 þ 19.18 + 0.15 3.51 + 0.34 GLR1238 Grave 52 0.708937 þ 17.14 + 0.05 7.93 + 0.11 GLR1328 Grave 57 0.710584 þ 17.97 + 0.14 6.13 + 0.30 GLR1340 Grave 104 0.709606 þ 17.91 + 0.15 6.26 + 0.34 GLR1360 Grave 133 0.709021 þ 19.01 + 0.17 3.88 + 0.38 GLR1364 Grave 104 0.711405 þ 18.23 + 0.10 5.57 + 0.22 GLR1518 Pit 72 0.713005 þ 17.91 + 0.15 6.27 + 0.33 GLR1520 Pit 130 0.709732 þ 18.64 + 0.07 4.67 + 0.15 GLR1539 Pit 66 0.708805 þ 17.50 + 0.04 7.16 + 0.08 GLR1541 Pit 51 0.712200 þ 17.05 + 0.08 8.14 + 0.19 GLR1544 Pit 102 0.709969 þ 18.18 + 0.08 5.69 + 0.19 GLR1546 Pit 177 0.710889 þ 19.05 + 0.06 3.80 + 0.14 GLR1553 Pit 84 0.709269 þ 17.44 + 0.20 7.28 + 0.43 GLR1560 Pit 114 0.709486 þ 18.71 + 0.21 4.53 + 0.46 GLR1561 Pit 67 0.713468 þ 18.70 + 0.10 4.54 + 0.23 GLR1565 Pit 104 0.709773 þ 19.01 + 0.14 3.87 + 0.31 GLR1596 Pit 105 0.709067 þ 17.71 + 0.13 6.71 + 0.28

152 Appendices

cannot rule out possible European origins from these results, speciﬁcally regions where the geology and oxygen combine to give the same range of values. However, the simplestinterpretation is that they are from the UK.

The smallerGroup 2 These samples have oxygen isotopes values that suggest awarmer climate of origin than canbefound in the UK and the clustering of the Sr would suggest a common region or dietary origin. Thestrontium valuesare close to modern seawater ( , 0.709) and may possibly be caused by adiethigh in marine components such as fish, shellfishand seaweed. Certaingeological bedrocks can be ruledout, assuming theygenerate the overlyingsoil, Rb-rich granites, basalt terrains,pure limestone and chalk, and the majority of pre-Mesozoic silicate terrains. The oxygen isotopes of this group exceed those Figure A3.1 Plot of enamel strontium and oxygen expected forthe UK, ranging from +18.65‰ to isotope results. +19.18‰ ( d 18Odw 4.67 to 3.51). Exceptionally high valuessuch as these might be expected along the extreme western or southern coastal Europe, at the 95% level. Similarly comparisonbysex also northern Africa or in extremely arid locations. Very showednosignificant difference at the 95%level. few valueshave beenpublishedfor these areas; however, Darling (2003) records arange of 4.0 to 5.0 for the OuterHebridesand small areas on the Discussion extreme westerly coast of Cornwalland Ireland, and Based on combined oxygen and strontium values, Longinelli (2003) cites values of > 5.0 for the eastern our sample set can be divided into two main groups Italiancoast south of Brindisi. (Fig. A3.1). Thefirst group is quite diverse and has a Two other individuals with oxygen isotopic values wide range of 87Sr/86Sr values and d 18Ovalues that that lie outside the UK range have higher strontium mirrorthe normal distribution of UK reference data isotope values, suggesting that they came from (Montgomery 2002; Montgomery et al. 2006; Evans somewhere different than group 2. et al. in press; Darling et al. 2003). The second group have d 18Ovalues outside those of the UK robust mean valuesand theyalso have restricted 87Sr/86Sr Conclusions valuesaround 0.7094. Twopoints are not easily The isotope data do not correlate withinany attributed to either group.They bothhave non-UK archaeological or osteological criteria, and in parti- oxygen but more radiogenic Sr than is typical of the cular, there is no difference between the bodies in the main group 2samples. mass grave and thoseinsingle graves, suggesting that they may well have come from the same population. This supportsthe interpretation based Group 1 on the osteologicalevidence (see Chapter 3) that the Groupone is compatible with arelatively diverse individuals buried in the mass grave were victims of UK derivedpopulation; there is awide range of an epidemicthat struck the populationofGloucester 87Sr/86Sr (0.7090 –0.7135) values but most of these randomly, rather than adistinct group of people that (0.7090 to 0.7099) can be derivedwithin a20km were selected for mass burial on account of their distancefrom Gloucester. The most radiogenic value place of origin, age, sex or social status. in thisgroup (0.7135) is consistent with data The results of this study suggest that Gloucester obtained from mineral waters from the Precambrian was atown in which individuals fromreasonably of the nearby Malvern Hills (Montgomery et al. 2006) diverseareas of the UK lived and died. However, an areaofRomanpottery and tile manufacture. The within thiscommunity the isotope data also identi- oxygen, however, suggested the origins of this group fied agroup whose origins lay in aregion with a mightbemore widely dispersed throughout Britain warmer climate than the UK.Additional archae- as their oxygen covers the full range of UK values. ologicalorosteological information on the indivi- While the strontium and oxygen isotope values duals analysed here would help to put these findings foundinthis group are consistent with lithologies into context and may aid in further interpretations of and drinking water zones found within the UK, we the data presentedhere.

153 Life andDeathinaRoman City

APPENDIX 4: DETAILED RECORDS OF Cremation burial1209 CREMATIONS Description: Several skull bones were presentinclud- Cremation burial1196 ing the frontal, parietal,occipital, temporal, maxilla, mandible and sphenoid bones. Landmarks includea Description: Theskull was represented by the frontal, portionofthe left supraorbital margin, the nuchal crest parietal and occipital bones only. The area of the and agonion of the mandible. Dental fragment external occipital protuberance was present. Very included roots and crowns, fragmentspossibly be- few rib and vertebra fragmentswererecovered as longing to an upper first premolar, an upper second well as few fragmentsofthe body of the ilium. premolar and an upper molar.Both the shoulder and Amongstthe long bones, all were represented. The pelvic girdles were represented, followed by vertebrae largestfragment measured43mmfrom ahumerus. (cervical), ribs and the long bones: humerus, radius, MNI:Onlythe remainsof1individual were ulna, femur,tibia and fibula. The patella was also identified. present. Hand bones included metacarpals and Age: Thedimensions of the skull and long bone phalanges whilst foot bones included atalus, meta- fragments indicated an adult individual. The coronal tarsalsand phalanges. The maximumfragment size and sagittalsutureswereopen. was that of atibia shaft measuring 59 mm. Sex: Sexcould not be determined due to the lack of MNI:Only the remains of 1individual were the landmarks requiredfor sex estimation. identified. Pathological lesions:Some pitting was presentinthe Age: Thedimensions of the skull and long bones ectocranial surface of one parietal bone (2 g). clearlyindicate an adult(>18 years) individual. The Colourand fracture pattern: Thebones reflected the complete formation of the premolar roots provides a typical fissure pattern characteristic of green bones safe estimate of over 12 yearsofage. The lambdoid when burnt. The colour was predominantlywhite and coronal sutures are open. (177 g) and white with shades of blue and grey Sex: The sex was difficult to determine and the (76 g), with some fragmentsfromthe upper and result possiblyreflectedamale individual, although lower limbs and even the skull predominantly grey the traits are verylimited. The supraorbital margin (1 g) and blue (10 g). In summary, ahomogeneous gave an ambiguous sex so the trait was intermedi- cremation fired at high combustion. ate. The posterior zygomatic arch provided a Animal Remains: Atotal of 20.8 gofburntanimal robusticity characteristic of amale individual. The bone. Five fragmentsofcalcinedprobablepig bone nuchal crest was also ambiguous although it did including cranial and long bone fragmentswere tend to be rather more marked as in males. One recoveredfrom the cremation deposititself.The hand phalanxsubjectively showedprobablytypical backfill contained eleven fragmentsofburntmedium male dimensions. and large mammal long bone and 14 fragmentsof Colourand fracture pattern: Thebones reflected the unburntlarge mammallong bone, recoveredfrom typical fissure pattern characteristic of green bones sieved residues and through hand collection. One when burnt. The colour was predominantlywhite indeterminate fragment was charred and the remain- (94 g) and white with shades of blue and grey mainly ing burnt fragments were fully calcined.None of (91 g), with some fragmentsfromthe upper and lower these fragments couldbeidentified to species and limbs and even the skull predominantly grey (18.5 g), some may be human. brown (1 g) in ahand phalanx and one unidentified Conclusion: One possibly adult individual. fragment, and finally black (0.5 g) for aproximal foot phalanx. In summary,ahomogeneous cremation fired at high combustion. The remaining is trabecular bone and mostly the colour is unidentifiable. Other observations: Five mandibular teeth (the first incisors, the second incisors andthe left canine)were CREMATIONBURIAL 1196 lost post-mortem. One unidentified bone fragment BONE ELEMENTWEIGHT (g) with corroded metal attached. Some slight (<1 mm) enthesophytes observed in the 30% that is present of CRANIAL VAULT (Neurocranium) 42.5 the patella. MANDIBLE 0.5 Animal remains: 11.9 gofburnt medium mammal RIBS 4 sized animal boneincluding cranial fragments, long VERTEBRAE 2 bone fragmentsand sheep sizedrib blade fragments. SHOULDERGIRDLE 0.5 Two fragments from alarge mammal skull fragment PELVIS 4 (possiblyhuman) and apig phalanxfrom ajuvenile UPPER LIMB BONES 37.5 animalwere recovered from the cremationdeposit LOWER LIMB BONES 51 (1202) and one indeterminate medium mammal FOOT BONES 0.5 sized fragment recovered from the fill of ancillary LONG BONE UNIDENTIFIED (>4 mm) 37 vessel 1199. The majority of the burnt animal bone UNIDENTIFIED FRAGMENTS 90 was recovered from the backfill (1198). TOTAL WEIGHT269.5 g Conclusion: An adult individual, probably male.

154 Appendices

CREMATIONBURIAL 1209 unidentified phalanx. The ribs were represented with 98 shaft fragments. The remaining bones BONE ELEMENTWEIGHT (g) correspondedtothe humerus, radius, ulna, femur CRANIAL VAULT (Neurocranium) 49.5 and tibia only. The maximumfragment was 38 mm FACIAL BONES (Splanchnocranium, 1.5 long, probably belonging to aradius. excluding mandible) MNI:Two subadult individuals were identified, MANDIBLE 1 according to the dental age givenbythe teeth as TEETH 1 well as by the different (and non-corresponding) RIBS 6 dimensions of the bones. VERTEBRAE 1 Age: The dental remainsclearlyreveal two subadult SHOULDERGIRDLE 1.5 individuals, oneofwhom diedaround 5years of age PELVIS 1 (category 2–6 years) and another who died in the age UPPER LIMB BONES 29 range 6–12 years. In addition, long boneand rib LOWER LIMB BONES 39.5 fragments indicatedtwo individuals of different HAND BONES 2.5 dimensions. Excludinganupper and alower molar FOOT BONES 4 unidentified with at least 3/4 of the root formed,the LONG BONE UNIDENTIFIED (>4 mm) 24 age was determined by the formationofthe teeth as PATELLAE 2 shown in the table below: UNIDENTIFIED FRAGMENTS 118 TOTAL WEIGHT281.5 g Permanent Location Stage of Estimated age tooth formation (Smith 1991 Cremation burial1227 (Moorrees for lower teeth) et al. 1963) Description: This is abackfill context and only four fragments survive and range in size between 10 and P1 Lower left Cr1/2 4.1 4mm. While one fragment belongs to the skull and right fragment, the other three were not identified. P2 Lower left Cr1/4 5.0 MNI:Onlythe remains of 1individual were Molar Upper >R3/4 (unidentifiedmolar) identified. Molar Lower >R3/4 (unidentifiedmolar) Age: Unknown. M2 Lower Cr1/2 5.1–5.4 years Sex: Sex could notbedetermineddue to the lack of M2 Upper right Cr1/2 Upper teeth the landmarks requiredfor sex estimation. and left (Ubelaker’s 1989 Colourand fracture pattern: All the bones were chart = c 5years) white in colour indicating acremation fired at high M3 Upper left Cr1/4-Cr1/211.3 combustion. Conclusion: One individual, age and sex unknown. In addition, the mandibular fragment revealeda CREMATIONBURIAL 1227 socketfor an erupted deciduousincisor and below it asocket for an uneruptedpermanent incisorwithan BONE ELEMENTWEIGHT (g) estimatedformation grade of R1/4 (Moorrees et al. 1963), which reveals an age-at-death range of 2–6 CRANIAL VAULT (Neurocranium) 0.5 years. Theteeth, therefore,indicate at least one UNIDENTIFIED FRAGMENTS 1 individual between the ages of 2–6 years. Taking into TOTAL WEIGHT1.5 g account, the identified third molar with its stage of formation, there is clearly an older individual in the Cremation burial1266 age range >6–12 years. Finally,some trabecular bonefrom metaphysis Description: The frontal, parietal, occipital and with unfused epiphyses were identified. possibly the temporal bones have been identified. Sex: Sex unknown. Some fragments includethe different cranial sutures. Colourand fracture pattern: The bones reflectedthe An anterior fragment of the mandible was also typical fissure patterncharacteristicofgreen bones present. Several dental remainswere recovered when burnt. Thecolour was predominantly white including threerootsand twelve crown fragments. (86.5 g) and white with shades of blue and grey These dental remainscorrespond to permanent (36 g), with some fragments from the upper and upper second molars, an identified upper molar lower limbs, neurocranium and teeth darkblue (root only), alower second molar,one lower molar (10 g) with some shades of grey. Theremaining root and the lowerleft and right premolars. There fragments, mainly trabecular bone fragments, were was also an unidentified third molar and clearly an not classified into any colour category. In summary, upper left third molar.There were only four small ahomogeneous cremation fired at high combustion. vertebral fragments, 1fragment of clavicle, ashaft Other observations:Some corroded metal adhered of either ametacarpal or ametatarsal and an on two (1.5 g) cranial fragments.

