Romano-British Urban Defences 9780860542643, 9781407317373


176 52 83MB

English Pages [214] Year 1984

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Front Cover
Copyright
Table of Contents
Acknowledgements
Key to Figures
Abbreviations
Chapter I: Introduction
Chapter II: Previous Approaches and the Development of the Framework
Chapter III: The Nature and Problems of the Evidence
Chapter IV: The Reassessment of the Evidence in the Light of New Data and Criteria
Chapter V: Conclusion
Chapter VI: Desiderata for the Future Study of Civil Settlements
Gazetteer
Bibliography
Table I
Table II
Table III
Table IV
Table V
Recommend Papers

Romano-British Urban Defences
 9780860542643, 9781407317373

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

Romano� British Urban Defences

Julie Crickmore

BAR British Series 126 1984

--B.A.R.

5, Centremead, Osney Mead, Oxford OX2 0ES, England.

GENERAL EDITORS A.R Hands, B.Sc., M.A., D.Phil. D.R Walker, M.A.

BAR 126, 1984

©

'Rrnnano-Brltish Urban Defences'

Julie Crick111ore, 198/l

The author’s moral rights under the 1988 UK Copyright, Designs and Patents Act are hereby expressly asserted. All rights reserved. No part of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any form of digital format or transmitted in any form digitally, without the written permission of the Publisher. ISBN 9780860542643 paperback ISBN 9781407317373 e-book DOI https://doi.org/10.30861/9780860542643 A catalogue record for this book is available from the British Library This book is available at www.barpublishing.com

List of Contents I. Introduction II.Previous Approaches and the Qevelopment of the Framework Early Defensive Circuits Late 2nd �entury Earthwork Defences Stone Defences Burgi, Bastions and Other Developments in the Late Roman Period

Page l 4

III. The Nature and Problems of the Evidence The'Type' Sites- -The Reinterpretation of the Evidence from Early Excavations Methods of Excavation and the Quality of the Evidence The Dating Evidence : its Nature and Problems The Incompatability of the Archaeological and Historical Evidence

21

IV.

45

The

Reassessment of the Evidence in the Light of New Data andCriliria --1.Isit possible to regard the provision of timber and earthwork circuits in Britain as anything other than a re­ stricted phenomenon adopted only as an emergency measure? 2. Is there any evidence for the more widespread existence of timber and earthwork circuits in the 1st and early 2nd centuries? 3. Did some centres begin to construct defences in stone prior to the emergency situation of the late 2nd century? 4. Is it possible to regard the cons­ truction of successive earth and timber circuits at certain centres in the 1st and 2nd centuries as attempts to defend these centres on a continuous basis? 5. What are the other reasons which might account for the construction of these early defensive circuits? 6. Was the provision of stone walls a sep­ arate scheme or maintenance and up­ dating? 7. Is there any evidence to suggest that the construction of a walled circuit may have been a prolonged process? 8. Is the evidence for the provision of bastions consistent with the inter­ pretation of a single scheme initi­ ated by Theodosius and with a new role for civil defences in the late

Roman period? 9. What were the effects of burgi and wider military considerations in the provision of walled circuits for many civil settlements? 10. What were the other developments i late Roman civil defences? V. Conclusions

96

VI. Desiderata for the Future Study of Civil Def nee

98

Gazetteer 'Lesser' Towns, Villages and Other Civil Settlements Civitas Capitals and Coloniae Bibliography

103

173

181 I. 'Lesser' Towns, Villages and Defended V1 ; II. The Civitas Capitals 182 III. The Coloniae 183 IV. Classification of the Evidence for 'lesser 184 towns, villages and defended Vici Table V. Classification of the Evidence from the Civitas 198 Capitals and the Coloniae

Table Table Table Table

Acknowledgements This s tudy w as p repared i n 1 979 u nder t he a uspices o f the U rban R esearch Committee of t he U niversity of B irmingham and t he Department o f t he E nvironment. I t a rose o ut o f work o n a s urvey of R oman u rban s ettlement i n t he West M idlands. The t ext w as u pdated a nd s ubstantially r evised f or p ublication by the end of 1 982 a nd a number of l ater r eferences have b een i ncluded. A s ection dealing w ith t he i nteraction b etween settlement and defences, i ncluded a s part of t he o riginal r esearch has b een omitted f rom t he publication t ext a nd t he reader's attention i s d rawn t o A .S. Esmonde C leary's f orthcoming b ook Towns and U rbanisation i n Roman B ritain : The Evidence f rom Outside t he Defences ( British A rchaeological Reports, f orthcoming). I a cknowledge g ratefully t he a ssistance a nd s upport o f members of t he U rban R esearch Committee and, i n particular, extend t hanks t o J .P. R oberts f or e ncouragement a nd d iscussion during t he p reparation of t he t ext. Comments and c orrections by P .C. Bucklar id, C .C. Dyer, J .S. Johnson a nd G . Webster a re also a cknowledged. I owe special t hanks t o A .S. E smonde C leary for s uggesting a lterations t o t he l ayout a nd f or many corrections and f or d iscussion during t he r evision o f t he t ext f or p ublication. T he f inal manuscript was t yped b y J .M. Buckland.

J .

C rickmore

K ey t o F igures

P age

1 .

E arly D efensive C ircuits

1 0

2 .

D efensive C ircuits e rected f rom t he l ater 2 nd C entury o nwards

1 3-14

3 .

B astions

7 5

4 .

4 th C entury d efensive a ctivity ( excluding w ork o n g ates)

A ic A ich A id A lf A nc B F B H B an B ath B ra C F C N C W C Y C ae r C am C ant C arl C arm C at t C he C hi Ci r C l C ol C or D T D on D or D orn D w E x F S G a G Ca G Ch G b G od H a H or Ii ch I lk ' p Ir c

A lcester A ichester A ldborough A lfoldean A ncaster B rough-on-Fosse B rough-on-Humber B anna v ent a B ath B rampton C hesterton-on-Fosse C aistor-by-Norwich C aistor-on-the-Wolds C aistor-by-Yarmouth C aerwent C ambridge C anterbury C arlisle C armarthen C atterick C helmsford C hichester C irencester C lause nt ur n C olchester C orbridge D orchester-on-Thames D oncaster D orchester D orn D roitwich E xeter F enny S tratford G atcombe G reat C asterton G reat C hesterford G loucester G odmanchester H ardhar n H orncastle I lchester I lkley I ping I rchester

K T K en K y L M L ei L et L in L itt L on L tw M C M an M d M du M ild M rg M t N ea P en R oc R och S C S il S i T d T N T ow T ri U C U xa V er W N W an W hit W in W ix W k W or W x W y Y ork

1 02

K irkby T hore K enchester K elvedon L ong M elford L eicester L etocetum L incoln L ittlechester L ondon L eintwardine M elandra C astle M anchester M aldon M anduessedur n M ildenhall M argidunum M alton N eatham P en noc ru c iu r n R ocester R ochester S choebury C amp S ilchester S lack T iddington T horpe-by-Newark T owc e ster T ripontium U phall C amp U xacona V erulamium W ater N ewton W an b orough W hitchurch W inchester W ixoe W ickford W orcester W roxeter W ycomb Y ork

Abbreviations Ant. Ant. J . A rch. Arch. Cant. A rch. J . A rch. NL BAR BBCS

A ntiquity Antiquaries Journal Archaeologia Archaeologia Cantiana Archaeological Journal Archaeological Newsletter B ritish A rchaeological R eports Bulletin o f the Board o f Celtic Studies T ransactions o f t he Cumberland a nd Westmorland Antiquarian a nd A rchaeological Society Derbyshire Archaeological Journal Journal o f t he B ritish Archaeological Association Journal of Roman Studies

cw

Derbys. JBAA

Arch

J .

JRS K eele A rch. Grp.NL LSSAHS

Med. Arch. Norfolk A rch. NSJFS P roc.Camb. Antiq.Soc. P roc.of the D orset Nat. Hist. a nd Arch.Soc. P roc.Hants.F.C.

Keele A rchaeological G roup Newsletter L ichfield and South Staffordshire A rchaeological a nd H istorical Society Medieval Archaeology Norfolk A rchaeology North Staffordshire Journal of F ield Studies P roceedings of the Cambridge Antiquarian S ociety P roceedings of t he D orset Natural H istory a nd A rchaeological S ociety P roceedings o f t he H ampshire F ield C lub and Archaeological Society

RCHM

Royal

Sa

Monuments Shropshire Newsletter

NL

Sussex TBAS

A rch.Col1.

TSSAHS

TWAS TWNFC T rans.Thoroton

Commission

H istorical

Sussex A rchaeological Collections T ransactions o f t he Birmingham and Warwickshire A rchaeological Society T ransactions o f t he S outh Staffordshire A rchaeological a nd H istorical Society T ransactions o f t he Worcestershire Archaeological Society T ransactions o f t he Woolhope Naturalists' F ield C lub

Soc.

T ransactions o f t he Thoroton Society f or Nottinghamshire

VCH

Victoria

WAM

Wiltshire Natural Yorkshire

YAJ

on

County

H istory

A rchaeological a nd H istory Magazine A rchaeological J ournal

R OMANO-BRITISH A R EVIEW OF

I .

C IVIL D EFENCES THE EVIDENCE

I ntroduction

S ince t he 1 950s t he s tudy o f R omano-British d efences h as been i nfluenced by t he desire t o e stablish a common pattern o f development t hroughout t he p rovince. The n eed t o p lace t he widespread p rovision of defences w ithin t he general c ontext of imperial p olicy h as b een t he f ormative e lement i n e stablishing this pattern and i n t he i nterpretation o f t he e vidence ( Frere 1 965, 1 38 9 ; 1 978, 2 84 5 ). C urrently t he s tudy i s b ased upon the p remise t hat t he p rovision o f d efences f or a l arge number o f s ettlements i n B ritain i n c ontrast t o o ther provinces i s t o b e e xplained i n t erms of e xtraordinary c ircumstances a ffecting t he s ecurity o f t he p rovince. T hese c ircumstances a re h eld t o b e:1 ) t he c laim o f A lbinus t o t he Empire i n AD 1 93-196 which l ed to t he withdrawal o f t roops, 2 ) i ncreasing i nsecurity f rom t he l ate 3 rd c entury o nwards caused by barbarian p ressure on t he f rontiers o f t he empire. I n t he f irst i nstance, t he i nsecurity o f t he p rovince w as dealt with by t he w idespread c onstruction o f t imber a nd e arthwork defences; i n t he s econd i nstance, a g eneral p rogramme o f walling was i mplemented ( Frere 1 978, 2 83-9; Wacher 1 978, 9 61 00). This f ramework e xpects t hat t he majority o f R omanoB ritish c ivil d efences w ill f ollow t his c ommon p attern o f -

-

development. H owever, t he g eneral a cceptance o f t his f ramework, a nd o f the f undamental p remise on which i t i s based, h as l ed t o a r igid and l imited a pproach t o t he e vidence. I t h as b ecome t he central e lement t o s uch an e xtent t hat i ts basis and validity a re v irtually u nquestioned. This h as e ffectively l imited t he s cope of current i nquiry t o t he a ims which c an be e ncompassed by a general t heory a bout t he d efence o f t he p rovince a s a whole. Statements about t he dating o f defensive c ircuits have been f ounded n ot s o much on s ecure a rchaeological d ata a s o n i nferences d rawn f rom what i s e ssentially ambiguous o r s canty evidence. The p roblems o f d ating i nherent i n t his t ype o f a rchaeological f eature make i t v ery d ifficult t o b uild up a chronological f ramework f or t he p rovince a s a w hole. There a re many ' anomalies' and t oo many i nstances where t he e vidence cited i s, on more d etailed e xamination, t oo i nconclusive t o a llow a s atisfactory understanding o f t he h istory of t he c ircuit. On s ome s ites, t he e vidence w as a ctually s tretched o r obscured i n c ertain r espects, particularly i n t he a ssertion o f a b road ' late 2 nd c entury ' d ate f or e arthworks, i n o rder t o f it particular s ites i nto the a ccepted c hronology. ( c.f. Hartley 1 983). I n t his way, t he c onsideration o f o ther possibilities i n t erms o f date a nd c ontext h as b een effectively d iscouraged. The a ctual e vidence o n w hich t he f ramework i s b ased i s o ften meagre a nd i nsubstantial. O ften t he development— of a p articular c ircuit i s b ased o n e vidence f rom one o r two t renches; y et t he i nterpretation o ffered w ill

1

o ften l ead t he r eader t o s uppose t hat t he d efences h ave b een t horoughly i nvestigated a nd t hat t he e vidence i s c onclusive. Anyone u ndertaking a s tudy o f R omano-British defences, h owever, w ill b e s urprised a t h ow l ittle i nvestigation h as t aken p lace e ven o n t he ' type' s ites w hich f orm t he basis of t he current f ramework. I n v iew o f t he ambiguous a nd l imited n ature of t he e vidence which defences p roduce, particularly when examined only by t renching, t he i nherent s hortcomings o f t he e vidence s hould b e b orne i n m ind when c onsidering generalisations about t heir d evelopment. F urthermore, i n l ooking a t t he e vidence f rom i ndividual s ites, i t i s d ifficult t o s ee how t he development o f i ndividual c ircuits c an b e e xplained s imply i n t erms of a n a ct of government policy. E ven i f one a ccepts t hat, i n t he majority o f c ases, t he p rovision o f a d efensive c ircuit and c ertain s ubsequent major s tructural phases w ere i mplemented a s p art o f government s chemes f or t he p rotection of t he p rovince, t here a re other f actors which deserve c onsideration. Evidence s hows t hat i ndividual defensive c ircuits have a complex and, i n most c ases, a l ong s equence o f development w hich may h ave h ad l ittle t o d o w ith g overnment i nvolvement. There i s c onsiderable variation f rom o ne s ite t o another a nd l ocal f actors must h ave p layed a s ignificant p art i n t heir c haracter and development. O nce constructed, a defensive c ircuit would t ake o n a h istory a nd d evelopment which was c losely r elated t o t hat o f t he s ettlement w ith which i t w as a ssociated; s uch a spects h ave r eceived l ittle attention. T he s tudy o f R omano-British d efences i s i n i ts i nfancy; y et t he e xistence o f a h istorical f ramework h as c reated t he impression t hat i t i s w ell a dvanced. As a r esult, t he s tudy h as b ecome f ossilised, a lthough t here i s a c onsiderable b ody o f i nformation f rom r ecent e xcavations w hich h as y et t o b e a ssimilated i nto t he g eneral f ramework ( c.f. Hartley 1983). The ' Roman B ritain i n .. . . ' s ection o f B ritannia i ndicates t he volume of t his n ew e vidence f or t he s tudy o f R omanoB ritish d efences a nd i t i s c lear t hat a r econsideration o f t he current t heories on defences i n t he l ight of t his n ew i nformation i s o verdue. M uch o f t his r ecent w ork i s, h owever, u npublished a nd a vailable o nly i n s ummaries which a re i na dequate f or a p roper a ssessment s ince t hey p resent i nterpretation but l ittle o f t he a ctual e vidence. T he detailed e xamination of e vidence f rom e xcavations on t own a nd v illage d efences i s a l ong a nd c omplicated p rocess, but i t n eeds t o b e undertaken t o a ssess t he validity of t he i nterpretations c urrently o ffered. H owever detail t he

i t i s n ot t he a im o f t his p aper e vidence f rom a ll s ites a nd t he

t o e xamine i n p resent writer

a ttempts only t o e ngender a n ew s pirit o f i nquiry i nto t he i nvestigation o f R omano-British defences, f irstly by o utlining p revious approaches t o t he s ubject a nd, s econdly, b y suggesting n ew l ines o f i nquiry i nto t he n ature, p urpose a nd development o f d efences. T here a re a lso t hree t ables, t wo b reaking c ategories

down o f

t he

i nformation

d evelopment

a s

2

f rom

i ndividual

o utlined

i n

s ites t he

i nto

current

f ramework. F igures 1 , 2 a nd 3 s how t hose s ettlements p rovided with t imber and earthwork defences, t hose which were walled, and t hose where b astions a re k nown. The i nformation f or t hese f igures comes basically f rom t he n otes on r ecent work i n B ritannia a nd R odwell & R owley ( 1975). Unfortunately, i t h as not b een possible t o e xplore a ll l ocal s ources. F igures 1 a nd 2 t herefore r epresent o nly t he more g enerally a vailable i nformation and s hould be u sed only as a b asis f or f urther work. I t must be c onsidered t hat w e s hould n ot b e a ttempting t o i dentify a uniform pattern of development, b ut s hould r ather s ee d efensive c ircuits a s i ndividual ' organisms' I which, i rrespective of a c ommon o rigin, a re s o d ifferent i n c haracter as t o r ender u nconvincing a ny a ttempts t o e stablish c ategories on t he basis of s uch f eatures a s s hape, s tructural h istory, o r t he n ature o f a rea e nclosed. F urthermore, w e s hould a pproach t he q uestion of t he e xcavation o f defences n ot f rom t he narrow viewpoint o f e stablishing t he date, main s tructural p hases a nd a lignment of the c ircuit, but by r egarding t he d efences a s an e lement o f t he s ettlement w hich h as a c omplexity o f d evelopment a ssociated w ith t he h istory of t he s ettlement. By a dopting t his a pproach, t he s cope o f t he i nquiry w ill b e considerably b roadened a nd one c ould l et t he dating e vidence emerge a s p art o f a n o verall p icture r ather t han h aving i t desperately s ought a s t he p rime objective.

3

I I.

P revious

approaches

a nd

t he

d evelopment

o f

t he

f ramework

U ntil t he l ate 1 930s, l ittle work h ad been done on defences c ompared w ith t he a ttention w hich public b uildings had a ttracted. E xamination of defences was l argely c onfined t o c ivitas c apitals a nd c oloniae ( Collingwood and M yres 1 937, 9 15-206; C ollingwood and R ichmond 1 969, 1 00-8) . S ome twenty years l ater e vidence f rom e xcavation a nd a erial p hotography had a dded c onsiderably t o our k nowledge of Romano-British d efences. A lthough t here w as s till a b ias t owards c ivitas capitals a nd coloniae, t he defences of other settlements h ad r eceived more a ttention. I t was n ow p ossible t o a ttempt a n overall a ssessment and synthesis o f t his evidence. C order's ( 1955) p aper was t he f irst r eal a ttempt t o c onsider i n d etail t he e vidence f rom e xcavations c onducted i n t he f irst h alf o f t his c entury. H is c hief c oncern w as t o e stablish t he d ate a t which walled c ircuits were f irst e rected and to d iscuss t he date o f t he r eorganisation r epresented by t he a ddition o f b astions. H e was careful t o s tress t he hypothetical n ature o f h is c onclusions, o utlining i n s ome d etail t he s hortcomings o f t he e vidence f or dating of defences. H e u rged c aution i n a llowing i nterpretation t o b e i nfluenced by p revailing f ashion and p reconceived n otions. I n r eviewing t he evidence t hen a vailable f or t he dating o f walled c ircuits, h e c oncluded ' it s eems p robable t hat t he majority o f t own walls belong t o t he years AD 1 94-197' during t he p eriod w hen A lbinus w as p reparing f or h is c laim on t he E mpire. B ut he emphasised t hat t he p ottery e vidence ' though p ointing i n t his d irection, i s hardly yet p recise enough t o c onstitute p roof' and h e s tressed t hat h is s uggestion was s imply o ffered a s a working hypothesis. H e was careful t o e mphasise a lso that i t could n ot b e a ssumed t hat a ll t own walls would p rove t o b e contemporary s ince t owns o f d ifferent types might h ave h ad d ifferent h istories, a nd t he e xpense a nd manpower r equired would b e an i mportant f actor i n t he t iming of t heir p rovision. H e a rgued p articularly a gainst t he t hen c urrent t rend o f attributing s tone walls t o t he w ork o f S everus i n r estoring t he p rovince i n t he e arly 3 rd c entury. F rom t his, Corder s eemed t o b e i mplying t hat, while many t own walls were p rovided d irectly i n r esponse t o a t hreat i n t he l ate 2 nd century, o thers c ould b e s een a s a p roduct of civic p ride o r o f s pecial c ircumstances p ertaining t o t hat p articular s ettlement ( ibid., 2 2-4). I n

d ealing

w ith

t he

e vidence

f or

t he

s ubsequent

r eorganisation of t hese defences, C order concluded t hat, unl ike t he w alls t hemselves, b astions were more o r l ess c ontemporary, f alling w ithin t he p eriod AD 3 37-367, p robably c . H e

AD 3 50 a nd c ertainly b efore d ecided t hat, while t his

u rgently

c arried

r esponse

t o

o ut

i n

a specific

a n

t he b arbarian r aids t ask c an hardly

emergency,

t hreat.

H e

i t

was

s uggested

o f AD h ave

p erhaps t hat,

a s

3 67. been

a gain

a

a r esult

o f d isturbances on t he f rontier, a n i mperial d ecree w as i ssued o rdering t he t owns of t he p rovince t o o verhaul their e xisting d efences. The i ntroduction o f b astions r epresented a c hange of f rom

f unction

f or

t he

r aiders

a nd

r egulating

defences

f rom

t raffic

4

t hat of t o

protecting

a more

a ctive

t he

t own

d efensive

role associa ted with the introduction of artillery. In the construction of both walls and bastions, Corder was impressed by the immensity of the task involved. But whereas in the case of walls this influenced him in deciding that it was unlikely that all walled circuits would have been erected at the same time, the same point led to the opposite conclusion in the case of the bastions. These he regarded as the result of a deliberate act of imperial policy; such an undertaking could only have been achieved by a tribal levy and would have been implemented under the direction of military engineers. In many of the sites then excavated the wall and bank behind it were treated as contemporary and it was invariably the evidence from the bank which gave this late 2nd century date for the wall. Even though the existence of independent earthworks was already known at some sites, the implications of this, especially for the dating of walled circuits, were not fully realised. Wacher's (1964; 1966) two articles on earthwork defences were a great step forward and formed the basis of subsequent thinking. His first article was prompted by an increase. in the number oL�arth ramparts shown to pre-date the building of walls, and he reviewed the evidence from these sites, suggesting other cases where it was possible that a pre-wall rampart might be found. He also attempted to find historical contexts for their construction. In considering the purpose of these defences, Wacher noted the western and southern bias i n distribution, as far as was known at the time, and suggested that this bias was explicable only in terms of defensive measures against an impending threat to these areas. Civic pride was thought to be an unlikely consideration where earthworks were involved and it was felt that this type of defence must have been used either in an emergency or for reasons of economy. He decided in favour of emergency as the motive and developed this theme in his second article.· But, in rejecting the idea of economy, Wacher presented no real argument. His preference for emergency seems to result from his assertion of the illogicality of civic pride as a motive t o be associated with earthwork defences and from the necessity of obtaining permission from the central government to erect fortifications. The importance of the latter seemed to militate against the idea that permission would be readi-ly granted for reasons other than those connected with the security of the province. The decision in favour of emergency carried with it the necessity that these earthworks should be contemporary. 'It may be suggested that, although the dates provided for the ramparts of individual towns are often widely divergent, they are not so wide, remembering the na-ture of the dating -evidence, that they may not all have been the product of one particular occasion'. In 1964 he suggested that a date soon after the middle of the 2nd century would perhaps not be stretching the evidence too far. There were two circuits, Chichester and Dorchester-on-Thames, which had been assigned to the late 2nd century and, following the logic of Wacher's argument, this should have suggested a late 2nd century date 5

f or t he g roup a s a w hole. H owever, t he a cceptance o f Corder's hypothesis t hat s tone walls were p rovided at t his t ime made i t n ecessary t o a dopt a n e arlier c ontext f or e arthworks. Wacher d id n ot e xplain h ow t hese t wo s ites f itted i nto h is context s oon after t he mid 2 nd c entury. By 1 966, h owever, t he dating e vidence f or s ome walled c ircuits h ad b een s hown t o be t oo l ate t o s upport Corder's s uggestion t hat t hey belonged to t he l ate 2 nd c entury a nd t he way was n ow open f or e arthwork defences as a whole t o b e assigned t o t he l ater 2 nd century date s uggested by D orchester-on-Thames, w hose r ampart h ad b een g iven a t erminus post quem of c . AD 1 85. An obvious c ontext f or t he p rovision o f t imber a nd e arth defences i n t he w est and s outh i n t he mid 2 nd c entury w as d ifficult t o f ind a nd Wacher e ventually s uggested a p eriod o f unrest a long t he Welsh border, possibly c . AD 1 60. W ith r egard t o t he f ew t owns i n E ast Anglia a nd t he n orth-east which were t hen k nown t o have earthwork c ircuits, t he m ost l ikely e xplanation w as i nsecurity a long t he e ast c oast. L ike C order, Wacher s tressed t hat, b ecause t he evidence was f ragmentary a nd many c ircuits s till r emained t o b e i nvestigated, h is c onclusions were v ery t entative. He a lso emphasised t he n eed f or more s ections a cross t he d efences i n o rder t o obtain more a dequate dating ‚ and y et he c oncluded t hat t he c oncept o f a s ingle e vent i n t he mid 2nd c entury w as perhaps n ot s tretching t he e vidence t oo f ar. B y 1 966, more e arthwork c ircuits h ad b een i dentified a t R ocester and I rchester and a possible r ampart at T horpe-byNewark, a nd t hese e arthworks made a s ignificant difference i n r edressing t he b alance f rom a b ias t owards t he south and west. R elatively f ew e arthwork s ites were k nown i n East A nglia a nd s outh-east E ngland but t he f illing i n o f t he c entral a rea w as i mportant f or t he d evelopment o f t he t heory t hat t he m ost l ikely date and c ontext f or earthwork c ircuits were to b e f ound i n t he A lbinian e pisode. I t w as, h owever, t he f act t hat t here were a s ignificant n umber o f walled centres which a ppeared t o l ack a n e arlier e arthwork p hase w hich s uggested t o Wacher r aiding

t hat h e needed c ontexts i nvolving l ocalised t hreats o f o r u nrest a long t he f rontiers, i n Wales, t he N orth a nd

a long t he E ast Coast, r ather t han a m ore general c ontext f or t he p rovince a s a w hole. I t was o nly w hen t he n umber a nd d istribution of k nown earthwork c ircuits i ncreased t hat t he s earch f or a c ontext a ffecting t he p rovince a s a whole l ed t o t he choice o f t he c laim o f A lbinus i n AD 1 93-196. F urthermore, once t his more o verall d istribution h ad b een e stablished, t he apparent l ack of earthwork c ircuits at s ome c entres, n otably t he c ivitas c apitals o f Canterbury a nd L eicester, s tood out and l ed t o t he i dea t hat i n s uch cases t he e arthwork c ircuit would b e f ound o n a d ifferent a lignment t o t hat of t he l ater wall ( Wacher 1 974, 1 88-91; 3 512 ). I n a ssessing i n 1 964 t he r elative m erits o f emergency a nd e conomy a s t he most l ikely e xplanation f or t he earthwork t ype o f d efence, Wacher c onsidered t he e vidence f rom C irencester. H ere, a monumental s tone gate was f ound t o have been e rected

6

p rior t o t he building o f t he r ampart. O ne e xplanation o f t his would be t hat t he gate was o riginally e rected as a f rees tanding a rch. B ut Wacher, who c onducted t he e xcavations, thought that t he s tone gate and t he r ampart were l ikely t o belong t o t he s ame o peration. This s ituation p resented s omet hing of an enigma s ince i t s eemed i llogical t hat i n an emergency t he c ompletion o f t he r ampart would b e d elayed while a massive s tone gate was built. Y et i t was f elt t hat t he t own would h ardly h ave b een s urrounded w ith t imber a nd e arth defences i f t hey c ould h ave b een b uilt a t l eisure.(1964, 1 09 & ni) . I n 1966, Wacher t ook up this point, n oting i n a ddition s tone-built gates ' of d efensive c haracter' a t S ilchester a nd Verulamium which a lso p receded r amparts ( 1966, 6 0-2, 6 3, 6 56 ) . Wacher a ssimilated t hese e xamples i nto h is a rgument b y suggesting t he possibility t hat, i n t he f ace of a t hreat f rom one o f t he f rontiers, c ertain t owns r eceived o rders t o e rect f ortifications and t hat t he work was p lanned entirely i n s tone. But t hreat was s uddenly t ransformed i nto o utright danger and earthwork defences h ad t o b e e rected hastily. Wacher mentioned t he f act t hat, a lthough t he majority o f stone walls o ccupied v irtually t he s ame positions as t he earlier earthworks, t here w ere c ases w here t he a lignments differed, f or e xample a t M ildenhall. H e s uggested t hat s ince t he w all h ere i s k nown t o b e l ate, t hat i s l ate 3 rd/4th century r ather t han early 3 rd century, t he change o f a lignment might b e e xplained b y t he d isappearance o f t he e arlier l ine by the t ime the wall was e rected. While t here i s n othing w rong with t his a rgument, i t i s p erhaps p ossible t o d etect h ere, a nd i n t he suggestion o f earthworks on d ifferent a lignments f rom t he w alls a t Canterbury a nd L eicester, t he b eginnings o f t he approach to t he s tudy o f defences b ased on a ' two period' analysis. ' There i s a c lear-cut d ivision i nto t wo c lasses': t owns which possessed earth ramparts f ortified at a s pecified t ime i n t he l ate 2 nd c entury, others w hich h ad t o w ait u ntil a walled circuit was p rovided. The p rovision o f s tone d efences was n ow c learly u nderlined a s a s econd major p hase i n t he defending of t he c ivil s ettlements of t he p rovince. Wacher made i t c lear t hat h is c onclusions w ere t entative and t hat much more e vidence was r equired, y et t he f oundations o f t he chronological a nd h istorical f ramework h ad b een effectively l aid. F rom t hen onwards, i t b ecame more a matter o f r efining t han o f t esting t he h ypothesis. J arrett hypothesis t o to i ntroduce

( 1965) b e an

understood

h ow

e asy

i t

was

f or

a

working

t ranslated i nto a ccepted d ogma a nd h e t ried e lement of caution i nto what h e s aw a s

a ttempts t o s tereotype d efences. H e a rgued t hat t he e vidence was i n many c ases i nadequate t o s ubstantiate t he c laim f or t wo d istinct chronological periods i n t he c onstruction o f b ank a nd wall. He s uggested t hat i t was e qually c apable i n s ome c ases of b eing i nterpreted i n t erms o f a s ingle s cheme w ith t he earth bank r epresenting t he e arlier e lement o nly i n t he structural s ense. H e s tressed t he i nadequacies o f t he dating evidence a nd p articularly t he t erminus p ost q uem w hose n ature and

r elevance

f or

t he

dating

of

detail.