155 Life andDeathinaRoman City

Animal remains: Amedium mammal long bone the body was fleshedorgreen when cremated.Most fragment and five indeterminate fragments were of the bones were white (111.5 g) or white with recoveredfrom the backfill,and one medium shades of othercoloursbut mainly grey (12 g) and mammal or bird long bonefragment from asoil others blue (7 g) or grey (1 g) some of which are sample. None of this material was burnt and it may darker colours (eg some dental roots). This overall all be residual in the soilmatrixand not of funerary white colour indicateshigh and complete combus- origin. tion of all the organicparts. Two fragments were Conclusion: Two individuals with an age of 2–6 and black and one was dark blue amongstthe unidenti- >6–12 yearsrespectively. fied fragments. One long bone shaft had amix of blue and orangeshades. CREMATIONBURIAL 1266 Animal remains: 86.8 g(166 fragments) of burnt animalboneincluding fragments from at least two BONE ELEMENTWEIGHT (g) pigs. Thepigs were represented by cranial and tooth CRANIAL VAULT (Neurocranium) 32 fragments, aleft tibia, aleft astragalus, acalcaneum, MANDIBLE 0.5 aleft humerus, and two left ulnae. Epiphyseal fusion TEETH 3 indicatesthat one of the pigs was under30months RIBS 8.5 and one was over 12 months old at death.Further VERTEBRAE 1 medium or large mammal sized scapula,long bone, SHOULDERGIRDLE 0.5 tooth and indeterminate fragments may also be pig. UPPER LIMB BONES 12 It is not clear whether the pigs were burntascom- LOWER LIMB BONES 3.5 plete carcasses or disarticulated joints and no butch- HAND BONES or FOOT BONES 1 ery marks were identified. LONG BONE UNIDENTIFIED (>4 mm) 21.5 Conclusion: Aminimum of oneindividual, possibly UNIDENTIFIED FRAGMENTS 59.5 adult. TOTAL WEIGHT143 g CREMATIONBURIAL 1766

Cremation burial1766 BONE ELEMENTWEIGHT (g) Description: The most represented bones were those CRANIAL VAULT (Neurocranium) 14 from the cranial vault and the long bones. Apart FACIAL BONES (Splanchnocranium, ex- 4.5 from the frontal, parietal and temporal (petrous part) cluding mandible) bones, other cranial vaults such as the maxilla, MANDIBLE 0.5 sphenoid and arightzygomatic. Aleft condyle of a TEETH 2.5 mandible was also identified, alongside anumber of RIBS 5 dental roots, one belonging to lower incisors,aswell VERTEBRAE 2 as upper and lower canines, upper and lower SHOULDERGIRDLE 6 premolars and alower molar. Fragments of clavicle, PELVIS 0.5 humerus, radius,ulna, femur and possibly fibula UPPER LIMB BONES 18 were represented. Thelargestfragment measured LOWER LIMB BONES 1 47 mm andbelonged to aclavicle. Cervical and HAND BONES 0.5 thoracicvertebrae were identified. The remaining LONG BONE UNIDENTIFIED (>4 mm) 21 bone fragmentswere unidentified. UNIDENTIFIED FRAGMENTS 57 MNI:Aminimum of only 1individual was identi- TOTAL WEIGHT132.5 g fied. There were no repeatedbones and none from a different age. Age: The size, dimensions and texture of the bones, Cremation burial1767 especially the long bones, were representativeofan adult (>18 years) individual. The dental roots Description: Amongst the skull fragments, the belonging to premolars, caninesand incisors were frontal, parietal, occipital and temporal bones were complete providingasafe minimumage of 15 years present. Portions of the left and rightorbits were and over (upper premolars). Lambdoid and sagittal identified. Maxilla and mandible were present. sutureswere open. Dental fragmentswererecovered belonging to at Sex: Sexcould not be determined due to the lack of least four premolars and five molars. Cervical, the landmarks requiredfor sex estimation. thoracicand lumbar vertebrae fragmentswere Non-metric traits:The only trait availablefor identified, as well as over 250 rib shaft fragments. observation was the multi-foramina in the zygomatic Clavicles, scapulae, os coxae, humeri, radii, ulnae, which was present. femora, tibiae, fibulae, metacarpals, metatarsalsand Pathological lesions: No pathological lesions were phalanges were also present. The longest fragment observed, although pitting in palate was noted. measured 58 mm and belonged to the shaft of a Colourand fracture pattern: Warpingand transverse fibula. Therewere also numerous femora and tibiae fracture pattern was observed in the bones indicating shaft fragments that measured 55 mm in length.

156 Appendices

MNI:Three mandibular condyles (preservationover CREMATIONBURIAL 1767 50%) were found. All of these threewere identified as belonging to the rightsideofthe mandible. BONE ELEMENTWEIGHT (g) Therefore, the MNI assuming all siding is correct CRANIAL VAULT (Neurocranium) 151 (minimum two individuals) results in atotal of three FACIAL BONES (Splanchnocranium, 5.5 individuals in this assemblage. One comes from excluding mandible) context (1062) and the other two from backfill MANDIBLE 3.5 context (1377).Moreover, two portions of two TEETH 5 different axisvertebrae (the dens) were found in RIBS 55 the same context (1377) and indicating at least two VERTEBRAE 23.5 individuals withinthe same context. SHOULDERGIRDLE 15 Age: Through the size and dimensions of the bones PELVIS 45 none were clearly subadult. Therefore the three UPPER LIMB BONES 82 individuals were adult. Some of the sutures such as LOWER LIMB BONES 330 the lambdboid were openwhilst portions of sagittal HAND BONES 3.5 suture displayedminimalclosure. Thedental roots FOOT BONES 20 of molars were completely formed providing asafe LONG BONE UNIDENTIFIED (>4 mm) 115 minimumage of 12/15 years. UNIDENTIFIED FRAGMENTS 401.5 TOTAL WEIGHT1255.5 g Sex: one clavicle, onehumerus, two femora and one tibia fragmentssubjectively seemed to be of male dimensions. The size and robusticity of two man- Cremation burial1768 dibular fragments, belonging to two different individuals, also indicated possible male traits. Description: Only 73 fragments were foundinthis Pathological lesions: There were no signs of osteoar- deposit. Afew skull vault fragments and parts of thritis in alowerright facet of vertebra, in two rib humeri, radii and ulnae were identified in addition facets,inamandibular condyle, nor in asuperior to arootofanuppercanine. The rest was uniden- right thoracicvertebrae facet. Therewas, however, tified.The maximumfragment size was 28 mm clear periostitisinfive shaft fragments of tibiae (2 g, measured on aportionofshaft of aradius. 2/120 =1.6%) and in two shaft fragments of fibula MNI:Onlythe remains of 1individual were (1 g, 1/17.5 =5.7%) characterised pitting and striae. identified. It was difficult to establish whether the lesions were Age: According to the dimensions and size of the active or healed at the time of death. Pitting in palate bones, these were typical of an adult individual. The was also noted. Slight enthesophytesinthe spinous full formation of the canine root provided asafe processofTVorLVwere also recorded. Joint disease minimumof12years of age. was present at the base of aproximal foot phalanx Sex: Sex could notbedetermineddue to the lack of for ray 1which had markedosteophytosis and the landmarks requiredfor sex estimation. deformity as well as pits that may be reflective of Pathological lesions: No pathological lesions were osteochondritis dissecans. observed, although two fragmentsofskull vault Colourand fracture pattern: The bones reflected appeared to be abnormal,asifa‘bulge’ in the the typical fissure pattern characteristic of green endocraniallesions, perhaps similar to hyperostosis bones when burnt. The colour was predominantly frontalis interna but the small remains (<2.5 g) and white (251 g) andwhite with shades of blue and the lack of otherparts limited this diagnosis. grey (830 g), with some fragmentsfrom the upper Colourand fracture pattern: The bones reflectedthe and lower limbs and even the skull predominantly typical fissure patterncharacteristicofgreen bones blue (42 g), grey (22 g) and lastlybrown (18 g). when burnt. Allthe fragments were white in colour. The rest are fragmentsoftrabecular bone. In Animal remains: Five fragments (1.8 g) of burnt summary, ahomogeneous cremation fired at high medium mammal sized long bone and indetermi- combustion. nate bone fragmentswere recoveredfrom the Animal remains: Ten fragments(48.6 g) of cremated cremation deposit in this burial. animalbone were recovered from the cremation Conclusion: Afew fragmentsbelonging possibly to deposit, comprising apossible cattle rib and a one adultindividual. medium mammal sized long bone and cranial CREMATIONBURIAL 1768 fragments, as well as an unburntfish rib/process and an unburnt indeterminate fragment.The fill of BONE ELEMENTWEIGHT (g) ancillary vessel 1063 contained amedium mammal long bonefragment and threeindeterminate bone CRANIAL VAULT (Neurocranium) 6 fragments weighing atotal of 8g,all of which were TEETH 0.5 unburnt. RIBS 0.5 Conclusion: Aminimum of three adult individuals UPPER LIMB BONES 3.5 were identified, two possibly male. Periostitiswas LONG BONE UNIDENTIFIED (>4 mm) 5 presentinthe lower limb bones. UNIDENTIFIED FRAGMENTS 2 TOTAL WEIGHT17.5 g

157 Life andDeathinaRoman City

Cremation burial1769 mammal sized cranial, vertebral, rib,longbone and indeterminate fragments whichmay alsobepig. Description: Amongstthe skull bones, only one Fusion of the radius fragments indicatesthatthe pig fragment of maxilla and one of parietal,temporal was between12and 42 months old at death. The burnt and occipital were identified. Theoccipital bone stateofthe bonesnegates analysis of osteometric data, presented aportion of the lambdoidsuture whilst but the boneswerefrom alarge animal, similar in size the maxilla presented three alveoli (incisors, canine) to thatofa40 month female wildboarheldinthe where the teeth had beenlost post-mortem. Amongst English Heritage referencecollection (CfA 3289). The the 29 dental fragmentswere afirstand two second faunal pyregoodscan be interpreted as acompletepig lower incisors, an upper canine,four upper molars skeleton, and possibly afurtherbirdpyregood. including asecond upper molar,and possibly two Conclusion: Theremains of one adultindividual, lower molars minimum. 47 rib shaft fragments were likely to be youngerthan 45 yearsofage. presentand amongstthe few vertebralfragments, the atlas vertebra was identified. With regard to long CREMATIONBURIAL 1769 bones, the humerus, radius,ulna, femur,metacarpal, metatarsal and hand and foot phalanges were repre- BONE ELEMENTWEIGHT (g) sented.The maximum size of afragment was 44 mm CRANIAL VAULT (Neurocranium) 27 measured on ashaft fragment of aradius. TEETH 7 MNI:Only the remains of 1individual were RIBS 3.5 identified. VERTEBRAE 1 Age: The dimensionsand texture of the bones PELVIS 0.5 indicatedanadultindividual. More importantly, UPPER LIMB BONES 22 the full formation of the first and second molars LOWER LIMB BONES 16 indicatedaminimum of 15 yearsofage. The HAND BONES 1.5 observation of no dental wear (no dentine exposed, FOOT BONES 1 cusps unworn) in aportion of molar may indicate a LONG BONE UNIDENTIFIED (>4 mm) 45 young adultindividual. Thesagittaland lambdoid UNIDENTIFIED FRAGMENTS 136.5 sutureswerealso open. This may tentatively indicate TOTAL WEIGHT261 g that thisindividual was not of old age, bearing also in mind that this individual was notedentulous. Cremation burial1770 Sex: Sexcould not be determined due to the lack of the landmarks requiredfor sex estimation. Description: Only atotal of 23 fragments(8.5 g): 5 Pathological lesions: Twoskull vault fragments (one from the skull vault, 1rib fragment, 1humerus from occipital bone)havemarked pitting in the ecto- fragment, 1radiusfragment and the rest unidenti- cranial surface (1.5g/26.5g=5.6%);two tibiafrag- fied. The maximumfragment size was 39 mm from a ments(8g)withperiostitis;one metacarpal shaft with long bone, perhaps humerus. activeperiostitis(0.5g)and one unidentifiedshaft MNI:Only the remains of 1individual were fragmentwithactiveperiostitistoo (0.5 g); one frag- identified. ment unidentified(either skull or pelvis) withabnor- Age: All the bones seemed to indicate adult dimen- mal andhighdensitypitting; one metatarsalfragment sions. Thereforethe remainspresenthere belonged (1 g) alsowithperiostitis; 14 long bone shaft frag- most probablytoanadult. mentswithactiveperiostitis (8 g) and similarwithone Sex: Sexcould not be determined due to the lack of fragmentofilumbody, same typeofperiosteal reac- the landmarks requiredfor sex estimation. tion, pitting andsomestriae. If we take allthe long Pathological lesions:Nopathological lesionswereob- bone fragments includinghandand foot bones, the served,althoughtwo long bone fragmentsmay have percentage of fragments withperiostitiswithregard hadsomeboneremodelling butthe striae presentis to weight is 21%(18/85.5). There seemstobeasys- unsure if it is duetopost-mortem disturbances. temicinfection in this skeleton.Finally onlyone joint Colour and fracturepattern: The bonesreflected the surface wasobserved, thefacet for dens articulation typicalfissure patterncharacteristicofgreen bones in atlas, butthere werenodegenerative changes. when burnt. All but two fragments (<0.5g)which Colourand fracture pattern: Thebones reflected the weredark blue,were whitewith shades of lightgrey. typical fissure pattern characteristic of green bones Conclusion: One individual, possibly adult. when burnt. The colour was predominantlywhite (22 g) and white with shades of blue and grey CREMATIONBURIAL 1770 (152 g), with some fragments from the upper and BONE ELEMENTWEIGHT (g) lower limbs and even the skull predominantly grey (15 g), blue(6g)and lastly brown (1 g). In summary, CRANIAL VAULT (Neurocranium) 1 ahomogeneous cremationfired at high combustion. RIBS 0.5 Animal remains:Cremation burial 1769containedthe UPPER LIMB BONES 1 greatestquantityofburnt animal bone (325 fragments, LONG BONE UNIDENTIFIED (>4 mm) 5 119 g). These fragments comprised pig skull, right UNIDENTIFIED FRAGMENTS 1 radius and lefttibia fragments, and further medium TOTAL WEIGHT8.5 g

158 Appendices

APPENDIX 5: PHOTOMICROGRAPHS OF THE TOMBSTONES by Kevin Hayward Photomicrograph of tombstone1: Martialis (Plate A5.1) Rock Type: Oosparite (Folk 1959; 1962) Geological Source: Painswick Stone –Lower Jurassic (Aalenian), Painswick Hill, Gloucestershire GR SO 865 115 Note:Horny, angular ooid edges, spalledrims where the edges of the Ooids (round grains) fracture. Cement between grains coarse (poikilotopic).

Plate A5.2 Photomicrograph of tombstone 2(Field of view 2.4 mm plane polarised light).

Comparative photomicrograph of Painswick Stone (Plate A5.3) Geological Source: Catsbrain Quarry, Painswick Stone –Lower Jurassic (Aale- nian) Painswick Hill, Glouces- tershireGRSO865 115

Plate A5.1 Photomicrograph of tombstone 1(Field of view 2.4 mm plane polarised light).

Photomicrograph of tombstone2: Lucius Octavius Martialis (Plate A5.2) Rock Type: Oosparite (Folk 1959; 1962) Geological Source: Painswick Stone –Lower Jurassic (Aalenian), Painswick Hill, Gloucestershire GR SO 865 115 Plate A5.3 Comparative photomicrograph of Painswick Note:Horny, angular ooid edges, Stone (Field of view 2.4 mm plane polarised light). spalledrims wherethe edges of the Ooids(round grains) fracture.Cement betweengrains coarse (poikilotopic).

159 Life andDeathinaRoman City

APPENDIX 6: THE PLEISTOCENEVERTIBRATE REMAINS by Dr Danielle Schreve Introduction An assemblage of vertebrate remains comprisingthe very fragmentary partialremains of asingleadult hippopotamus ( Hippopotamusamphibius), with addi- tionalmaterial from asingleindividual each of probablebison (cf. Bison priscus)and elephant (Elephantidae sp.) were recovered during the excavation. The remains were found in sediments attributed to the Wotton hillock gravelcap, which had previously been mappedasforming part of the fourth terrace of the River Severn (eg Wills 1938; Worssam et al.1989). They lay withinarestricted area of afew square metres within the lowerpartof the gravel, and included anumber of elements that were found in closeassociation and may have been articulated, most notablyafemoral shaft (sf 151, Plate Plate A6.2 Comminuted upper canine of Hippopotamus A6.1),distal femoral condyle (sf 187) and tibia (sf amphibius. 186). undertake as much refitting and stabilisation work as possible during the analysis in order to maximise the amountofidentifiable material. Although timeconsuming, thisapproach proved extremely worth- whilebyfacilitating the identification of both dental and post-cranial remains that would otherwisehave been impossible to recognise. The clay blocks were washedthrough a500m nsieveinorder to check for additional large bone or tooth material preserved within them and to recover microvertebrate and invertebrate remains. Again, the approach proved justifiedwhen avirtually complete molar of hippo- Plate A6.1 Partial femoral shaft of Hippopotamus am- potamus (sf186, Plate A6.3) was recoveredfrom phibius. insideone of the blocks, together with long bone fragments of probable hippopotamustibia. No small vertebrates were noted in any of the residues. Measurements, where possible, were taken with Methodology electronic callipers according to the methodologyof Although some specimens were cleanedprior to von den Driesch (1976). study, the faunal material was presented predominantly in field damp condition, with several specimens lifted in blocksofsoil. Preservation of the material is extremely poor, both in terms of the overall surface condition and the high degree of fragmentation. This is particularly noticeable with elements of the dentition, especially caninesand incisors, which are exceptionally comminuted and are represented by many hundreds of small fragments(Plate A6.2).None of the material extracted is complete and in some cases, the original surface of the bone has been completely destroyed, exposing the cancellous bone below. Virtually all the material is heavily crushed, although it is not possible to establish whether thisis the result of past activities such as tramplingbylarge mammals, sediment pressure or more recent factors such as damage by heavy machinery at the ground surface. Acombination of old breaks and recent excavation damageisalso apparent on most of the material. However some conjoining bonefragments Plate A6.3 Virtually complete ?LM3 of Hippopotamus were observed and it was therefore decided to amphibius.

160 Appendices

Species list The list of identified species is given below, with Number of Identified Specimens (NISP) and Mini- mum Numbers of Individuals (MNI)inbrackets. PROBOSCIDEA Elephantidae Elephant undet., indeterminate elephant(NISP =1; MNI =1) ARTIODACTLYA Hippopotamidae Hippopotamus cf. amphibius Linne´ ,1758, hippopotamus (NISP =21; MNI =1) Bovidae Bovini sp., large bovid,cf. Bison priscus Bojanus, 1827, bison (NISP =4;MNI =1) Plate A6.4 Reconstructed cervical vertebra of Hippopo- tamus amphibius. Description of the finds The most diagnostic elements within the assemblage tibia. Separation of the postcranialremains of are dental remainsofhippopotamus, attributedto aurochs(Bos primigenius)and Pleistocene bison Hippopotamus amphibius ,the extant African species. ( Bison priscus)isnotoriously difficult (Gee 1993) but The anterior-projecting lower incisor tusks are large evaluation of the morphology of the specimens with asmooth surface and blunted points, whereas against recent and fossil comparative material the upper incisors are muchsmaller.Refitting of suggested closer resemblance to the latter. Addi- specimensf1052, the verycrushed ivory fragments tionallarge long bonefragments (sf 151) have been of aleft lower 1stincisor, allowedthe bluntedwear tentatively attributed to an indeterminate elephant facet to be clearly discerned. The upper and lower (Elephantidae sp.) on account of the compactand caninesconsist of large, curving tusks,withfine massive aspect of the material. grooves running lengthwise along their surface; in the largest modern adults, these may reach up to Taphonomy 700 mm overall (300 mm above the gum) (Skinner and Smithers 1990). The canines, likethe incisors, From the remains recovered,itappears that amajor grow continuously. They are triangular in section part of the skull and dentition, with some vertebrae and aresharpened against each other, thereby and postcranial elements, of asingle adulthippopo- creatingapronouncedflat wear facet. The cheek tamus was presentatthe site. Although these are in teeth (premolars and molars) are also highly poor conditionand demonstrate evidence of abra- characteristic and readilyidentifiable. They are sion, the presence of associated remains provides a relatively low-crowned when compared to other strong indication that they have not beentrans- herbivores, areflectionofthe soft dietofthe animals. ported agreatdistancesince death. In particular, the Premolars have asinglelarge cusp that risestoa presence of amandibular process, fragmentsof rounded point, with apronouncedcingulum (ridge) mandibular ramus and isolated teeth suggests that and smaller accessory cusps nearthe base. The acomplete lower jaw with in situ dentition once molarshave four cusps, which rise to high peaks. occurred at the site. It is possible that more of the Once in use, these form ahighly recognisable, almost skeleton was initially presentbut only afew, ‘flower-like’ pattern of wear. Three partial molars incomplete postcranial elements have been identi- have beenidentified withinthe assemblage (sf 185, fied. The majorityofthe postcranial skeletonmay 186 and 194). therefore have been destroyed previously or more The postcranial remains of hippopotamusfrom widelyscattered afterdisarticulation, beyond the the site consist of apartialatlas vertebra, onecervical limits of the site under investigation. vertebra(Plate A6.4), two partialthoracic vertebrae In additiontothe high degree of fragmentation, and othervertebral fragments. Major anterior ele- most of the material is heavily iron-stained, from the ments such as the humeriand scapulae are appar- post-depositional uptake of iron oxides, and bears ently missing, although distal parts of the limbs are evidence of moderateabrasion from transportation. presentinthe form of aproximal rightradius, a Asmall number of specimens also show the fine proximal left 2nd metacarpaland afirst phalanx. tracery of acid etchingleft by roots. This indicates The only otheranimal identified to species level is that eitherthe decomposing carcass lay exposed on alarge bovid, identified on the basis of four the land surface foraconsiderable period prior to its postcranial elements: adistal humerus condylar burial in the fluvial sediments, or that the bones fragment, an incomplete calcaneum, afemoral becamere-exposed at alater date, thereby allowing diaphysis fragment and afragmentary proximal left them to become root-damaged before being buried

161 Life andDeathinaRoman City asecond time. No evidence of carnivore, human or episodes alike and inhabit awide range of environ- othermodification, such as gnawmarks or cutmarks, ments at the presentday, including both closed was noted although the fragmentary state of the woodland and more openhabitats. material may have obscuredsuch evidence. The vertebrate assemblage from the site therefore suggests that the remains were laid down under temperate conditions with mean summertempera- Evidence of local environment and climate from tures several degrees warmer than at presentin the vertebrate remains proximity to asubstantial body of water. In termsof Although the condition of the bones is poor, the the surrounding vegetation, amosaic of environ- different elements form acoherent assemblage that ments with some regional temperateforest and may credibly be interpretedasasingleentityin locally open grassland is inferred. terms of its palaeoecology. The modern hippopotamus (H. amphibius)istoday restricted to sub-Saharan Biostratigraphyand the date of the Wotton Africa although in historical times it extended from gravel cap the Nile Valley into the Mediterranean. The smaller pygmy hippopotamus ( Choeropsis liberiensis )ofWest The utility of mammalian assemblages as indicators Africa is not closely related to the Pleistocene species. of relative age for Pleistocene sediments has been In Britain, hippos were restricted to interglacial repeatedly demonstrated(eg Currantand Jacobi occurrences during the Pleistocene and were extre- 2001; Schreve 2001; Stuart and Lister2001). The mely widespread when present, extending as far vertebrate assemblage from the site, although small, north as North Yorkshire and as far west as south is considered to be particularlyage-diagnostic on Wales during the Last Interglacial. As amphibious accountofthe presence of hippopotamus. Rare early mammalstheyare foundinclose association with MiddlePleistocene (‘Cromerian Complex’) occur- lakes or slow-flowing rivers sufficiently deep enough rences of an extinct hippopotamus ( Hippopotamus to allow total immersion, spending most of the day in amtiquus = H. major)are known from anumber of the water and feeding nocturnally. Notableexce- sites in East Anglia,for example Norton Subcourse ptions to immediate proximity to water are the (Lewis et al. 2004). However, the modern species, numerous fossilrecords from Last Interglacial age Hippopotamus amphibius,isdistinctive in that it makes cave sites(Marine Oxygen Isotope SubStage [MIS] only asingleappearanceinthe British Pleistocene 5e), particularly thosefrom North Yorkshire such as record, during the Last Interglacial. It is not known Kirkdale Cave and Victoria Cave.Atthe latter site, from any other late Middle Pleistocene or Late the cave is located more than 2kmfrom and about Pleistocene temperateepisode (Schreve 2001) and is 290 mabovethe floodplain of the RiverRibble, the thus an excellent biostratigraphical indicator. The nearest source of water. It is assumed that the mammalian fauna of the Last Interglacial climatic animals left the water at night to graze on the rich optimum(the Joint Mitnor Cave Mammal Assem- limestone herb flora (Stuart1982a); even today blage-Zone of Currantand Jacobi 2001), of which hipposmay travelsubstantial distances at night in hippopotamus is the most characteristic component, search of food. has been dated on the basis of consistent radiometric During the Last Interglacial, aperiod for which datingofdeposits clearlyassociatedwith thisfauna there are abundant palaeoenvironmental records to around 125 000yearsBP. Uranium-series dating of available, hippopotamusapparently occurred during speleothem sealing deposits containing aHippopo- the climatic optimumofthe interglacial (pollenzones tamus assemblage at VictoriaCave has produced II-early III) but notlater in the stage(Stuart 1982b). age-estimates of 120 + 6kaBP(Gascoyne et al. 1981) Thereare no records of this animal from later parts of and these have since been corroboratedatother Last MIS 5, the early Devensian (Currant and Jacobi2001). Interglacial sites by further U-series age-estimates, During MIS 5e mean summertemperatures were for example 129–116 ka BP obtained from stalagmite approximately 4 – Cwarmer than in southern Britain fragments at Bacon Hole,Gower(Stringer et al. 1986). today(Coope2001). Temperate mixed-oak forest In the light of this the presence of hippopotamus with Quercus, Pinus, Corylus, Alnus and Carpinus was would suggest that the deposits from 120–122 widespread (eg Sparks and West 1970) but in many London Road contain material of Last Interglacial cases the floodplain vegetation was locally defor- (MIS 5e) age, although thismay have been reworked ested and dominated by herbaceous vegetation such into youngerdeposits (see below). as plantain ( Plantago sp.), the likely result of Adiscrepancy therefore exists with the current age extensive grazing and trampling by large herbivores. interpretation of the Wotton gravelspread, which At Barrington, Cambridgeshire, where hippopota- has been mapped by the British Geological Surveyas mus remains were extremely abundant, the sedi- part of the fourthterrace of the River Severn ments had anotably high mineragenic content sequence (Wills 1938), and the presence within it of implying in-washing of soils from bare trampled remains of H. amphibius.The fourth terrace is underground, and the pollen spectra consisted of 90% herb lain by sands and gravels of the Kidderminster pollen, implying clear modification of the vegetation Station Member, generally located 20–8 mabovethe by the megaherbivores (Gibbard and Stuart 1975). modern River Severn (Maddy et al. 1995). Although Bison are foundincold-climate and temperate no direct geochronologicalinformationisavailable

162 Appendices the KidderminsterStation Member has beencorre- Summertown-Radley gravelinthe Upper Thames lated with the Ailstone Member of the Avon Valley (Bridgland 1994). Formation (Maddy et al. 1991) and associated Depositscontaining other reworkedremainsof deposits at Strensham in Worcestershire (deRouf- Last Interglacial hippopotamus have been recognised fignac, et al. 1995), both of which have been nearbyatthe base of the Holt Heath Member in attributed to MIS 7–6 ( c 245–128 000 years BP) on tributary valleys of the Severn, at Stourbridge in the the basis of amino-acidgeochronology (Maddy et al. Stour valley(Boulton 1917). However no actual Last 1995; Bowen 1999). This would suggest asubstan- Interglacial deposits have yet been identified in the tially older agethan the Last Interglacial. However, SevernValley itself, the closest being the NewInn hippopotamus has not been recognisedinany MIS7 Member at Cropthorne(Strickland in Jardine 1858) interglacial deposits despite the availability of and Eckington Railway Cutting (Keenand Bridgland numerous sites and many thousandsofspecimens 1986) in the Avon valley. It is therefore likely that the (Schreve 2001). That observation strongly suggests hippopotamus came from now-vanished Last Inter- that the current age interpretation forthe Wotton glacial deposits which were reworked during the gravelasfourth terrace is incorrectand shouldbere- early part of the ensuingDevensian cold stage(MIS examined. 5d/b or MIS 4) when the fans would have been The hippopotamus skeleton from London Road activated. Reworking of the hippopotamus and other is clearlyincompleteand has suffered substantial faunal remains intoanearly last cold stage context breakage and some abrasion,attesting to adegreeof would also fit well with the report of woolly transportation after death.Nevertheless the presence rhinoceros (Coelodonta antiquitatis) bones found of ahippopotamusmandible with teeth still in situ within sand and gravel deposits at adepth of 1.9 m mitigates against transport over any substantial on aplotofland adjacent to the site (Sermon1996), as distance. Anumber of fan gravel deposits have been the presence of this animal would otherwisebe identified to the north of the locationofthe site palaeoecologically incompatible with hippopotamus. (D. Maddy pers.comm. 21/4/06) anditispossible The megafaunal remainsfrom 120–122 London Road that the Wotton gravelrepresents alater fan gravel are therefore of greatsignificance in demonstrating deposited on top of the Kidderminster Station the erstwhile presence of Last Interglacial deposits in gravel, similar to the deposition of the hippo- this area and also in calling into question the potamus-bearing Eynshamgravel on top of the published age of the Wotton gravelcap.