7

defences

h e

d iscussed

i n

s ome

Jarrett, h owever, d id n ot d eny t hat a r ampart a nd w all c ould r epresent two entirely s eparate phases of d efence o r t hat a c onsiderable p eriod o f t ime m ight h ave e lapsed b etween t he building of t he bank and t he a ddition of t he wall. H e d iscussed i n s ome detail ways i n w hich one could establish t hat a s tone wall was s ignificantly l ater t han t he b ank b ehind i t. I n t his, h e emphasised t he n eed t o i nvestigate t he physical and s tructural r elationship of t he bank and wall r ather t han s imply r elying on t he i nadequate dating of t he t wo e lements. ' It was n o part of my purpose t o s uggest that t here a re n o t owns w ith e arthwork d efences s ignificantly e arlier t han t heir s tone walls. I s ought r ather t o ensure t hat, w here t his was b eing c laimed ( and i mportant h istorical c onclusions t he e vidence f or t wo d istinct d efences was d rawn f rom i t) , a dequate, a nd h ad b een c onsidered i n t he l ight o f various p ossible i nterpretations'. H e f elt t hat t he e vidence f rom o ne s ite was b eing e xtrapolated t o o ther s ites where i t was i nconclusive. B y t his t ime, i t h ad b ecome obvious t hat t own walls could not be r egarded as c ontemporary i n the s ame way as t he majority o f e arthwork d efences, b ecause many h ad by n ow p roduced dating e vidence l ater t han t he l ate 2nd c entury w hich C order h ad s uggested ( Corder 1 955, 2 4; Jarrett 1 965, 27; 2 Wacher 1 966,67). The l ogic of t his h ad of course been to e xpose e arthwork d efences a lso t o t he p ossibility o f different dates. ' Each t own must be c onsidered on the e vidence i t p roduces t he a ttempt t o i llumine one t own b y r eference t o another only s erves t o darken t he whole' ( Jarrett 1 965, 27). 2 But Jarrett's p oint, understood by b oth F rere w hich s howed t hat t hey s aw

a lthough c learly s tated, and Wacher, who r eplied h is a rticle s imply as a n

was m isin t erms a ttempt t o

d iscredit t he i dea of two d istinct chronological phases i n t he c onstruction o f b ank a nd wall. Their r eplies w ere t herefore devoted t o v indicating t he c oncept o f a s eparate i ndependent e arthwork p hase. I n many w ays, t his s erved t o d raw attention t o t he i nadequacies and t he d iversity o f t he e vidence, a nd s howed

i n

p articular

h ow

i mpossible

i t

w as

t o

date

earthworks

and walls a nd t o a ssess t he amount o f t ime which h ad e lapsed b etween t he t wo p hases s imply o n t he b asis o f t he t erminus post quem p rovided by t hese f eatures ( Frere 1 965, 1 38; Wacher 1 965, 2 26). These t hree a rticles by J arrett, F rere a nd Wacher h ave a l asting value i n t heir d iscussion of t he s tructural phases a nd f eatures o f b ank a nd w all i n t he l ogistics o f defence c onstruction ( Jarrett 1 965, 5 7-9; F rere 1 965, 1 37-8; Wacher 1 965, 26), b ut i t i s c lear f rom t he a rguments u sed by F rere 2 and Wacher t hat t hey h ad i n f act a dopted t he very approach a gainst w hich Jarrett h ad b een a rguing. Their r eplies t o Jarrett s how h ow f ar t he c oncept o f a g eneral f ramework a nd i ts bas ic p remises a s o utlined b y W acher h ad a lready p assed i nto a ccepted t hinking. Jarrett's e vidence, an c hange

i n

t he

c all f or a n o bjective a ppraisal o f t he a cceptance of i ts l imitations and t herefore a w ay

i n

w hich

t he

8

w hole

s tudy

o f

t he

defences

should b e a pproached, was i gnored. S ince Wacher's s econd article ( 1966) , t here has been a c hange only i n t he details of the b asic c hronological a nd h istorical f ramework a nd n o c hange at a ll i n t he a ctual approach t owards t he s tudy of d efences ( c.f. Hartley 1 983). The a lterations c oncern t he dating a nd historical contexts f or t he various phases of widespread defensive a ctivity. I n i ts f inal f orm, t his f ramework h as b een s ummarised by F rere ( 1967, 2 48-56) . This e ffectively marked t he end of any f urther d evelopment o r r einterpretation o f t he p urpose a nd nature of defences and, s even y ears l ater, i n t he s econd edition of F rere ( 1974, 2 83-92), t he f ramework r emains unchanged e xcept f or a n a djustment i n t he date s uggested f or the b eginning o f t he p rogramme o f w alling ( Frere 1 967, 2 53; 1978, 2 88-9) . I n t his way, hypotheses about c ivil defences have t ended t o h arden i nto ' facts' w ith a l ate 2 nd c entury date b eing a ccepted f or t he majority o f earthwork defences, a 3rd c entury date f or t he p rovision o f s tone c ircuits w ith t he occasional walling i n t he 4 th c entury, and a Theodosian context, a fter t he b arbarian c onspiracy o f AD 3 67, f or t he introduction of b astions.

Early Defensive C ircuits

( ist

a nd

e arly

t o m id

2 nd

c entury)

D efences were p rovided i n t he i st c entury and f irst half of t he 2 nd c entury a pparently only f or a small n umber o f s ites ( fig. 1 ). S o f ar, t he s ites where t his has b een conclusively demonstrated i nclude t he c oloniae a nd a f ew o f t he c ivitas capitals and ' there i s a s y et n o e vidence t hat any v illages were s o t reated' ( Wacher 1 975, 5 1). These e arly d efences usually i nvolve earthwork c ircuits a nd s tone walls were only provided a t a n e arly date f or t he t hree c oloniae. E arthwork

c ircuits

century have b een Winchester, and a t

a ttributable

i dentified E xeter t he

t o

s ome

t ime

i n

t he

i st

a t S ilchester, Verulamium a nd l egionary f ortress defences were

apparently r etained f or t he c ivitas c apital. There i s a certain amount of e vidence f or f ortifications i n s tone i n t he early 2 nd c entury a t t he t hree c oloniae Colchester, Gloucester a nd L incoln ( Wacher 1 974, 7 2; 1 978, 9 6; F rere 1 978, 2 83). There a re p recedents f or t he w alling o f c hartered c ities such as t he c oloniae and i t i s a rgued t hat t heir special status as s ettlements o f R oman c itizens w ould e nable t hem t o qualify more r eadily t han o thers f or i mperial permission t o erect d efences ( Wacher 1 974, 7 2-5; F rere 1 978, 2 83). B ut, apart f rom t he coloniae, where i t i s a ccepted t hat a n e lement of p ride may h ave b een i nvolved i n t he p rovision o f w alled circuits at an e arly date, t he p rovision o f defences i n t he -

ist c entury a nd e arly p art o f t he 2 nd c entury i s r egarded a s exceptional. I t h as b een s uggested t hat t he c hartered s tatus of V erulamium a s a municipium would e xplain i ts e arly d efences ( Frere 1 964a, 6 5 & 8 0; Wacher 1 974, 2 05). S ilchester i s l ess easy s ince i ts s tatus was t hat o f a p eregrine c ivitas c apital and t he e xplanation most often quoted f or i ts early d efences

9

0

1 50K m

e arthwork 0

w alled c ircuit



e arthwork a nd w all e arthwork w ith p ossible e arly w all

F IGURE

1

E arly d efensive c ircuits

1 0

:

l OOm i e s

i s i ts position w ithin t he c lient k ingdom o f Cogidubnus. T he same i s thought t o apply t o Winchester where t here i s evidence f or a defensive c ircuit i n t he l ater 1 st c entury. I t i s t hought necessary, however, f urther t o e xplain these defences i n t erms o f s pecific t hreats a nd t here a re a number of occasions during t he 1 st c entury when t he t hreats t o t he r ecently a nnexed l owland z one h ave b een r egarded a s sufficient j ustification f or t he c onstruction of d efences at certain important a nd d eveloping c entres. The e arliest defences s eem t o have b een t hose at Silchester ( Inner E arthwork) a nd Verulamium (1 1 955' d itch) which have been a ttributed t o t he C laudian period a lthough with l ess c ertainty a t Verulamium w here a c lose d ate f or t he construction of t he earthwork was n ot e stablished ( Frere 1 964b, 1 04; 1 978, 2 83; Boon 1 974, 4 5). Boon c oncluded t hat the S ilchester I nner E arthwork was e rected s oon a fter AD 4 3 and c ertainly b efore AD 5 0. H e t hought t hat t he e xplanation for i ts p rovision m ight be f ound i n t he p roximity o f Silchester t o t he n ew f rontier w hich was b eing t hreatened by the activities o f Caratacus ( Boon 1 974, 4 5) . This was s uggested a s a l ikely c ontext t oo f or t he p rotection o f Verulamium

( Wacher

1 974,

2 04-5

& 2 56;

1 978,

9 6)

One e vent w hich might b e t hought t o b e r esponsible f or the p rovision of defences i s t he B oudiccan r ebellion o f AD 6 0/61 o r i ts a ftermath a nd F rere s uggests t hat t his i s t he possible a lternative c ontext f or t he defences at Verulamium ( Frere 1 964b, 1 04). A lso a ttributed t o t he Boudiccan r ebellion, o r r ather t o i ts a ftermath, a re t he r emains o f what could p rove t o b e a h itherto u nsuspected e arly d efensive circuit r ecently d iscovered a t C olchester. The e vidence c ons ists of a d itch s een s o f ar only o n t he w est a nd r unning i n f ront of t he Balkerne G ate. U nfortunately, i ts r elationship t o a n immediately p ost-Boudiccan c ontext i s n ot s ecure s ince t he r emains o f buildings which were possibly burnt during t he Boudiccan r ebellion w ere n ot d irectly c ut b y t he d itch ( Crummy 1 977, 9 6 & 1 01, f igs. 2 0 & 2 1) . Until r ecently, i t had b een assumed t hat C olchester r emained u ndefended u ntil a w alled circuit was e rected s ome t ime i n t he 2 nd c entury, s ince t he defences o f t he l egionary f ortress s eem t o h ave b een l evelled when the c olonia was f ounded i n AD 4 9 and Tacitus' s tatement h as b een t aken t o mean t hat C olchester h ad n o r ampart i n AD 6 0 I t i s n ot e asy, ( Crummy 1 977, 7 6 & 9 6; F rere 1 978, 2 83) . h owever, t o e xplain w hy Verulamium a nd S ilchester s hould h ave been p rovided w ith defences while Colchester r emained u nprotected. Wacher h as s uggested t hat a t Winchester, w here t he e vidence p resently a vailable s eems t o i ndicate a d efensive c ircuit o f l ate N eronian o r e arly F lavian d ate ( terminus p ost quem c . AD 7 0), an e xplanation f or t he c onstruction of defensive c ircuits w ithin t he k ingdom o f Cogidubnus m ight b e f ound in t he c ivil war i n t he empire i n AD 6 8-70 and t he u nreliability o f t he a rmy i n B ritain a t t his t ime ( Wacher 1 974, 2 57-60; 1 978, 9 6) a nd h e h as a lso s uggested t hat t hree minor s ites A lfoldean, H ardham a nd I ping which a lso s eem -

t o t his

have s ame

b een

-

f ortified

p rogramme

i n

f or

t he

Neronian

t he

d efence

1 1

period o f

were

t he

part

of

k ingdom

o f

Cogidubnus ( Wacher 1 978, 9 7). date, h owever, i s questionable.

The

e vidence

f or

a

Neronian

I t i s more d ifficult t o e xplain t he apparent r etention o f t he l egionary f ortress defences f or t he c ivitas c apital a t E xeter f rom c .AD 8 0 u ntil t he Antonine p eriod i n t erms e ither of a special s tatus f or t he s ettlement o r o f a specific k nown t hreat. H owever, f ollowing t he e xplanation offered i n t he he r etention c ases mentioned above, B idwell s uggested t hat ' t o f t he f ortress defences a t E xeter t hus s upplies s ome evidence t hat t he population was j udged t o b e s trongly p hilo-Roman. Their r etention may a lso s uggest t hat t he t own was l iable t o attack, a lthough w hether by t he D umnonii, or a t l east a s ection o f t hat t ribe, o r by m ore d istant e nemies i s uncertain' ( Bidwell 1 980, 2 3 4 6 & 6 6). A lthough i t i s p ossible t o s uggest t hat t here was a t radition of defence i n t he i st c entury f or major c entres, a t l east i n t he s outh, t hese d efences a re s een a s a d irect r esult of specific t hreats o r periods of i nsecurity f ollowing t rouble w ithin t he p rovince. There i s n o a cceptance of t he possibility t hat t hey might h ave b een p rovided ' as a matter o f c ourse' ( Wacher 1 978, 9 6). With t he c oloniae i t was a d ifferent matter and i t h as b een s uggested, i n the c ase of t wo o f t hem G loucester a nd L incoln which were f ounded i n t he l ater i st c entury ( AD 9 0s) , t hat i t may have been i ntended t o p rovide t hem w ith d efences f rom t he b eginning. F or, unlike t he c olonia f ounded h alf a c entury earlier a t Colchester, t he l egionary f ortress d efences a t G loucester a nd L incoln w ere l eft i ntact ( Frere 1 978, 2 83) -

-

We t herefore h ave a model w hich r egards e arthwork defences of t he 1 st c entury a s emergency fortifications e rected i n r esponse t o t hreats t o t he n ewly a nnexed c ivil z one ( Wacher 1 978, 9 6) . The s tone wall c ircuits s ubsequently p rovided o r b egun a t e ach o f t he t hree c oloniae s tand apart f rom t hese measures ( Frere 1 978, 2 88; Wacher 1 978, 9 7).

Late

2 nd

c entury T imber

a nd Earthwork

D efences

' Mass p rovision of earth r amparts f or t owns i s s omething q uite w ithout p arallel i n t he R oman empire' ( Frere 1 978, 2 85). I n t his context t he hypothesis t hat A lbinus' b id f or t he imperial t hrone i n AD 1 96/197 n ecessitated a w idespread p rog ramme of c ivil d efence f or t he p rovince h as r eceived s ome c redence.So s uitable d oes t his e xplanation s eem t hat t he evidence f rom t he majority o f earthwork c ircuits h as b een i nterpreted i n i ts l ight. R ampart c ircuits w ith widely divergent t ermini post quos i n t he m id and l ate 2 nd century a re a ssumed t o b elong t o t his c ontext a nd o f c ourse t his dating e vidence i s n ot i nconsistent w ith t his i nterpretation. S imilarly, t he d iscovery o f e lements i n s tone, u sually gates, a ssociated w ith earthwork c ircuits a t a n umber of t he civitas c apitals h as l ed t o t he d evelopment o f t he h ypothesis t hat work

w as

o riginally

b egun

i n masonry b ut

1 2

i ncreased

u rgency

a nd

7D 'KV C he o X /

( uc

0

0

E

a x

e arthwork w alled c ircuit e arthwork a nd w all e arthwork w ith p ossible e arly w all

1 3

D WM kd

_‚ — X Sc

1 50K m

1 00 m i les

F IGURE 2 :

A le A ich A id A lf A nc

D efensive c ircuits e rected f rom t he

A lc est e r A lchester A ldborough

K T K en K y

A lfoldean A ncaster

L M

B F

B rough-on-Fosse

B H

B rough-onHumb e r B ann a venta

B an B ath B i B ra

B ath

C F

C hesterton-on-Fosse

C N C W C Y C aer C am

C aistor-by-Norwich

C ant C arl C rr m C att C he C hi C ir C ol C or D l D on D er D orn D w

L ei L et L in

B itterne B rampton

K irkby T hore K enchester K ei vedon L ong M elford L eicester L etocetum L incoln

L itt L on L tw

L ittlechester L ondon

M C

M elandra C astle M anchester M aldon

M an M d M du

C aistor-on-the-Wolds C aistor-by-Yarmouth C aerwent C ambridge C anterbury

l ater 2 nd c entury o nwards

L eintwardine

M ild M rg M t

M anduessedu j M il denhal 1 M argidunum M alton

N ea

N eath a m

P en

P ennocrucium

R oc R och

R ocester R ochester

S C S il S i

S hoebury C amp S iichester S lack

D orchester D orn

T d T N

T iddington

D roitwich

T ow T n

C arlisle C armarthen C atterick C helmsford C hichester C irencester C olchester C orbridge D orchester-on-Thames D oncaster

T horpe-by-Newark T ow ce s te r T n p ont i um

E >

E xeter

F S

F enny S tratford

U C U xa

U phali C amp U xacona

G a

G atcombe

G Ca G Ch G b

V er

V erulamium

G reat C asterton G reat C hesterford G loucester G odmanchester

W N

W ater N ewton

W an W hit W in

W anborough W hitchurch W inchester W ixoe

G cd H a

H ardham

H er

H orncastle

l ich I lk

Ii chester I lkley I ping I rchester

' p I re

1 4

W ix W k W or W x W y

W ickford W orcester W roxeter W ycomb

Y ork

Y ork

p ressure scheme i n 1 966, 6 6;

o n r esources n ecessitated t he c ompletion c heaper a nd more r eadily a vailable materials 1 974, 7 5; 1 975, 5 1; 1 978, 9 7-8).

o f t he ( Wacher

I t i s f elt t hat t he coloniae, i f n ot a lready p rotected, and t he t ribal c apitals would h ave b een d efended a s a matter of p riority and s o i ndications of t he i nitial u se of s tone here, but a t f ew o f t he ' lesser' c entres ( but c .f. L ittle Chester, Webster 1 975b, 5 3), f its i n w ith t his i nterpretation ( Wacher 1 975, 5 1). H owever, t he a bsence o f a ny i ndication o f an i ndependent e arthwork c ircuit a t t hree major c entres Canterbury, L eicester a nd L ondon i s hardly c onsistent w ith priority having b een g iven t o t he major s ites. The explanation i s t hought t o l ie i n an earthwork phase on a different a nd a s y et u ndiscovered a lignment f rom t hat o f t he walled circuit ( Wacher 1 974, 1 88, 3 51-3; 1 978, 9 8; F rere 1 978, 2 83 & 2 86). More r ecently h owever Wacher h as s uggested a n alternative e xplanation f or t he l ack o f an earthwork c ircuit at L ondon. H e c onsiders t hat t he i ndications o f work b egun i n masonry at other major centres may a ctually have b een f ully r ealised h ere a t t he p rovincial c apital i n t he k nown w alled circuit which has b een a ttributed i n t he past t o t he early 3 rd century ( Wacher 1 974, 7 5, 9 4-5; 1 978, 9 8; F rere 1 978, 2 86). This possibility w as f irst r aised by Merrifield who n oted t hat a c onsiderable a mount o f p ottery f rom w ithin a nd b eneath t he internal b ank o f t he wall on f our d ifferent s ites ( Cooper's Row, Trinity P lace, t he O ld Bailey a nd t he T ower o f L ondon) indicated t hat t he wall was p robably built w ithin a decade o f AD 2 00. The p ottery a ntedating t he b ank r anged d own t o t he late 2nd c entury b ut d id n ot i nclude a ny material which could be d efinitely a ttributed t o t he e arly 3 rd c entury. H e suggested t herefore t hat, s ince i t was unlikely t hat L ondon was l eft u ndefended i n t he l ater 2 nd c entury, a masonry w all was p rovided f or t he c apital w hile l esser c entres w ere p rovided with earthworks ( Merrifield 1 969, 1 18-20) -

-

Although t he p rovision o f e arthwork d efences was extensive, i nvolving many s ites b esides t he t ribal capitals and t he c oloniae, t here w ere many s ites w hich w ere l eft u ndefended. I n considering t he b asis f or t he s election of t hese ' lesser' s ites, t here i s a greement t hat c onsiderations o f s ize or s tatus a re i rrelevant ( Wacher 1 975, 5 1; Webster 1 975b, 5 3 & 5 9). I t w as n ot t he s ettlement w hich was b eing d efended b ut an o fficial f unction ( Todd 1 969, 4 9-53; Wacher 1 975, 5 1; 1 978, 9 8; Webster 1 975b, 5 3, 5 6 & 5 9). The w ithdrawal o f t roops which Albinus' c ampaign i nvolved would h ave l eft t he p rovince vulnerable t o a ttack a nd t o t he p ossibility o f t he b reakdown of l aw and o rder ( Wacher 1 974, 7 5; 5 1; 1 978, 9 7; Webster 1 975b, 5 3; F rere 1 978, 2 85). Webster ( 1975b) o ffers t he i nterpretation t hat t he a im o f t his w idespread s ystem of earthwork defences was t o c reate a s ystem o f f ortified p oints a long t he major routes t o p rotect t he communications n etwork and enable Albir ius t o r ecover c ontrol o f t he p rovince q uickly. I n t his respect, t he a im w as s imilar t o t hat o f t he l ater b urgi which, i t i s a rgued, w ere e stablished by C onstantius Chlorus a s p art of h is ( Webster

campaign f or t he 1 975b, 5 3 & 5 9).

r ecovery o f

1 5

t he

p rovince

i n

AD

2 96

This i nterpretation i nvolves t wo e lements i n t his l ate 2 nd c entury system. F irstly t he major t owns c ivitas c apitals, c oloniae a nd t he p rovincial c apital w ere g iven p riority and defences were i ntended t o e nclose as m uch of t he s ettlement a s was f easible. Webster h as s uggested t hat, a part f rom t he obvious i mportance of t hese c entres in t he a dministrative f ramework o f t he p rovince, A lbinus p robably r elied on t he t ownspeople f or f inancial s upport i n h is b id f or t he purple. As a r esult, t heir i nterests i n t he matter o f defence and i n t he c ircuit t aken by t he defences h ad t o b e c onsidered ( pers. c omm. G . Webster, F ebruary 1 979). -

-

The s econd e lement i nvolved t he p rotection of g overnment i nstallations a nd t he e stablishment o f a f ortified r oad n etwork. U nlike t he major c entres, t he p rotection of t he s ettlement a s s uch w as n ot c onsidered ( Webster 1 975b, 5 6).

Stone Defences

The s ubsequent p rovision o f s tone w alls f or s ettlements h itherto p rovided only w ith t imber and e arth d efences and f or s ome p reviously u nfortified c entres w as n ot s atisfactorily e xplained. F or i t meant t hat t he emergency earthworks e rected u nder t he v ery s pecific c ircumstances o f t he withdrawal o f t roops by A lbinus i n AD 1 96/197 were b eing g iven ' permanence' a t a t ime w hen t he s ecurity o f t he p rovince h ad been r estored. F urthermore, t he p rovision of walls f or c ertain p reviously undefended c entres a nd t heir a bsence f rom s ome of t he c entres which had r eceived earthwork c ircuits meant t hat i t was n ot s imply a q uestion o f making p ermanent t his s ystem o f t imber and earthworks a s h as b een i mplied ( Wacher 1 975, 7 5; F rere 1 978, 2 88) b ut o f modifying a nd e xtending i t, t hat i s assuming t hat t he c oncept of a ' system' o f f ortified centres can b e r egarded a s t he r esult o f a n a ct o f g overnment policy. Various contexts have b een s uggested f or t he provision o f d efences i n s tone. I nitially t hey w ere t hought t o b e t he w ork of S everus ( AD 1 97-211) and, when Corder s ( 1955) h ypothesis t hat t he majority o f w alled c ircuits w ere t he work o f Albinus was d isplaced and earthworks a scribed t o t his event, a Severan context f or w alls a gain s eemed l ikely ( Wacher 1 964, 1 11; 1 966, 6 7; H artley 1 966, 5 7; F rere 1 978, 2 86). I t s eemed l ogical t o a ssociate t his w ith S everus' r estoration o f t he f rontier defences of t he p rovince ( Frere 1 967, 2 53) , but i t h as s ince b een t hought u nlikely t hat S everus w ould h ave concerned h ims elf more t han was n ecessary w ith a p rovince which h ad s hown anti-Severan s ympathies, a nd h is i nterest i n t he p rovince would have b een c onfined t o r e-establishing t he s ecurity o f i ts f rontiers. I ndeed, Webster s uggests t hat S everus may h ave i nflicted punitive measures on t he p rovince a nd on t he t owns i n p articular ( pers. c omm. G . Webster, F ebruary 1 979; Saiway 1 965, 1 87-8) . T he c urrent i nterpretation offers the s uggestion t hat t he p rogramme o f walling may h ave b egun u nder Severus' successor, Caracalla ( AD 2 11-217), e ither t o match t he p rog ramme o f m ilitary i mprovement i n t he n orth o r a s a c orollary of h is g rant o f R oman c itizenship i n AD 2 12 to a ll f ree

1 6

i nhabitants o f t he empire ( Rivet 1 964, 9 3; Merrifield 1 965, 5 1; F rere 1 978, 2 88). These t wo c ontexts i nvolve t wo very d ifferent c oncepts o f t he p urpose o f t hese walled c ircuits, t hat i s a s a matter o f p rotection f or t he p rovince o r a s a p rivilege t o b e v iewed i n c onnection w ith c ivic p ride. I n

the

development of

various

t heories

concerning

t he

c ircumstances w hich s urrounded t he p rovision o f s tone walls, t he i nitial concept h as persisted t hat t he walling o f various c entres r epresented a s eparate p hase o f d efensive a ctivity which i s l ikely t o b e part o f a s ingle s cheme f or t he p rovince a s a whole a nd t herefore b roadly c ontemporary a t a ll t he c entres i nvolved ( Corder 1 955, 1 24; R ivet 1 964, 9 2; F rere 1 978, 2 86-7). B ut b y t he m id 1 960s a n umber o f t own w alls w ere f ound to have a t erminus post quer n i n t he l ater r ather t han t he e arly 3 rd c entury a nd t his c ast s ome d oubt o n t hose walls which had b een dated p reviously t o t he l ater 2 nd o r early 3 rd c entury, ( for e xample, A lchester, A ldborough, G reat Casterton) a nd p rompted a r e-examination o f t he pottery f rom t he early e xcavations a t t hese a nd o ther s ites. This l ed Hartley ( 1966, 5 7) t o conclude t hat ' towns with one o r two published s ections w hose walls c an b e d ated e arlier t han t he c losing y ears o f t he 2 nd c entury a re n ow c learly s uspect. At l east we may n ow be s ure t hat s ome t own w alls were b uilt v ery much l ater t han u sed t o b e thought'. I n emphasising a l ate date f or s tone walls, H artley mentioned C anterbury ( AD 2 70-290), G reat Chesterford ( perhaps early 4 th c entury) , a nd Caerwent, D orchester-onThames and S ilchester w hich w ere l ikely t o b e a fter t he m id 3 rd c entury ( Hartley 1 966, 5 7). H owever, i t b ecame c lear t hat a s a g roup t own w alls c ould n ot b e s hown t o b e a s n early c ontemporary as e arthwork defences w ere a rgued t o b e. Alt hough a c onsiderable n umber o f w alled c ircuits h ad a t erminus p ost quem i n t he l ate 3 rd c entury, t here were s ome s ites which s eem t o have b een p rovided w ith walls v ery much e arlier. Apart f rom L ondon and t he t hree c oloniae, Verulamium and possibly C irencester s eem t o b elong i n t he c ategory w here a n e arly 3 rd c entury date i s i ndicated ( Wacher 1 974, 9 4-5, 13-15, 1 22, 1 1 44,

1 47-9,

2 13-15,

2 65-6,

3 02;

F rere

1 978,

2 87-8).

W ith r egard t o t he dating o f walled c ircuits, t here i s i n r eality very l ittle e vidence. I t i s b ecause o f t he a bsence o f c onclusive dating e vidence f or wall c ircuits t hat F rere, Wacher and o thers h ave a ttempted t o e stablish a c hronology f or s tone defences b ased on a rchitectural s tyles, u sually of g ates, based o n c omparison w ith d evelopments i n m ilitary a rchitecture. On t he b asis of t his r easoning, F rere a rgues t hat m ost o f t he w alls b elong t o a s tyle w hich i s e arlier t han t hat of t he l ate 3 rd c entury s eries o f Saxon Shore f orts a nd e arlier t han t he t own w alls o f G aul. I t i s a ssumed t hat t he majority of walled c ircuits i n t his p rovince were b uilt b efore t he r eign o f P robus ( AD 2 76-282), w hen t he main s eries o f S axon Shore f orts i s t hought t o have b een b egun ( Frere 1 978, 2 86). I t i s a rgued t hat a s mall n umber Catterick, G reat Chesterford a nd Thorpe-by-Newark can b e a ssumed t o b elong t o t he e arly 4 th c entury s ince t hey c onform t o t he n ew m ilitary a rchitecture, having walls w ithout banks b ehind t hem ( Frere -

-

1 978,

2 88;

Wacher

1 978,

9 9).

1 7

On t he b asis o f t his e vidence, F rere ( 1978) c onsiders t here t o be two e lements i n t he c onstruction o r r econstruction o f d efences i n s tone :F irstly a p rogramme o f w all b uilding w as b egun i n t he e arly 3 rd century w ith t he i ntention of making permanent t he t imber and e arth d efences a lready i n e xistence. I n t his p rogramme, L ondon and t he c oloniae were g iven p referential t reatment b ecause o f t heir e levated s tatus w ithin t he p rovince ( Hartley 1 966, 5 7; Wacher 1 975, 5 1; F rere 1 978, 2 88). i n

Secondly a n t he l ater

a cceleration i n t he p rogramme o f wall 3 rd century was perhaps a r esponse t o

b uilding t he i n-

c rease i n b arbarian p ressure o n t he f rontiers of t he empire and specifically i n r aiding a long t he c oasts. The s ituation c an b e p aralleled i n t he p rovision o f w alled c ircuits f or many s ettlements i n G aul i n t he l ate 3 rd c entury f ollowing t he barbarian i nvasion a cross t he R hine ( Frere 1 978, 2 84). This a spect has b een s tudied by Johnson ( 1976) who considers t hat a n e xplanation f or t he w alling o f many c ivil s ettlements a nd f or t he t iming of t hat walling at v arious c entres m ay b e f ound i n t he w ider c ontext o f t he d evelopment o f a s ystem o f c oastal defence f rom t he early 3 rd c entury a nd t he acceleration o f t his development i n t he l ater 3 rd c entury ( Johnson 1 976, 1 52 0, 1 00-6, 1 41-2, 1 44). F rere ( 1978, 2 89) a lso a rgues t hat t he i ncreased c oncern f or s ecurity during t he 4 th century r esulted i n t he p rovision of defences f or s ome apparently u ndefended c entres Cambridge, Catterick a nd G reat C hesterford a nd a n ew c ircuit a t M ildenhall. -

-

The q uestion i s l argely o ne o f e mphasis i n v iew o f t he i nadequate i nformation. The p rovision o f defences i n s tone c an b e s een a s a f airly c ontinuous p rocess t hroughout t he 3 rd century, which t ook some t ime t o i mplement, and which may h ave b een a ccelerated i n t he l ate 3 rd c entury. Alternatively i t could b e s een a s a p rocess which i nvolved a deliberate policy o f w alling f or t he major c entres o f t he p rovince i n t he e arly 3 rd century and a s ubsequent government decision t o p rovide s tone defences f or ' lesser' c entres w hen t he s ecurity o f t he p rovince w as a gain t hreatened i n t he l ate 3 rd century. The c onstruction o f d efences i n s tone i nvolved s ome s ettlements which had p reviously b een u nprotected. The basis f or t heir s election i s r egarded a s b eing t he s ame a s t hat f or t he p rov ision of t imber and earth defences i n t he l ate 2 nd c entury, t hat i s, t he p resence o f g overnment i nstallations. T he p rovision of defences at s ome o f t hese n ew s ites i s t herefore s een a s t he r esult o f a dministrative r eorganisation i n t he 3 rd and early 4 th centuries which p roduced, by a process o f s ubd ivision o f t he p rovince a nd t he g rowth o f b ureaucracy, a n i ncrease i n t he n umber of c entres w ith a n o fficial f unction o f s ome k ind ( Rivet 1 975, 12; Wacher 1 975, 5 1; 1 978, 9 8-9; 1 Webster 1 975b, 5 9 & 6 4; F rere 1 978, 2 91-2)

1 8

Burgi, Period

Bastions

a nd

Other

Developments

i n

t he

Late

Roman

T here a re a n umber o f d evelopments i nvolving d efensive activity at c ivil s ettlements i n t he l ater R oman p eriod, namely t he c onstruction o f a s ystem o f b urgi, t he p rovision o f bastions and r epairs o r construction work on defensive c ircuits. The s ystem o f b urgi a nd t he p rovision o f b astions a re regarded as t he d irect r esult of t he application of government policies a t s pecific t imes. S imilarly, a ny r emodelling o r major r econstruction o f a c ivil c ircuit i nvariably p rompts t he s earch f or a r elevant h istorical c ontext t o e xplain s uch a n i nitiative. T he c ase f or a s ystem o f s o-called b urgi h as b een presented by Webster. Chains of burgi o r small f ortified posts a re k nown i n o ther p arts o f t he empire, e .g. a long t he R hine and t he D anube f rontiers ( Webster 1 971, 4 5). I t has b een s uggested t hat t hey were u sed a s f rontier o r b order p osts i n these a reas. They were manned by small garrisons i n o rder t o control i llegal c rossings a nd w ere o ften c onnected w ith f rontier t rade ( Chevallier 1 976, 1 96-7, 2 91, 2 93, 3 22-3) These, however, a re r ather d ifferent f rom t hose s ites i n Britain t o which Webster h as applied t he t erm. This s ystem i nvolves f ive small r ectangular d efended e nclosures l ying a long and b isected by Watling S treet b etween Wroxeter and Towcester. They a re r egarded a s q uasi-military s trong p oints. T he

s ites

i dentified

by Webster

1 .

U xacona

2 .

P ennocrucium

3 .

L etocetum

4 .

£ 4 anduessedum

5 .

T ripontium

6 .

Bannaventa

-

0 .

7 5a -

6 a

-

-

-

-

4 a

a re

:-

( o.29ha) ( 1.67ha)

( 2.4ha) 5 a

2 .5a

( 2.lha) ( 1.08ha)

a nd

1 2a ( 5ha) h as b een a dded ( Webster 1 975b, 6 4n3)

t o

t he

l ist

T he enclosures vary i n s ize a nd i n t he n umber of d itches p rovided a nd t here i s n o c onclusive e vidence i n e very c ase f or a s tone wall i n a ddition t o t he r ampart, but t heir general character i s r egarded a s b eing s o s imilar t hat t hey m ay h ave been a s ingle system ( Webster 1 971, 4 2; 1 975a, 4 6). A lthough t here i s -l ittle e vidence, Webster r egards t hese s ites a s t he survivors o f a more e xtensive system. H e s uggests t hat o ther burgi may b e e xpected s outh o f T owcester a t F enny S tratford and D unstable ( Webster 1 975b, 6 4n3). No d efences a re k nown i n the l atter c ase a nd t he defences a t F enny Stratford do n ot s eem t o b e r eadily c omparable o n p resent e vidence w ith t he known burgi. H e s uggests a lso t hat t wo f urther p ossible s ites i n t he s outh may b e p art o f t he s ystem e xtending t o D over o r Bosha r n.

There

a re

t hree

small

1 9

r ectangular

enclosures

i n

Sussex a t A lfoldean, H ardham a nd I ping, t he f irst t wo lying o n t he Stane Street b etween Chichester a nd L ondon and t he t hird on t he r oad f rom Chichester t o S ilchester. Webster ( 1975b) considers t hat t hey a re s o l ike t he b urgi on Watling S treet t hat i t i s n ot i nconceivable t hat t hey r epresent p art of t his f ortified system. I n r eviewing t he dating e vidence a nd considering t he Possible context f or s uch a system of defended strong points, Webster was o riginally i nclined t o t he v iew t hat t hey w ere l ikely t o b e a ssociated w ith t he work of C ount Theodosius i n r estoring t he p rovince a fter t he b arbarian i nvasion o f A D 3 67 a nd specifically w ith t he t ask of c learing t he c ountry of b rigands ( Webster 1 971, 4 2 & 4 5). S ubsequent e vidence f rom Manduessedum p rompted a re-examination of t he evidence f rom t he other s ites a nd l ed h im t o c onclude t hat t hese b urgi w ere l ikely t o be c onsiderably earlier i n d ate ( late 3 rd/early 4 th c entury) , performing t he s ame f unction b ut i n t he c ontext o f C onstantius Chlorus' r ecovery of t he p rovince f rom t he u surper A llectus i n AD 2 96 ( Webster 1 975b, 5 3 & n 3, 5 9). The p rovision of b astions f or a n umber of civil s ettlements i s a n otable d evelopment o f t he l ater Roman perIod. With r egard t o t heir dating and t he context f or t heir p rovision, i t i s g enerally t hought t hat, a s w ith e arthwork defences, b astions a re t o b e r egarded a s t he r esult o f a s ingle p olicy. The c ontext n ow p referred i s t he r eorganisation of t he defences of t he p rovince u nder C ount T heodosius i n t he y ears f ollowing t he b arbarian conspiracy o f AD 3 67 ( Todd 1 973, 1 27; Wacher 1 974, 7 5; 1 978, 1 00; Webster 1 975a, 1 05-6; F rere 1 978, 2 91 & 3 97). This would mean t hat t he p rovision of b astions o ccurs considerably l ater i n Britain t han i n G aul w here t he walls o f t he l ate 3 rd a nd e arly 4 th c enturies were p rovided w ith b astions f rom t he f irst ( Frere 1 978, 2 86) . Certainly i n B ritain, b astions usually s eem t o h ave b een a dded t o t he walled c ircuit a nd i t i s t herefore a ssumed t hat t heir p rovision post-dates t he main p eriod o f w alling i n t he l ate 3 rd c entury ( Frere 1 978, 2 86) Apart f rom t he p rovision of b astions, t here i s evidence f or o ther d efensive a ctivity, s ome o f i t o n a major s cale a nd most n otably i nvolving attempts t o p rovide more s olid a nd s ubstantial w alls a t a n umber o f c ivil s ettlements i n t he l ater R oman period. A Theodosian context h as been s uggested f or much o f t his work a nd w here t he d ating e vidence m akes t his unlikely, various a lternative c ontexts a re c ited u nder w hich s uch d efensive a ctivity m ight h ave b een u ndertaken a s part of a general policy f or t he p rovince a s a whole ( Webster 1 975a, 1 13-14, 1 18; F rere 1 978, 2 92, 3 87-9; Merrifield 1 978, 2 6; Marsden 1 980, 1 78-9)

2 0

I II.

The

T he

Nature

' Type'

a nd P roblems

o f

t he Evidence

S ites

C urrent t hinking a bout d efences, t heir d ate a nd c ontexts is b ased i n f act on t he e vidence f rom a very small n umber o f ' type' s ites, n amely Caerwent, C olchester, S ilchester a nd Verulamium and t o a l esser e xtent B rough-on-Humber, Cirencester, L incoln a nd L ondon. These a re t he s ites most often mentioned i n a ny g eneral d iscussion o f t he p attern o f Romano-British defences ( see e .g. Wacher 1 974, 7 5-8; 1 978, 9 51 00; F rere 1 978, 2 83-92). O f t he o ther c ivitas c apitals a nd coloniae, A ldborough, Caistor-by-Norwich, Canterbury, E xeter, Gloucester a nd L eicester a re mentioned. A lthough a c ertain amount of i nvestigation has t aken p lace, t he s equence i n each case h as hardly b egun t o b e u nderstood. At Carlisle a nd York very l ittle i s k nown b eyond t he e xistence, o r p robable existence, o f a d efensive c ircuit. I n t he c ase o f Wroxeter, we a re still no n earer t o an understanding of t he defensive s equence despite a n umber o f s ections a cross t he c ircuit. T he s o-called ' minor' t owns f igure l ess p rominently i n general discussions. D orchester-on-Thames and G reat Casterton a re t he o nly c entres w hich h ave b een r egularly i ncluded. Otherwise t here h as been comparatively l ittle attention t o t he defences o f t he ' lesser' c entres u ntil t he l ast t en y ears. There a re n ow a s ignificant n umber where t here i s e vidence which ought t o b e c onsidered i n a ny g eneral d iscussion, s uch as Ancaster, B rampton, Godmanchester, Margidunum a nd Mildenhall. I n t he l ast t wenty y ears, a n umber o f h itherto unknown circuits have b een discovered, namely B rampton, Cambridge, Carmarthen, F enny S tratford, N eatham, T iddington, Wanborough a nd possible Wycombe. There a re other c entres where the e xistence o f d efences h as l ong b een k nown b ut w here only very l imited i nvestigation h as been possible, at A lchester, Bath, Dorn, G reat C hesterford, M ildenhall, T owcester a nd Water Newton. I n a ddition, c ertain a reas have r eceived l ittle attention, s uch a s E ssex, where defences a re k nown a t Chelmsford, G reat Chesterford, K elvedon a nd Wickford, o r t he West Midlands, w here a l arge n umber o f d efended n uclei a re known. T he answered i nclusion

f act t hat t here a re s till many q uestions t o b e a bout t hese c ircuits c annot i nvalidate t heir i n a ny g eneral d iscussion, f or t he e vidence f rom t he t y p e s i t e s i s hardly more complete o r conclusive. There h as a lso b een a s teady a ttention i n r ecent y ears t o t he e xcavation of t he d efences o f s uch major c entres a s C irencester, Coichester, L incoln and L ondon, b ut t his i nformation has y et t o b e a ssimilated i nto t he ' general p attern o f d evelopment'. '

'

I f we i solate f rom amongst t he ' type' s ites t hose where the ' normal' d evelopment a ttributed t o R omano-British t own a nd village defences t ook p lace, t he f ollowing ' anomalies' must be excluded

:-

2 1

London

The u nusual s ituation h ere earthwork c ircuit on t he e lsewhere, c oupled w ith a n 2 nd i ts

o r early s tatus a s

3 rd t he

o f n o i ndependent l ine of t he wall o r e arly d ate ( late

c entury) i s p rovincial

attributed capital.

t o

C olchester} Here t he ' abnormal' p rovision of a defensive G loucester} c ircuit b efore t he g eneral p rogramme o f t he L incoln } late 2nd century is attributed to the special s tatus o f t hese c entres a s coloniae. E xeter

The r ecent d iscovery o f t he p robable r etention o f t he l egionary f ortress d efences until t he Anton ine period i s a n u nusual development f or a c ivitas c apital.

S ilchester} Verulamium} Winchester}

Again d efensive c ircuits e rected p rior t o t he l ate 2 nd c entury a re attributed, i n the c ase o f S ilchester a nd Winchester, t o t heir position i n t he k ingdom of Cogidubnus, and f or t he s ame r eason Chichester i s t hought t o b e a l ikely candidate f or special t reatment; t he l ikely s tatus o f Verulamium a s a municipium o f t he i st c entury h as ation f or i ts

b een s een a s s ufficient a nomalous development.

j ustific-

B rough-on-Humber This s ite h as a lso p roduced a n ' abnormal' s equence but here s ome doubt has b een c ast o n t he s tatus o f t he s ettlement; a n e xplanation f or i ts u nusual d evelopment has b een f ound i n t he f act t hat i t may b e a n aval b ase r ather t han a c ivil s ettlement. T his l eaves only Caerwent a nd C irencester f rom amongst t he main g roup of s ites a s e xamples o f t he ' normal' p attern o f d evelopment. The n ature o f t he e vidence f rom t hese t wo s ites will b ecome apparent i n t he f ollowing s ections.

The

Reinterpretation

o f

t he

Evidence

f rom Early

Excavations

The e vidence f rom s uch ' type' s ites a s B rough-on-Humber, Caerwent a nd S ilchester c omes f rom e xcavations c onducted twenty o r more years a go; a t Verulamium, t he l atest p rogramme o f e xcavations f inished e ighteen y ears a go; a nd, a part f rom S ilchester, t here h as b een l ittle o r n o i nvestigation s ince. At Caerwent a nd S ilchester, t here h as b een s ome r einterp retation of t he evidence ( Hartley 1 983). At t he f ormer, a r evised date f or t he r ampart a nd l ater w all w as b ased on r ee xamination of t he pottery a nd a questioning of t he o riginal i nterpretation of t he c onstruction o f t he r ampart. T he r ampart f rom 1 925.

was h is H e

a ttributed t o t he p eriod AD 7 5-100 by Nash-Williams t wo s ections a cross t he s outhern d efences i n 1 923based

h is

dating

o n

2 2

pottery derived

f rom h is

primary

bank a nd, on t he e vidence o f pottery o f c . AD 7 5-200 a ssociated w ith w hat h e i dentified a s t he cutting b ack a nd h eightening o f t he r ampart when t he w all was b uilt, h e c onc luded that t he wall was built n ot before t he l ate 2 nd c entury ( Nash-Williams 1 930, 8 0, 2 68-70, 2 74 & 2 76, f ig. 1 1). T he date of AD 7 5-100 f or t he i nitial earth r ampart was a ccepted by C raster ( 1954), i n h is e xcavations on t he e astern d efences i n 1 949-1950, i n w hich he i dentified t hree s uccessive h eightenings during t he 2 nd c entury o f t his p rimary b ank. The h eightenings were r epresented by a lternating l ayers o f s and a nd c lay c ontaining p ottery o f t he e arly 2 nd c entury t o t he e arly Antonine p eriod a nd were f ormed, he suggested, by t he c learance o f s ilt f rom t he i st c entury d itch. Hartley ( 1983) n otes that one s herd, S 96, f rom t he p rimary bank i n NashWilliams' e xcavations on t he s outhern c ircuit c ould n ot b e earlier t han AD 1 30 and might be Antonine. F rere has c ommented t hat t he bank i n t his part o f t he c ircuit c ontained pottery down t o t he end o f t he 2 nd c entury and t hat t he s ection has b een m isinterpreted ( Frere 1 978, 2 99 n10). S imilarly w ith t he date of t he wall at Caerwent, C raster ( 1954,57) s uggested t hat t he p ottery f rom t he s and w hich w as u sed to f ill t he gap b etween t he wall and t he p re-existing r ampart was c onsistent w ith t he wall h aving b een c onstructed s ometime a fter c . AD 1 80. Nash-Williams ( 1930,276) h ad a scribed t he p ottery f rom t he s econdary r ampart on t he s outhern c ircuit t o t he period c . AD 7 5-200. H artley, h owever, r e-examined t his material a nd f ound p ottery o f a t ype d ated AD 2 40-320 a s weil a s other 3 rd c entury pottery. H e t herefore p roposed a r evised date o f a fter AD 2 40 f or t he wall ( Hartley 1 966,

5 7n39)

I n v iew o f t hese r einterpretations o f t he c rucial dating evidence f rom t he Caerwent defences and t he s tructural s equences o f t he r ampart, i t must b e a ccepted t hat t he e vidence p resently a vailable i s i nadequate f or a p roper u nders tanding o f t his c ircuit. Nor c an t he s ubsequent r einterpretation o f t he e vidence be a ccepted a s p roviding a s ecure b asis w ithout f urther p roof s ince t he c ontexts o f t he r edated p ottery n eed t o b e f irmly e stablished. I ndeed, i t i s d ifficult t o s ee how a r einterpretation of C raster's ( 1954, f igs. 1 4 & 1 5) p ublished s ection a cross t he e astern d efences i s p ossible. I t s eems l ikely t hat t he e xcavator f ailed t o s ee d istinctions w hich must h ave e xisted i n t he s and a nd c lay l ayers behind t he wall. Only f urther work on t he defences can r esolve

t he

d ifficulties.

Casey's ( 1983) r ecent r e-examination o f t he e vidence f rom Nash-William's e xcavation of t he s outh-east bastion h as p resented t he p ossibility o f a n e ven l ater date f or t he c onstruction o f t he walled c ircuit i n t he f orm of a t erminus p ost quem o f AD 3 30-335. The d ating o f t he r ampart a nd t he wall at Caerwent have b een i mportant i n t he g eneral f ramework f or R omano-British defences b ut t his n ew i nterpretation p laces t he walled c ircuit very much l ater t han the 3 rd century. The and

t his

d ating a llows

e vidence t he

h as

walled

a lso

b een

c ircuit

2 3

r evised

h ere

t o

a t

f all

S ilchester w ithin

t he

' normal'

p attern

o f

development.

A d ate

a fter

AD

1 80

for

t he

c onstruction of t he wall was s uggested i n t he excavations o f 1 938-1939 a long t he n orthern p art o f t he c ircuit. T his w as b ased on samian f rom t he wall t rench a nd t he secondary b ank w hich a ccompanied t he c onstruction o f t he w all, a lthough t he c oarse p ottery w as ' consistent w ith t hat i n u se a t t he t urn o f t he 2 nd a nd 3 rd c enturies'. A p rovisional date o f A D 1 90-210 w as considered w ith a n emphasis o n t he e arly 3 rd century h orizon ( Cotton 1 947, 1 30, 1 32). B ut a gain a re-examination o f t he p ottery l ed Hartley t o s uggest t hat t he w all w as u nl ikely t o be earlier t han t he m id 3 rd c entury ( Hartley 1 966, 5 7 & n 39). More r ecent work a t S ilchester h as p roduced material f rom t he s outhern defences i ndicating a p robable c onstruction date f or t he wall i n t he m id-late 3 rd c entury ( Britannia ( 1976) 7 , 3 68) Although t he dating o f R omano-British p ottery h as b een r efined s ince t hese e xcavations, i t was at l east p artly t he d esire f or u niformity w hich l ed t o a r e-examination of t he pottery f rom the early e xcavations on t he defences a t Caerwent a nd S ilchester, w hose d ating w as o therwise a t variance w ith t he chronological f ramework f or R omano-British defences. I n c haracteristic f ashion, t he d ate s uggested f or t he w all i n t he early e xcavations at A ldborough ( soon a fter AD 1 50) was t oo e arly t o f it i n w ith o ther s ites a nd, a s w ith C aerwent a nd S ilchester, i t was s uggested t hat t he e xplanation f or t his c ould

p robably

b e

f ound

i n

t he

d ating

o f

t he

pottery

( Frere

1 978, 2 86 & n 15) . But t here s eems t o be v ery l ittle pottery t o s upport t he 3 rd c entury date n ow f avoured on t he e vidence of a worn c oin f rom one o f t he early e xcavations ( Charlesworth . However, a r edating o f t he material f rom 1 971, 1 60-162 ) t hese s ites only r einforces t he r eservations about t he r esults o f e arly e xcavations a nd t he f eeling t hat f urther work i s n eeded t o c larify t he i nterpretation o f t he basic d efensive s equence. I n many c ases, t he c ontext o f t he d ating m aterial i s r ecorded i n i nsufficient detail and, i n a ssuming t hat t he r ampart was n ot a s eparate e ntity wall, e xcavators often f ailed p ottery f rom t he p rimary b ank a nd r ampart h eightening.

f rom

At Verulamium, t he 1 955-1961 by F rere

b ut s imply a backing f or t he t o d istinguish b etween t he t hat f rom t he wall t rench o r

p rogramme o f i ncluded some

e xcavations c onducted i nvestigation of t he

d efences. The e xcavations a nd magnetometer s urveys a dded t o a nd r evised Wheeler's r esults of t he 1 930s and f orm t he basis o f our p resent u nderstanding o f t he d efensive chronology a t Verulamium ( Wheeler 1 936, 4 9-63; F rere 1 956, 1 962). N o further e xcavation h as t aken p lace on t he d efences h ere s ince t he p rogramme was completed i n 1 961 and t he r esults of t renching a cross t he l ine o f t he t wo p re-wall e arthwork c ircuits ( the ' 1955' d itch and F osse e arthwork) a nd t he walled c ircuit, w hich t his p rogramme i nvolved, l eft m uch s till open t o i nterp retation ( Hartley 1 983). The s ame i s t rue o f B rough-on-Humber w here Wacher c onducted e xcavations f rom 1 958-1961 a cross t he n orthern a nd western d efences t ogether w ith a c ertain amount o f l imited i nvestigation

t o

establish

t he

2 4

c iruit on

t he

s outh.

These

e xcavations b uilt u pon t he e arlier ones b y Corder i n t he 1 930s on t he eastern defences ( Corder 1 942; Wacher 1 969). Here a gain t he b asic s equence i s t hought t o h ave b een f airly w ell established as a r esult o f t his work and n othing more has b een done

s ince.

B itterne i s a nother c entre w here e xcavations i n 1 951, 1 954 and 1 959 have s ubsequently been c hallenged i n t erms o f t he dating a nd s tructural s equence a nd w here t he a bsence o f any f urther work has l eft unresolved t he questions o f t he dating, development a nd t he purpose o r c ontext o f t he d efences ( Cotton and Gathercole 1 958, 6 -11, 3-43) . 3 The c onclusions d rawn were t entative s ince t he d ating e vidence w as s carce b ut i t w as s uggested t hat t he I nner F osse r epresented a p alisade o r s tockade and d itch defence of t he Antonine period and t he walled c ircuit was e rected t o e nclose a s maller a rea s ometime i n t he l ate 4 th c entury ( Cotton and G athercole 1 958, 4 0-1, 4 3). Wacher h as o ffered a r einterpretation o f t he e vidence which would b ring B itterne more i nto l ine with t he ' normal' pattern of d evelopment ( Wacher 1 964, 1 05). The d ifficulties o f i nterpreting t he defensive s equence a re n ot l essened by t he f act t hat t he c haracter o f t he a ssociated s ettlement h as n ot been e stablished. At Caistor-by-Norwich, t here h as b een n o i nvestigation o f t he defences s ince t he work of Atkinson i n t he 1 930s on t he n orthern a nd s outhern p arts o f t he c ircuit w here h e c oncluded that the r ampart and t he wall were c ontemporary ( Atkinson 1 931, 2 32; 1 935, 2 13). I t w as n ot u nusual i n e arly e xcavations f or t he bank and t he wall t o be r egarded a s c ontemporary. Atkinson's e xcavations p roduced n o d ating e vidence f or t he wall a nd were unsatisfactory i n t erms of e lucidating t he q uestion o f a p re-wall c ircuit w hich i s s uggested by t he appearance on a erial photographs o f two p arallel e ast/west d itches s outh o f t he walled c ircuit ( St. Joseph 1 966, 2 4, p 1. II; Wacher 1 966, 12 and 1 13; 1 974, I 1 2 34). D r. A . S . E smonde C leary h as s uggested t hat t hese may be m odern f ield boundaries ( pers. comm., 1 980). G reat Casterton, A ldborough, E xeter a nd C olchester a re f urther e xamples w here t he r esults o f e arlier w ork h ave b een questioned. At G reat Casterton, t he i nformation comes f rom a s eries of e xcavations i n t he 1 950s a nd t his i s a nother c ase where the e xcavator f ound t he wall and bank t o h e contemporary. P ottery f rom b eneath a nd w ithin t he r ampart a nd f rom beneath t he f ootings of t he wall gave a t erminus post quem o f c . AD 1 50-160 f or i ts c onstruction ( Hartley 1 983). There has b een n o s ubsequent work on t he defences but t here i s d isagreement o ver t he i nterpretation o f t he p ublished s ection d rawing. T odd ( 1973) s uggests t hat t here a re i ndications t hat t he w all was a dded t o a p re-existing e arthwork ( see a lso Hartley 1 983) . F urther doubts about t he G reat Casterton s equence a rose f rom t he l ate 2 nd o r e arly 3 rd c entury date p roposed f or t he w all, t his date b eing r ather e arlier t han t hat normally e xpected f or t he e rection o f a walled c ircuit, a t l east f or a ' minor' t own s uch a s t his. The n eed f or f urther i nvestigation t o r esolve these p roblems of dating a nd s tructural

development

h as

a lready

2 5

b een

s tressed

( Hartley

1 966,

5 7;

Todd

1 973,

3 8).