163 Bibliography

Adams,BJand Byrd, JE,2006 Resolution of small- Berry, RC,and Berry, RJ,1967 Epigeneticvariation scale commingling: acase report from the Vietnam in the human cranium, Journal of Anatomy 101, War, Forensic Science International 156,63–9 361–79 Allason-Jones, L, 1989a Ear-rings in Roman Britain, Biddle, M, 1967 Two Flavian burials fromGrange BAR Brit Ser 201,Oxford Road, Winchester, Antiquaries Journal 47,224–50 Allason-Jones, L, 1989b WomeninRoman Britain, Biddulph, E, 2005 Last orders: choosing pottery for London funerals in RomanEssex, Oxford JArchaeology 24, Anon, 1990 Verulamium, Current Archaeology 120, no. 1, 23–45 410–17 Biddulph, E, 2006 The Romancemetery at Pepper Arie` s, P, 1981 The Hour of Our Death ,New York Hill, Southfleet, Kent, CTRL integrated site report Atkins,M,1987 FoundatRichard Cound: the King- series,inCTRL digital archive,Archaeology sholm dig,Gloucester Data Service, http://ads.ahds.ac.uk/catalogue/ Atkins,Mand Garrod, AP(eds), 1987 Archaeology projArch/ctrl in Gloucester 1986, Trans Bristol Gloucestershire Birck, JL,1986 Precision K-Rb-Sr isotopic analysis - Archaeol S oc 105,233–41 application to Rb-Sr chronology, Chemical Geology Atkins,Mand Garrod, AP(eds), 1988 Archaeology 56 (1–2),72–83 in Gloucester 1987, Trans Bristol Gloucestershire Bland, Rand Burnett, A, 1988 Normanby, Lincoln- Archaeol S oc 106,209–18 shire, in RBland and ABurnett (eds), The Aufderheide, ACand Rodrı´ guez-Martı´ n, C, 1998 Normanby hoard and other Roman coinhoards, The Cambridge Encyclopedia of Human Paleopathol- Coin hoards from RomanBritain VIII, London, ogy,Cambridge 114–215 Bailey, CJ,1967 An early Iron Age/Roman-British Blockley, K, 1985 Marshfield: IronmongersPiece site at Pins Knoll. Litton Cheney.FinalReport, Excavations 1982–3: an Iron Age and Romano- Proc Dorset Natural History andArchaeol S oc 89, British settlementinthe South Cotswolds, BAR 147–59 Brit Ser 141,Oxford Barber,Band Bowsher,D,2000 The EasternCemetery Boaz, RJand Behrensmeyer, AK,1976 Hominid of Roman London: excavations 1983–1990 ,Museum taphonomy: Transportofhuman skeletal parts in of London Archaeology Service Monograph 4 an artificial fluvial environment, American J Barnes, E, 1994 Developmental defects of the axial Physical Anthropology 45,53–60 skeleton in palaeopathology, Colorado Bocquet-Appel,Jand Masset, C, 1982 Farewell to Bass, WM,1987 Human Osteology. Laboratory Field paleodemography, JHuman Evolution 11,321–33 Manual (3rd edition),Special Publication Number Bodel, J, 2000 Dealing with the dead, in VMHope 2 ,MissouriArchaeologicalSociety, Columbia and EMarshall (eds), Death andDisease in the Bateman, Cand Williams, J, 2002 The South Gate Ancient City,London, 128–51 Cemetery of Romans Gloucester: evidence from Booth,P,2001Late Romancemeteries in Oxford- excavations off Parliament Street, 2001, Glevensis shire: areview, Oxoniensia 66,13–42 35,25–8 Booth,P,Crockett, AD,Fitzpatrick, AP,Powell, AB Bauchhenss, G, 1988 Corpus Signorum Imperii Romani. and Walker, K, forthcoming The archaeology of the Deutschland III.1 GermaniaInferior. Bonn und M6 Toll,Oxford-Wessex Archaeol Joint Venture Umgebung.Milita¨ rische Grabdenma¨ ler,Bonn Monograph Bayley, Jand Butcher,S,2004 Roman brooches in Boppert,W,1992 CorpusSignorum Imperii Romani. Britain:atechnologicaland typological study based on Deutschland II.5 Germania Superior. Milita¨ rische the Richborough collection,RRCSAL 68,London Grabdenma ¨ lerausMainz und Umgebung,Mainz Bello, Sand Andrews, P, 2006 The intrinsic pattern of Boulton, WS,1917 Mammalian remains in the preservation of human skeletons and its influence glacial gravels at Stourbridge, Proc Birmingham on the interpretation of funerary behaviours,inR Natural History and Philosophical Society 14,107–12 Gowland and CKnu¨ sel, Social Archaeology of Bowen,DQ(ed.),1999 Arevised correlation of Funerary Remains,Oxford, 1–13 Quaternary deposits in the British Isles,Geological Bello, SM,Thomann, A, Signoli, M, Dutour, Oand Society Special Report 23,London Andres,P,2006 Age and sex bias in the recons- Boylston, Aand Roberts, C, 1996 The Romano- truction of past population structures, American British Cemetery at Kempston, Bedfordshire. JPhysical Anthropology 129,24–38 Reportonthe Human Skeletal Remains, Unpub- Bennett, J, 1985 Sea Mills: the Roman Town of Abonae. lished report, University of Bradford Excavations at Nazereth House 1972,Bristol Boylston, A, Wiggins, Rand Roberts, C, 1998 Bermu´ dez de Castro, JM,1989 Third molar agenesis Material evidence. Human skeletal remains, in in human prehistoric populations of the Canary GDrinkalland MForeman (eds), The AngloSaxon Islands, American JPhysical Anthropology 79,207– Cemetery at Castledyke South, Barton-On-Humber, 15 Sheffield Academic Reports 6 ,Sheffield, 221–36

165 Life andDeathinaRoman City

Boylston, A, Holst, Mand Coughlan, J, 2000 Physical Chenery, CA,2005 The analysis of 18 O/ 16 Oratios anthropology, in VFioratoand CKnu¨ sel (eds), of biogenic phosphates, NIGL Report Series Blood Red Roses. The Archaeology of aMass Grave No. 195 from the Battle of TowtonAD1461, Oxford, 45–59 Clarke,G,1979 The Roman Cemetery at Lankhills, Brickley, Mand McKinley, JI,2004 Guidelines to the Winchester Studies 3 :Pre-Roman and Roman Standards for Recording Human Remains, IFA Winchester. PartII, Oxford PaperNo 7 Clarke,S,1996 Acculturation and continuity, in J Brickley, M, Mays, Sand Ives, R, 2007 An Investiga- Websterand NCooper (eds), RomanImperialism: tion of Skeletal Indicators of Vitamin DDeficiency post-colonialperspectives,LeicesterUniversityMono- in Adults: EffectiveMarkers for Interpreting Past graphs 3 ,71–84 Living Conditions and Pollution Levels in 18th Clough, S, 2003 Report on 56 Skeletons, Disarticulated and 19th Century Birmingham, England, Amer- Remains and 18 Cremations from London Road, ican JPhysical Anthropology 132,67–79 Gloucester,The Rudyard Consultancy, unpub- Bridgland, DR,1994 Quaternary of the Thames,Geo- lished report logical Conservation Review volume 7 ,London Cohen, Aand Serjeantson, D, 1996 AManual for the Bronk Ramsey,C,1995 RadiocarbonCalibration and identification of Bird Bones from Archaeological Sites, Analysis of Stratigraphy: The OxCal Program, London Radiocarbon 37, 425–30 Collingwood,RGand Wright, RP,1965 Roman Bronk Ramsey, C, 1998 Probability and dating, Inscriptions of Britain 1. Inscriptions on Stone,Oxford Radiocarbon 40,461–74 Conheeney, J, 2000 The inhumation burials, in BE Bronk Ramsey, C, 2001 Development of the radio- Barber and DBowsher,277–96 carboncalibration program, Radiocarbon 43,355– Cook,S,2004 Archaeological Field Evaluation 118– 63 120 London Road, Gloucester,Unpublished client Bronk Ramsey, C, Higham, TFG,Owen, DC,Pike, report AWGand Hedges, REM,2002 Radiocarbon Cool, HEM, 1983 Astudy of the Roman personal dates fromthe Oxford AMS System; Datelist 31, ornaments made of metal, excluding brooches,from Archaeometry 44(3) Suppl. 1, 1–149 southern Britain, unpubl.PhD Thesis, Univ. Wales Bronk Ramsey, C, Higham, T, Bowles, Aand Cool, HEM, 2003 TheMetal SmallFinds, in Hedges, R, 2004a Improvements to the Pretreat- Foundations Archaeology 2003 ment of Bone at Oxford, Radiocarbon 46,155–63 Cool, HEM, 2004 The Roman Cemetery at Brougham, Bronk Ramsey, C, Higham ,Tand Leach,P,2004b Cumbria: Excavations 1966 and 1967,Britannia Towards High-oresicion AMS: Progressand Monograph 21,London Limitations, Radiocarbon 46,17–24 Cool, HEM, 2006 Eating and Drinking in Roman Brooks,Sand Suchey, JM,1990 Skeletal age Britain,Cambridge determination based on the os pubis: acompari- Cool, HEM, forthcoming Finds from funerary son of the Acsa´ di-Nemeske´ ri and Suchey-Brooks contexts, in LAllason-Jones (ed.), Artefacts in method, Human Evolution 5 ,227–38 Roman Britain:Their Purpose andUse, Cambridge Brothwell, DR,1981 Digging Up Bones ,Oxford Cool, HEM, Lloyd-Morgan, Gand Hooley, AD, Brothwell, Dand Zakrzewski, S, 2004 Metric and 1995 Findsfrom the Fortress.Archaeology of York non-metric studies of archaeological human bone, 17/10,York in MBrickley and JMcKinley (eds), Guidelines to Cool, HEMand Philo, C(eds.), 1998 Roman the Standards for Recording HumanRemains,IFA Castleford Excavations 1974–85. Volume I: the small paper 7 ,28–33 finds. Yorkshire Archaeology 4 ,Wakefield Buckberry, JLand Chamberlain, AT,2002 Age Coope,GR, 2001 Biostratigraphical distinctionof estimation from the auricularsurface of the ilium: interglacial coleopteran assemblages from south- arevised method, American JPhysical Anthropology ern Britain attributed to Oxygen Isotope Stages 5e 119,231–39 and 7, Quaternary Science Reviews 20,1717–22 Buikstra, JEand Ubelaker, DH,1994 Standards Costa, RLJr, 1980 Age,sex, and antemortem loss of for Data Collection from Human Skeletal Remains, teeth in prehistoric Eskimo samplesfrom Point Arkansas Archaeological Survey Research Series Hope and Kodiak Island,Alaska, American J No 44 Physical Anthropology 53,579–87 Cartwright,FFand Biddiss, MD,2004 Disease and Cox, M, 2000 Ageingadults from the skeleton, in M History,Stroud Cox and SAMays (eds), Human Osteologyin Chamberlain, A, 2000 Problems and prospectsin Archaeology and Forensic Science,Greenwich Med- palaeodemography, in MCox and SMays(eds), ical Media, London, 61–81 Human OsteologyinArchaeology and Forensic Crawford Adams,J,1983 Outline of Fractures (8th Science,Greenwich Medical Media, London, ed.), NewYork 101–15 Crowfoot, E, 1981 Textile Impressions, in CSGreen, CgMs, 2003 Archaeological Desk-based Assessment MPaterson and LBiek,ARomanCoffin-Burial 118–120 LondonRoad,Gloucester, Unpublished from the Crown Buildings Site, Dorchester: with client report particular referencetothe headofwell-preserved

166 Bibliography

hair, Proc Dorset Natural History andArchaeol Soc Emery, GT,1963 Dental pathology and archaeology, 103,67–100 Antiquity 37,274–81 Crummy, N, 1983 The Roman small finds from Erdal, YS,2006 Apre-Columbian case of congenital excavations in Colchester 1971–9,Colchester Archa- syphilis from Anatolia (Nicaea, 13th centuryAD), eological Report 2 ,Colchester International JOsteoarchaeology 16,16–33 Crummy, N, 1992 The Romansmall finds from the Esmonde Cleary, S, 2000 Putting the dead in their Culver Streetsite, in PCrummy, Excavations at place:burial locationinRomanBritain, in JPearce Culver Street, the Gilberd School, and other sitesin et al., 127–42 Colchester 1971–85, Colchester Archaeology Re- Evans,Eand Maynard, D, 1997 Caerleon Lodge Hill port 6 ,Colchester, 240–5 cemetery: the Abbeyfieldsite 1992, Britannia 28, Crummy, N, Crummy, Pand Crossan, C, 1993 169–243 Excavations of Roman and latercemeteries, churches Evans,JA, Tatham, S, Chenery, SRand Chenery, and monastic sitesinColchester, 1971–88,Colchester CA,inpress Anglo Saxon animal husbandry Archaeological Trust techniques revealed though isotope and chemical Crummy, P, Benfield, S, Crummy, N, Rigby,Vand variations in cattle teeth, AppliedGeochemistry Shimmin, D, 2007 Stanway, Britannia Monograph Farwell, DEand Molleson, TI,1993 Poundbury, 24,London vol. 2: The cemeteries,Dorset Natural Historyand Cunliffe, B, 1971 Excavations at Fishbourne1961–1969. Archaeol Soc Monograph 11 Volume II: The Finds,RRCSAL 27, Leeds Fe´ de´ ration Dentaire Internationale (FDI), 1982 An Currant, APand Jacobi, RM,2001 Aformal mam- epidemiological indexofdevelopmental defects of malian biostratigraphy for the Late Pleistocene of dental enamel (DDE Index), International Dental Britain, Quaternary Science Reviews 20,1707–16 Journal 32,159–67 Darling, WG,Bath AHand Talbot, JC,2003 The O Finnegan,M,1978 Non-metric variation of the &Hstable isotopic composition of fresh waters in infracranial skeleton, Journal of Anatomy 125,23–37 the British Isles. 2: Surface waters and ground Fiorato, V, Boylston, Aand Knu¨ sel, C(eds), 2000 water, Hydrology and Earth System sciences 7 (2), BloodRed Roses. The Archaeology of aMass Grave 183–95 from the Battle of TowtonAD1461,Oxford de Rouffignac, C, Bowen,DQ, Coope, GR,Keen, Folk, RL,1959 Practical petrographic classification DH,Lister, AL,Maddy, D, Robinson, JE,Sykes, of limestones,American Association of Petroleum GAand Walker, MJC, 1995 Late Middle Geologists Bulletin,v. 43,1–38 Pleistocene interglacial deposits at Upper Stren- Folk, RL,1962 Spectral subdivision of limestone sham, Worcestershire, England, Journal of Quater- types: Classification of carbonateRocks, Ameri- nary Science 10,15–31 can Association of Petroleum Geologists Memoir 1 , Dennehy, E, 2001 Children’s burial ground , in 62–84 Archaeology Ireland 15,20–3 Foundations Archaeology, 2003 124–130 London Road, Dias,Gand Tayles, N, 1997 ‘Abscess cavity’ –amis- Gloucester: Archaeological excavation report and post- nomer, International JOsteoarchaeology 7 ,548–54 excavation assessment,http://www.foundations. Dobney, KM,and Reilly, K, 1988 Amethodfor co.uk/reports/gloucestershire/lrg.shtml recording archaeological animal bones: the use of Frere, SS,1991a Britannia ,3rd edition, London diagnostic zones, Circaea 5 ,79–96 Frere, SS,1991b RomanBritain in 1990: sites Down,Aand Rule, M, 1971 ChichesterExcavations 1, explored, Britannia 22,256–308 Chichester Frere, SS,1992 RomanBritain in 1991: sites explored, Drancourt M, Roux V, Dang LV,Tran-Hung L, Britannia 23,222–92 CastexD,Chenal Francisque V, OgataH,Fournier Fulford,M,1999 Veteran Settlement in 1st century PE,Crubezy Eand RaoultD,2004 Genotyping, Britain and the foundations of Gloucester and Orientalis-like Yersinia pestis, and plague pan- Lincoln, in HHurst (ed.), 177–80 demics, Emerg Infect Dis. 10,1585–92 Fullbrook-Leggatt, LEWO, 1933 Glevum, Trans Driel-Murray,Cvan, 1999 And did thosefeet in Bristol GloucestershireArchaeol S oc 55,55–104 ancienttimes ... Feet and shoes as amaterial Galloway, A, 1999 Broken Bones. Anthropological projection of the self, in PBaker (ed.), TRAC98 , analysis of blunt forcetrauma ,Springfield Leicester, Oxford, 131–40 Garrod, Pand Heighway,C,1984 Garrod’sGloucester: Duday, H, 2006 L’arche´ othanatologie ou l’arche´ o- archaeological observations 1974–81,Western Archa- logie de la mort (Archaeoethnoanatology or the eological Trust, Gloucester Archaeology of Death),inRGowland and C Gascoyne, M, Currant, APand Lord, T, 1981 Knu¨ sel (eds), Social Archaeology of Funerary Re- Ipswichian fauna of Victoria Cave and the marine mains ,Oxford, 30–56 palaeoclimaticrecord, Nature 294,652–4 Duncan-Jones, RP,1996 The impact of the Antonine Gee, H, 1993 Thedistinction between postcranial Plague, JRoman Archaeol 9 ,108–36 bones of Bos primigenius Bojanus, 1827 and Bison Duray, SM,1996 Dental indicators of stress and priscus Bojanus, 1827 from the British Pleistocene reducedage at death in prehistoric native Amer- and the taxonomicstatusof Bos and Bison, Journal icans, American JPhysical Anthropology 99,275–86 of Quaternary Science 8 ,79–92

167 Life andDeathinaRoman City

Gejvall, NG,1963 Cremations, in DBrothwell and Heighway, CM,1980 Romancemeteries in Glouce- EHiggs (eds), Science in Archaeology,London, ster District, Trans Bristol GloucestershireArchaeol 379–90 S oc 98,57–72 Genove´ s, S, 1962 Introduccio´ nalDiagno´ stico de la Henig, M, 1974 Acorpus of Roman engravedgemstones Edad ydel Sexo en RestosO´ seosPrehisto´ ricos, from British sites,BAR 8 ,Oxford Universidad Nacional Auto´ noma de Me´ xico, Henig, M, 2004 CorpusSignorum Imperii Romani. Mexico Great Britain I.9 Roman Sculpture from the North Gibbard, PLand Stuart,AJ, 1975 Flora and West Midlands,British Academy, Oxford vertebrate fauna of the BarringtonBeds, Geological Herbert, NM,1988 The Victoria County History of Magazine 112,493–501 Gloucestershire, Volume IV: Gloucester City Gilliam,JF, 1961 The Plagueunder Marcus Aurelius, Hillson,S,1996a Mammal bones and Teeth. An American Journal of Philology 82(3),225–51 Introductory Guide to Methods of Identification, Gloucester Archaeology, 1993 ArchaeologicalEva- London luationofFormer Trent-Severn Depot, London Hillson,S,1996b Dental Anthropology,Cambridge Road, Gloucester, Unpublished client report Hillson,S,2001Recording dental caries in archae- Gloucester Archaeology, 2002 Gloucester Archaeol- ologicalhuman remains, International JOsteoarch- ogy Unit Annual Report, 2002, Glevensis 36,55–63 aeology 11,249–89 Gowland, R, 2001 Playing dead: implications of Hillson,S,Grigson,Cand Bond, S, 1998 Dental mortuary evidence for the social construction of defects of congenital syphilis, American JPhysical childhood in Roman Britain, in GDavies, A Anthropology 107,25–40 Gardnerand KLockyear (eds), TRAC 2000: Holden, JL,Phakley, PPand Clement,JG, 1995 Proceedings of the Tenth Annual Theoretical Roman Scanning electron microscopeobservations of Archaeology Conference London 2000,Oxford, 152– heat-treated human bone, ForensicScience Interna- 68 tional 74,29–45 Grauer, A, 1991 Patterns of life and death:the Hope, VM,2000 Contempt and respect: the palaeodemography of medieval York, in HBush treatment of the corpse in ancient Rome, in VM and MZvelebil(eds), Health in Past Societies. Hope andEMarshall (eds), Death and Disease in Biocultural Interpretations of Human Skeletal Re- the Ancient City,London, 104–27 mainsinArchaeological Contexts, BAR Int Ser 567, Hopkins, K, 1983 Deathand Renewal,Cambridge Oxford, 67–80 Hurlburt, S, 2000 The taphonomy of cannibalism: a Grauer, ALand Roberts, CA,1996 Paleoepidemiol- review of anthropogenic bonemodification in the ogy, healing, and possible treatment of trauma in American Southwest, International JOsteoarchaeol- amedieval cemetery populationofSt. Helen-on- ogy 10,4–26 the-Walls, York, England, American JPhysical Hurst, H, 1972 Excavations at Gloucester, 1968–71: Anthropology 100,531–44 first interim report, Antiquaries Journal 52,24–69 Greenburg, J, 2003 Plagued by doubt: reconsidering Hurst, H, 1974 Excavations at Gloucester, 1971– the impact of amortalitycrisis in the 2nd c. AD, J 1973: second interim report, Antiquaries Journal Roman Archaeology 16,413–25 54,8–52 Greep, S, 1998 The bone, antler andivory artefacts, in Hurst, H, 1985 Kingsholm,Gloucester Archaeological HEMCool and CPhilo, 267–85 Reports Vol. 1 Haglund, WDand Sorg, MH(eds), 1997 Forensic Hurst, H, 1988Gloucester ( Glevum), in GWebster Taphonomy. The Postmortem Fate of Human Remains, (ed.), Fortress into city,London, 48–73 New York Hurst, H(ed.),1999a The Coloniae of Roman Britain, Harman, M, Molleson, Tand Price, JL,1981 Burials, Journal of Roman Archaeology Supplementary bodies and beheadings in Romano-British and Series 36 Anglo-Saxon cemeteries, Bull British Museum Nat Hurst, H, 1999b Topography and Identity in Glevum Hist Geol 35(5),145–88 Colonia, in HHurst (ed.), 113–35 Hassall, M, 2000 Pre-Hadrianic Legionary Disposi- Hurst, H, 2005 RomanCirencester and Gloucester tions in Britain, in RJBrewer, Roman Fortresses and compared, Oxford JArchaeology 24(3),293–305 their legions,Soc Antiq And Nat Museum Wales, Huskinson, J, 1994 Corpus Signorum Imperii Romani. Cardiff, 51–67 Great BritainI.8 Roman Sculpture from Eastern Hassall, Mand Hurst, H, 1999Soldier and Civilian: a England ,BritishAcademy, Oxford debate on the bank of the Severn, in HHurst (ed.), Isaksen, L, Loe Land Saunders M, forthcoming, 181–90 X-Bones: Anew approach to recording skeletons Hassall, MWCand Tomlin, RSO, 1984 Roman in 3D, in GEarl (ed.), Proceedings of the 2006 &2007 Britain in 1983: II. Inscriptions, Britannia 15,333–56 ComputerApplications and Quantitative Methods in Hauser, Gand De Stefano, GF,1989 Epigenetic Archaeology UK Conferences, Southampton Variants of the Human Skull,Schweizerbart, Stutt- Jacobi, F, Bramanti, B, Dresely, Vand Alt, KW,2006 gart Why did they have to die? An Iron Age mass grave Hayes, J, 1991 Second century grave of awealthy from Westerhausen, Germany,Poster presented at citizen, The Searcher 71,18–21 the 16th Paleopathology AssociationEuropean

168 Bibliography

meeting 28th Aug-1st September2006, Santorini, Lewis, Mand Roberts, C, 1997 Growing pains: the Greece interpretation of stress indicators, International J Jardine, W, 1858 Memoirs of Hugh Edwin Strickland, Osteoarchaeology 7 ,581–6 London Lewis, SG,Parfitt, SA,Preece, RC,Sinclair, J, Jones, R, 1987 Burial Customs of Rome and the Coope,GR, Field, MH,Maher, BA,Scaife,RG Provinces, in JWacher (ed.), The Roman World , and Whittaker, JE,2004 Age and palaeoenviron- London and New York, 812–44 mental setting of the Pleistocene vertebrate fauna Jurmain, RD,1999 Stories from the Skeleton. Beha- at Norton Subcourse, Norfolk, in DCSchreve viouralReconstruction in Human Osteology.Inter- (ed.), The Quaternary Mammals of Southern and preting the Remains of the Past 1, Amsterdam EasternEngland ,Quaternary Research Associa- Kausmally, T, 2007 East Smithfield Black Death tion, London, 5–17 Cemetery,Centre forHuman Bioarchaeology, Littleton, Jand Frohlich, B, 1993 Fish-eaters and Museum of London, http://www.museumoflon- farmers:dental pathology in the Arabian Gulf, don.org.uk/English/Collections/OnlineRe- American JPhysical Anthropology 92,427–47 sources/CHB/Database/Medieval+cemeteries/ Littman, RJand Littman, ML,1973 Galen and the ESmithfieldBlackDeath.htm Antonine Plague, American JPhilology 94,243–55 Keen, DHand Bridgland,DR, 1986 An interglacial Loe, Land Cox, M, 2005 Peri-and post mortem fauna from Avon No.3 Terrace at Eckington, surface features on archaeological human bone: Worcestershire, Proc Geologists’ Association 97, why theyshould not be ignoredand aprotocolfor 303–7 their identification and interpretation, in SR Kelley, MA,Levesque, DRand Weidl, E, 1991 Zakrzewski and MClegg (eds), Proceedings of the Contrasting patterns of dental disease in five early fifth annual conference of the British Association northern Chilean groups, in MAKelleyand CS for Biological Anthropology and Osteoarchaeology, Larsen (eds), Advances in Dental Anthropology, Oxford, 11–21 New York, 203–13 Mackreth, DF,1995 Pre-Roman and Roman Kennedy, KAR,1989 Skeletal markers of occupa- brooches, in KBlockley,MBlockley, PBlockley, tionalstress, in YMI˙ s¸ can and KARKennedy SSFrereand SStow, Excavations in the Marlowe (eds), Reconstruction of Life From the Skeleton,New car park and surrounding areas,Archaeology of York, 129–60 Canterbury V ,Canterbury, 955–82 Key, CA,Aiello, LCand Molleson, T, 1994 Cra- Maddy, D, Green, CP,Lewis, SGand Bowen, DQ, nial suture closure and its implications for age 1995 Pleistocene geology of the Lower Severn estimation, International JOsteoarchaeology 4 , Valley, UK, Quaternary Science Reviews 14,209–22 193–207 Maddy, D, Keen, DH,Bridgland, DR,Green, CP, King, T, Hillson, Sand Humphrey, LT,2002 A 1991 Arevisedmodel forthe Pleistocene devel- detailed study of enamel hypoplasia in apost- opment of the RiverAvon, Warwickshire, J medieval adolescent of known age and sex, Geological Society of London 148,473–84 Archives of Oral Biology 47,29–39 Manning, WH,1985 Catalogue of Romano-British Iron Kooijmans, LP,2005 Bronze war, Acollective burial at Tools, Fittings and Weaponsinthe British Museum, Wassenaar,Amsterdam London Langsjoen, O, 1998Diseases of the dentition, in AC Manning, WH,2000 The Fortresses of LegioXX, in R Aufderheide and CRodrı´ guez-Martı´ n, 393–412 JBrewer (ed.), Roman Fortresses and their Legions, Larsen, CS,1997 Bioarchaeology. InterpretingBeha- Soc Antiquaries of Londonand Nat Museum and viour from the Human Skeleton, Cambridge Studies Galleries of Wales,London in Bioanthropology, Cambridge Maresh, MM,1972 Measurements from roentgen- Larsen, CS,Shavit, Rand Griffin, MC,1991 Dental ograms, in RWMcCammon, (ed.), Human Growth caries evidence for dietary change: an archaeolo- and Development,Springfield,IL, 157–200 gical context, in MAKelleyand CSLarsen (eds), Margerison, BJ,1997 AComparison of the Palaeo- Advances in Dental Anthropology,New York, demography of Catastrophic and Attritional 179–202 Cemeteries,Unpublished Ph.D. thesis, University Lasker, GW,1951 Genetic analysis of racial traits of of Bradford the teeth, Cold Spring Harbor Symposia on Quanti- Margerison, BJ,and Knu ¨ sel, CJ,2002 Paleodemo- tative Biology 15,191–203 graphic comparison of acatastrophic and an Lauwerier, RCGM,1993 Bird remains in Roman attritional death assemblage, American JPhysical graves, Archaeofauna 2 ,75–82 Anthropology 119,134–43 Levinson, AA,Luz, Band Kolodny, Y, 1987 Martin-Kilcher, S, 2000 Mors immatura in the Roman Variations in oxygen isotopic compositions of world—a mirror of societyand tradition,in humanteeth and urinarystones, AppliedGeochem- Pearce et al.,63–77 istry 2 ,367–71 Mattingly, D, 2006 An Imperial Possession: Britain in Lewis, ME,2004 Endocranial lesions: their distribu- the RomanEmpire, London tion and aetiology, International JOsteoarchaeology May, T, 1930 Catalogue of the Roman Pottery in the 14,82–97 Colchester and Essex Museum,Cambridge