At A ldborough, e xcavations i n 1 924 a nd 1 934-1938 f ound t he wall and bank t o be c ontemporary a nd t o h ave b een c onstructed s oon a fter AD 1 50 ( Charlesworth 1 971). A f urther s ection i n 1 960 was n ot i nconsistent w ith t his v iew of t he s tructural s equence. I t was o nly i n t wo s ections a cross t he eastern defences i n 1 965 t hat a f eature which could b e i nterp reted a s a p re-existing r ampart w as f ound. T he earlier e xcavations were r e-examined a nd i t w as s uggested t hat n one o f t hem p recluded t he e xistence o f a n e arlier bank a nd t hat i t h ad n ot b een o bvious i n t he p revious s ections b ecause o f t he u nstable n ature of t he p rimary ( sand) bank a nd i ts c onsequent e rosion. I t s eemed p ossible t herefore t o i nclude A ldborough i n t he ' normal' pattern of development. H owever, t he observations r ecently made b y B idwell o n t he e vidence f rom A ldborough a rgue f or t he c ontemporaneity of t he w all a nd r ampart ( Bidwell 1 980, 6 0 a nd 6 2). At u ntil

E xeter, r ecently

t he as

b ank b ehind t he w all h ad b een r egarded r epresenting a n i ndependent e arthwork

c ircuit. R e-examination o f t he e vidence f rom p revious e xcavations l ed B idwell t o t he c onclusion t hat the i nitial l ow gently s loping b ank may r epresent m erely upcast f rom t he i nitial s tages of t he construction o f t he wall a nd d itch. ' The s light d imensions o f t he b ank a nd t he a bsence o f a t urf l ine on t he s urface s hift t he burden o f p roof onto t hose w ho would c ontend t hat t he b ank c onstituted a n i ndependent p hase of defences' ( Bidwell 1 980, 6 0 and 6 2) . B idwell a lso c asts doubts o n t he e xistence o f i ndependent e arthwork p hases a t B ath, K enchester, Chichester and p ossibly Caerwent on t he g rounds t hat t he b anks s tood n o h igher t han 6 f eet ( 2m.). At

Colchester,

s even

s ections

w ere

c ut

at

various

t imes

a cross t he defences i n t he y ears u p t o 1 951. B ut t he e vidence was n ot u seful, by i tself, i n t erms o f c onstructing a c oherent p icture of t he defensive s equence. T he l atest s ection a cross t he wall a nd b ank c lose t o t he n orth-west c orner i n 1 951 w as t he most comprehensive. Here t he wall and bank were f ound t o b e c ontemporary a nd t here was n o e vidence f or a p re-existing earthwork ( Hull 1 958, 2 5-7) . While H ull could n ot d etect a ny c lear e vidence f or a n e arlier b ank i n a ny o f t he o ther s ections, Wacher f elt t hat i t was possible t o a rgue t he c ase f or a p re-wall r ampart i n t wo o f t hese s ections ( sections V I & V II on t he east and s outh n ear t he s outh-east corner) where Hull h ad n oted a c lay b ank b ut c ould o ffer n o e xplanation f or i t ( Hull 1 958, 4 6-63, 5 4-5; Wacher 1 964, 1 02) . I n a ddition, Wacher q uestioned t he dating o f t he w all w hich w as a gain earlier t han t hat a ccepted f or walled c ircuits e lsewhere. I n s ummarising t he r esults o f t hese e xcavations, Hull r egarded t he wall as Antonine on t he b asis o f t he p reponderance o f Antonine p ottery i n t he b ank a nd, o n t he e vidence f rom s ection V II, a date n ot earlier t han AD 1 40 w as p roposed f or t he w all a t t his s outhern p oint. However, Wacher n oted t hat a l ater date was i ndicated by s ome of t he published pottery f or t his s ection ( Hull 1 958, 5 4; Wacher 1 964, 1 06 t he bank a long t he western c ircuit p roduced o f

C .

AD

1 50

a nd

t he

l atest

2 6

p ottery

f rom

n 2). s ome t he

F urthermore, l ater pottery

f inal

s tage

o f

t he b ank here was o f a t ype d ated AD 1 40-170 ( section l A, H ull 1 958, 2 7) . This enabled a r ather l ater date t o b e p roposed f or t he walled c ircuit. But t he e vidence w as s till i nsufficient t o e nable Colchester t o f it i n w ith a ' normal' p attern of a ' late 2 nd c entury' e arthwork w hich w as r eplaced o r r einforced by a walled c ircuit s ometime i n t he 3 rd c entury. U nlike

t he

o ther

s ites

mentioned

a bove,

t he

d efences

a t

Colchester have b een subject t o more r ecent e xcavations and, despite attempts t o r einterpret t he e vidence f rom t he e arly excavations i n a way which would make i t conform t o a ' normal' pattern, a more r ecent s tudy h as s hown t hat t he s equence i s both u nusual and complex. This s tudy i s b ased on f urther s ections a cross t he w estern d efences a t S t. Mary's R ectory i n 1 967 and on t he s outhern defences a t L ion Walk i n 1 971-1973. In t hese t wo s ections, t he wall h as b een d ated t o t he e arly 2nd c entury and t he bank has been s hown t o be an a ddition o f t he l ater 2 nd o r e arly 3 rd c entury t o a n o riginal f reestanding wall ( Crummy 1 977, 9 1-2, 9 8) . This p rovides a v ery early date f or t he wall a nd o ne w hich i s s o f ar u nparalleled e xcept a t t he o ther two coloniae ( Gloucester and L incoln) where walled c ircuits were b eing i nserted i nto t he f ront o f t he l egionary f ortress r amparts a t about t his t ime ( Crummy 1 977, 1 02). Despite t his e arly date f or t he w all, e xcavations i n t he 1 960s and 1 970s i n other parts o f t he colonia have s uggested t he e xistence o f h itherto u nsuspected d efensive c ircuits w ith t he construction of t he wall r epresenting a s tage of development f ollowing a l ong s equence o f defensive s ystems o n d ifferent a lignments. The d efensive h istory o f Colchester i s p roving t o be m ore c omplicated t han t hat o f a w all r eplacing a n e arlier rampart on t he s ame c ircuit. The h istory of t he wall may, i n i tself, p rove t o b e a c omplex matter f or t here i s n ow t he problem of the l ack of c orrespondence b etween t he s ections north and s outh o f t he B alkerne G ate on t he w estern c ircuit. The 1 967 s ection s outh o f t he Balkerne Gate ( St. Mary's Rectory) s howed t he wall t o h ave b een e rected h ere a s a f reestanding structure i n t he early 2 nd c entury and w ith t his t he 1 971-1973 s ection a long t he s outhern c ircuit ( Lion Walk) agreed, but t he 1 951 s ection n orth o f t he Balkerne Gate f ound t he c ondition o f t he r ear f ace o f t he wall c onsistent w ith i t never having been e xposed t o t he weather. I t must t herefore have b een p rotected f rom t he f irst. The w all a nd b ank w ere contemporary and t he material f rom t he bank p rovided a t erminus post q uem i n t he l ate 2 nd c entury f or t he w all ( Hull 1958, 2 5 & 2 7; s ee a lso C rummy 1 977, 9 2 & 1 02). S o t hat, a s well a s t he d ifference i n s tructural s equence b etween t hese two s ections, t here i s a divergence i n t he dating. These p resent v ery r eal p roblems o f i nterpretation f or t he development of t he c ircuit a s a whole. One s uggestion i s t hat the f reestanding w all s outh o f t he Balkerne Gate a nd i n L ion Walk n ever e nclosed t he w hole c olonia b ut was p art o f a n unfinished f ortification, w ith much of t he r est o f t he walled circuit b eing a dded i n t he 3 rd c entury a nd p ossibly e ven a s 1 late as t he l ate 3 rd c entury ( Wacher 1 974, 14); a lternatively, F rere h as s uggested t hat t he f reestanding w all could represent a n early p rogramme f or defence which was i nterrupted

a nd

( Frere

2 88)

1 978,

c ompleted

o nly

2 7

i n

t he

c risis

o f

AD

1 93-196

I t s eems p ossible t hat t he d ifferent dates a nd s tructural s equences r eflect e ither t he l engthy i mplementation o f a s ingle p rogramme a nd/or t he s ubsequent modification a nd r epair of d ifferent parts o f t he c ircuit a t d ifferent t imes. I t s hould b e r emembered t hat t he more r ecent i nterpretation o f t he main defensive s equence o ffered b y C rummy ( 1977) is b ased o nly o n t wo s ections a t w idely s paced p oints a long t he c ircuit and more work i s needed t o c onfirm o r modify these r esults f or t he c ircuit a s a w hole.

Methods

o f

The c ircuits

Excavation

quality a nd e ven

of f rom

and

t he Quality o f

t he Evidence

t he e vidence f rom d ifferent defensive d ifferent p oints a long t he s ame c ircuit

varies c onsiderably. I n Verulamium a nd Wroxeter,

s ome c ases, t he w all h as

a s a t B rough-on-Humber, b een a lmost c ompletely

r obbed f or most o r a ll of i ts c ircuit, while at C aerwent a nd S ilchester, t he wall a nd i ts r ampart s till s tand a c onsiderable h eight. At L incoln a nd t o a l esser d egree C irencester, t he p reservation o f t he w all i s very good f or s ubstantial s tretches and t he defences can b e t raced for m uch o f t he c ircuit by e arthworks a t A ichester, Caistor-by-Norwich and K enchester. I n o ther c ases where t he s ite i s d eserted b ut t he d efences h ave b een l argely l evelled, t he c ircuit h as o ften b een e stablished by a erial photography and there a re m any e xamples w here t his method h as made a c onsiderable c ontribution t o t he s tudy o f R omano-British defences ( St. Joseph 1 966, 2 1-30; Wilson 1 975, 9 -49). I n s ome c ases, f or e xample Bath, Carmarthen a nd p ossibly G odmar ichester, i t h as b een suggested t hat t he outline o f t he defended a rea may b e p reserved i n t he modern S treet p lan.

t o

Where t he whole c ircuit can b e t raced, i n whatever f orm, t his h as a d irect b earing o n t he i nterpretation o f t he s tructural s equence. There i s a t endency f or the ' type' s ites o f Caerwent, C irencester, Colchester, Silchester a nd Verulamium t o f all i nto t his c ategory, and here t he evidence f rom a f ew t renches i s s omehow g iven g reater weight by t he f act t hat t he whole c ircuit c an b e t raced as a n e xtant s tructure o r a c ropmark. I n s ome c ases, n otably S ilchester where t he wall s eems t o b e u niform a nd o f one build, this m ay b e j ustified. The s urvival o f t he c ircuit p rovides a ' context' o r f ramework f or t he r esults f rom a n umber o f t renches a cross i ts a lignment. This perhaps e xplains t he c oncern t o e stablish t he c ircuit i n t hose c ases w here i t i s n o l onger v isible. Survival does o f c ourse materially a ffect t he q uality a nd r ecovery of t he e vidence and t he a bility t o e stablish t he s tructural s equence. Defences a re m ost vulnerable i n t his r espect. The l evelling o f t he bank p resents s erious p roblems, p articularly w ith dating, s uch a s a ffects t he i nner a nd O uter E arthwork c ircuits a t S ilchester, s ince t he material f rom t he d itch

i s

l ess

t han

a dequate

f or

2 8

t he

d ating

o f

t he

defences.

Robbing o f t he w all c an l ead t o t he l oss o f much o f t he e vidence r elating t o i ts c onstruction, s uch a s a t C irencester, where the s ection cut i n 1 952 i n t he s outh-east corner o f t he c ircuit c ould n ot b e a dequately i nterpreted b ecause o f t he damage caused by t he e xtensive r obbing of t he wall; but i t b ecame i ntelligible i n t he l ight o f t he more c omplete r emains e xcavated i n 1 960 a long t he n orth-east c ircuit ( Wacher 1 961, 4 1, 6 3-4). With r egard t o t he s urvival o f d eposits, t he v ery construction of t he wall c ould have r esulted i n t he destruction o f much o f t he e vidence r elating t o t he c ondition of t he rampart at t he t ime of t he i nsertion of t he wall a nd e ven, with i ts a ssociated d itches, c ould h ave l ed t o t he obliteration of e arlier l ess s ubstantial d efensive c ircuits ( see e .g. Jarrett 1 965, 5 8-9; C rummy 1 977, 9 2 & 1 02). I n all c ases where defences have been e xamined, i ncluding t he ' type' s ites, e xcavation h as b een v ery l imited, t he evidence c onsisting of t he r esults f rom a n umber of t renches a cross t he defences r ather t han a ny s ystematic e xamination o f i ndividual c ircuits. The t renches t hemselves a re o ften r estricted t o a s hort l ength o f t he w all a nd/or r ampart. A complete s ection a cross t he wail, r ampart and t he a ssociated d itch system i s r arely t hought n ecessary a nd i n many c ases i s n ot p ossible. E ven w here a f air n umber o f s ections h ave b een e xcavated i n a single c ircuit, a s at C irencester a nd L incoln, t hese r epresent o nly a s mall s ample o f t he w hole. I n s ome c ases t he number of s ections i s f ew, amounting t o t wo o r t hree t renches s upported o n o ccasions by observations during r edevelopment at o ther points a long t he c ircuit. Godmanchester i s one s uch e xample. Four s ections h ave b een c ut: a t t he s outh g ate ( 1959-1961) ‚ t he n orth gate ( 1972-1973) ‚ t wo s ections a long t he w estern c ircuit a nd o bservations a t t he n orth-west c orner, a ll i n the 1 950s ( Green 1 975, 2 04) . A v ery s ubstantial part o f t he c ircuit c onsisting o f t he e ast a nd n orth-east a lignments has yet t o be e xamined and t hese a re a ssumed on t he b asis of t he p roperty b oundaries o n t his s ide o f t he t own. F rom t he e vidence of t hese f our s ections, t he development o f t he d efences i s c onfidently p resented a s r easonably w ell understood and t he c ircuit o f t he wall well-established ( Green 1 975, 2 04). F or o ther e xamples s uch a s l ichester, Margidunum o r T owcester where i nvestigation has b een v ery l imited i ndeed, s ee

t he

Gazetteer.

A t S ilchester, t he development o f t he d efensive c ircuits has b een c onfidently d iscussed i n a way which belies t he f act t hat t he e xamination o f t hese c ircuits c onsisted o f a v ery l imited n umber o f s ections. The evidence was i n f act i nconc lusive a nd t his c auses s ome d isquiet a bout t he i nterpretation of t he Silchester defences which became generally a ccepted ( Boon 1 974, 4-6, 6 5-70; Wacher 1 974, 2 56-61, 2 64-6; F rere 4 1 978, 2 83-6) . F ortunately, r ecent work by M . F ulford on t he Outer Earthwork a nd t he I nner D efence h as emphasised t he l imitations of t his e vidence ( Fulford 1 983) . The s ame l imited i nvestigation f orms t he b asis f or t he w ell-established i nterp retation of ( Wheeler 1 936,

t he various defensive 4 9-63; F rere 1 964a,

2 9

c ircuits a t 6 1-7; 1 964b,

Verulamium 1 04, 1 08-9;

1 978,

2 83,

2 86-7;

Wacher

1 974,

2 04-5,

2 13-15).

I n virtually e very case, t he i nformation a vailable f or t he s tudy o f a defensive c ircuit i s c umulative evidence f rom a number of i ndividual s ections, often at widely s eparated p oints. This i s u nsatisfactory s ince d ifferent s ections o f t he s ame c ircuit i nvariably p roduce varied e vidence, which may e ven c ontradict t hat f rom o ther s ections. N one o f t he s ections a lone i s c apable o f p resenting t he w hole p icture a nd, e ven when t he e vidence f rom a number i s t aken, t he r esults r arely b ear a ny r elationship t o t he p roblems a nd c omplexity o f t he c ircuit a s a whole. An e xample o f t he dangers o f u sing t he e vidence f rom o ne s ection as t he basis f or e stablishing t he h istory o f t he w hole c ircuit c omes f rom L incoln l ower c olonia. E xcavations o n t he western c ircuit i n 1 971 ( Motherby H ill/West Parade) s howed t hat t he wall a nd b ank i n t his part o f t he c ircuit w ere o f o ne b uild and e rected i n t he l ate 2 nd o r e arly 3 rd century ( Colyer 1 975, 2 32, 2 34, 2 48-50 & f ig 8 ). H owever, a s ection i n 1 973 o n t he eastern part o f t he c ircuit ( Broadgate) demonstrated t hat t he wall i n t his a rea h ad b een i nserted i nto a p re-existing b ank w hich t he e xcavator p laced i n t he l ater 2 nd c entury ( Britannia ( 1974) 5 , 4 21-2 & f ig. 8 ; Wacher 1 974, 1 32) . The e vidence f rom e ach s ection t aken on i ts own o ffers a d ifferent s equence. A lthough t he c urrent method o f e xamination o f a defensive c ircuit, c onsisting of a n umber of t renches, h as p resented p roblems o f i nterpretation i t i s s till r egarded a s t he n ormal means of dealing w ith defences. This i s partly b ecause o f t he d ifficulties o f e xcavating l ong s tretches o f a f eature o f t his type, and partly b ecause of t he f ailure t o appreciate t heir c omplexity, a nd a lso b ecause o f t he v ery l imited a ims, n amely defining t he basic s tructural s equence, dating t he major f eatures a nd e stablishing t he a lignment. Obviously i ndividual t renches d o h ave a value b ut t hey n eed t o b e r egularly r epeated and t o be s een i n t he context o f what could b e achieved by i nvestigation on a l arger s cale. At v irtually a ll s ites t he case can be made f or r egarding t he i nvestigation which h as t aken p lace n ot a s a c omprehensive s tudy b ut a s a p reliminary t o a p rogramme o f f urther work w ith much b roader a ims a nd u sing d ifferent methods o f e xcavation f rom t hat o f s imply cutting narrow t renches a t w idely s eparated i ntervals. Only a t L incoln l ower c olonia ( The Park 1 968, 1 970-1972 and West P arade/Motherby H ill 1 971; Ant. J . ( 1975) 5 5, 2 27-59; Colyer 1 975, 1 5-22). a nd C irencester ( north-east d efences) h as t here b een a ny a ttempt t o b reak away f rom t he t renching method. I n e ach c ase t he r esults s howed h ow much m ore c ould be l earnt by t he e xposure of l ong s tretches of t he c ircuit. C ircencester h as o therwise f ollowed t he u sual pattern o f i nv estigation w ith r estricted s ections at widely s paced i ntervals a nd n o i nvestigation o f t he d itch s ystem. T he excavation of a 1 60 f eet ( 49m ) s tretch t ogether w ith part o f t he d itch s ystem n orth o f t he Verulamium Gate ( 1960, 1 9651 968) n ot i ndividual

only made s ections

s ense o f s ome of t he d iscrepancies i n t he by d efining more c learly t he i ndividual

3 0

phases but a lso r evealed a c omplexity o f development w hich h ad not b een shown i n t he narrow t renches ( Wacher 1 970) . I n t his s tretch, t he w idth o f t he wall c hanged f our t imes, a nd f ive major construction phases f or t he d efences were i dentified. At l east t wo o f t hese w ere i n t hemselves c omplex a nd c ould n ot be r egarded as s ingle c lear-cut s tages of development but were perhaps carried o ut i n a p iecemeal f ashion o ver a l onger period of t ime t han had h itherto been s upposed ( Wacher 1 967, 1 88-91; Wacher 1 970, 27-8, 2 39). The e xamination o f a s ub2 stantial s tretch was t he only way t o appreciate t he complexity o f t he defences a nd t o l ocate t he c rucial p oints w here c hanges occurred i n t he c haracter o f t he wall. E ven s o, t his e xcavation d id n ot p rovide a ll t he a nswers e ven f or t his particular s tretch o f t he defences. There was n o dating evidence f or a ny o f t he p hases a nd t he r elationship b etween some of the major f eatures, namely t he outer d itch, various t ypes of w all a nd b astions, h as s till t o b e e stablished ( Wacher 1 970, 2 31 & 2 39) . Certainly f urther work i s n eeded b efore t he f ive major p hases o f c onstruction c an b e s hown t o apply to a l arger a rea o r t o t he c ircuit as a whole. At L incoln, a c omplicated s equence o f development a nd t he differential t reatment a ccorded t o a djacent s ections o f t he defences was a lso d emonstrated i n t he e xcavations on t he western defences o f the l ower c olonia. At West P arade a lmost 1 80 f eet ( 54m ) o f t he d efences w ere e xcavated b ut t he a ctual section a cross t he defences was r estricted by modern buildings, s o t hat t he e xcavation o f t he d itch s ystem w as n ot possible n or c ould a c omplete s ection b e o btained t hrough t he rampart ( Ant. J . ( 1975) 5 5, 2 32 & 2 47). H owever, u nusual c ircumstances a t T he P ark made p ossible t he e xcavation o f a t otal width of t he defences f rom t he outer l ip o f t he d itch t o t he t ail o f t he r ampart ( Ant. J . ( 1975) 5 5, 28-43; Colyer 2 1 975, 1 8) . I n t his e xcavation, a l ength of t he defences o f c . 1 85 f eet ( 56m ) w as e xposed a nd i t was d iscovered t hat t he o riginal colonia wall and i ts contemporary rampart were e rected i n t he l ate 2 nd o r e arly 3 rd c entury. Subsequently, a stone i nterval t ower was a dded t o t he i nternal f ace o f t he wall p robably s ome t ime i n t he f irst h alf o f t he 3 rd c entury; this was demolished f or t he construction of a gateway which was t hen r ebuilt i n a monumental s tyle s ome t ime i n o r a fter t he m id 4 th c entury. F ollowing the r ebuilding o f t he gate, a major r emodelling o f t he d efences t ook p lace w hich i ncluded the d ismantling o f t he o riginal colonia wall ( 5 f eet/1.5m w ide) north o f t he gateway a nd i ts r eplacement by a w ider w all ( 10 f eet/3m),while s outh o f t he gate t he o riginal wall h ad b een r etained a nd was t hickened by t he c onstruction o f a n additional wall t o t he r ear. This c ontinued a long a s tretch e xtending 4 9 f eet ( 14.9m ) s outh o f t he gate, a t w hich p oint t he o riginal wall h ad a gain b een demolished and r eplaced by a w ider wall. Major work was d one o n t he r ampart a t t his t ime and t here were i ndications t hat t he d itch was a lso r ecut at s ome s tage i n t he l ate R oman p eriod. The r emodelling m ay h ave occurred

i n

As a t where the understand

o r

a fter

t he mid

4 th

c entury.

C irencester, t he i nvestigation character o f t he wall changed t he o verall d efensive s equence

3 1

o f t he c rucial a reas made i t possible t o a nd t his e stablished

t he c ontext i n c ould be better

w hich t he more r estricted s ections e lsewhere understood ( Colyer 1 975, 1 4, 1 8 & 2 2).

I n a dvocating g reater a ttention t o t he d etailed s tudy o f i ndividual c ircuits, t he opening up o f l arger a reas of t he d efences i s e ssential. Even a s tretch o f 1 60 f eet a t C ircencester, which i s n ot a ctually v ery l arge in t erms of t he t otal c ircuit o f c . 2 m iles ( 4 km ) h ad a much g reater c ontrib ution t o make t o t he understanding o f t he defences t han t he multiplication f or a ssessing s ections.

The

o f small s ections. I t a lso t he r esults a nd c onclusions

Dating Evidence:

I ts Nature

a nd

p rovided a context f rom t hese o ther

P roblems

I t i s often i mpossible t o date R oman defences a ccurately b ut, b ecause o f t he emphasis p laced u pon t he i dea t hat t he appearance of defences was t he r esult of t he ' application o f a s ingle p olicy i n a s ingle c ontext', t he dating o f RomanoB ritish defences has b ecome a major p re-occupation ( Frere 1 978, 2 84-5). C order ( 1955) a imed t o e stablish t he date o f s tone defences and t he a ddition of bastions a nd Wacher ( 1964; 1 966) p resented t wo s uccessive p apers i n w hich h e a ttempted t o i dentify a nd e stablish a c ommon date a nd c ontext f or t imber and e arth defences. I nitially, h e concluded t hat t he majority o f e arthen r amparts w ere p robably b uilt i n t he mid 2 nd c entury b ut, i n r eviewing subsequent e vidence, he decided t hat a l ate 2 nd c entury date was more l ikely. U nfortunately t he dateable m aterial f rom d ifferent defensive c ircuits p roduces widely divergent d ates a nd e vidence i s r arely c apable o f p roviding a p recise d ate f or a particular phase of construction. This i s t rue of d efences i n a ny p eriod, b ut f or t he medieval p eriod, t o w hich t he R oman period may be compared, i t i s at l east i n t heory m uch easier t o t ie i n t he a rchaeological e vidence f or t he phases o f c ons truction and r econstruction w ith t he dating provided f or t hese e vents by h istorical s ources. S cholars of t he R oman period have t ried t o r emedy t he l ack of c ontemporary documentation by t rying t o f ind h istorically dateable e vents t o which t he construction of d efences might b e r elated. H aving apparently f ound, f or e xample, s uch a c ontext i n A lbinus' a ttempt t o c laim t he empire i n AD 193-196, t he r eluctance t o abandon t he d ate t hus p rovided i s a ll t he m ore k een ( Frere 1 978, 2 85) ' Since u rban defences a re n ot r andom p henomena b ut w ere s ubject t o c entral decision, t hey f orm a c lass i n w hich t he date o f a ny o ne w ill a ssist t he dating o f t he others' ( Frere 1 965, 1 38). Having established t he b asic p remise t hat t he g reat majority o f earthwork defences f orm a s ingle chronological group, t he l atest d ate p rovided by a ny o ne member o f t he c lass w as t hen u sed t o date t he g roup a s a whole. This has r esulted, f or e xample, i n t he d isplacement o f a d ate f or e arthworks i n t he m id 2 nd c entury by o ne i n t he

3 2

l ate 2 nd c entury a nd, more r ecently, back a gain ( Hartley 1 983), as more e vidence f rom other c ircuits was b rought i nto t he p icture. While t he p rovision o f a t imber a nd e arth r ampart should be an event o ccurring more o r l ess contemporaneously a t a ll t he s ettlements i nvolved, t he d ating evidence f rom i ndividual members o f t he g roup of i ndependent earthworks i s i nadequate f or t he p urposes o f p roving contemporaneity, which c an only be a sserted as a l ikely possibility. I n a ddition, while t he i nsistence t hat a s ingle historical context i s r esponsible f or t he majority o f earthwork d efences i n R oman B ritain meansthat t he date o f their construction had been p laced w ithin t he narrow b racket of A D 1 93-196, by i ts v ery n ature t he a rchaeological e vidence i s i ncapable of p roviding s uch a p recise date ( Frere 1 976, 1 94-5). The t enuous n ature o f t he dating e vidence f or defences

Timber

i s

a f actor

which

cannot be

emphasised

t oo much.

and Earth D efences

E arthwork c ircuits u sually r ely f or t heir d ating o n pottery derived f rom t he r ampart; c oin e vidence i s l ess commonly a vailable f or r ampart c ircuits. S ome o f t he p ottery cannot be p recisely dated, belonging t o a type which spans a l ong p eriod. F or e xample a t Caerwent t he l atest s herds i n t he s condary r ampart were o f a type dated AD 2 40-320 ( Hartley 1 966, 5 7n39); a t L incoln West P arade t he r elevant s herd f or dating the r ampart and contemporary wall was dated AD 1 60-190 ( Ant. J . ( 1975) 5 5, 2 50). F urther c omplications a re c aused b y regional variations i n t he l ife span o f a particular type o f pottery a nd t he u se o f p ottery o f l ocal manufacture w hich may be d ifficult t o date ( see H artley 1 966, 5 7n41 r e L incoln) C onsiderable emphasis provided by c oins a nd

i s p laced o n i n certain

t he dating e vidence c ircumstances t heir

deposition c an b e c losely dated ( Casey 1 983). B ut p roblems o f coin supply and coin l oss m ean t hat t his t ype of dating evidence i s u sually s ubject t o much t he s ame r estrictions a s pottery ( Reece 1 973, 27-51) . 2 I t i s only f rom t he t hird quarter o f t he 3 rd c entury t hat c oins a re s ufficiently c ommon and h ave a s ufficiently s hort l ife t o b e u sed f or dating; i n t he 1 st and 2 nd c enturies s amian i s more u seful ( Hartley 1 966, 5 3). The shortage of c oins f or most of t he 3 rd c entury has contributed t o t he d ifficulty o f dating t he c onstruction o f most walled c ircuits. Where c oins a re f ound i n a ssociation with o ther dateable material, t hey c an p rovide a u seful m eans of a ssessing t he e vidence of t he pottery. But t he t emptation t o p lace t oo much r eliance o n c oins, e specially w here d ating of t he pottery i s i nsecure, can p roduce m isleading r esults. The dangers mind ( Webster

o f t he u se o f 1 977, 3 21-2)

c oin

e vidence

s hould

b e

b orne

i n

F urthermore, material f rom c ertain c ontexts may b e more useful f or dating t han f rom others. The n ature o f t he c onstruction o f e arthwork d efences i s i n i tself r esponsible f or the d ifficulties o f dating. The material o f which t he r ampart

3 3

i s c omposed i s r esidual, b eing d erived o ften f rom t he u pcast of t he d itches and/or f rom t he s craping t ogether o f occupation material i n t he v icinity. Occasionally a dditional material was t ransported f rom e lsewhere, a s h as been s uggested at A ldborough ( Charlesworth 1 971, 1 56 & 1 63). The c onstruction of defences t herefore i nvolves a g reat d eal of disturbance of p re-existing l evels a nd t his w ill i nvariably r esult i n t he deposition, i n t he body of t he r ampart, of material o f an e arlier date t han t he date o f t he r ampart i tself. T his e xplains why t he r ange of pottery w ithin t he r ampart may s pan a c onsiderable p eriod. The pottery a nd o ther d ateable m aterial i n t he r ampart i s t herefore r elated t o material o f which t he r ampart i s composed r ather t han t he a ctual date o f t he r ampart's c onstruction; e ven t he l atest pottery f rom t he rampart will not n ecessarily r epresent t hat c urrently i n u se a t t he t ime of i ts construction. S o t he l atest s herd f rom t he rampart provides o nly a t erminus p ost q uem o f t he most t entative k ind. A c ase i n point i s G reat Casterton where, i n t he f irst s ection a cross t he d efences, e very s herd f rom t he b ank was o f t he i st c entury b ut a small h earth i mmediately b elow t he r ampart w as a ssociated w ith s herds s imilar t o a t ype dated AD 1 40-160 a nd two other s herds which c ould have b een i n u sed f rom c . AD 1 602 00. S ix s ubsequent s ections a cross t he r ampart p roduced n o pottery l ater t han t he early 2 nd century ( Webster 1 977, 3 21) However, t his does n ot m ean t hat dating e vidence f rom a c ont ext s ealed beneath a r ampart i s i ntrinsically more r eliable o r c loser t o t he t ime o f c onstruction t han dating evidence f rom w ithin a r ampart. At S ilchester I nner D efence a nd Dorchester-on-Thames t he material f rom t he r ampart p rovided a t erminus post quem l ater t han t hat provided by material s ealed d irectly b eneath t he r ampart ( Cotton 1 947, 1 29-30; F rere 1 965, 1 38). Much depends on t he o rigin of t he material a nd whether t he material s ealed b y t he b ank w as d eposited during construction o r a s t he r esult o f o ccupation i mmediately , p rior t o, o r s ome t ime p rior t o, t he r ampart's c onstruction. G iven t hat t he c ircuit o f t he d efences cuts a cross o r e ncompasses a reas w here t he c haracter o f t he occupation i s b ound t o vary, i t i s l ikely t hat d ifferent parts o f the c irc uit w ill b e c omposed o f d ifferent material. In s ome a reas, t he c ircuit t aken by t he defences may b e peripheral t o, a nd n ot d irectly a ssociated w ith, t he f airly good o ccupation s equences i n t he n ucleus of t he s ettlement. I n t hese i nstances, t he r ampart w ill p ossibly b e c omposed o f material f rom a reas where t here i s only s cattered o r much earlier a ctivity. The p ottery may h ave b een l ying a round i n t hese a reas a s r ubbish f or much l onger t han i n a reas where development w as more r apid. S o s ome p arts o f t he c ircuit m ay be l ess u seful t han others i n p roviding e vidence of date w hich is in '

a ny way c lose t o t he a ctual d ate o f c onstruction. Short o f e xcavating t he whole c ircuit, t here w ill a lways be t he problem t hat a f urther s ection t hrough t he d efences may y ield a t erminus post quem o f a l ater o r more s ecure date. The been

d itches

r emoved,

a re

t hemselves, e ven

l ess

e specially w here r eliable

3 4

f or

dating

t he

r ampart

purposes

h as

s ince

t he m aterial i n t he d itch f ill i s derived f rom t he e rosion o f occupation l evels a t t he l ip o f t he d itch o r f rom t he d umping o f m aterial f rom e lsewhere. This material n eed h ave v ery l ittle r elevance t o t he date a t w hich t he d itch w as d ug a nd will a ctually be d eposited at s ome t ime a fterwards.