169 Life andDeathinaRoman City

Mays,SA, 1993Infanticide in RomanBritian, Miles, D, Palmer, S, Smith, Aand Jones, GP,2007 Antiquity 67,883–8 Iron Age and Roman settlement in the UpperThames Mays,S,1998 The Archaeology of Human Bones, Valley. Thames Valley Landscapes Monograph 26, London Oxford Archaeology Mays,S,2000 The archaeologyand historyof Millett,M,1986 An early Romancemetery from infanticide, in JSofaer-Derevenski (ed.), Children Alton, Hampshire, Proc Hampshire FieldClub and and Material Culture,London, 180–90 Archaeol Soc 42,43–87 Mays,S,2002 The relationship between molar wear Millett,M,1990 The Romanization of Britain: an essay and ageinanearly 19th century AD archae- in archaeological interpretation,Cambridge ologicalhuman skeletal series documented age at Mills, JM,1993Hobnails, in DEFarwelland TL death, JArchaeological Science 29,861–71 Molleson1993 McKern, TW,1970 Estimation of skeletal age: from Molleson, Tand Cox, M, 1993 The Spitalfields Project puberty to about 30 yearsofage, in TDStewart Volume 2–The Anthropology –The Middling Sort, (ed.), Personal Identification in Mass Disasters, CBA Res Rep 86,York National Museum of Natural History, Washing- Montgomery, J, 2002 Lead andStrontium Isotope ton D.C., 41–56 CompositionsofHumanDental Tissues as an McKinley, JI,1993 Bone fragment size and weights Indicator of Ancient Exposure and Population of bone from modern British cremations and it Dynamics, Unpublished PhD thesis, Universityof implications forthe interpretation of archaeologi- Bradford cal cremations, International JOsteoarchaeology 3 , Montgomery, J, Evans, JAand Wildman, G, 2006 283–87 87Sr/86Sr isotope composition of bottled British McKinley, JI,1994 Bone fragment size in British mineral waters for environmental and forensic cremation burials and its implications for pyre purposes, AppliedGeochemistry 21,1626–34 technology and ritual, JArchaeological Science 21, Moore, T, 2006 Iron Age Societies in the Severn- 339–42 Cotswolds: developing narratives of social and land- McKinley, JI,1997 Cremated human bone from scape change,BAR Brit Ser 421,Oxford burial andcremation-related contexts, in AP Moorrees, CFA, Fanning,EAand Hunt, EEJr, Fitzpatrick, Archaeological Excavations on the Route 1963 Age variation of formation stages forten of the A27 Westhampnett Bypass, West Sussex, 1992. permanent teeth, JDental Research 42,1490–1502 Volume 2: The Cemeteries,WessexArchaeology and Moyniham, PJ,1998 Update on the nomenclature of the Highways Agency, Salisbury, 244–52 carbohydrates and their dental effects, JDentistry McKinley, JI,2000a Cremation burials, in BBarber 26,209–18 and DBowsher (eds), 264–77 OA, 1992 Fieldwork Manual, (ed. DWilkinson, first McKinley, J, 2000b The analysis of cremated bone, in edition, August 1992) MCox and SMays (eds), Human osteology in O’Neil, JR,Roe, LJ,Reinhard, Eand Blake, RE,1994 archaeology and forensic science ,London, 403–21 Arapid and precise method of oxygen isotope McKinley, J, 2000c Phoenixrising: aspects of crema- analysis of biogenic phosphate, IsraelJEarth tion in RomanBritain, in JPearce et al. (eds), 38–44 Science 43,203–12 McKinley, J, 2004 Compilingaskeletal inventory: Ortner, DJandErikson, MF,1997 Bone changes in disarticulated and co-mingled remains, in M the human skull probably resulting from scurvy Brickley and JMcKinley (eds), 14–17 in infancy and childhood, International JOsteoarch- McKinley, J, forthcoming Ryknield Street, Wall (Site aeology 7 ,212–20 12), in Booth et al. forthcoming Ortner, DJand Putschar,WGJ,1981 Identification of McKinley, JIand Roberts, CA,1993 Excavation and Pathological Conditions in Human Skeletal Popula- Post-Excavation Treatment of Cremated and Inhumed tions,Washington DC Human Remains,IFA Technical Paper No 13, Ostendorf Smith, M, 1997 Osteological indications of Birmingham warfare in the Archaic period of the Western McWhirr, A, Viner, Land Wells, C, 1982 Romano- Tennessee Valley, in DLMartin and DWFrayer British Cemeteries at Cirencester,Cirencester Ex- (eds), Troubled Times. Violence andWarfare in the cavations Committee, Cirencester Past,Amsterdam, 241–66 Meindl,RSand Lovejoy, CO,1985 Ectocra- Parker Pearson, M, 1999 The Archaeology of Death and nial suture closure: Arevised method for the Burial,Stroud determination of skeletal age at death based on Parkhouse, Jand Evans, E, 1996 Excavations at the lateral-anterior sutures, American JPhysical Cowbridge, South Glamorgan 1977–88, BAR Brit Anthropology 68, 29– 45 Ser 245,Oxford Miles, A, 1962Assessment of ageofapopulation of Pearce,J,1998 From death to deposition: the Anglo-Saxons from their dentition,Proc Royal sequence of ritual in cremationburialsofthe Society of Medicine 55,881–6 Romanperiod, in CForcey, JHawthorne and R Miles, AEW,2001 The Milesmethod of assessing Witcher(eds), TRAC 97: Proceedings of the seventh age from tooth wear revisited, JArchaeological annual Theoretical Roman Archaeology Conference, Science 28,973–82 Oxford, 99–111

170 Bibliography

Pearce,J,2001 Constructions of infancy—mortuary Fairbanks, RG,Friedrich, M, Guilderson, TP, rituals for infants in the Romanprovinces’, in A Hogg,AG, Hughen, KA,Kromer, B, McCormac, Gardner et al. (eds.), Proceedings of the 9th Theoretical G, Manning, S, Bronk Ramsey,C,Reimer, RW, Roman Archaeology Conference,Oxford,125–42 Remmele, S, Southon, JR,Stuiver,M,Talamo, S, Pearce,J,Millett,Mand Struck, M(eds), 2000 Burial, Taylor, FW,van der Plicht, Jand Weyhenmeyer, Society and Context in the Roman World ,Oxford CE,2004 IntCal04 Terrestrial radiocarbon age Philpott,R,1991 Burial practices in Roman Britain, calibration, 0–26 Cal Kyr BP, Radiocarbon, 46, BAR Brit Ser 219,Oxford 1029–58 Pindborg, JJ,1982 Aetiology of developmental Resnick, Dand Niwayama, G, 1995 Osteomyelitis, enamel defects not related to fluorosis, Interna- septic arthritis, and soft tissueinfection: mechan- tional Dental Journal 32,123–34 isms and situations, in DResnick, Diagnosis of Polfer,M,2000Reconstructing funerary rituals: the Bone and Joint Disorders,3rd ed. vol. 4, London, evidence of ustrina and related archaeological 2325–418 structures, in JPearce,MMillett andMStruck Richmond, IA,1946 The four coloniae of Roman (eds), 30–7 Britain, ArchaeolJ103,57–84 Pope, EJ,O’Brian,MAand Smith, C, 2004 Rhodes, M, 1980 Leather Footwear, in DMJones, Identification of traumatic injury in burned cranial Excavations at Billingsgate Buildings ‘Triangle’, bone: an experimental approach, Journal of Foren- LowerThames Street, 1974,London and Middle- sic Science 49(3), 431–40 sex Archaeol Soc Special PaperNo. 4 ,99–128 Potter, TW,1979 Romans in north-westEngland, Roberts,CA, 1989 The human remainsfrom 76 Cumberland and WestmorlandAntiquarian and Kingsholm, Gloucester, Unpublished report, Cal- Archaeological Society Research Series 1 ,Kendal vin Wells Laboratory, UniversityofBradford, Powell, K, forthcoming The metal finds, in PBooth et Bradford al. forthcoming Roberts,C,1991 Traumaand treatment in the British Powell, ML,1985 Theanalysis of dental wear and Isles in the historic period: adesign for multi- caries fordietary reconstruction,inRIJr Gilbert disciplinary research, in DJOrtnerand AC and JHMielke (eds), The AnalysisofPrehistoric Aufderheide (eds), Human Palaeopathology: Current Diets,New York, 307–38 Synthesis and Future Options,Smithsonian Inst. Price, Jand Cottam, S, 1998 Romano-British glass Press,Washington, D.C., 225–40 vessels: ahandbook. CBA Practical Handbookin Roberts,CAand Manchester, K, 1995 The archaeology Archaeology 14,York of disease,2nd edition, New York Price, Jand Cool, HEM, 1985 Glass (includingglass Roberts,Cand Cox, M, 2003 Health and disease in from 72 Dean’s Way), in HRHurst, Kingsholm, Britain from prehistory to the present day,Stroud Gloucester Archaeol. Rep. No. 1 ,Cambridge, 41–54 Roberts,Cand Connell, B, 2004 Guidanceon Rawes, B, 1972 Romanpottery kilns at Gloucester, recording palaeopathology, in MBrickley and JI Trans Bristol GloucestershireArchaeol S oc 91,18–59 McKinley (eds), 34–9 Rawes, B1983 (ed.),Archaeological Review No. 7 Rogers,Jand Waldron, T, 1995 AField Guide to Joint 1982, Trans Bristol GloucestershireArchaeol S oc 101, Disease in Archaeology,Chichester 189–96 Rogers,J,Rogers,J,Watt, Iand Dieppe, P, 1981 Rawes, B(ed.),1991 ArchaeologicalReview No. 15 Arthritis in Saxon and Medieval skeletons, British 1990, Trans Bristol GloucestershireArchaeol S oc 109, Medical Journal 283,1668 223–38 Rosten, J, 2007 Identities in life and death in Roman Rawes, B(ed.),1992 ArchaeologicalReview No. 16 Britain: the caseofBaldock,inBCroxford, NRay, 1991, Trans Bristol GloucestershireArchaeol S oc 110, RRoth and NWhite (eds), TRAC 2006: Proceedings 219-224 of the Sixteenth Annual Theoretical Roman Archae- Rawes, B(ed.),1993 ArchaeologicalReview No. 17 ology Conference,Oxford, 172–82 1992, Trans Bristol GloucestershireArchaeol S oc 111, Rousham, EKand Humphrey, LT,2002 The 215–35 dynamics of child survival, in HMacbeth and Rawes, JAand Wills, J(eds), 1998 Archaeological PCollinson (eds), Human Population Dynamics: Review No. 22 1997, Trans Bristol Gloucestershire Cross-Disciplinary Perspectives,14th Biological Archaeol S oc 116,191–212 Society Symposium Series, Cambridge, 124–40 RCHME, 1962 An inventory of the historical monuments Rowbotham, FW,1978 The River Severnin in the city of York: Volume 1: Ebvracvm,London Gloucester, Glevensis 12,4–10 Reece, R, 1962 TheOakleyCottage Romano-British Rudder, S, 1779 ANew History of Gloucester, Cemetery, Cirencester, Trans Bristol Gloucestershire Cirencester Archaeol S oc 81,51–73 Sauer,EW, 2005 Inscriptions from Alchester: Reece, R1999 Colonia in context: Glevumand the Vespasian’s base of the Second AugustanLegio, civitas Dobunnorum, in HHurst (ed.),73–85 Britannia 36,101–33 Reimer,PJ, Baillie, MGL,Bard, E, Bayliss, A, Beck, Saville,A,1990 Hazelton North. The Excavation of a JW,Bertrand,CJH,Blackwell, PG,Buck, CE, Neolithic Long Cairn of the Cotswold-Severn Group, Burr,GS, Cutler, KB,Damon,PE, Edwards, RL, EnglishHeritage Archaeol Rep 13,London

171 Life andDeathinaRoman City

Scheuer, Land Black,S,2000 Developmental Juvenile Staelens, Y, 1982 The Birdlip Cemetery, Trans Bristol Osteology ,London GloucestershireArchaeol S oc 100 ,20–31 Schmid, E, 1972 Atlas of Animal Bones ,Amsterdam Stead,IMand Rigby, V, 1989 Verulamium. The King Schreve,DC, 2001 Differentiation of the Britishlate Harry Lane site,English Heritage Archaeol Rep 12 MiddlePleistocene interglacials: the evidence from Steckel, RH,1995 Stature and the standard of living, mammalian biostratigraphy, Quaternary Science Journal of Economic Literature 33,1903–40 Reviews 20,1693–1705 Stewart, TD,1979 Essentials of Forensic Anthro- Scott, E, 1991 Animal and infant burials in Romano- pology, Springfield British villas: arevitalisation movement,inP Stirland,AJ, 1998 Musculoskeletal evidence for Garwood, DJennings, RSkeates and JToms (eds), activity: problems of evaluation, International J Sacred and profane: proceedings of aconference on Osteoarchaeology 8 ,341–62 archaeology, ritualand religion, Oxford 1989,OUCA Stringer, CB,Currant, AP,Schwarcz,HPand Monograph 32,Oxford, 115–21 Collcutt, SN,1986 Age of Pleistocene faunas from Seeley, F, 1995 Romandoorbells, Roman Finds Group Bacon Hole, Wales, Nature 320,59–62 Newsletter IX,5–6 Stuart,AJ, 1982a Pleistocene vertebrates of the British Sermon, R, 1995 Gloucester Archaeology Annual Isles,London Report, 1995 http://www.gloucester.gov.uk/Con- Stuart,AJ, 1982b The occurrence of Hippopotamus tent.aspx?cindex=149andcitem=6749andurn=1009 in the BritishPleistocene, Quarta ¨ rpala¨ ontologie 6 , Sermon, R, 1996 Gloucester Archaeology Annual 209–18 Report, 1996 http://www.glos-city.gov.uk/Con- Stuart,AJand Lister, AM,2001 The mammalian tent.aspx?urn=1010 faunas of Pakefield/Kessingland and Corton, Suf- Sermon, R, 2003 Gloucester Archaeology Unit folk, UK: evidence for anew temperate episode in Annual Report 2002, Glevensis 36,55–63 the Britishearly Middle Pleistocene, Quaternary Shennan, S, 1997 Quantifying Archaeology,Second Science Reviews 20,1677–92 edition, Edinburgh Stuart-Macadam,P,1985 Porotichyperostosis: rep- Shipman, P, Foster, G, and Schoeninger,M,1984 resentative of achildhoodcondition, American J Burntbones and teeth:anexperimental study of Physical Anthropology 66,391–8 colour,morphology, crystal structure and shrink- Stuart-Macadam,P,1991 AmaemiainRoman age, JArchaeological Science 11,307–25 Britain: Poundburycamp, in HBush and M Signoli et al., 2004 Discovery of amass grave of Zvelebil(eds), Health in Past Societies. Biocultural Napoleonic period in Lithuania (1812, Vilnius), Interpretations of Human Skeletal Remains in Archae- ComptesRendus Palevol 3 ,Issue 3, 219–27 ological Contexts, BAR Int Ser 567,Oxford, 101–13 Sjøvold, T, 1984 Areport on the heritability of some Stuiver, Mand Kra,RS, 1986 Editorial comment, cranial measurements and non-metric traits, in G Radiocarbon, 28(2B), ii Nvan Varkand WWHowells (eds), Multivariate Stuiver, Mand Polach, HA,1977 Reporting of C 14 Statistics in Physical Anthropology,Dordrecht,The data, Radiocarbon, 19,355–63 Netherlands, 223–46 Sutherland, T, 2000 Recording the grave,inVFiorato Skinner, JDand Smithers, RHN, 1990 The Mammals et al., 36–44 of the SouthernAfrican Subregion (New Edition), Swift,E,2000 Regionality in Dress Accessories in the Pretoria late Roman West,Mongraphie Instrumentum 11. Skinner, Mand Goodman, AH,1992 Anthropolo- Millau. gical usesofdevelopmental defects of enamel, in Sykes, NJ,2001 The Norman Conquest: azooarch- MAKatzenberg and SRSaunders (eds), Skeletal aeological perspective,unpublished Ph.D. thesis, BiologyofPast Peoples: Research Methods,New University of Southampton York, 153–74 Sykes, NJ,forthcoming The Rabbit, in TPO’Connor Smith, BH,1991 Standardsofhuman tooth forma- and NJSykes(eds), The Extinct and Introduced tion and dental age assessment, in MAKelley and FaunaofBritain CSLarsen (eds), Advances in Dental Anthropology, Tacitus,Annals of Tacitus: ‘Annals’ 1.1–54 Books1–6 New York, 143–68 (Cambridge Classical Texts and Commentaries), Smith, Pand Kahila, G, 1992 Identification of Cambridge infanticide in archaeologicalsites:acase study Tatham, S, 2004 Aspects of health and popula- from the late Roman—early Byzantine periods at tion diversityin10th–12th centuryNorthern Ashkelon, Israel, JArchaeological Science 19,667–75 Europe, Unpublished Ph.D. thesis, University of Sparks,BWand West, RG,1970 Late Pleistocene Leicester deposits at Wretton, Norfolk. 1. Ipswichian inter- Taylor, A, 2001 Burial Practice in Early England, glacial deposits, Philosophical Transactions of the Stroud Royal Society of London B258, 1–30 Thomas, A, Holbrook, Nand Bateman, C, 2003 Later Spry, N, 1971 TheNorthgate Turnpike, Journal Prehistoric and Romano-British Burial andSettlement of Gloucestershire Society for Industrial Archaeology, at Hucclecote, Gloucestershire,Bristoland Gloucester- 1–58 shire Archaeol Rep 2