Stone Defences

T he d ating o f s tone walls p resents a dditional p roblems. The f act t hat t here i s i n most c ases n o r eally s atisfactory dating evidence f or t he c onstruction o f w alled c ircuits h as l ed t o attempts t o e stablish s ome s ort o f c hronology b ased on a rchitectural s tyles. This i nvolves a t ypology n ot s o much o f t he w alled circuit i tself but o f s uch f eatures as gates, i nternal t owers a nd b astions w hich f orm p art o f t he c ircuit ( Wacher 1 974, 7 5-7, 14; 1 978, 9 9-100; F rere 1 978, 2 86-91). 1 Rarely does t he t ype o f wall c onstruction l end i tself t o typological analysis, a lthough i t has b een s uggested t hat t he f ree-standing w alls o f Catterick a nd G reat Chesterford r epresent a s tyle m ore i n k eeping w ith c ontemporary ( early . 4 th c entury) military a rchitecture ( Todd 1 973, 4 3; Wacher 1 974, 5 1). However, t he s ection of f ree-standing wall s outh of t he Balkerne G ate a t C olchester r epresents a much e arlier e xample of t his type of c onstruction ( Wacher 1 974, 14; C rummy 1 977, 1 9 1-2). There a re o ther c entres, e .g. Cambridge ( late 3 rd/early 4 th c entury) , M ildenhall ( terminus post quem ) a nd London r iverside w all ( AD 3 50-370) w here a 4 th c entury date has b een s uggested f or t he walled c ircuit b ut where t he wall, a lthough e rected d e n ovo a s a t Catterick a nd G reat Chesterford, i s p rovided w ith a r ampart ( see G azetteer). There i s o ne s pecific e xample, H orncastle, where T odd f eels t hat t he p resence o f a r ampart b ehind t he wall i ndicates t hat t he w all c ould b elong t o t he 3 rd c entury r ather t han a ter. But i n t his c ase b astions were p rovided w hich were of one l

build w ith t he w all a nd t he t ypological a rgument w hich r elates t o t he p rovision of b astions would p lace i t i n t he period a fter AD 3 69 ( Todd 1 973, 4 3). N earby Caistor, w here t he w all and b astions a lso appear t o b e c ontemporary, h as n o earth r ampart behind t he wall a nd t his l ack o f c onsistency makes i t unrealistic t o a ttempt a ny t ypological a ssessment. This t ype of s tudy h as o bvious d isadvantages e specially a gainst a b ackg round of c omparatively s canty i nvestigation o f s uch f eatures a s g ates, i nternal t owers a nd b astions, w hich m ay n ot i n a ny case be c ontemporary w ith t he wall a nd s o n ot d irectly r elevant t o dating i ts c onstruction. The i ntroduction o f s uch evidence underlines t he unsatisfactory n ature of t he dating evidence f rom e xcavations on w alled c ircuits. I t i s n ot u ncommon f or s ections a cross t he defences t o p roduce n o dating e vidence f or t he c onstruction o f t he w all o r other associated f eatures; f or e xample i n t he l ong s ections a cross t he defences a t C irencester a nd a t L incoln, West P arade, t he dating e vidence f or t he wall c onsisted only o f a small amount o f p ottery f rom t he c ontemporary b ank ( Britannia ( 1979) 1 , 2 31 & 2 39; Ant. J . ( 1975) 5 5, 2 50, 2 54 & 2 56).

3 5

A n umber o f t ypes o f wall c onstruction w ere i nvolved a nd t his i s r elevant more t o t he potential o f d ifferent methods of c onstruction f or d irect dating e vidence t han t o a ny typological s tudy. Where a s tone w all was i nserted i nto t he f ront o f a n e xisting e arthwork c ircuit t he f ollowing methods of c onstruction could b e u sed :1 ) a s imple i nsertion i nto t he r ampart w ith t he w all r esting on a nd possibly i n t he r ampart material. I n these cases, t he r ampart a nd d itch p rofile might h ave b ecome f airly e roded a nd t he wall could b e built a lmost a s a f ree s tanding structure, w ell i n f ront o f t he s urviving o uter p rofile of t he r ampart, a s f or e xample a t Caerwent, where t he wall was apparently b uilt o ut o ver t he d itch on t he n orth a nd east at l east ( Wacher 1 974, 3 82) a nd p ossibly a lso A ldborough ( Charlesworth 1 971, f ig. 2 1). The r esulting gap b etween t he wall and p ree xisting r ampart would b e f illed b y t he d eposition o f m ore r ampart material. 2 ) t he r ampart c ould b e c ut b ack a nd t he w all i nserted i n f ront of t he r esulting rampart f ace b ut i n a position more w ithin t he main b ody o f t he r ampart t han i n f ront o f i t, a s a t C irencester. The cutting of t he c onstruction t rench f or t he wall would a lso n ecessitate t he d eposition o f material i n t he gap b etween t he wall and r ampart. 3 ) t he r ampart c ould b e c ut b ack t o a v ertical f ace a nd t he wall b uilt up a gainst t he r ampart f ace. I n t his c ase, t here would b e n o g ap b etween wall a nd r ampart a nd t he d eposi tion o f a ny dateable deposits would be m inimal. I t i s v ery r are f or a ny r eally u seful d ating m aterial t o be d irectly a ssociated w ith t he c onstruction levels f or t he w all, e ither i n t he w all i tself o r s ealed i n t he c onstruction l evels. The dating evidence f or t he wall u sually c onsists o f t he p ottery a nd o ccasional c oins c ontained i n t he a dditional material deposited b etween t he wall a nd t he e xisting rampart. This w ill u sually b e a bsent w here t he t ype of c onstruction i nvolves ( 3) a bove. Where t he c onstruction of t he wall i nv olved t he h eightening o f t he e xisting r ampart, t his would p rovide a f urther possibility o f dating e vidence. For t he p urposes o f dating, t his i s n ot v ery h elpful s ince i t i nvolves n ot o nly t he s ame p roblems a s t hose a ssociated w ith t he c ons truction of t he p rimary rampart b ut i t a lso t he f urther c omplication o f t he i ntrusion o f r esidual material f rom t he e xisting r ampart caused by d isturbance i n t he p rocess of w all c onstruction. I n n one o f t he m ethods o f c onstruction i nvolving t he deposition of f urther r ampart material i s t he a ssociation o f t he d ate o f t his material w ith t he d ate o f t he wall n ecessarily very d irect. F or dating purposes, t he h eightening o f t he r ampart h as a lso t o b e c learly s hown t o b e contemporary w ith t he c onstruction of t he wall and p art of t he s ame p rocess r ather t han a l ater r efurbishing. Where e arthwork were u sed

t he wall was d efence, a gain

n ot i nserted i nto a n umber o f m ethods

:-

3 6

a of

p re-existing c onstruction

1 )

t he

wall

structure w ith n o G reat Chesterford

c ould

b e

r ampart. h ave b een

c onstructed

a s

I n t his c lass, i dentified.

a

f ree

s tanding

only Catterick

a nd

2 ) the wall c ould be constructed o riginally as a f ree s tanding s tructure a nd t he r ampart b uilt u p s ubsequently b eh ind t he wall. S o f ar t his t ype of c onstruction h as b een i dentified o nly a t C olchester a long t he western c ircuit s outh of t he Balkerne G ate ( St. Mary's R ectory) and a long t he s outhern c ircuit ( Lion Walk) . B ut i t does n ot s eem t o a pply t o t he whole c ircuit, at l east i n i ts f inal f orm. 3 ) t he wall c ould b e p rovided w ith a r ampart w hich w as b uilt up b ehind t he wall a s t he l atter was c onstructed. This i s t he method which was i dentified by t he e xcavator i n 1 951 o n t he s ite n orth o f the B alkerne Gate a t Colchester and i t h as been i dentified more r ecently o n t he e astern d efences a t A lchester. Further s ites which may belong i n t his c lass i nclude: Aldborough, Ancaster, Caistor-by-No rwich, Cambridge, Canterbury, Godmanchester, G reat Casterton, E xeter, L eicester, London, Verulamium a nd Water Newton ( Frere 1 965, 1 37). This type of c onstruction will o ften c ontain mortar spreads at i ntervals i n t he b ank s howing t he p rogressive h eightening o f t he wall and bank a s a t Colchester where mortar s preads o ccur i n t he l ower l evels o f t he b ank a nd a t A lchester ( Hull 1 958, 6 3, f ig. 6 ; Oxoniensia 1 975 ( 1976) 4 0, 1 39). Jarrett ( 1965, 5 7) a nd Bidwell ( 1980, 6 0-2) n ote a v ariation o n t his w hich i nvolves the c onstruction of a bank o n t he l ower l evels o f t he r ampart before b ut a s p art o f t he w all b uilding. This m ethod could lead t o t he m isinterpretation of t he i nitial deposition o f r ampart material a s a n i ndependent e arthwork t o w hich a wall was l ater a dded. L ittle more c an b e a dded w ith r egard t o t he dating o f t his l ast t ype of c onstruction. Where a r ampart was p rovided a s i n ( 3) a bove, t his c arries t he s ame l imitations f or d ating as t hose a ssociated w ith t he c onstruction of an i ndependent earthwork c ircuit o r t he a ddition o f material t o a p ree xisting earthwork. I n u sing a t erminus post quem p rovided by t he l atest material i n t he r ampart w hich i s b uilt u p a s t he wall i s constructed, i t i s i mportant t o e stablish w hich o f t he successive l ayers a re part of t he s cheme f or t he c onstruction of t he wall a nd w hich r epresent l ater a dditions. I n the case o f ( 2) above, t he e vidence f or dating p rovided by r ampart material a dded s ubsequently t o t he r ear o f a f ree standing wall will b e of e ven l ess u se f or t he dating of t he wall a s e ven t he t erminus post q uem f or t he r ampart does not c ertainly p rovide a t erminus a nte quem f or t he wall. W here t he wall w as c onstructed a s a f ree-standing s tructure w ith n o rampart a s a t Catterick a nd G reat Chesterford, dating e vidence would b e p rovided b y o ccupation material b eneath t he wall and i n i ts c onstruction l evels, a lthough t his has y et t o b e p ublished. The s ame d ating e vidence m ight b e f ound i n ( 3) where a c ontemporary r ampart w as p rovided w ith o ccupation

3 7

l evels

s ealed

b eneath

Terminus post quem and

w all

a nd

Terminus

r ampart. ante q uem

The dating e vidence f or t he c onstruction o f t imber a nd earth and s tone defences t akes t he f orm of a t erminus post q uem. A t erminus a nte q uem f or a ny p hase o f defensive a ctivity i s r are but has been a rgued f or a number of walled c ircuits. This t erminus a nte q uem i s most c ommonly p rovided by f eatures associated d irectly w ith t he wall, t hat i s g ate t owers o r i nternal t owers c ontemporary w ith t he wall o r, l ess s ecurely, b astions, gate t owers o r i nternal t owers a dded s ubs equently t o t he w all. F or dating p urposes, t hese s tructures h ave t he a dvantage, w hich t he w all i tself l acks, of b eing l ikely t o contain o ccupation debris w hich might p rovide a date f or t he c ontemporary u se o f t he s tructure a nd a t erminus a nte quem f or i ts c onstruction. Where t hese s tructures have p rov ided g ood dating s equences, t hese i nvariably i nvolve a c oin o r coin hoard a ssociated w ith i nternal o ccupation o f t he s tructure. I t h as b een a rgued t hat t he walls a t Verulamium o r L ondon a re p rovided w ith a t erminus a nte quem of t his nature by c oin h oards w ithin i nternal t owers w hich w ere c ontemporary w ith t he wall. A date f or t he wall at Verulamium was suggested o n t he b asis o f a h oard o f f ive c oins e nding w ith o ne minted i n AD 27-229, concealed i n t he f loor of one o f t he i nternal t owers. 2 I t w as c onsidered t hat t he h oard w as u nlikely t o h ave b een b uried l ater t han AD 2 40 and t hat t his t herefore p rovided a t erminus a nte q uem f or t he c onstruction o f t he t ower a nd t he wall o f which i t f ormed a part ( Wheeler 1 936, 6 2-3; F rere 1 978, 2 87). I t i s d ifficult, h owever, t o r econcile t his w ith t he o f

i nformation t hat a h ouse demolished t he defences o n t he n orth c ontained

f or t wo

t he c onstruction b owls dated n ot

earlier t han AD 2 65-270 ( Frere 1 983). A t erminus a nte q uem f or t he w all a t L ondon a lso t akes t he f orm o f a h oard o f c oins a nd f orger's c oin moulds i n one o f t he i nternal t owers, a gain c ontemporary w ith t he wall, on t he w estern c ircuit n ear Newgate. The moulds w ere made f rom g enuine u nworn c oins o f AD 2 01-210, 2 10-212 a nd 2 15; t ogether w ith a n ew denarius o f AD 2 13-217 a nd s everal o ther c oins t hey w ere d eposited u nder t he s tairway of t he t ower where a f air amount of r ubbish had a lready a ccumulated s ince t he t ower h ad b een built ( Merrifield 1 969, 19; 1 978, 1 9-20; Marsden 1 980, 1 26) . 1 F rere suggests t hat t his c oin e vidence i ndicates a d ate o f c . AD 2 10-220 f or t he t ower and t he wall a t t his point, while Wacher considers 2 t hat a s lightly l ater d ate o f c . A D 25-230 i s i ndicated ( Wacher 1 974, 9 4; F rere 1 978, 2 87)

part

I n t hese of t he

t wo i nstances, a s tructure w hich wall and c ontemporary w ith i t,

i s a n i ntegral p rovides t he

t erminus a nte q uem f or t he c onstruction o f t he wall. A d ate f or t he c onstruction of t he wall a rgued on t he basis of s tructures a dded s ubsequently t o t he wall w ill b e questionable. There may b e a t emptation t o u se t he t erminus p ost q uem d ate b astions, which

f or t he c onstruction o f s uch i nvariably s eem t o p ost-date t he

3 8

f eatures d efences,

a s t o

provide a t erminus a nte q uem f or t he w all i n much t he s ame w ay as t he evidence f rom comtemporary f eatures s uch a s t he i nterval t owers a t L ondon a nd Verulamium. There a re f ew o ther examples where a t erminus ante quem i s a rgued on t he b asis of dating e vidence a ssociated n ot w ith f eatures w hich f orm a n i ntegral p art o f t he walled c ircuit, b ut w ith f eatures c ut i nto t he defensive c ircuit and a gain post-dating i ts c onstruction. At Ancaster a nd Canterbury t he r ampart s eems t o b e contemporary w ith the wall and a t erminus ante quem f or t he wall i s p rovided i n t hese c ases by f eatures c ut i nto t his rampart. At Ancaster, a t erminus a nte quem o f AD 2 80 given f or t he r ampart i s p rovided by f ive c oins o f t he d ecade AD 2 70-280 sealed beneath s uccessive f loors o f a b uilding e rected over t he t ail o f t he r ampart ( Todd 1 981). At Canterbury, a date f or t he wall a nd i ts rampart was a rgued on t he basis o f a small hoard o f c oins o f t he e arly 4 th c entury f ound i n a p it cut i nto t he t ail of t he r ampart a long t he s outh-east c ircuit ( Wacher 1 974, 1 88). C oins f ound i n t he c onstruction l evels a re i n a m ore secure and d irectly r elevant context f or dating t he construction o f a bastion o r g ate t ower t han t hose f ound i n the o ccupation l ayers w ithin i t. A c oin, f or e xample, of AD 3 60 i n a l ayer w ithin, means o nly t hat t he b astion was b uilt before the o ccupation debris was d eposited. This debris could have been deposited i n o r a fter AD 3 60 a nd o n i ts own s ays l ittle about t he construction of t he bastion, s ince t he p resence o f t he c oin i n t his c ontext c ould mean t hat i t w as built either b efore t he c oin was minted o r a t a bout t he t ime o r a fterwards. T he date o f t he c oin n eed n ot b e c lose t o t he date o f i ts deposition w ithin t he bastion. A lthough t he p otential f or e stablishing a r easonably secure t erminus a nte quem does e xist f or s uch s tructures a s bastions a nd gate t owers, a t p resent t here h as n ot b een e nough excavation of t his type of s tructure t o e nable good dating s equences t o b e e stablished. T he c ontext a nd n ature o f t he e vidence i n e ach i nstance will affect t he degree o f p robability i n f avour o f t he ' terminus a nte q uem' a rgument s o t hat, i n t he r ight c ircumstances, i t i s possible f or a c onvincing a rgument t o be c ons tructed. The a ttempt t o date a p rimary w all b e means o f a date p rovided by a s econdary f eature s uch a s a b astion i s only possible i f t he b astion i tself i s c losely d ated. Casey ( 1983) has a rgued t hat t his was possible a t Caerwent where t he p recise n ature o f c oin e vidence a pparently p rovided a t erminus ante quem o f AD 3 50 f or t he c onstruction o f t he bastion and t herefore o f t he w all t o w hich i t h ad b een a dded. B ut i n most cases the e vidence r elating t o t he c onstruction of s uch f eatures t akes t he f orm o f a t erminus p ost q uem a nd t his does not p rovide a t erminus ante q uer n o f a ny k ind f or t he wall s ince t he wall may a lso h ave b een b uilt a fter t he t erminus post quem p rovided f or t he b astion. The s ame i s t rue of t he t erminus a nte q uem f rom Ancaster a nd Canterbury, d iscussed above, a lthough t he Ancaster e xample i s on balance more s ecure s ince t he c oins w ere s ealed b eneath s uccessive f loor l evels. Obviously a s ingle c oin on i ts own i s o f l ess u se f or t he

3 9

c onstruction o f a ' terminus a nte q uem a rgument' t han a g ood s equence o r hoard t ogether w ith s upporting e vidence f rom p ottery. As well a s a g ood s tratigraphic s equence, t he c ontext o f t he dating e vidence i s a lso of i mportance i n a ssessing what i t i ndicates a bout t he c onstruction a nd/or o ccupation o f t he bastion, t ower, etc. One needs t o a ttempt to e stablish ( i) h ow l ong a fter t hey w ere minted w ere t he coins deposited, ( ii) how l ong before t his deposition was t he structure b uilt. E ven w here t his c an b e s ecurely e stablished, i t s hould b e emphasised t hat t his i s s till an ' argument' w ith a greater o r l esser d egree o f p robability d epending on t he context t he p recision w ith which t he deposition can be dated.

a nd

on

The s earch f or a t erminus a nte q uem r esults f rom t he f act t hat a t erminus post quem w ill n ot n ecessarily be v ery c lose t o t he a ctual d ate o f c onstruction. The s ame p roblem a pplies t o a t erminus ante quem, t hat i s how l ong before t he t erminus a nte q uem i s t he a ctual date o f c onstruction and h ow can one begin t o a ssess t his? A t erminus a nte quem i s u sed with most e ffect w hen i t c an b e r elated t o o r o ccurs i n association w ith a s ecure and s tratigraphically r elevant t erminus p ost quem, i n w hich c ase t he t erminus a nte q uem t hen p rovides a n u pper l imit f or t he ' post quem' and v ice versa. At Caerwent, t he terminus ante q uem o f AD 3 48-349 f or t he wall o n t he n orth-east c ould be t aken i n a ssociation w ith H artley's t erminus post quem of AD 2 40 f or t he w all a long t he s outhern p art o f t he c ircuit a nd t hese two dates t ogether have b een s een as an indication t hat t here may h ave b een n o g reat d ifference i n date b etween t he wall and t he b astions ( Hartley 1 966, 5 7; Boon 1 974, 1 02). Casey ( 1983) h as r ecently a rgued t hat t he u nworn c oin of I J rbs R oma ( AD 3 30-335) f ound by Nash-Williams i n h is e xcavation of t he s outh-east b astion p rovides a t erminus p ost q uem f or t he wall s ince i t was s ealed by a s andy l evel a ssociated with t he c onstruction o f t he w all. As w ell a s t he t erminus ante q uem of c . AD 2 10-220/225-230 f or t he western part of t he walled c ircuit a t L ondon, t here i s a lso a t erminus p ost q uem p rovided by a worn c oin of AD 1 83-184 which was f ound i n t he ' accumulation o f mixed s oil' b elow t he t hickening o f t he C ripplegate f ort wall; t his t hickening i s ' reasonably assumed t o b e c ontemporary w ith t he t own wall d efence' ( Hartley 1 966, 5 7; Merrifield 1 965, 5 2; 1 969, 1 18-19; 1 978, 1 9; G rimes 1 968, 5 0-1; Wacher 1 974, 9 4). Coins a nd p ottery i n p re-wall deposits and deposits c ontemporary w ith t he wall a lso indicate t hat i t c ould n ot h ave b een b uilt e arlier t han c . AD 1 90 ( Merrifield 1 969, 18-19; 1 978, 1 9; Marsden 1 980, 1 25) . 1 T his p rovides a d ate o f u nusual p recision f or t he wall at L ondon and t he s ame i s t rue o f t he walled c ircuit a t Ancaster w here dating e vidence i ndicates t hat t he w all a nd r ampart w ere c onstructed b etween AD 2 25, o r more p robably AD 2 50, a nd AD 2 80 ( Todd 1 981). At Canterbury, t he t erminus ante q uem i n t he early 4 th c entury may b e t aken i n c onjunction with t he e vidence f rom a nother s ection a cross t he d efences j ust t o t he n orth ( Burgate L ane) where a c oin of TetricuS I ( AD 2 70-273) f rom b eneath t he r ampart p rovided a t erminus p ost q uem of t his date f or t he wall and t his was supported by the associated p ottery ( Wacher 1 974, 1 88). H owever, i t h as a lready b een s tressed t hat a date, whether a t erminus post q uem o r a t erminus a nte q uem may apply o nly t o t hat p articular s ection

4 0

of t he defences f rom w hich i t c ame a nd n ot t o t he whole circuit. D ifferent parts o f a c ircuit o ften p roduce d ifferent dating evidence a nd s ections c ould h ave b een b uilt o r r ebuilt at d ifferent t imes. I n g eneral, i t i s i mpossible t o a ssess h ow l ong ' after which' the a ctual o r t rue date i s. There i s a lways t he possibility t hat a nother s ection may p rovide a t erminus post quer n of a l ater d ate. Wacher ( 1965,226) c omments, ' given enough s ections, s uch t ermini post q uos may a pproach t he t rue date of construction, w ithout our b eing a ble t o s ay conclusively t hat t his d ate h as b een r eached s imply b ecause w e a re never quite c ertain t hat s omething a good deal l ater may not t urn u p i n t he n ext s ection t o b e dug'. H owever, t he concept t hat t he more s ections dug t he c loser t he e vidence approaches t o t he c orrect date i s dubious s ince t he r ampart need not contain anywhere a long i ts c ircuit a ny material contemporary w ith o r c lose t o t he d ate o f i ts c onstruction a nd one can n ever p rove t hat i t does. ' It i s d ifficult t o k now how f ar a ny p articular f inds c an b e r elied u pon t o i ndicate a close date' ( Boon 1 974, 6 5) I t i s n oticeable t hat a n umber o f t imber a nd e arth circuits have a t erminus post quem c entred on t he mid 2 nd century ( Hartley 1 983). The r amparts a t A ldborough, Bath, Exeter and Wroxeter a ll contain a l arge b ody o f material o f the Hadrianic Antonine p eriod; y et e ven w here t here i s n o later material, a n Antonine date i s only a t erminus p ost quem and t his n eed n ot b e a nything n ear t he date a t w hich t he b ank was b uilt. H owever, B oon ( 1974, 5 3) was s till i nclined t o accept a n Antonine d ate f or t he I nner D efence r ampart a t Silchester o n t he b asis o f t he p ottery o f t his p eriod f rom t he rampart i n t he e xcavations o f 1 938-1939. S imilarly, a lthough t he Antonine material f rom t he F osse e arthwork a t V erulamium has been r egarded as a t erminus post quem f or w hat was i n r eality a l ate 2 nd c entury e arthwork, F rere h as s uggested t hat there a re other f actors which point t o t his e arthwork having -

been under c onstruction during t he Antonine p eriod ( Wacher 1 974, 2 13; F rere 1 978, 2 85-6) I n t he c ase o f B rough-onHumber, t he f irst e arthwork c ircuit c ontained H adrianicAntonine material a nd t his r ampart has a lso b een a ttributed t o t his period b ecause a l imiting f actor f or t he t erminus p ost quem provided by t his material e xists i n t he f orm o f a s econd r ampart ( period I V) w hich h as a t erminus post q uem i n t he l ate 2 nd/early 3 rd c entury. This s econd e arthwork c ircuit a nd t he rampart h eightening which a ccompanied t he l ater wall ( period V ) a lso c ontained H adrianic-Antonine material b ut i n t hese i nstances i t was t he p resence o f a f ew l ater s herds which p rovided t he t erminus p ost q uem r ather t han t his s ubstantial body of earlier material ( Wacher 1 969, 2 9, 3 2 & 3 8; Ramm 1 978, 3 9-42, 5 1-62). D espite t he u se o f t he t erminus p ost q uem o f each successive c ircuit a t B rough-on-Humber t o p rovide a ' control' o n t he d ating o f t he p revious p hase, t he a ctual d ate of e ach phase c ould b e s een a s much l ater t han t hat p rovided i n e ach c ase by t he t erminus p ost q uem f or t hat p hase, s o t hat t he w hole chronology c ould b e e xtended; namely, t he i nitial r ampart may b e l ater t han t he H adrianic-Antonine p eriod j ust as t he period Iv r ampart could b e l ater t han t he l ater 2 nd o r