172 Bibliography

Thomas, M, Gilbert, P, Bandelt, H-J, Hofreiter,M Walker, PL,1997 Wife beating, boxing and broken and Barnes, I, 2005 Assessing ancientDNA noses:skeletal evidence for the cultural patterning studies, Trends in Ecology and Evolution 20,541–3 of violence, in DWFrayer and DCMartin(eds), Timby, JR,1990 SevernValley wares:areassess- Troubled Times: Osteological andArchaeological Evi- ment, Britannia 21,243–52 dence of Violence,Amsterdam, 145–79 Timby, JR,1991 The Berkeley Street kiln, Gloucester, Walker, PL,Dean, Gand Shapiro, P, 1991 Esti- JRoman Pottery Studies 4 ,19–31 mating agefrom tooth wear in archaeological Timby, JR,1999 Pottery supply to Gloucester colonia, populations, in MAKelleyand CSLarsen in HHurst (ed.), 33–44 (eds), Advances in Dental Anthropology,New York, Timby, JR,forthcoming The pottery (from London 169–78 Road). Publication report prepared for Founda- Warwick, R, 1968 The skeletal remains, in LP tions Archaeology (2004) Wenham, 111–216 Tomlin, RSOand Hassall, MWC, 2005 Roman Webster, G, 1993 The Roman invasion of Britain , Britain in 2005. II. Inscriptions, Britannia 36,473–97 London Toynbee, JMC, 1971 Deathand Burial in the Roman Wells, C, 1982a Thehuman burials. South of the World ,Baltimore and London Fosse Way, in AMcWhirr et al.,135–96 Trinkhaus, E, 1985 Cannibalism and burial at Wells, C, 1982b The human burials. North of the Krapina, JHuman Evolution 14,203–16 Fosse Way, in AMcWhirr et al., 197–202 Trotter, M, 1970 Estimation of stature from intact Wenham, LP,1968 The Romano-British Cemetery at long limb bones, in TDStewart (ed.), Personal Trentholme Drive, York,HMSO, London Identiﬁcation in Mass Disasters. [Report of aSeminar Wheeler, H, 1985 The racecourse cemetery, Derby- Held in Washington, D.C., 9–11 December 1968, by shire Archaeol J 105 ,222–80 Arrangement Between the Support Services of the Whimster, R, 1981 Burial Practices in Iron Age Department of the Army andthe Smithsonian Britain, BAR Brit Ser 90,Oxford Institution],National Museum of Natural History, White,TD, 1992 PrehistoricCannibalism At Mancos Washington D.C, 71–83 5MTUMR-2346,Oxford Trotter, Mand Gleser, GC,1952Estimation of White,TD, 2003 Human Osteology ,London stature from long bones of American Whites and White,TDand Folkens, PA,1991 Human Osteology , Negroes,American JPhysical Anthropology 10, 463– New York 514 Whiting, W, 1925The Romancemetery at Ospringe. Trotter, Mand Gleser, GC,1958 Are-evaluation of Description of the ﬁnds continued, Archaeologia estimation based on measurements of stature Cantiana 37,83–96 taken during life and long bones after death, Wigg, A, 1993 Zu Funktion und Deutungder American JPhysical Anthropology 16,79–123 Aschengruben,inMStruck (ed.) Ro¨ merzeitliche Tyrell, A, 2000 Skeletal non-metric traits and the Gra¨ ber als Quellen zurReligion, Bevo ¨ lkerungsstruktur assessment of inter-and intra-population diver- und Sozialgeschichte,Johannes Gutenberg Institut sity: past problems and futurepotential, in MCox fu¨ rVor- und Fru¨ hgeschichte, Mainz and SMays (eds), Human Osteology in Archaeology Wills, J(ed.),2000 Archaeological Review No. 24 and ForensicScience,London, 289–306 1999, Trans Bristol GloucestershireArchaeol S oc 118, Ubelaker, DH,1989 Human Skeletal Remains. Exca- 213–34 vation, Analysis, Interpretation,2nd edn, Taraxa- Wills, LJ,1938 The Pleistocene development of cum,Washington the Severn from Bridgenorth to the sea, Quarterly Vennemann, TW,Fricke, HC,Blake, RE,O’Neil, JR Journal of the Geological Society of London 94,161– and Coleman,A,2002, Oxygen isotope analysis 242 of phosphates: acomparison of techniques for Wols, HDand Baker, JE,2004 Dental health of analysis of Ag3PO4, Chemical Geology 185,321–36 elderly confederate veterans:evidence from the Villa, Pand Mahieu E, 1991 Breakage patterns of Texas Statecemetery, American JPhysical Anthro- humanlong bones, JHuman Evolution 21,27–48 pology 124,59–72 von den Driesch, A, 1976 AGuide to the Measurement Worley,F,forthcoming a. Taken to the Grave: An of Animal Bones from Archaeological Sites,Peabody Archaeozoological Approach Assessing the Role Museum Bulletin 1 ,Harvard of Animals as Crematory Offerings in First Wacher, J, 1995 The Towns of RomanBritain,2nd Millennium AD Britain edition,London Worley,F,forthcoming b. Cremated animal bone, in Waldron, HAand Cox, M, 1989 Occupational arthro- PBooth et al.forthcoming pathy:evidence from the past, British Journal of Worssam, BC,Ellison, RAand Moorlock, BSP, Industrial Medicine 46,420–22 1989 Geology of the Country Around Tewkesbury , Waldron, T, 1994 Counting the Dead. The Epidemiology Memoir of the BritishGeological Survey, Sheet of Skeletal Populations,Chichester 216 (England and Wales), London Waldron, T, 1998 Anote on the estimation of height Wright, R, 2003 Collecting Metric Data from Four from long-bone measurements, International J HumanLong Bones for an Experimental Com- Osteoarchaeology 8 ,75–7 mingling Study, Unpublished draft manuscript

173 Index

A Abscesses ...... 45, 77 Afterlife ...... 130, 133, 135, 139 Alveoli...... 43–5, 47–8, 50, 76, 78, 147, 158 Ancillary vessels ...... 19, 100, 103, 129–30, 137, 154, 157 Animal bone...... 12–3, 26, 94, 97, 119–21, 136–7 Ante-mortem toothloss ...... 40–1, 44, 49, 68, 77–8, 147 Antonine Plague...... 70, 140–1 Arm position ...... 83–6, 88, 90–5, 131 parallel ...... 81, 84–5, 87–93, 95 right...... 16–7, 23, 81,83–4, 86–95 Arms...... 15, 21–4, 60, 70, 81, 85, 89–95, 131–2, 134, 143 left ...... 15, 21–3, 26, 81, 83–4,86–95, 108, 112, 131, 134, 151

B Bead, melon ...... 105–8, 110 Bedfordshire ...... 34, 53, 55, 71–2 Bell ...... 108 Birds ...... 26, 97, 119–22, 135, 156 Body fluids ...... 70–1 Bone counters ...... 105–7 fish ...... 121–2 fragments ...... 39, 72–3, 76, 149, 156 indeterminate ...... 97, 120–1, 157 surfaces ...... 29–31, 50 Bones adult ...... 53, 79 alveolar ...... 44, 50–1, 66 calcined ...... 19, 98, 105 cortical ...... 50, 56, 59, 61 disarticulated ...... 24, 29, 38, 70 epipteric ...... 35, 40 facial ...... 74, 155–7 frontal ...... 39, 40, 61, 79 hand ...... 36, 154–8 mammal ...... 26, 89, 137 parietal ...... 50, 52, 61, 73, 148–9, 154–6, 158 pig...... 96, 154 sphenoid ...... 51–2, 154 Bracelets ...... 13, 18, 24, 26, 112, 114, 125, 133–4, 143 Brooch ...... 18, 113–4, 140, 145 Wroxeter ...... 113–4, 140 Burial pit...... 114, 121 plots ...... 126–7 practices ...... 123, 127, 130, 138 process ...... 73, 133 rites ...... 103, 108, 111, 125, 127, 130, 138–9, 144 traditions ...... 116, 141 Unurned ...... 74 Burials coffined ...... 24, 116, 133

175 Life andDeathinaRoman City

crouched ...... 12, 21, 125, 130, 135, 144–5 decapitated...... 131, 133 distribution of ...... 21, 126 prone ...... 12, 16, 18, 21,24–5, 81, 84, 86–7, 95, 101, 130–5, 138–9, 151 supine ...... 21–3, 132 urned ...... 19, 121 Burnt...... 19, 73, 78, 121, 128, 130, 154–8 animal bone...... 121, 154, 156 bone grave assemblages ...... 121 bones ...... 9, 73–4,119–21, 138

C Catherine’s Church ...... 5, 129–30 Cattle ...... 119–21 Cemetery population...... 62–3, 65, 71 urban ...... 123, 125–7, 130, 132, 135 Wotton ...... 4–7, 123–30, 134–6, 138 Cenotaph...... 25–6, 120, 134, 137 Children ...... 19, 29, 59, 68, 79, 107–8, 111, 138, 143–4 burials of ...... 20, 138 Cinerary urns ...... 5, 6, 11–2, 19, 21, 72,96–9, 105, 115,122, 129–30 Cirencester ...... 6,32, 34, 46–7, 51,53–7, 62–3, 65, 71, 121, 132, 141 Cist ...... 24, 94, 133–4 Claudio-Neronian ...... 99, 116–8, 136 Clavicle ...... 40, 54–6, 63–4, 68,149, 155–7 Cofﬁn ...... 5, 7, 13, 24–5, 85–93, 95, 115–6, 132–4, 138–9, 151 nails ...... 8,88 Coins ...... 5–7, 86, 103,130, 132, 137, 139, 145 Colchester ...... 1, 108, 117, 125 Collagen...... 28, 73 Colonia Nerva ...... 4, 144 Congenital syphilis ...... 45, 48, 71 Consumption ...... 43, 122, 128–9, 136–7 Containers, organic ...... 24, 73–4 Copperalloy...... 13, 26, 110, 112–4, 134 Counters ...... 105–10 bone ...... 105–7 Cranial ...... 30, 33, 41,52, 63, 96–7, 154, 156 vault ...... 50, 52,60–2, 154–8 fragment of ...... 148–9 Cremated bone ...... 8, 19, 28, 72–4, 76, 78–9, 115, 119, 127–9, 137 deposits of ...... 73, 128–9 samples...... 28 Cremation ...... 5–9, 19,72–3, 75, 78, 81, 96, 119, 122, 126, 128, 136–8, 144, 154–5 burial deposits ...... 119 burials ...... 5, 6, 9, 10, 12, 19–21, 73–7, 85, 96–7, 99, 102–6, 108–10, 115–6, 120–5, 127–30, 137–9, 141, 154–8 urned ...... 6–9, 12, 19, 73, 79, 96–7, 99, 100, 103, 105, 120, 129 deposit ...... 8,19, 96–8, 121, 129, 154, 157 process ...... 72–3, 79, 121 Crest, nuchal ...... 76, 148–9, 154 Cribra orbitalia ...... 46, 52–4, 62–3, 81,86–90, 94–5 Crossbones ...... 31–2, 36–7, 70, 31–2, 37, 70 application of ...... 31, 36

176 Index

D DenmarkRoad ...... 5,6,125, 127–30 Dental caries ...... 41, 43–6, 49, 50, 77, 81, 84–90, 92–6, 147 anterior ...... 49, 50 conditions ...... 50, 66, 77 formation ...... 75–7 fragment ...... 76–7, 154, 156, 158 roots ...... 75, 156–7 Dice ...... 97, 105–8, 138 Disease...... 44–5, 52, 57–8, 60–1,63, 66–8, 71–2, 88, 140 degenerative disc ...... 65, 88, 94–5

E EH (Enamelhypoplasia) ...... 41, 43, 45–6, 48–50, 68, 70–1, 81, 84, 86–90, 92–6, 147 Enamelhypoplasia, see EH Enthesophytes ...... 81, 85–92, 94–5, 154, 157 Ermin Street...... 1, 6, 124–6, 1, 6, 124–6