4 1

e arly 3 rd c entury a nd s o o n, b ut h ow m uch l ater i n e ach c ase i s v ery d ifficult t o e stablish. W ith r egard t o t he p recision o f t he t erminus p ost q uem i t w ould b e i nstructive t o s ee h ow f ar s hort o f t he h istorical d ates o f AD 1 93-196 a re t he v arious d ates p rovided b y d ifferent c ircuits ( see T ables F our a nd F ive). T he i nadequacies o f t he d ating e vidence a nd t he w ide r ange o f d ates p rovided b y d ifferent c ircuits do n ot a llow t he p recise c hronology w hich i s n ecessary t o s upport a ny t heory a bout t he d ate a nd h istorical c ontext f or p hases o f d efensive a ctivity. I n a ttempting t o d ate d efences a s a g roup, i ndividual e xamples m ay v ary c onsiderably. T his a pplies t o t imber a nd e arth d efences a nd t o b astions w hich a re a lso r egarded a s t he r esult o f a s ingle p olicy. I n t he c ase o f t he f ormer, t he d ates d o v ary, w hile b astions a re as y et i nsufficiently i nvestigated. W hile o ne c annot p rove t hat t hose c ircuits w ith a t erminus post q uem i n t he m id 2 nd c entury a re n ot i n f act c ontemporary w ith t hose c ircuits w hose d ating l ies i n t he l ate 2 nd c entury, t he e vidence f rom B rough - o n-Humber c ited a bove i s i nstructive i n t he c ase o f t hose r amparts w ith a m id 2 nd c entury h orizon. T here i s n o r eason w hy a t erminus p ost q uem i n t he m id 2 nd c entury s hould b e ' dismissed' s imply a s i ndicating a d ate i n t he l ate o r l ater 2 nd c entury; n or s hould a t erminus p ost q uem o f w hatever d ate b e s o u sed t hat i t l oses i ts r elevance. A c ommon t ype o f s urvey o f t he d ating e vidence s tates t hat a p articular s ettlement ' appears t o h ave b een p rovided w ith a n e arth r ampart a nd d itch i n t he l ater 2 nd c entury' o r ' the m aterial i n t he r ampart i ndicated a t erminus p ost q uem f or i ts c onstruction i n t he l ater 2 nd c entury. T he e vidence u sed t o s upport a l ate 2 nd c entury d ate v aries v ery c onsiderably i n i ts n ature a nd p recision a nd i t i s i mportant t o k now w hether t he e vidence c onsists o f p ottery o f t he m id o r l ater 2 nd c entury o r e ven a s ingle s herd a mongst a p red ominantly e arlier b ody o f m aterial, a nd w hether c oins a re i nvolved. I n t he l ast r esort, w hatever t he t erminus p ost q uem i nvolved, t he c hoice o f a l ater 2 nd c entury d ate w ill b e b ased o n ' analogy w ith o ther s ites' t hus p erpetuating a c hronology w hich i s b ased n ot o n t he p recision o f t he a rchaeological e vidence b ut o n t he s earch f or a s uitable h istorical c ontext f or t he c onstruction o f d efences. I mprecision i n p resentation o f e vidence r elating t o t he d ating o f d efences i s f airly c ommon a nd t here i s a t l east a s uspicion t hat t he s earch f or a n e mphasis o n s uitable h istorical c ontexts i s l argely r esponsible f or t his. C onsequently, t he t erminus R 2st q uem d ating h as b een u sed r ather t o r einforce t he t heory c oncerning a s ingle l ate 2 nd o r m id 2 nd c entury c ontext f or e arthworks a nd a s ingle c ontext a fter t he m id 4 th f or b astions. B ut t he v ery i mprecision o f t his t ype o f d ating d oes m ake it e qually i mpossible t o p rove a ny a lternative t heory. '

4 2

The I ncompatibility Evidence

o f

t he

Archaeological

and

H istorical

The a ttempt t o r elate a rchaeological e vidence t o s pecific h istorical c ontexts i nvolves a g reat deal o f speculation. The t wo t ypes o f e vidence a re c ompletely d ifferent a nd i t i s u nrealistic t o attempt t o r econstruct t he complete p icture on t he basis o f f ragmentary h istorical a nd a rchaeological evidence. To a c onsiderable e xtent, i t was t he i nsistence on t he n ecessity f or i mperial p ermission f or t he e rection o f d efences w hich w as r esponsible f or t he c oncept t hat s ettlements in B ritain w ere p rovided w ith d efences as a r esult of major t hreats t o t he s ecurity o f t he p rovince. Subsequent phases of defensive a ctivity were a lso t hought of a s r esponses t o s pecific t hreats. The i nsistence o n i mperial p ermission c reated t he compulsion t o f ind a c ontext f or t he c onstruction o f i ndividual c ircuits i n t he p olitical s ituation w ithin t he p rovince and t his i n t urn r einforced t he t endency t o ' stretch' t he d ating e vidence t o f it t he h istorical c ontexts. Doubts have been e xpressed about t his emphasis on i mperial perm ission. The D igest, c ontaining t he e vidence o f t he n eed f or i mperial permission, i s essentially a document of t he l ater 2 nd c entury ( Wacher 1 964, 1 07). H ow l ong t he n ecessity f or permission c ontinued after t hat t ime i s n ot c lear. E ven i f i t w as n ecessary, i t may h ave b een a f oregone c onclusion o r i t may e ven have b een possible t o obtain permission r etros pectively. Hassall c omments ' that t he o ften c ited r escript o f Marcus t o a p rovincial governor s tating t hat i mperial p ermission s hould b e s ought b efore walls a re e rected s hows only that c ities h ad been e recting walls w ithout permission. E ven i f t his measure w as e nforced a fter t he t ime o f Marcus, w e have no r eason t o s uppose t hat, g iven t he general i nsecurity o f B ritain i n t he l ater 2 nd c entury, p ermission, i f ' sought, would have been r efused' ( Hassall 1 977, 1 28). P erhaps more emphasis s hould b e p laced o n t he f act t hat p rudent a uthorities had only t o s eek p ermission t o e rect f ortifications and, i f t he t own f ulfilled c ertain r equirements, t here m ight b e a r easonable e xpectation t hat t his would be g ranted ( Boon 1 974, 6 5-6). I n a ddition, i n s eeking h istorical c ontexts f or a ll major phases of d efensive a ctivity, t here has b een a t endency n ot t o make t he d istinction b etween t he i nitial p rovision o f d efences and a ny subsequent d efensive a ctivity. R epair, maintenance o r u pdating was l ess l ikely t o n eed p ermission a nd t here i s n o n eed to s earch f or h istorical contexts t o e xplain s uch a ctivities. F inally, t he h istorical c ontexts s uggested f or d efensive a ctivity a re questionable. I t i s t empting t o r elate t he p rovision o f d efences t o a d ocumented e vent, i f o nly t o g ive p recision t o t he i mprecise a rchaeological dating. But h istorical e vidence i s meagre a nd a mbiguous, a nd h ow r elevant a re t he suggested c ontexts t o t he c ircumstances under which d efences m ight b e p rovided? I n s earching f or c ontexts f or t he

4 3

p rovision o f defences f or s pecific c entres at p articular t imes, there i s a danger t hat t he r elevance of t hese c ontexts may n ot b e p roperly a ssessed. The a ctivities o f C aratacus, t he Boudiccan r ebellion o r i ts a ftermath a nd civil war i n t he empire i n AD 6 8-70 h ave b een s uggested a s p ossible o ccasions f or t he construction o f d efences during t he i st century ( Wacher 1 974, 2 57-60, 2 80; 1 978, 9 6). A lternative c ontexts, e specially f or i ndividual defensive c ircuits on t he f ringes o f t he p rovince, c an b e f ound i n t he t roubles a t various t imes a long t he n orthern o r western f rontiers ( for example Wacher 1 964, 11-12). B ut t he s uitability o f t hese contexts t o t he 1 type of defence u sed n eeds t o b e t horoughly examined; f or e xample, d efences e rected a s t he B oudiccan r ebellion m ateriali sed would have b een hastily c onstructed and t he n ature o f t he c onstruction must s uit s uch a c ontext. A lternatively, t he Boudiccan r ebellion may have b een o ver t oo q uickly a nd defences may h ave b een p rovided b ecause o f t he i nsecurity i t c aused. Under t hese c ircumstances, s peed would have b een l ess v ital, t he a im b eing r ather t o p rotect i nvestment i n developing c entres s uch a s Colchester. B ut i t i s equally p lausible t hat t he p ossibility o f i nsurrection, e specially i n t he a ftermath of t he Boudiccan r ebellion, would make t he c onstruction o f d efences, e ven f or t he most l oyal c ommunity, a potential t hreat t o t he general s ecurity. The government may h ave p referred, e specially i n t he e arly y ears after t he c onquest, t o r ely on t he a rmy t o p rotect t he c ivil z one r ather t han t o a llow a ny c ommunity t o a ssume r esponsibility f or i ts own defence ( Rivet 1 964, 7 8). S imilarly, t here a re t hose w ho c hallenge t he s uitability of t he A lbinian episode f or t he p rovision of defences. i t may b e wondered w hether, i n t he game o f p ower politics t hat h e was p laying, A lbinus would h ave spared much thought f or t he B ritons a nd t he d efensive n eeds o f t he t owns i n t he a bsence o f t he B ritish garrison and whether, i f h e h ad, h e w ould h ave a dvertised s o o bviously h is i ntention o f t aking t roops o ut o f t he p rovince by u ndertaking a vast p rogramme o f u rban defence'

( Hassall

1 977,

The point i s t hat h istorical c ontexts f or

1 28; t he t he

s ee

a lso

J ohnson

1 983).

r elative merit o f various p ossible p rovision o f d efences i s a matter

of conjecture. The A lbinian episode i s t he only e pisode on w hich t o f ocus w idespread d efensive a ctivity i n t he 2 nd century; i n t he 3 rd c entury i t i s t he t hreat f rom b arbarian r aids. L ater, t he r ecovery a nd r eorganisation of t he p rovince under Constantius Chiorus i n the l ate 3 rd and e arly 4 th c enturies, a nd t he e xpeditions o f Constans i n AD 3 43, L upicinus c . AD 3 60, Theodosius c . AD 3 68 a nd Stilicho c . AD 3 96-398 a re t he h istorical e vents w hich a re v ariously suggested a s p roviding t he c ontext f or t he p rovision o f b astions o r f or o ther p hases o f d efensive a ctivity ( Casey 1 983) . A ll were d isasters o r e vents a ffecting t he s ecurity o f t he p rovince. B ut a ll a re s uspect a s c ontexts f or d efensive a ctivity and t he l imitations of b oth t he h istorical and t he a rchaeological e vidence make i t u nrealistic t o d ate e very earthwork, walled c ircuit and bastion on t he b asis o f a h istorical r eference w hose r elevance i s u ncertain.

4 4

IV. T he Reassessment and C riteria

o f

t he

Evidence

i n

t he

L ight

o f New Data

T he a dherence t o a h istorical f ramework h as meant t hat, i nstead of r eviewing t he actual f ramework and a ssessing t he validity o f t he b asic p remises o n w hich i t i s b ased, i n t he l ight of n ew e vidence t oo much emphasis has b een p laced on r efining t he d ates a nd h istorical c ontexts f or t he major phases of defensive a ctivity t imber a nd earth, walls and bastions a nd t he e vidence f rom i ndividual s ites h as b een largely submerged i n t his p rocess. I nstead, t his e vidence s hould b e a nalysed i ndependent o f t he f ramework a nd i n a n objective way a s a piece of i nformation about a particular defensive c ircuit. I n v iew o f t he l imitations o f t he a rchaeological evidence and t he uncertainties a bout t he r elevance o f various h istorical e vents, i t i s v alid t o e xamine o ther factors i nfluencing t he c onstruction o f defences. -

-

1 ) I s i t p ossible t o circuits i n B ritain a s phenomenon a dopted o nly a s

r egard t he p rovision anything o ther t han a n emergency m easure?

o f a

e arthwork r estricted

T he current f ramework i s b ased o n t he a rgument t hat, w ith the e xception of t he coloniae and other chartered s ettlements, c ivil s ettlements i n R oman B ritain would n ormally h ave r emained open, r elying p rimarily on t he m ilitary p resence a long t he f rontiers f or t heir p rotection a gainst b oth i nternal a nd external t hreats ( Frere 1 978, 2 83). I t h as b een a rgued f urther t hat i n t he i nitial s tages o f c onquest a nd Romanisation, t he t hreat was f elt t o be more l ikely t o be ' internal' a nd t his i s r eflected i n t he d isarming o f t he t ribes behind t he f rontier i n AD 4 7-48 ( Frere 1 978, 9 2-3) A lso s uch w idespread p rovision o f t imber a nd e arth defences as e xists f or t he t owns and v illages o f Roman B ritain i s r egarded a s b eing w ithout p arallel i n t he R oman e mpire a nd the u se of t his p articular type of defence r ather t han s tone i s s een a s r einforcing t he c oncept t hat t he p rovision o f defences was i ntended t o b e a r estricted phenomenon a ssociated e ither with s pecial s tatus, o r w ith t he vulnerability o f particular c entres. Johnson ( pers. comm.), however, n otes t hat earthen defences a re f ound e lsewhere i n t he Empire, a lbeit n ot on t he same s cale, s uch a s L epcis Magna a nd possibly a t certain o f t he G erman s ites i n t he e arly 3 rd c entury, for e xample, L adenberg and H eddernheim. T hese c onclusions a re b ased o n a nalogy w ith G aul w here the m ajority of c ivil s ettlements appear t o have r emained undefended u ntil l ate i n t he 3 rd c entury w hen i ncreased barbarian p ressure on t he f rontiers apparently made defences a necessity ( Butler 1 959; v on P etrikovits 1 971; Johnson 1 973; F rere 1978, 2 84) . B ut t he a nalogy w ith G aul h as l ittle r elevance f or a nalysing d evelopments i n R oman B ritain. There is a clear c ontrast b etween t he r ealities o f t he w idespread

4 5

p rovision o f defences i n Roman B ritain i n t he i st and 2nd centuries ( see f igs. 1 & 2 ) and t he s ituation i n G aul ( Butler 1 971, f ig. 1 ). E ven w hen defensive a ctivity o ccurs i n t he t wo p rovinces at much t he s ame t ime, t he d ivergence b etween t he t wo p rovinces i s n oticeable. T owns i n Gaul were r eceiving s tone defences i n t he l ate 3 rd c entury a t much t he s ame t ime, i t i s a rgued, a s many o f the t owns a nd v illages of R oman B ritain w ere a lso p rovided w ith w ailed c ircuits. T he walls o f G aulish t owns c losely r esemble t he l ate 3 rd century military a rchitecture r epresented by t he main s eries of S axon S hore f orts, while t hose i n B ritain d o n ot a dopt t hese n ew t echniques ( Frere 1 978, 2 86).

which

F urthermore, i t i s p ossible s uggests t hat earthworks i n

t o a dopt a n B ritain d id

a lternative v iew not r epresent a n

outdated method o f defence w hich w as r elied u pon o nly a s a n emergency measure. A well-established t radition o f earthwork defences i s s een i n t he e arthwork c ircuits e rected a t Broughon-Humber, Colchester, S ilchester, V erulamium and W inchester i n t he i st a nd e arly 2 nd c enturies a nd r etention of t he l egionary f ortress defences a t E xeter. At L incoln a nd G loucester, work b egan on f acing t he l egionary f ortress defences i n s tone i n t he early 2 nd century. The c onstruction o f d efences i n s tone f or t he c olonia a t C olchester a lso s eems t o have b egun i n t he early 2nd c entury. This reconstruction i n s tone w ould b ring t he c oloniae o f B ritain i nto l ine w ith developments i n G aul ( Frere 1 978, 2 84; Wacher 1 978, 9 7) Much more w ork i s n eeded o n t hese s ites b efore a p roper understanding of t hese earthworks c an b e a chieved b ut i t i s possible t hat t hey may r epresent c ontinuity o f defence over a l ong period f ications.

of

and were

not

s imply

2 ) I s t here a ny e vidence e arthwork c ircuits i n t he

f or i st

s hort-lived

t he a nd

emergency

f orti-

m ore w idespread e xistence e arly 2 nd centuries?

I t i s quite p ossible t hat t he w ide d iversity i n t he t erminus p ost quem f or various e arthwork c ircuits w hich a re a ssumed t o b elong t o t he l ate 2 nd c entury does a ctually r ef lect t he f act t hat not a ll were e rected a t t he s ame t ime o r f or t he s ame r easons. Boon ( 1974, 5 3, 6 5-6) s uggested s ome t ime a go t hat perhaps n ot a ll 2 nd century earthwork defences were t he p roduct o f t he A lbinian e pisode. The possibility o f a more w idespread t radition of earthwork d efences t han w hich p laces t he main b ody o f e arthwork c ircuits i n t he t ext o f A lbinus' p reparations s hould n ot b e i gnored. Apart f rom t he i st c entury t imber a nd e arth evidence f or t heir c onstruction p rior t o t he mid i s a t p resent d iverse a nd f ragmentary.

t hat c on-

c ircuits t he 2 nd century

At Verulamium, an Antonine date was n ot r egarded a s i nappropriate f or t he F osse e arthwork ( Hartley 1 983). B oon was p repared t o a ccept a n Antonine date f or t he I nner Defence r ampart

a t

S ilchester

s uggesting

4 6

t hat

t he

i mpetus

f or

t his

m ay

have been p rovided by u nrest i n Wales a nd t he n orth during this period ( Boon 1 974, 5 3, 6 5-6). This context i s n ot v ery convincing s ince S ilchester l ay well a way f rom t hese a reas. At A ldborough, a t erminus post q uem i n t he H adrianic-Antonine period suggested t o Charlesworth t hat an earthwork c ircuit might have been e rected a s a r esult o f t he t roubles i n t he north in AD 1 69 o r AD . 1 80 ( Charlesworth 1 971, 1 59) . The p roximity o f A ldborough t o t he a rea o f u nrest c ould g ive a plausible r eason f or defences being p rovided p rior t o the Albinian episode. However, t here h as b een v ery l ittle i nvestigation of t he r elevant earthwork c ircuits at S ilchester, ' lerulamium a nd A ldborough a nd t here i s t he o bvious p ossibility that t hey were constructed at a l ater date. I ndeed, t he l ate 2 nd c entury h as b een a ccepted a s t he most l ikely o ccasion f or the c onstruction o f t he c ircuits at l east at A ldborough and Silchester ( Wacher 1 974, 2 13, 2 64-6, 4 01; F rere 1 978, 2 85-6). Brough-on-Humber i s one o f t he f ew centres where a date p rior t o t he l ate 2 nd c entury h as l ong b een a ccepted f or t he c onstruction o f a n e arthwork. I n t he l ate Hadrianic/Antonine period t he developing v icus P etuarensis w as e nclosed by a n earthwork defence ( Wacher 1 960, 6 2; Wacher 1 969, 2 9; 1 971, 1 66; 1 974, 3 95; Ramm 1 978, 3 8-40). The r eason f or a ssigning this circuit t o t he mid r ather t han t he l ate 2 nd c entury i s t he p resence o f quem ( c. AD 1 80) early 3 rd c entury

a s ubsequent e arthwork w hose t erminus p ost s uggests t hat i t belongs t o t he l ate 2 nd o r ( idem.).

There a re a n umber o f other c entres f or w hich t here i s some evidence f or t he possible e xistence o f a d efensive circuit p rior t o t he l ate 2 nd c entury ( Hartley 1 983). I n t he case of Chelmsford, t he earthwork c ircuit shows e vidence o f two p hases o f c onstruction b efore i t w as l evelled c . AD 2 0020. The only dating e vidence f or t he a ctual construction of 2 the c ircuit s eems t o c onsist o f material s ealed b eneath t he rampart which ' suggests a c onstruction date of c . AD 1 60-75' ( Drury 1 975, 1 70), a lthough t he e xcavations a re a s y et u npublished. A s econd minor phase i s r epresented by a s hort l ength o f outer d itch on e ither s ide o f t he s outhern gate. This is r egarded as a n attempt t o s trengthen t he defences later i n t he 2 nd c entury w hich was h ardly b egun w hen, i t i s argued, Chelmsford a long w ith s everal other t owns and r ural s ites i n s outhern a nd e astern E ssex s uffered c onsiderable destruction by f ire ( Rodwell 1 975, 9 3). At Manduessedum, t he p ossibility o f a n e arly e nclosure defining a s lightly smaller a rea t han t hat of t he b urgus a rises f rom t he d iscovery o f a d itch a t t wo p oints ( north-east and s outh-east) b eneath the r ampart o f t he burgus d efences. Pottery f rom t he s ilt o f t his d itch i ndicated t hat i t h ad b een filled i n by t he e nd of t he 1 st/early 2 nd c entury. This ditch w as n ot l ocated i n a s ubsequent s ection a cross t he e astern defences o f t he burgus and i t i s d ifficult a t p resent t o assess i ts s ignificance ( Webster 1 971, 2 5; Todd 1 973, 3 8). Apart f rom t hese e xamples, t he o nly o ther s ite w hich might be mentioned h ere i s Dorchester-on-Thames where a l ate 1st c entury defensive d itch a ligned p arallel t o t he w estern defences was f ound i n t he n orth-west c orner of t he s ettlement

4 7

( Britannia ( 1973) 4 , 2 97). I t i s i nteresting t o note i n c onnection w ith t hese l ast two e xamples t hat a ll t he o ther s ites p reviously mentioned i n a ssociation w ith early earthwork defences E xeter, S ilchester, Winchester and Verulamium a re c ivitas c apitals. Even B rough-on-Humber a nd Chelmsford a re r egarded as p robable capitals o f t heir r espective c ivitates e ven i f t hey a re s een a s h aving ' failed'. The e vidence f rom D orchester-on-Thames and Manduessedum, a lbeit i nconclusive, e xtends t he possible u se o f e arthwork c ircuits p rior t o t he mid 2 nd c entury t o so-called ' lesser' centres. -

-

On t he i nconclusive e vidence p resently a vailable f or t he e xistence of t imber and earth c ircuits p rior to t he mid 2 nd c entury, l ittle more c an u sefully b e s aid. B ut t hese e xamples do at l east b roaden t he s cope o f t he i nquiry by opening up t he p ossibility o f a more g eneral o r a t l east l ess r estricted construction of defences. No consistent pattern e merges b ut i n f act, i n c onsidering t he p ossibility o f a more f lexible o r a d hoc u se of earthwork c ircuits i n t he 1 st and 2 nd centuries one s hould n ot e xpect t hat s uch a p attern would e xist. I ndeed, i t has a lready b een mentioned t hat those who h ave c onsidered t he b ackground t o t he p rovision o f t hese e arly c ircuits have f ound explanations, n ot i n a policy a pplied t o t he p rovince a s a w hole, b ut i n t he l ocal c ircumstances affecting t he s ecurity o f i ndividual c entres. The r elevance o f t his s ort o f a pproach i s perhaps s hown most c learly i n t he c ase o f B rough-on-Humber w here t he p rovision o f d efences m ight b e c onvincingly e xplained w ithin t he context of i ts p roximity t o t he n orthern f rontier. Webster ( 1981) h as d eveloped t he a rgument t hat H adrian's policy of consolidating t he f rontiers o f t he Empire involved t he s ettlement o f a m ilitary s quirearchy amongst t he c ivilians i n t he n orth and w est where t he l and w as n ow b eing h anded b ack t o t he t ribes. There a re i ndications t hat, w hen t erritory i n e ast Yorkshire was h anded b ack t o t he P arisi and t he B rigantes, special quasi-military a rrangements of t his k ind w ere made f or t he p rotection o f t hese c ivitates w hich l ay uncomfortably c lose t o t he permanent m ilitary z one ( Hadrian's Wall a nd t he P ennine z one b ased o n C hester i n t he w est a nd York i n t he e ast) ( Frere 1 978, 1 47 & 1 82). There i s s ome e vidence t hat t hese a rrangements i ncluded t he d eliberate s ettlement of r etired s oldiers on small f arms and p ossibly t he s ettlement o f v eterans i n t he v icus P etuarensis at B rough-onHumber ( Ramm 1 978, 3 8-40, 7 3, 7 7). I n a ddition, a p ermanent f ort w as t erritory o f t he P arisi a t Malton and here

maintained i n t he a f lourishing v icus

d eveloped i nto a n i mportant u rban c entre f or t he r egion. I t t oo was p rovided w ith d efences apparently a t much t he s ame t ime a s r ough-on-Humber r eceived i ts s econd e arthwork c ircuit ( late 2 nd/early 3 rd century ) ( Ramm 1 978, 4 1, 5 3, 6 3-4 & f ig. 2; R obinson 1 978, 6 -7). The p rovision o f d efences f or t he 2 n ewly established u rban n ucleus/possible c ivitas c apital a t B rough-on-Humber a nd t he s ubsequent p rovision of d efences t he very s uccessful v icus a t Malton m ight be s een a s i ntegral p art o f t he n eed f or g reater s ecurity of ' frontier' c ivitates. I f a specific t hreat f or

4 8

f or a n t he t he

construction o f t he defences a t B rough-on-Humber n eeds t o sought t hen a ttention h as b een d rawn t o t he i ndications serious t rouble amongst t he B rigantes c . AD 1 54-158 and

b e o f the

fact t hat t his r esulted i n n ew a rrangements b eing made f or t he long t erm s ecurity o f t he a rea ( Wacher 1 969, 2 9; F rere 1 978, 1 76-7; Ramm 1 978, 4 0-1). Although t he c onsideration o f ' local' factors i nvolves a whole r ange o f c ircumstances which may b e r elevant t o t he p rovision o f d efences, i t i s p erhaps against a ' local' background t hat t he defensive s equence o f Brough-on-Humber a nd t hat o f many other c entres a re b est understood. T here i s a lso t he s uggestion t hat, e ven a t t his e arly date, the s iting o f B rough on t he Humber e stuary m ight have made i ts p rotection e xpedient ( Ramm 1 978, 3 9-40). I n f act, this whole question o f t he early development of a system o f coastal defences m ay p rovide t he r elevant c ontext f or a n umber of t imber and earth c ircuits ( Johnson 1 976, 1 5-20) . The e xplanation w hich h as b een o ffered by D rury ( 1975) f or t he defences at Chelmsford c oncerns t he question o f i ts vulnerable position c lose t o t he e ast c oast a nd t he p ossibility t hat e ven before the end o f t he 2 nd c entury E ssex, K ent and E ast Anglia were b eing t hreatened by t he a ctivities o f s ea r aiders. I t was t hese a reas t hat t he coastal f orts of t he early 3 rd c entury were d esigned t o p rotect a nd i t may h ave b een t he materialisation o f t his t hreat which p rompted t he p rovision of defences a t C helmsford a nd p ossibly a lso a n e nigmatic e arthwork c ircuit a t Wickford e nclosing a v ery small a rea, w hich was ' begun i n t he 2 nd c entury and n ever completed' ( Rodwell 8 1 975, 8). Unfortunately, v ery l ittle i s k nown a bout t he defences i n t he t erritory o f t he T rinovantes and i t i s difficult t o p lace t he Chelmsford c ircuit i n a s ecure a nd dateable

context

( Rodwell

1 975,

8-92) 8

The a rguments q uoted a bove i ndicate t hat l ocal f actors a y b e relevant t o t he p rovision o f t imber and earth c ircuits particularly w here t hese d efences p re-date t he l ate 2 nd

,

century. earthwork parallels

The e xamples mentioned make a l ong t radition d efences s eem v ery p lausible d espite t he l ack f rom o ther p rovinces ( Frere 1 978, 2 83).

o f o f

There i s a f urther p ossibility t o b e c onsidered i n onnection w ith t he a rgument f or a l asting and widespread t radition o f e arthwork d efence. U ntil Hartley's ( 1983) r ecent paper, the f ramework r egarded t he majority o f e arthwork c ircuits as t he p roduct o f t he A lbinian episode. The a rgument that t hese were e rected only a s emergency measures r ests partly on t he f act t hat t hereafter defensive c ircuits f or t owns and v illages were c onstructed i n s tone ( Wacher 1 964, 1 05; 1 974, 7 2; 1 978,96; F rere 1 978, 2 86 & 2 88). B ut a t erminus p ost quem i n t he l ate 2 nd century would a llow t he possibility t hat s ome e arthwork c ircuits may b elong t o a date later than t he A lbir iian episode. S o f ar t here i s only one c lear case o f a n e arthwork c ircuit w hich c ertainly p ost-dates this c ontext. This o ccurs a t t he Bays Meadow v illa complex, D roitwich, which h as r easonably b een i nterpreted a s t he r esidence of a government official c onnected w ith the a dministration o f t he s alt i ndustry t here. This v illa n ucleus

4 9

r emained u ndefended u ntil a double d itch s ystem w as e rected o n two s ides and a r ampart on a t hird s ide s ometime b etween c . AD 2 70 a nd 2 90. The c onstruction o f defences w as possibly connected w ith t he r eplanning o f t he . complex at t his time. I ts c onstruction may p erhaps h ave b een n ecessitated by t he general i nsecurity of t he p rovince i n t he l ate 3 rd c entury o r more s pecifically during t he u surpation o f C arausius ( AD 2 862 93) ( WMNS ( 1975) 1 8, 4 8; ( 1976) 1 9, 4 2-4; F reezer 1 977, 5 -7). Of c ourse, i t s hould b e n oted t hat t he p rovision o f defences f or a government i nstallation i s a r ather d ifferent matter f rom t heir p rovision f or c ivil s ettlements b ut t he u se o f t imber and earth defences here i n t he 3 rd c entury and t he e arlier e arthwork c ircuits a t B rough-on-Humber, C olchester, E xeter, S ilchester, Verulamium and W inchester show t hat t he t radition o f t his p articular t ype o f d efence w as w ell established and widely u sed. I n t he l ight o f this e vidence, t he i dea t hat e arthwork c ircuits w ere s imply t emporary measures needs t o b e r e-examined.

3 ) p rior

D id s ome c entres b egin t o t he l ate 2 nd c entury?

t o

c onstruct

d efences

i n

s tone

There i s a c ertain amount o f e vidence which c ould b e i nterpreted, i n t he p resent s tate o f k nowledge, a s i ndicating t hat p rior t o t he l ate 2nd c entury s ome centres were p rovided w ith

d efensive

c ircuits

w hich

i ncluded

e lements

i n

s tone.

Wacher h as d rawn a ttention t o a n umber o f centres where s tone gateways w ere b uilt p rior t o t he w alled c ircuit and t he f act t hat, i n s ome cases at l east, i t can b e a rgued t hat these e xisted i n a ssociation w ith t he e arlier t imber a nd e arth c ircuit. The e xamples s o f ar a re :C irencester

-

Gloucester

S ilchester

-

-

Verulamium ( north-east) Gate ( Wacher . 1 09; 1 966, 6 1-2 & 6 6; 1 974, 3 02) North 2 73)

a nd

e ast

North, s outh, ( Wacher 1 966,

g ates

( Britannia

e ast, w est 6 1-2, 6 3

B ritannia ( 1976) F ulford 1 983).

7 ,

1 964,

( 1975)

6 ,

a nd s outh-east gates & 6 6; 1 974, 2 64-5;

3 68;

( 1977)

8 ,

4 18-19;

P ossibly a lso Caerwent s outh gate ( Arch. Camb. ( 1954) 1 03, 5 4; Wacher 1 974, 3 83). At Verulamium t he gates s eem to b e s tructurally e arlier t han t he w alled c ircuit a nd i t h as b een suggested t hat t hey may b e contemporary w ith t he F osse earthwork ( Frere 1 964a, 6 9-71; 1 964b, 1 08-9; Wacher 1 966, 6 1-2, 6 3 & 6 6; 1 974, 2 13-14); and, a t E xeter, t he s outh gate w as built -

b efore t he w all b ut i t h as b een s uggested t hat t here m ay have b een a n i ndependent earthwork p hase ( Wacher 1 974, B idwell 1 980, 6 0 & 6 4). At t he

C irencester

s tone

gate

was

i t

i s

p ossible

t o

s ay

w ith

s tratigraphically e arlier

5 0

certainty t han

t he

n ot 31; 3

t hat

rampart

s ince t he l atter s ealed i ts f ootings ( Ant. J . ( 1961) 4 1, 6 8-9 & 7 1; Wacher 1 964, 1 05-6 & 1 09; 1 974, 3 02). E xcavation o f t he s outh gate a t S ilchester h as a lso s hown t hat i t was s tanding when the r ampart was c onstructed. The wall overrode t he f ootings o f t his g ate a nd a butted t he s outh-east g ate a nd Fulford ( 1983) has s uggested t hat t his gate a nd t hose on t he north, west a nd e ast , w ere p robably a lso s tanding w hen t he rampart was c onstructed. I t has been a rgued t hat t hese stone gates p erhaps b elong t o t he s ame phase o f c onstruction a s t he earthwork c ircuit o f t he l ate 2nd century. There i s a lso evidence f or t he e xistence o f s tone t owers i n a ssociation w ith t imber and earth c ircuits at Cirencester and G loucester. At Cirencester, a t l east o ne h as s o f ar b een i dentified ( Ant. J . ( 1964) 4 4, 1 6; ( 1966) 4 6, 2 42-3; ( 1967) 4 7, 1 90; B ritannia ( 1970) 1 , 27), w hile a t G loucester, t here a re f our s tone 2 i nterval t owers ( Ant. J . ( 1972) 5 2, 2 9-31 & 3 6; B ritannia ( 1970) 1 , 2 93; ( 1971) 2 , 2 75). At D orchester ( Dorset), a Chalk and f lint f oundation observed i n two p laces s ealed beneath t he t ail o f t he r ampart i s r egarded by Wacher a s a possible i ndication o f an i nitial attempt t o p rovide a masonry curtain wall ( Wacher 1 974, 3 21). Wacher h as a rgued t hat t he e xplanation f or i n s tone l ies i n t he f act t hat t he p rogramme o f implemented by A lbinus i n t he emergency o f AD

t hese f eatüres f ortifications 1 93-196 must

have i nitially e nvisaged t he c onstruction o f d efences i n s tone but i ncreased u rgency meant t hat banks and d itches had t o b e p rovided i nstead ( Wacher 1 966, 6; 1 974, 7 5; 1 975, 5 1; 1 978, 6 9 7-8) The i mplications o f t he p rovision o f gateways i n s tone need to b e f ully a ssessed i f one i s t o accept t he a rgument t hat t hey i ndicate t he i ntention t o c onstruct d efences i n stone. I t has t o b e a ssumed f or i nstance t hat t hese f eatures would have b een c onstructed a s a matter o f c ourse b efore t he curtain wall r ather t han a s a n i ntegral part o f i ts cons truction. The p rovision o f gateways i n s tone n eed n ot i n i tself i mply t he i ntention t o b uild t he e ntire c ircuit i n stone. As t he most vulnerable points i n t he c ircuit the u se o f s tone r ather t han t imber may h ave b een e xpedient e ven w hen t he r est of t he defences were built i n earth and t imber. However, t he n umber o f c ases w here gateways w ere c onstructed i nitially i n t imber and r ebuilt i n s tone only when t he main c ircuit was r econstructed i n s tone t ends t o s upport Wacher's r easoning t hat t he c onstruction o f s tone gateways i ndicates t he i ntention t o e rect a c urtain wall. But t here i s a v ery considerable amount o f e vidence t o s uggest t hat e ven where t he gates a re r egarded a s b elonging t o t he s ame p hase o f c onstruction a s t he w all, t he gates were often b uilt s eparately a s a t Caerwent e ast a nd w est gates ( Wacher 1 974, 3 83). G ateways were a lso s usceptible t o considerable r ebuilding and a lteration s o t hat t heir i ndependence o f t he main s tructural sequence i s a natural phenomenon. I t cannot be a ssumed t hat t he o nly e xplanation f or t he c onstruction o f s tone g ates i n t imber and earth c ircuits i s t hat t hey r epresent a n i nitial i ntention t o b uild a walled c ircuit w hich h ad t o b e h astily completed i n cheaper and more r eadily a vailable materials.

5 1

At S ilchester, t he a bsence o f a f ront r evetment o r t imber s tructure w ithin t he body o f t he I nner Defence r ampart, a t l east i n t he e xcavation o n t he e ast s ide i n 1 974, w as i nterp reted by F ulford a s a possible i ndication o f u rgency. I t i s i nteresting t hat t he S ilchester e vidence f or t he b uilding o f s tone gates prior t o, but possibly a s part o f, t he s ame c ons truction p hase a s a h astily e rected r ampart w ould f it Wacher's hypothesis. H owever, F ulford r aised the q uestion o f w hether t he b uilding o f gates i n s tone r ather t han t imber a t S ilchester at t his stage might n ot h ave b een the r esult o f woodland c learance i n t his a rea a f eature which h as b een demonstrated by pollen analysis. Of c ourse, this m ay account a lso f or t he a bsence o f t imber i n t he r ampart c onstruction ( Fulford 1 983) -

Even w here t he s tone gates c an b e s hown t o b e s ealed b y t he r ampart and t herefore r epresent t he p rimary element i n t he s ystem, a lternative e xplanations may b e f ound. I t i s possible t hat s ome of t hese stone gates, s uch a s t he Verulamium Gate a t C irencester, may r epresent f reestanding g ates d emarcating entrances i nto t he t own comparable p erhaps w ith t he Balkerne Gate a t C olchester a nd w ith t he t wo m onumental a rches e rected at Verulamium a t points where t he Watling S treet h ad passed t hrough t he i nitial e arthwork c ircuit ( '1955' ditch) ( Frere 1 974b, 1 04 & 1 09; Wacher 1 974, 2 05 & 2 15; C rummy 1 977, 9 7). Wacher a rgued t hat t he c onstruction o f t hese g ates a t C irencester and Verulamium c ould n ot b e r egarded a s f rees tanding s tructures b uilt s imply t o p rovide i mpressive entrances i nto t he t own s ince t he character o f these g ates w as c learly d efensive ( Wacher 1 966, 6 3, 6 5-6). Wacher ( 1964, 1 09) considered t he possibility t hat t he a nomalous p lan of t he Verulamium Gate a t C irencester m ight b e e xplained b y t he p rior e xistence of a monumental a rch which was s ubsequently i ncorporated i nto t he c ore o f t he g ate. There i s n o r eason however why l arge s tone gates i n a d efensive s tyle s hould n ot h ave b een e rected a s f reestanding e ntrances. On t he other h and, t he p rovision of i nternal t owers i n s tone might b e t hought t o b e a r easonably c onclusive i ndication o f t he i ntention t o p rovide, at l east u ltimately, a curtain w all. Two c entres a re k nown s o f ar t o h ave e rected s tone i nterval t owers w ithin t imber a nd earth circuits. A t Cirencester, t he s tone t ower o n t he n orth-east c ircuit j ust n orth of t he Verulamium G ate was c onstructed b efore t he r ampart w as p iled u p a round i t ( Ant. J . ( 1967) 4 7 , 190; Britannia ( 1970) 1 , 2 27) . I t i s t herefore r easonable t o assume t hat t he c onstruction o f t he Verulamium Gate m ay w ell h ave b een c ont emporary w ith t he construction o f t he i nterval t owers a nd part o f a p rogramme w hich i s l ikely t o h ave envisaged t he entire c ircuit i n s tone. At G loucester, t he e vidence s uggests t hat t wo , s tone i nterval t owers a nd a p ossible f urther t wo e xamples were constructed a s part o f t he massive r efurbishing o f t he e xisting l egionary f ortress r ampart. The b ody o f t his s econdary r ampart c ontained p redominantly m id 2 nd century pottery b ut apparently s ome l ate 2 nd c entury p ottery a lso, a nd a date i n t he l ate 2 nd century was t herefore a ccepted f or t his r efurbishing. The s ituation h ere i s c omplicated t hat i t i s not e ntirely c lear whether a walled

5 2

b y t he c ircuit

f act w as

a lready i n t he p rocess o f b eing b uilt a t t his t ime ( Ant. J . ( 1972) 52, 2 9-31 & 3 6; Britannia ( 1970) 1 , 2 93; ( 1971) 2 , 2 75; Wacher 1 974, 1 45-6). Stone gates on n orth a nd e ast a d t he a djacent s hort s tretches o f curtain wall would s eem t o h ave b een e rected a t a date e arlier i n t he 2 nd c entury a nd t he c onstruction o f i nterval t owers i n s tone, whatever t he specific occasion, would b e a l ogical part o f t his p rogramme ( Britannia ( 1975) 6 , 2 73). T hese a re t wo i nstances w here t he a dditional p resence o f stone i nterval t owers l ends s ome s upport t o t he hypothesis o f a n o riginal i ntention t o b uild i n s tone. H owever, i t s hould be r emembered t hat t here i s n o e vidence t hat s tone i nterval t owers would n ecessarily b e t he p rimary e lement i n t he c onstruction of a walled c ircuit. The e vidence i s d ifficult because t here h as b een i nsufficient i nvestigation o f t hose circuits where a wall was c onstructed de n ovo t o demonstrate whether s tone t owers w ere c onstructed b efore o r a s a n i ntegral part o f t he building o f t he w alled c ircuit. Where a w alled c ircuit with s tone t owers was a dded t o a p re-existing r ampart t he r elationship b etween wall a nd i nterval t owers i s p erhaps unlikely t o b e t he s ame a s i n t hose cases where there was n o p re-existing r ampart. I n t he f ormer c ase, i t would s eem more logical t o i nsert t he s tone t owers i nto t he r ampart o nce the wall had b een b uilt s ince a ny d isturbance t o t he r ampart which was i nvolved i n t he c onstruction of t he t owers could b e r etained by t he w all. This s equence h as b een o bserved a t L incoln ( JRS ( 1965) L5, 1 81; Ant. J . ( 1975) 5 5, 2 37 & 2 54). Also, while Wacher's a rgument w ith r egard t o t he work b egun i n masonry may perhaps b e t entatively a ccepted i n t he c ase o f Gloucester a nd C irencester, other c entres w here s tone g ateways h ave been f ound i n c onjunction w ith e arthwork c ircuits w ill have to be e xamined t o establish t he c lose c ontemporaneity o f gateway a nd r ampart a nd t o d emonstrate t hat t here i s n o possibility of t he gateway h aving b een e rected as a f rees tanding

e ntrance.

I t i s a nother matter e ntirely t o a ccept t hat w ork i n m asonry can b e a ttributed t o A lbinus. F irst o f a ll, difficult t o a ccept t hat A lbinus m isjudged t he the e xtent t hat t ime and r esources f ell s o f ar

b egun i t i s

s ituation short of

t o t he

t ask t hat i t p roved i mpossible t o c onstruct a nything b ut g ates and i nterval t owers i n s tone. Wacher a rgues t hat t he major t owns were g iven p riority i n A lbinus' s cheme, t here b eing n o i ndication t hat w ork was begun i n masonry a t a ny o f t he ' lesser' c entres. He s uggests t hat t he w alled c ircuit a t London may have b een e rected a s part o f t his p rogramme s ince the d efence o f t he p rovincial ' capital' would h ave b een t he f irst concern ( Wacher 1 974, 9 4-5; 1 975, 5 1; 1 978, 9 8) E ven i f i t i s a ccepted t hat w ork i n s tone w as b egun a t t he major centres a s a matter o f p riority, o r e ven t hat i t was only i n t hese c ases t hat s tone d efences w ere i ntended, t he progress made by t he t ime t he s ituation b ecame s ufficiently u rgent t o n ecessitate a c hange o f p lan i s h ardly c onsistent with there ever h aving been a t hreat a t a ll. S ince A lbinus knew i n AD 1 94 t hat h e was l ikely t o h ave t o w ithdraw t roops f rom t he p rovince t o s upport h is c laim, he c annot h ave t hought

5 3

t hat h e would h ave a g reat deal o f t ime i n w hich t o make h is p reparations f or t he d efence o f t he t owns a nd villages o f t he p rovince ( Frere 1 978, 1 94). F or t he work o n t he c onstruction of stone defences t o have begun early i n AD 1 94 a nd t o h ave p rogressed s o l ittle by t he t ime i t b ecame o bvious t hat e arthworks would have t o be constructed i nstead, it w ould mean e ither t hat t he t owns i nvolved w ere v ery s low t o i mplement t he s cheme, perhaps f inding i t t oo c ostly a n enterprise t o f inance o ver a s hort p eriod, o r A lbinus m iscalculated t he i mmensity o f t he manpower and r esources which such a p roject would i nvolve. Wacher i s i nclined t o t he l atter v iew b ut i t s eems h ardly c redible t hat s uch a b asic miscalculation was possible ( Wacher 1 978, 9 8). I t m ight b e a rgued t hat t he work i n m asonry may have b egun early i n AD 1 93 f ollowing t he death of t he emperor C ommodus a t t he e nd o f AD 1 92 a nd t hat i t w as t he d efeat o f another c ontender f or t he i mperial t hrone e arly i n AD 1 94 by S eptimius S everus, t he t hird c ontender, w hich n ecessitated e xtreme u rgency and t he change t o e arthwork c ircuits. I n w hich case, a y ear would h ave e lapsed during which t ime w ork i n s tone m ight have been i mplemented. I t s eems much more f easible t o s uggest t hat, a ssuming A lbinus had w ished t o i mplement a w idespread p rogramme of d efences f or t he t owns a nd v illages o f t he p rovince, h e w ould have b een aware t hat i n a s hort t ime a nd w ith t he a vailable r esources t his would h ave t o b e a chieved b y t he c onstruction of earthwork defences. This method o f defence, f ar f rom b eing u nusual i n t he p rovince, was a lready i n u se a t l east a t s uch c entres as B rough-on-Humber, S ilchester, Verula inium a nd p ossibly C olchester. The i ndications a t t he various c entres m entioned a bove o f work begun i n masonry might b e t hought t o undermine t he i dea t hat a n emergency s ituation w as r esponsible i n e very c ase, apart f rom t he c oloniae, f or t he i nitial e rection o f a defensive c ircuit. Wacher a rgues t hat t he c ircumstances o f t he A lbinian episode c an b e i nterpreted i n a way w hich e xp lains t he c onstruction o f e arthworks i n a ssociation w ith t hese e lements i n s tone ( Wacher 1 974, 7 5; 1 978, 9 7-8). B ut i t i s e qually valid t o s uggest t hat s ome c entres may h ave b egun t o e rect defences i n s tone at a date p rior t o t he l ate 2 nd c entury a nd f or r easons u nconnected w ith t he A lbinian episode. I n which case, t he i nitial c onstruction of gates and e ven t owers i n s tone a t f ixed p oints a long t he p lanned a lignment would have p rovided a u seful demarcation f or t he c urtain w all. A lthough i t might b e a rgued t hat t he a im i n t hese i nstances must s urely have b een t o c onstruct a c ircuit i n s tone r ather t han a t imber a nd e arth d efence, t his may h ave b een o nly t he u ltimate a im. A t imber and earthwork c ircuit may h ave b een r egarded a s a s uitable i nterim measure u ntil r esources w ere a vailable t o complete t he work i n s tone, and i n t he meantime s tone gates would s erve a s u seful m arkers a nd p rovide a n e lement of c ivic p ride. The e xample o f G loucester i s perhaps a c ase i n p oint. H ere t he n orth a nd t he e ast gates w ere b uilt apparently a t s ome t ime i n t he early 2 nd century b ut t here i s a s trong p ossibility t hat t he walled c ircuit w as n ot c ompleted u ntil much l ater, perhaps even a s l ate as the e arly 3 rd c entury ( Britannia ( 1975), 6 , 2 73) a nd t he s ame e xplanation

5 4

may a pply t o Colchester ( Crummy 1 977). Jarrett's ( 1965, 5 8) comment t hat considerations of f inance a nd manpower would s uggest t hat t he b uilding o f t own walls was p lanned t o t ake a number of years i s certainly valid. Therefore i t can hardly be t hought l ikely t hat A lbinus would h ave u ndertaken s uch a s cheme of walling e ven i f he i ntended t he p rovision o f s tone c ircuits t o b e r estricted t o t he major t owns a lone. The possibility t hat an earthwork c ircuit was b uilt a s t he f irst s tep i n t he c onstruction o f a c ircuit i n s tone w as suggested by Jarrett ( 1965, 5 7-8) . F rere a nd Wacher cons idered t his u nlikely o n s tructural g rounds s ince t he a lterations which t he construction of t he wall i nvolved, e ither i n t he c utting b ack o f t he r ampart o r i n t he i nfilling of t he ditch, s uggested t hat t he p rovision o f a wall was n ot envisaged i n t he o riginal s cheme ( Frere 1 965, 1 37-8; Wacher 2 1 965, 25). Nevertheless, Jarrett's point i s worth c ons ideration i f o nly b ecause h e was a ttempting t o d emonstrate t hat each case s hould b e c onsidered o n i ts own merits ( see a lso B idwell 1 980, 6 0-2) I t i s a lso worth n oting t hat t here a re a n umber o f walled circuits, apart f rom t hose o f L ondon and possibly t he l ower colonia a t L incoln, which c ould c onceivably b elong o n p resent evidence t o t he l ater 2 nd century r ather t han t he 3 rd. There i s t he walled c ircuit a t E xeter, w hich was o riginally t hought to c onsist of a n i ndependent earthwork t o which a wall had s ubsequently b een a dded. B ut B idwell h as r ecently s uggested that wall and bank may be contemporary; material a ssociated with both f eatures p rovides a l ater Antonine t erminus post quem, a lthough t he a ctual construction date c ould have b een l ater i n t he 2 nd c entury ( Bidwell 1 980, 4 6, 6 0-2, 6 6). The o ther e xamples a re ( Alchester, G reat Casterton f ramework does n ot a llow f or

a ssociated w ith ' lesser' centres a nd Water N ewton) a nd t he c urrent t he possibility t hat t heir walled

c ircuits may h ave b een p rovided b efore t he seems hardly l ikely t hat walled c ircuits

3 rd c entury. would have

I t b een

completed a t t hese c entres a s p art o f A lbinus' s cheme w hen they were abandoned at s uch an early s tage at t he major centres. A l ate 2 nd c entury date f or t he w alled c ircuits o f these ' lesser' c entres i s t herefore n ot a cceptable w ithin t he current f ramework ( Wacher 1 975, 5 1; 1 978, 9 7-8). F urther w ork i s n eeded a t each o f t hese t hree s ettlements t o establish more s ecure dating e vidence. At b oth A lcester a nd Water N ewton there was n ot e vidence f or an i ndependent earthwork p hase and t he s ame h as b een a rgued t o b e t he c ase a t G reat Casterton. I n e ach c ase t herefore t he wall and bank a re r egarded a s contemporary. T he r esults f rom G reat Casterton a nd Water Newton a re i n n eed of c onfirmation s ince t hey were obtained s ome t ime a go a nd a t Water N ewton t he i nvestigation h as b een very l imited i ndeed. The b est a nd most r ecent e vidence c omes f rom Aichester w here, i n t he e xcavations o n t he e astern defences i n 1 974, ' pottery f rom t he r ampart dated i ts construction t o t he e nd o f t he 2 nd c entury' ( 1976, 1 39-40; Britannia ( 1975) 6 , 2 56). S imilarly, a t Water N ewton, pottery under t he r ampart e rected b ehind t he w all ' dated b oth t o t he

5 5

2 nd c entury' ( JRS ( 1958) 4 8, 1 39). P ottery s ealed b eneath t he f ootings o f t he wall at G reat Casterton was d ated n ot l ater t han AD 1 70 a nd t he l ater 2 nd c entury . p ottery s ealed beneath t he r ampart a lso s uggested a date f or t he wall ( JRS ( 1951) 4 1, 1 27; ( 1954) 4 4, 9 2). The

dating

e vidence

f or

t hese

w alled

c ircuits

i s

only i n

t he nature of a t erminus post quem and t here i s no e vidence a t p resent which would p revent t hem f rom b eing l ater i n d ate t han t he 2 nd c entury. H owever, t his e vidence i s worth b earing i n mind i n v iew o f c urrent t hinking a bout t he p rovision o f s tone walls i n t he l ate 2 nd century and l ater contexts. F or, i f a l ate 2 nd c entury date i s e stablished f or t hese t hree c ircuits, i t may be necessary t o consider t he p ossibility t hat t hese c entres w ere walled a s a r esult o f e vents u nconn ected w ith t he A lbinian episode. The i ndications o f work b egun i n m asonry a t s everal o f t he major c entres may t herefore a lso b elong t o a s cheme s eparate f rom t hat i mplemented by A lbinus. L eaving a side t he e vidence f rom A lchester, G reat Casterton and Water N ewton, i t i s f easible t o suggest, g iven t he p resent s tate o f k nowledge, t hat d efensive e lements , i n s tone at t he major centres may point t o e ither the d eliberate s trengthening o f c ertain vulnerable f eatures w ithin a circuit which was i ntended otherwise t o b e c onstructed o f e arth a nd t imber o r t he i ntention t o c onstruct a c ircuit i n s tone. I f t he l atter a lternative i s accepted t hen t he c onstruction o f a r ampart c ould b e e xplained e ither b y t he n eed t o r etain t he well-established method of c onstruction u ntil f inances p ermitted c ompletion i n s tone, o r t he n ecessity o f completing t he c ircuit i n earth and t imber i n t he emergency of t he l ate 2 nd c entury. The p ossibility t hat s ome c entres may h ave b egun t o e rect defences at a date p rior t o t he A lbinian episode and a s a n i ndependent a ction c onnected p erhaps w ith t he a spirations of t he i ndividual s ettlement o ffers a f undamentally d ifferent e xplanation f or t he p rovision o f d efences. The i mplications of t his w ill b e e xamined b elow when v arious p ossible reasons f or t he c onstruction o f d efences a re d iscussed. At p resent, t he r easons and contexts put forward for t he p rovision o f d efences i n R oman B ritain a re very r estricted and, i nvolve r esponses t o specific t hreats and except in t he c ase of t he c oloniae, t hey a re r egarded a s h aving n othing t o do w ith t he i nterests o r aspirations o f t he i ndividual settlement. H owever, t here a re o ther f actors w hich a re w orth c onsidering.

4 ) I s i t p ossible t o r egard t he c onstruction o f s uccessive earthwork c ircuits at c ertain c entres i n the i st a nd 2 nd c enturies a s a n a ttempt t o d efend t hese c entres o n a c ontinuous b asis?