F Females...... 29, 30, 33–4, 40–1, 43–51, 53–4, 57–8, 60–3, 65, 67–9, 71–2, 83–4, 86–7, 95–6, 112, 141, 143–5 Femora ...... 39, 40, 86–91, 94,120, 156–7 Femur ...... 30, 33–4, 36, 38, 40, 51, 54–5, 62, 64, 75–6, 92, 148–9, 154–6, 158 left ...... 24, 28, 38–9, 57, 59, 83, 86–7, 94–6, 120, 148–9 right...... 28, 33, 38–9,83, 87, 91–2, 96, 148–9 Fibula ...... 36, 54–5, 59, 61–4, 87–8, 90–1, 148–9, 154, 156–7 left ...... 63, 148–9 Fibulae ...... 40, 51, 54, 85, 156 Finger ...... 26, 92, 108, 111–4, 134, 143, 149 Flagons ...... 12–3, 19,21, 24–5, 89, 97, 99–103, 120, 129–30, 134–5, 137 oxidised ...... 19, 96, 99, 100 ring-necked ...... 23, 87,89, 101, 103, 132, 151 Flavian ...... 101 Flavian-Trajanic ...... 12, 19, 99–101, 103, 136 Foetus...... 40 Food ...... 43, 122, 134, 136–8, 162 Foot ...... 1,52, 63, 85–8, 90, 115–7, 135, 149 left ...... 26, 87, 89, 91, 95, 115, 117 right...... 87, 89, 95, 115 Footwear ...... 18, 25–6, 86–7, 89, 115–6, 130, 135, 138–9 hobnailed ...... 25, 135 Foundations Archaeology ...... 5, 7, 104, 123–6,128–33, 141 Fovea capitis ...... 38–9, 75 Fowl, domestic ...... 26, 88, 92, 119–22, 135, 137, 139

G Gaming counters ...... 105–6 Gaming sets ...... 12, 97, 105–8, 138–9 Glass ...... 5, 104–5, 107 blue/green ...... 89, 96, 104 counters ...... 97, 105, 107–8 vessels ...... 6, 104, 127, 129 white ...... 106–7, 109–10 Gloucester ...... 1,3,4,6,7,28, 43, 45, 55, 67,99, 102–3, 108, 111–2, 117, 141–2, 144–5, 153 Gloucester Archaeology Unit ...... 7 Gloucester burials...... 108, 121 Gloucester City Council...... 1 Gloucester City Museum ...... 5, 6, 8

177 Life andDeathinaRoman City

Goose ...... 119 Gout ...... 59, 60, 72 Grave goods ...... 8, 12, 24–5, 73, 81, 86–90, 93–7, 99,108, 115–6, 120, 122, 127–8, 130, 138–9, 143 markers ...... 95, 131, 135, 138 wooden ...... 136 Grave furniture ...... 81,88–91, 93–4 Graveyard ...... 5, 6 Grey ware ...... 99, 101–3

H HFI (Hyperostosis frontalis interna) ...... 61, 87, 89, 90, 157, 61, 87, 89, 90, 157 Hip ...... 57–9, 65, 68 left ...... 18, 37, 58, 60, 114 Hippopotamus ...... 160–3 Hobnails ...... 25–6,81, 83, 86–90, 92–3, 95–7, 104, 115–6,129–30, 135 complete ...... 115 Hoop...... 111, 114 Hospital ...... 6, 7, 136 Humerus ...... 33–4, 36–7, 40, 46, 54–5, 60, 64, 75, 77, 120, 148–9, 154–8, 161 right ...... 39, 86–7, 90, 94,121, 148–9 Hyperostosis frontalis interna, see HFI Hypoplasias ...... 41, 43, 45–6

I Incisors ...... 48, 66, 147, 154, 156, 158, 160–1 Infanticide ...... 69, 142, 144 Infants ...... 7, 33, 39, 40, 49, 61, 67, 69, 144 Inhumation burials...... 5–9, 12–3, 19, 20, 22, 25,81, 83–95, 103–4, 108, 110–3, 115–6, 120–2, 124–8, 130–2, 141–3, 145 Inscription ...... 18, 116–8 Intaglios ...... 26, 87, 111, 134 Iron ...... 26, 86, 110–1, 114–5 nails ...... 85–97 Ivory ...... 105, 110

J Jars ...... 26, 102–3, 128–9, 131, 135, 140 black burnished ware ...... 25, 102–3 necked ...... 96–7, 99, 100, 103 Jewellery ...... 25–6, 134, 138, 143 Joint disease ...... 50, 57, 59, 60, 65, 72, 76–7, 157

K Kempston ...... 34,51, 53, 55,62–3, 71–2 Kingsholm ...... 1,3,6,45, 47, 51,55, 58, 62–3, 72, 99, 101–104, 117, 123–4, 126, 130, 144–5 Knee ...... 23, 57–8, 60, 65, 93

L Lankhills ...... 112–3, 125–6, 132, 134–5, 137, 141 Legion, Twentieth ...... 1,6,19, 118 Legions ...... 3–6, 108, 117–8 Legs ...... 12, 15, 18, 21–2, 25–6, 81, 83–93, 95, 110, 118, 134, 151 left ...... 36, 87–8, 92 right ...... 87–8, 95, 151 Lesions ...... 41, 43–5, 47, 49, 50,52, 56, 59–61, 64, 67, 70, 147, 157 pathological ...... 46, 48, 71–2, 121, 154, 156–8

178 Index

Limb bones ...... 51, 70 lower...... 30, 32,37, 54, 59,63, 68, 70,72–3, 154–8 upper ...... 37–8, 54, 63, 75, 154–8 London Road ...... 1, 5–9, 28–9, 32, 34, 51, 53–7, 69, 71–2, 77–9, 102–4, 121–36, 138–9, 141, 143–5, 162–3 Long bones ...... 29, 30, 36, 40, 65, 74–9, 84, 96–7, 121, 149, 154–8 fragments of ...... 75, 78, 96–7, 148–9, 154, 156–8, 160 sized...... 97, 157

M Male adolescent ...... 21, 56, 60, 65, 132 adult ...... 13, 24, 51, 53, 56–7,59, 65–6, 114, 145, 148 Mandibles ...... 35, 37, 45, 50, 75, 147 Mastoid foramen...... 34–5, 42 Maxilla...... 45, 74, 154, 156, 158 Medieval ...... 7,8,26, 99, 118, 120 assemblage ...... 119–20 Metacarpals ...... 36, 51, 55–6, 61, 65, 75, 88, 90, 148–9, 154–6, 158, 161 Metatarsals ...... 51, 59, 61–3, 75–6,78, 86, 149, 155–6, 158 Molars...... 45, 48, 50–1, 75, 147–8, 155–8, 161 lower...... 148, 155–6 third ...... 45, 48, 75, 155 Mourners ...... 78–9, 127, 131, 134, 136

N Nails ...... 13, 24, 84, 86–90,95, 97, 115–6, 133, 135 Fe ...... 151 Native population...... 123, 125, 144–5 Necrosis ...... 59, 60 Neronian ...... 9, 12, 99, 100 Neurocranium ...... 154–8

O OA (Osteoarthritis) see osteoarthritis Occipital bone...... 39, 40, 50, 52, 56, 73, 77, 86, 148–9, 154–6, 158 Offerings ...... 128–9, 139 Orbits ...... 52–4, 73–4 Osteoarthritis ...... 8, 54, 57–60, 63, 65, 68–70, 77, 84, 86–92, 94–5, 157 rib ...... 86, 88, 95 Osteomyelitis ...... 45, 50–1, 62, 67, 70, 88 Osteophytosis ...... 57, 59,77, 84, 87–8, 91–2, 94 Oxidisedware ...... 101 Oxygen isotopes ...... 150, 152–3

P Painswick Hill ...... 116, 159 Painswick Stone ...... 159 Patella ...... 75, 148–9, 154 Pelvis ...... 15, 18, 21–3, 30, 32, 40,60, 74–5, 84–9, 92–3, 95–7, 120, 131, 154–6, 158 Periodontal disease ...... 44–5, 47, 147 Periodontitis ...... 42,45, 47–50, 68, 84, 86, 88–9, 96, 148 Musculo-skeletal markers ...... 81, 86–8, 90, 95 Periostitis ...... 50–2, 59, 61–2, 65, 67–8,71, 76–7, 86, 88, 157–8 active ...... 83, 87, 89, 91–3, 158 healed ...... 83–92, 94–5 Phalanges ...... 36, 65, 75, 88, 91, 149, 154, 156 Phalanx, hand ...... 88, 149, 154 Pigs ...... 97, 119–22, 128, 156, 158

179 Life andDeathinaRoman City

Pins ...... 109–10, 114 split ...... 110 Plague...... 69, 140 Population ...... 29, 34, 41, 43, 45, 50, 53–4, 60, 67–9, 71–2, 123, 125, 127, 140–1, 143–5, 152–3 archaeological ...... 43, 50, 57–8, 67 modern ...... 45, 57, 60 Porotichyperostosis ...... 52–3, 63, 68, 77 Post-medieval ditch ...... 14, 148–9 Pottery ...... 6, 12, 15, 26, 28, 74, 93, 99, 101–4, 123, 130, 135–8, 140, 144, 151 Roman ...... 27, 99, 100, 102, 153 sherds of ...... 93, 130, 137 Pre-Flavian ...... 9, 74, 97 Pyre ...... 19, 73–4, 78, 104, 108, 115–6, 127–30, 137, 139 debris, redeposited ...... 74 goods ...... 73–4, 81, 96–7, 104–5, 107–8, 114, 121, 128 sites ...... 74, 128

R Radius...... 34, 36, 40, 54–5, 62, 64, 75, 120–1, 148–9, 154–8 fragments...... 148, 158 right ...... 39, 87, 97, 121, 148–9, 158, 161 Ribs ...... 35, 51, 55, 62, 64, 74–5, 84, 86, 96–8, 148–9, 154–8 River Severn...... 1, 3, 6, 124, 160, 162, 163 Roman period ...... 1, 3, 6, 7, 60, 69, 72, 99, 116, 127, 130–2, 134, 137, 139, 143 pottery ...... 27, 99, 100, 102, 153 RomanBritain ...... 48, 121–2, 125–6, 129, 142–3, 150 Romancity ...... 1 RomanGloucester ...... 1, 123, 141, 144 cemeteries of ...... 6, 141 RomanInscriptions of Britain ...... 117–8 RomanItaly ...... 136 RomanLondon...... 121, 132 Romano-British cemeteries ...... 6,7,26, 101, 104, 108, 115–6, 118–9, 122–3, 126, 133, 141 populations ...... 29, 32, 43, 45–7, 51, 55, 61, 65, 71, 143 Rome ...... 4, 122,125, 139–41

S Sacrum ...... 57, 60, 95, 149 Schmorl’s nodes ...... 58, 65, 86, 89, 92, 95 Scurvy ...... 50–2, 72 Severn, River...... 1, 3, 124, 160, 162, 163 Sheep...... 18,119–20, 154 Sheep/goat ...... 119, 121 Shoes ...... 26, 116, 135 Shouldergirdle...... 74–5,96–7, 154–7 Shrouds...... 83, 133–4, 138 Skeletons adult ...... 30, 36, 40, 67–8, 70, 148 articulated ...... 17, 31, 34–8, 40, 62–3, 65, 69, 70 complete ...... 35, 39, 40 female ...... 143 male ...... 33, 71 subadult ...... 35, 42 unsexed ...... 68 Skull ...... 15, 18, 21, 32, 39, 40, 50, 56, 61, 69–71, 74, 77, 81, 96–8, 148–9, 154, 157–8 bones ...... 39, 51, 148, 154, 158 fragments...... 74, 77, 121, 148–9, 155–6

180 Index

vault ...... 50,61, 63, 73,78–9, 157–8 fragments ...... 78, 157–8 Smallpox...... 70, 140–1 Soldiers...... 4, 6, 19, 117–8, 136, 144–5 Spines ...... 56–8, 60, 65 lumbar ...... 58, 65 thoracic ...... 58, 65 Spondylolisis ...... 56, 58–9 St Catherine’s Church ...... 6, 125–6 St Margaret’s Hospital ...... 6,125–7,129–30, 134 St Mary Magdalen ...... 5–7, 136 St Mary Magdalene’s Chapel ...... 7 Stele ...... 116–7 Stomach...... 21, 23, 90–3, 95 Stone cists ...... 5,6,24–5, 129, 134, 138 Subadults ...... 30, 32–3, 35–6, 38, 40, 42, 46–7, 50,52–4, 61–2, 66–8, 71–2, 75, 77, 96, 143 Supine ...... 23, 81,83–96, 151 Syphilis, congenital ...... 45, 48, 71

T Teeth anterior ...... 43–4, 47, 77, 147 deciduous ...... 42, 45, 49 molars ...... 45, 48, 50–1, 75, 147–8, 155–8, 161 lower ...... 148, 155–6 third ...... 45, 48, 75, 155 permanent ...... 42, 45–6, 49, 50 posterior ...... 43–4, 46–8, 147 premolars ...... 66, 147–8, 154, 156, 161 Tibia ...... 30, 35–6, 42, 51, 54–5, 61–4, 68, 71, 75–8, 90–1, 120, 148–9, 154–5, 160–1 fragments ...... 148, 157–8 left ...... 39, 59, 65, 67, 88, 90–1, 97, 148–9, 151, 156 right...... 39, 54, 59,81, 84, 87–8, 91–2, 149 Toga ...... 116–7, 131, 143 Tombstones ...... 4, 6, 18–9, 116–8, 122, 127, 136–8, 145, 159 Trajan...... 5, 102, 116 Trauma...... 43, 45,49–51, 53, 55, 57–64, 71, 77, 88, 91 healed ...... 61, 64, 85–7, 94 Tuberculosis ...... 50–1, 61, 72

U Ulna ...... 36, 54–5, 62, 64, 75, 78, 120–1, 148–9, 154–6, 158 left ...... 38–9, 56, 88, 149 right...... 39, 92, 94, 148–9 Underworld ...... 128, 135, 137, 139 Unguent bottles ...... 104, 112, 137 tubular ...... 89, 96–7, 104, 110–1 Urn ...... 8, 19, 73–4, 79, 99, 100, 105–7, 115, 129

V Vertebra clupeid ...... 121 lumbar ...... 56–7, 149 Vertebrae ...... 35, 46, 48, 57–8, 65–6, 74–5, 84, 86–9, 92, 94–7, 121, 154–8, 161 Vertebrate assemblage ...... 160, 162 Verulamium ...... 6, 103, 107–8 Vessel glass...... 104, 110–1 Vessels...... 12, 19, 25, 79, 92, 96–7, 99–103, 117, 120, 128–30, 133–5, 137, 151

181 Life andDeathinaRoman City

W Winchester ...... 106–9, 112, 125–6, 132, 134–5, 141 Wotton...... 1, 4–7, 123–31, 134–6, 138, 141, 145, 160, 162–3 cemetery ...... 4–7, 123–30, 134–6, 138 Wroxeter ...... 3, 4, 113–4, 140 brooch ...... 113–4, 140

182