The

a rgument

a continuous c onstruction

and i n

a ccepting

more

a

h as

b een

p ut

f orward

i n

2 )

a nd

3 )

a bove

f or

c hronologically d iverse h istory o f defence t he p rovince. A f undamental difficulty i n c ontinuous

5 6

d efensive

h istory

f or

centres

where e arthwork c ircuits were e rected i n t he i st a nd e arly 2 nd centuries i s t he f act t hat t imber and earth c ircuits a re more d ifficult t o k eep i n a d efensible s tate t han s tone walls a nd r equire f airly continuous maintenance. Being c omparatively i nexpensive t o c onstruct, t hey a re more s uitable a s s hort-term expedients t han wails but a s l ong-term f ortifications t hey a re l ikely t o r equire g reater e ffort i n maintenance t han s tone wails. H owever, considered:

t here

a re

f our

p oints

which

n eed

t o

b e

i ) I t h as a lready b een mentioned t hat t he c onstruction s tone walls may n ot h ave b een v iable u ntil r esources manpower and f inance c ould be f reed f rom t he c onstruction t he m ajor public b uildings.

o f o f o f

i) A t imber a nd e arth c ircuit c ould b e more e asily a djusted i or p laced on a c ompletely n ew a lignment t han c ould a s tone wall. Earthwork d efences w ere t herefore much b etter s uited t o the e arly s tages o f development when e xpansion o r c ontraction o f t he s ettlement would b e h ampered by a more p ermanent stone f ortification. ii) i

Earthworks

which which

the native population was t hey w ere h ighly c ompetent.

r epresented

a

t radition very

o f

d efence

f amiliar

and

w ith i n

i v) A lthough a t imber and earth c ircuit r equired very considerable maintenance, i t was p robably n ot a s e xpensive t o maintain a s a walled c ircuit. F rere ( 1978, 2 86 & 2 88) and Wacher ( 1978, 9 6) have t ended t o a rgue a gainst t he i dea o f a ny l asting a ttempt t o maintain early defensive c ircuits a s permanent f eatures, w ith t he possible e xception o f t he c oloniae. The e arly t imber a nd earth defences a re s een as e ssentially s hort-lived f eatures r esponding t o particular t hreats. This emphasises t he p oint that permanent d efences i n t he f orm o f s tone walls may have been r egarded a s i mpractical f or developing c entres, a lthough some s ort o f enclosure was t hought t o b e expedient. H owever, i t i s d ifficult t o d ecide w hich i s t he more r elevant; w hether the t imber and e arth c ircuits go out of u se comparatively quickly b ecause t hey r epresent h asty work r elated t o p eriods of i nsecurity o r whether t hey a re s oon outgrown by t he developing n ucleus, i n w hich c ase t hey may b e r eplaced b y a new e nceinte. This l atter i dea h as been considered b ut w ithout i ts i mplications b eing e xplored ( Wacher 1 978, 9 6). Unfortunately, only i n t he case o f E xeter ( Bidwell, 1 980, 4 6) i s t here s ufficient e vidence t o e stablish t hat a ttempts w ere made t o maintain t hese early c ircuits. T he

s equence

o f

e arly d efences

t entatively p roposed

b y

C rummy ( 1977) f or C olchester p resents a p icture of s uccessive c ircuits e nclosing a s ettlement w hose s ize was i n a s tate o f f lux. The i nitial enclosure i s r epresented by a d itch dug p robably sacked.

a fter This

t he Boudiccan r ebellion d itch was f ound on t he

5 7

w hen west

t he c olonia w as b eyond t he l ater

c olonia wall ( Crummy 1 977). No t race o f a r ampart s urvived but t his may have been destroyed by l ater a ctivity. The d itch h as n ot b een l ocated e lsewhere b ut wor .k on t he development o f t he s treet p lan s uggested t o C rummy t hat t he n orthern a nd e astern a lignments o f t he e arly d efensive c ircuits may b e f ound w ithin t he l imits of t he l ater walls. This d itch w as b ackfilled c . AD 7 5; e xpansion t ook p lace t o t he w est a nd a n ew d itch may have been dug f urther o ut t o enclose i t f or an e ast/west d itch was d iscovered o n t he s outhern l ine o f t he l ater c olonia wall b ut e xtending f urther t o t he w est o f t he walled c ircuit. Another ditch r unning east/west w as f ound i n t he n orth-western part o f t he s ettlement n orth o f i nsula 9 b. No d irect dating evidence was r ecovered b ut a deposit dated t o AD 6 5-120 s ealed i t, a nd a l ate F lavian o r T rajanic date w as o riginally s uggested f or i ts c onstruction. This w ould t end t o p lace t his d itch i n t he context o f t he s econd defensive c ircuit which, C rummy s uggests, was e rected c . AD 7 5-100, but t he e vidence i s i nconclusive a nd C rummy i nclines t o t he v iew t hat i t b elonged t o the i nitial enclosure e rected c . AD 6 0-61. H e s uggests t hat t his e nclosure was l ikely t o h ave e xtended a s f ar east as i nsula 1 4 where t he s hape o f t his i nsula could b e e xplained by t he e xistence o f d efences o f t he p ost-Boudiccan period. This enclosure was backfilled c . AD 75 t o a llow t he e xpansion o f t he s ettlement t o t he n orth a nd w est and t he d istribution of f inds of t he F lavian and T rajanic periods ( late 1 st/early 2 nd c entury) s eems t o s uggest t hat t his h ad perhaps t aken p lace by c . AD 1 00. This e xpansion was t hen consolidated o n t he n orth w hen t he c olonia wall w as e rected i n t he early part of t he 2 nd c entury, a lthough the d ate of t he n orthern part o f t he walls h as y et t o b e e stablished. W hile on t he west t he l ine of t he defences c ontracted b ack t o t he l ine o f t he Balkerne Gate w ith t he c onstruction o f t he w alled c ircuit t hus l eaving t he western e xpansion o utside t he defences. C rummy a lso r aises t he possibility t hat t he Balkerne Gate was a lready i n e xistence a s a f reestanding monumental a rch when t he s tone wall w as e rected i n t he e arly 2 nd c entury a nd t hat t he o ccasion f or i ts c onstruction m ay have b een t he period c . AD 7 5-100 when t he c olonia w as e xpanding b eyond i ts o riginal l imits. I t m ight t herefore h ave b een e rected t o mark t he f ormer l egal l imits of t he c olonia i n much t he s ame way a s t he t wo monumental a rches at V erulamium. Apart f rom the i nitial c ircuit c . AD 6 0-61, t he s equence p roposed i nvolves t he c onstruction o f s uccessive n ew c ircuits t o a ccommodate t he changing t opography of the s ettlement r ather t han t o meet n ew t hreats. C rummy's hypothesis i nvolves one c ircuit f ollowing c losely on another, i mplying t hat f rom t he t ime t he i nitial c ircuit w as e rected c . AD 6 0-61 t he s ettlement was d efended c ontinuously a nd t he construction o f ' t he s tone wall i n t he early 2 nd c entury gave i t permanence'. The s equence, e rected i n c lose s uccession, i s i n k eeping w ith t he a cceptance of defence a s a permanent consideration and n ot a s a n e rratic emergency measure. At

Verulamium,

t he

F oss

e arthwork

l arger a rea t han t hat of t he o riginal p rojection o f t his l ater c ircuit o n s uggest t he n eed t o enclose an a rea o f

5 8

e nclosed

a v ery

m uch

d efences. T he angular t he s outh-west m ight s ome i mportance. H ow-

e ver, no e vidence h as b een f orthcoming t o s how w hether t his was t he case and when t he walled c ircuit was e rected t his a rea was l eft o utside a nd a n a dditional a rea e nclosed o n t he s outh ( Frere 1 964b, 1 08-9; Wacher 1 974, 2 14-15 & f ig. 4 7). U ntil f urther work h as b een done on t he e xpansion a nd c ontraction o f t he s ettlement i n r elation t o t hese c ircuits, i t i s d ifficult t o a rgue t hat s uccessive a lignments w ere r elated t o t he s ettlements i n t his way. F rere i s i nclined t o t he v iew t hat Verulamium was n ot c ontinuously d efended f rom t he t ime w hen t he f irst earthwork c ircuit was e rected and t his compares w ith S ilchester but c ontrasts w ith Colchester. F rere h as s uggested that t he i nitial earthwork ( '1955' d itch) a nd t he F osse r epresent c ircuits e rected i n r esponse t o s pecific t hreats a nd n ot a s part of an a ttempt t o p rovide c ontinuous defence f or a developing u rban n ucleus ( Frere 1 964a, 6 5 & 6 9). B ut h is more r ecent suggestion of an Antonine date f or t he F osse earthwork might a lter t his v iew ( Frere 1 983). A gain f urther work i s needed on t he defences at Verulamium t o confirm F rere's c onc lusions e specially r egarding t he date o f t he F osse e arthwork f or t he various s uggestions offered r ely on a r einterpretation o f t he l imited e xcavation o f t he e arthwork i n 1 932 ( Hartley, 1 983) I n h is e xcavations on t he s outh-west a ngle of t he Fosse e arthwork, W heeler ( 1936, 1 & 5 0) c oncluded t hat i t h ad been e rected i n t he t hird quarter o f t he i st c entury. The bank c ontained o nly n ative a nd i mported B elgic w ares a nd t he .

evidence on which Wheeler based h is F lavian date c ame f rom t he d itch s ilt w hich i ncluded B elgic p ottery a nd s amian o f t he m id i st c entury AD. F urther material i n t he s econdary d itch s ilting contained n ative a nd R oman wares o f t he s econd h alf o f t he i st century AD ( Wheeler 1 936, 5 1-4). The e vidence f or a F lavian earthwork c ircuit a t Verulamium a ccorded w eil w ith t he s ubsequent d iscovery by F rere of a n earthwork c ircuit (11 955' ditch) enclosing a much smaller a rea t han t hat o f t he F osse earthwork. The d ating e vidence f or t his c ircuit was s parse s ince t he b ank h ad n ot s urvived b ut F rere c oncluded t hat a C laudian date would n ot b e i nappropriate ( 1960, 2 -4; 1 961, 8 03 ; F rere 1 964a, 6 1 & 6 5; 1 964h, 1 04). on t he b asis o f t he material i n t he d itch s ilt, i t was was b eing a llowed t o s ilt u p i n t he i t h ad gone out o f u se by c . AD 1 25

s uggested t hat t he c ircuit p eriod AD 8 5-100 a nd t hat a fter which t hese defences

w ere l evelled. T he i nitial e arthwork c ircuit c ould t herefore be s een t o be going out of u se j ust a t t hat t ime when Wheeler postulated t he F osse e arthwork c ircuit was b eing c onstructed. Wheeler ( 1936, 5 1-2) a lso c onsidered t hat t he F osse earthwork was r efurbished s ubsequently by t he a ddition o f f urther material t o t he r ear of t he bank a nd t hat t his t ook p lace i n o r j ust before t he Antonine p eriod. H owever, F rere c onsiders that t he F osse e arthwork s hould be r egarded as a f eature o f t he 2 nd c entury r ather t han t he l ater i st c entury a nd h e suggested a t erminus post quer n i n t he Antonine p eriod on t he basis of t he p ottery o f t his date f ound i n w hat Wheeler r egarded as t he r efurbishing of t he o riginal bank. F rere considered t hat t hese ' additions' w ere i n f act p art o f t he o riginal c onstruction s ince t hey a dded n othing t o t he h eight a nd s trength o f t he b ank. H e a lso p ointed t o t he f act t hat the F osse earthwork i s r egarded a s u nfinished a nd a s s uch i t was u nlikely t o h ave b een r efurbished ( Frere 1 964a, 6 9; 1 964b, 1 08). This i nterpretation has been generally a ccepted and, i n

5 9

v iew o f s ubsequent t heories c oncerning t he w idespread c ons truction of earthwork c ircuits under Albinus, t he F osse earthwork was p laced i n t his l ate 2 nd . c entury context r ather t han i n t he Antonine context o riginally p roposed by F rere ( Wacher 1 974, 2 13; F rere 1 978, 2 85-6). According t o t his a rgument, Verulamium r emained u ndefended f rom t he t ime t he i nitial c ircuit w ent o ut o f u se i n t he l ate 1 st/early 2 nd century until t he F osse earthwork was e rected in t he l ate 2 nd c entury ( Frere 1 964a, 6 5; Wacher 1 974, 2 13). H owever, i t i s a lso worth n oting t hat F rere h as r ecently r eturned to h is o riginal s uggestion t hat t he F osse e arthwork may h ave b een under c onstruction during t he Antonine period ( Hartley 1 983). F urther work o n t he dating o f t he F osse e arthwork i s n eeded. I f Wheeler i s correct, t he s uccessive earthworks a t V erulamiuf f would p rovide t his u rban n ucleus w ith a c ontinuous defensive h istory f rom t he l ate 1 st century o nwards and one which i s c omparable w ith t hat s uggested f or Colchester. At B rough-on-Humber, E xeter, S ilchester and W inchester, t here i s l ess s cope f or d iscussion l argely b ecause t he s equences p roposed i nvolve only one t imber and earth c ircuit e rected p rior t o a s econd p hase o f d efence w hich i s a ttributed i n each case t o t he l ate 2 nd century A lbinian context ( Wacher 1 974, 2 64 & 2 83; F rere 1 978, 2 85-6). At B rough-on-Humber, a r elatively s hort period of t ime e lapsed b etween t he c onstruction o f t he f irst c ircuit s ometime a round t he mid 2 nd c entury and i t i s n ot c lear why i t w as necessary t o construct a n ew c ircuit. The d ecision t o c hange t he a lignment n o d oubt . had s omething t o do w ith t he n eed t o construct a n ew c ircuit . r ather t han r efurbishing t he e xisting d efences b ut t his d oe' n ot e xplain why a n ew a lignment was a dopted. This c hange o f a lignment i s a f eature a lso o f E xeter a nd o f t he l ater e arthwork c ircuits a t S ilchester ( Inner D efence) and V erulamiun ( Fosse e arthwork) . The e xplanation f or t hese c hanges m ay a gain be f ound i n t he e xpansion and c ontraction o f various p arts o f t he s ettlement. At b oth E xeter a nd S ilchester t h E ' n ew c ircuit enclosed a l arger a rea t han t hat of t he i nitia : . defence a nd, i n t he c ase o f S ilchester, t he l aying o ut o f t he s treet g rid i n t he l ate 1 st/early 2 nd c entury over t he I nne : E arthwork p resumably meant t hat t he s ettlement was i ntended t o e xpand b eyond t hese o riginal c onfines. This apparently i n .. volved r eturning t o a n u ndefended s tate s ince t he I nne D efence was n ot e rected a round t he l arger a rea u ntil t h Antonine p eriod a t t he e arliest. H owever, t he q uestion i . ; whether t he Antonine t erminus post q uem f or t his n ew c ircui: means t hat i t w as e rected p rior t o t he l ate 2 nd c entur/ ( Wacher 1 974, 2 64-5) At E xeter t here i s s ome e vidence t hat t he l egior iarj f ortress defences were r efurbished and r etained by t he developing c ivitas c apital u ntil t he Antonine p eriod a 1thoug -i t here w as ' clear evidence f or t he n eglect of t he e arly d efences p rior t o t heir demolition'. A n ew d efensive c ircuit ( with a t erminus post quem i n t he Antonine period) w as e rected e nclosing a n a rea t wo a nd a h alf t imes l arger t han t hat o f t he l egionary f ortress defences and i nvolving e ither a n i ndependent e arthwork o r a w all a nd t o know on p resent e vidence whether

6 0

r ampart. I t i s d ifficult t he i ndication o f neglect

o f t he defences a nd t he dating e vidence f or t he n ew c ircuit r epresent t he r eturn t o a n undefended s tate, and whether t his was i ntentional. There i s a lso i nsufficient e vidence t o s how how t he new c ircuit was r elated t o t he e xtension o f t he s treet g rid b ut c learly t he e nlargement o f t he a rea e nclosed w as a ssociated w ith t he e xpansion o r i ntentions f or t he e xpansion of t he s ettlement ( Bidwell 1 980, 4 6-7, 6 6-7). At Winchester t here was s ome e vidence f or t he decay o f t he d efences and i t was f elt t hat t hey were p robably s hortl ived. I n t he e xcavations on t he w est t he t urf s tack o f t he rampart had been c onsiderably e roded, a nd on t he s outh t he t imber gate h ad b een r emoved a t s ome s tage a nd a n ew s treet l aid a cross t he u nbarred gap i n t he r ampart. I t i s n ot k nown whether t he c ircuit e nclosed a l arger o r smaller a rea t han t hat of t he e arly defences b ut t he i nformation a vailable i ndicates a p eriod o f d ecay b efore r econstruction o n a n ew a lignment ( Ant. J . ( 1970) 5 0, 2 82; ( 1975) 5 5, 1 11-12). T he s light r eduction i n t he e nclosed a rea a t B rough-onHumber cannot be e xplained. Wacher s uggests t hat t he earlier c ircuit may n ever h ave b een c ompleted, o r t hat i t e nclosed a n i nconveniently l arge a rea, o r t hat i t w as perhaps t oo decayed t o b e a ny l onger d efensible ( Wacher 1 969, 2 9). The f irst t wo r easons a re n ot very convincing but s evere damage t o t he e xisting c ircuit m ay h ave b een a c onsideration e specially i f e ither of t he o ther t wo f actors was i nvolved. These c onsiderations o f e xpansion a nd c ontraction may b e t hought t o have been l ess p ressing i n an emergency when t he r efurbishing o f a n e xisting e arthwork c ircuit c ould b e a chieved more speedily t han t he c onstruction of a n ew c ircuit. An a cceptance o f o ther p ossibilities n ecessarily i nvolves a r econsideration of t he concept of defence a s a purely emergency measure, a nd a n e xamination o f t he f unction o f t hese defensive c ircuits i n t he developing t ownscapes o f t he i ndividual c entres.

5 .

What

r easons m ight

early defensive

a ccount

f or

t he

c onstruction

o f

t hese

c ircuits?

i ) Can t he p rovision o f and e arly 2 nd c enturies, t he u rban d evelopment?

d efences, e specially i n b e s een a s an i ntegral

t he part

1 st of

A lthough t he c urrent f ramework a ccepts t he e arly p rovision of defences f or t he coloniae, i t i s d ifficult t o e xplain t he p rovision o f t imber a nd e arth c ircuits a t E xeter, S ilchester, Verulamium and Winchester w ithin t he c ontext o f t he p remises h eld t o govern t he p rovision o f d efences ( Wacher 1 978, 9 6). E xplanations have been s ought i n t he i dentification o f a n u nusual s tatus f or t he s ettlement c oncerned and i n t he i nsecurity caused by a specific t hreat. But n one o f t hese e xplanations i s c onvincing a nd t he e vidence a llows different c ircumstances t o b e a rgued i n e ach c ase. The a rgument

f or

s pecial

s tatus

a nd

6 1

a g eneral

o r

s pecific

t hreat

i s u ndermined i n t he c ase o f E xeter w here t he e xplanation t he r etention of t he l egionary defences had t o be s ought p hilo-Roman s ympathies a nd a vague t hreat o r f eeling i nsecurity ( Bidwell 1 980, 6 6)

f or i n o f

The c oncept t hat s ome e arly d efensive c ircuits w ere p rov ided a s part o f official attempts t o develop certain centres a long R omanised l ines i s worth c onsideration. Vitruvius m akes i t c lear t hat d efences were s een as a n i ntegral part of u rban d evelopment a nd h e g ives t hem a h igh p riority i n h is a ccount of t own p lanning ( Boon 1 974, 6 5-6). W ith particular relevance t o R oman B ritain, Agricola ( AD 7 8-84) i s r eported b y Tacitus as h aving g iven official encouragement and f inancial a ssistance t o t he t ribes t o develop t heir t owns a long Mediterranean l ines ( Collingwood & M yres 1 937, 1 91, 1 94-5; F rere 1 978, 1 34). I t i s p ossible t hat A gricola may h ave e nvisaged t he p rovision o f d efences. I ndeed, F rere ( 1964a, 6 1-2; 1 964b, 1 04-7) s uggests, i n c onnection with t he f irst e arthwork c ircuit a t Verulamium, t hat t he i nitial p eriod o f u rban development under C laudius ( AD 4 3-54) may w ell h ave i ncluded t he c onstruction o f d efences. E ven i f i t i s a rgued t hat C laudius and A gricola r egarded t he p rovision o f defences f or d eveloping t owns a s a w ise p recaution i n a n ewly a nnexed p rovince whose f rontiers were s till i nsecure, the i mportance o f t his s tep s hould n ot b e o verlooked p articularly s ince i t i nvolves defensive c ircuits b eing p rovided not for w ell e stablished c entres w ith p roven e conomic p otential b ut f or embryonic t owns i n t he very earliest s tages of their growth. F rere's e xcavations a t Verulamium l ed h im t o t he c onc lusion t hat u rban development a long R omanised l ines was i n p rogress i n t he C laudian p eriod. A n umber o f s treets w ere l aid out i n a r ectangular g rid and a c olonnaded r ow o f t imber f ramed s hops w ere e stablished a long t he Watling S treet f rontage i n i nsula X IV. T races of b uildings of t his period h ave a lso b een f ound i n s ome o f t he o ther c entral i nsulae. A ll of t his w as destroyed by f ire during the Boudiccan r ebellion i n AD 6 0-61 ( Frere 1 964a, 6 3-5; 1 964b, 1 04-5). U nfortunately, t he i nitial earthwork c ircuit ( '1955 ditch') h as n ot b een p recisely dated b ut F rere a rgues t hat t hese defences p robably date f rom t he o riginal f oundation of t he t own c . AD 4 9 a s a municipium u nder C laudius. I ts s tatus a s a municipium i s s een as s ufficient r eason f or i t to q ualify f or d efences i n t he C laudian p eriod. At S ilchester t oo, t here i s s ome e vidence f or a n e arly phase which i nvolved a s cheme f or development a long Romanised l ines i n t he p eriod u p t o Cogidubnus' d eath i n c . A D 8 0. T he l aying out of t he k nown r egular S treet g rid i s attributed t o t he F lavian p eriod w hen i t i s t hought S ilchester w ould h ave b een c onstituted as a r egular c ivitas c apital following t he d eath o f C ogidubnus a nd t he amalgamation o f h is k ingdom i nto t he p rovince ( Boon 1 974, 4 5 & 5 3). L ittle i s known about t he e arly d evelopment o f S ilchester b ut t he c ase f or t he possible e xistence of a p re-Flavian system of s treets h as been made by F ox ( 1948, 1 72) b ased o n h er o bservations o f t he n on-alignment of t he

a c onsiderable n umber k nown s treet g rid a nd

of buildings i ncluding t he baths t he s ubsequent a daptation o f s ome

6 2

t o o f

these buildings t o t his r egular S treet l ayout. With t he discovery of t he I nner Earthwork defence, which Boon p laced i n t he p eriod s oon a fter AD 4 3, t his a rgument was t aken a s tage further. Boon ( 1974, 4 7) n oticed t hat, o f t hese u naligned buildings, s ome t hirty i ncluding t he b aths s eemed t o c orr espond t o a l ine d rawn between t he w est and t he e ast ( location o f l atter u nconfirmed) e ntrances t hrough t he I nner Earthwork and he s uggested t hat t his pointed t o a r ectangular partition o f t he s ite b ased o n t he l ine o f t he a ctual r oute f rom e ast t o w est. I ndications o f t wo s urfacings, u ndated, o f an e ast/west s treet on a s lightly different a lignment f rom t he visible main e ast/west s treet w ere s een i n a t rench a cross t he east c ircuit of t he I nner Earthwork ( Arch. ( 1969) 1 02, 9 -10) Otherwise very l ittle h as b een done t o i nvestigate t he existence of this postulated p re-Flavian s treet g rid. Boon considered t hat s uch a t own p lan c ould n ever h ave r eached t he stage of having metalled s treets between t he p lots; o therwise t here would h ave b een n o r eason t o c hange t he a lignment o f t he street plan l ater and, a lthough a r egular g rid could be postulated, t he v arying a lignment o f t he b uildings i mplies t hat t he o riginal l ayout was l argely h aphazard ( Boon 1 974, 4 7).

r oute

This model through

o f a n e arly t own p lan r elated t o t he e ast/west t he I nner E arthwork i s a u seful e xample of

defences b eing p rovided a s a n i ntegral p art o f e arly u rbanisation. The apparent a lignment of many of t he buildings, i ncluding t he b aths, i s on a l ine j oining t he e ast and w est entrances t hrough t he I nner E arthwork. B ut t he b aths a re i n f act u ndated, a nd p artly o verlie t he s outh-eastern l ine of t he I nner Earthwork ( Wacher 1 974, 2 57). I n a ddition, t he early east/west s treet p assed o ver t he d itch o f t he I nner Earthwork ( Boon 1 969, 1 0). This does of course make i t difficult t o a ccept t hat t he c onstruction o f t he I nner Earthwork c an have b een a n i ntegral e lement i n t he l ay-out. A r elevant

c ontext

f or

t he

p rovision

o f

t imber

a nd

e arth

defences f or B rough-on-Humbe r i n t he 2 nd century c ould be i n t he u rbanising p olicy o f Hadrian ( Ramm 1 978, 3 8-40).

s een

S imilar possible c onnections b etween t he p rovision o f early defensive c ircuits a nd o fficial e ncouragement o f u rban development can b e postulated i n t he cases o f Winchester a nd Exeter, a lthough m ore e vidence i s n eeded. I n t he f ormer c ase, the early earthwork ( terminus post quem c . AD 7 0) may b e connected w ith t he development o f Winchester a s a c ivitas capital a nd the l aying out of t he s treet g rid i n t he F lavian period, b ut s o l ittle i s k nown a bout t he e arly s tages o f development t hat s uch s uggestions a re pure speculation ( Wacher 1 974, 2 77-80). At E xeter, t he i nitial r etention o f t he l egionary defences and t he subsequent e xtension of t he defended a rea may have b een r elated t o s chemes f or u rban development. Certainly w ith r egard t o t he enlarged c ircuit, a lterations t o t he F orum a nd p ublic b aths a nd c hanges i n t he character of t he b uildings apparently t ook p lace a t about t he s ame t ime a s t he n ew c ircuit w as p rovided. These d evelopments and t he various i ndications of p rosperity a ssociated w ith t hem might

suggest

t hat

t he

d efences

6 3

w ere

c onnected

w ith

a

n ew

p hase

o f

u rban

g rowth

( Bidwell

1 980,

4 6-7,

There i s t herefore s ome e vidence, c onsidering t hat t hese e arly d efensive connection w ith u rbanising policies.

6 0-2,

6 7).

a lbeit tentative, c ircuits may h ave

f or s ome

F inally i t h as b een a rgued t hat t he n eed f or i mperial permission t ogether w ith t he e xpense i nvolved, especially i n t he b uilding o f d efences i n s tone, w ould make i t u nlikely t hat many of t he s ettlements o f R oman B ritain would b e able t o p rovide defences f or r easons o ther t han t hose o f emergency ( Wacher 1 974, 7 2; F rere 1 978, 2 84). B ut while t he c ons truction o f a walled c ircuit a t a n early s tage i n t he development of a t own might have been f inancially b urdensome w hen money w as b eing s pent o n major p ublic buildings, t he construction of an earthwork c ircuit would have been a c omparatively i nexpensive u ndertaking. The construction o f e arthwork d efences r ather t han s tone walls i n t he i nitial period of u rban development of a ny of t he centres of t he p rovince i s much more i n k eeping w ith t he w idespread u se o f t imber f or building, f or e xample at W roxeter, Cirencester a nd Verulamium w here t he e arliest s tructures a re a ll o f t imber ( Bushe-Fox 1 916, 1 -22; Ant. J . ( 1963) 4 3, 1 6-18; Wacher 1 974, 3 01; F rere 1 964b, 1 04-5). i) i

Were c ivic p ride a nd f inancial considerations f actors i n t he p rovision o f defences?

This i nvolves a n ew a pproach t o t he f unction o f e arthwork c ircuits; f or i f t hey were p rovided a s part of a g eneral p rog ramme o f u rbanisation, t here a re c learly o ther f unctions t han defence t o be considered. I t s eems a r easonable a ssumption t hat municipal d isplay w as i nvolved i n t he c onstruction o f major public buildings. H owever i t i s generally a rgued t hat s uch matters o f p restige w ere n ot i nvolved i n t he c onstruction of d efences, e xcept i n t he case of t he coloniae, a nd would c ertainly n ot h ave b een a f actor i n t he construction o f t imber and earth c ircuits; c ivic p ride would not e xplain why t he c ivitas c apitals o f Canterbury a nd L eicester w ere a pparently w ithout r amparts, a lthough t hey were built a round l esser c entres ( Wacher 1 966, 6 5; Webster 1 975b, 5 6). I t h as a lready b een s uggested t hat t imber c ircuits would h ave b een a ppropriate i n t he e arly s tages o f u rban d evelopment. I t i s neverthiess t rue t hat c onsideration of c ivic p ride a nd municipal d isplay s eem m ore r elevant t o t he c onstruction o f defences i n s tone. I n t his c onnection, t he p revious d iscussion on t he c onstruction o f e lements i n s tone, particularly gateways i n a ssociation w ith e arthwork circuits i s i mportant s ince i t r aised t he p ossibility t hat s ome centres may have b egun t o e rect defensive c ircuits p rior t o t he l ate 2 nd c entury p erhaps f or r easons o f c ivic p ride. T he s uggestion i s t hat t hese s tone gateways may i ndicate t hat c ertain t owns h ad e mbarked o n l ong-term s chemes f or defences i n s tone which were designed t o be carried o ut as f unds a nd l abour w ere a vailable. The p rovision o f walled c ircuits f or t he coloniae of Colchester, G loucester and L incoln i n the 2 nd c entury i s 1 978, 9 7) .

a ccepted a s a n e xpression o f c ivic p ride ( Wacher While i t may b e a rgued t hat t he s pecial s tatus o f

f ; 4

the c oloniae s et t hem a part f rom o ther t owns i n t his r espect, it i s n ot i nconceivable t hat t he a spirations of c ertain o ther centres i n matters o f municipal d isplay w ould h ave a t l east tended i n t his d irection. Merrifield ( 1969, 1 19-20) a nd Wacher ( 1978, 9 8) s uggested t hat t he walled c ircuit o f L ondon might b elong t o t he l ate 2 nd century, a nd i t i s possible t hat i t w as e rected n ot i n t he c ontext o f t he A lbinian e pisode b ut as a n expression of c ivic p ride. I n t his r espect, t he p rovincial c apital might h ave f ollowed t he c oloniae i n t he construction of a walled c ircuit i n t he 2 nd century. This may have a lso b een t he c ase a t C irencester w here t he d iversion a nd canalisation of t he r iver Churn f rom t he centre of t he t own t o a n ew course a long t he e ast s ide o f t he t own r epresents a civic achievement i n i tself, but i t i s n ot c lear whether i t was u ndertaken s imply i n c onnection w ith t he l aying o ut o f t he town o r whether i t was a ssociated i n s ome way w ith t he intention t o c onstruct a defensive c ircuit ( Ant. J . ( 1967) 4 7, 191; B ritannia ( 1970) 1 , 27; Wacher 1 974, 3 01-2). The 2 construction o f t he monumental Verulamium Gate a nd o f t he stone b ridge which carried t he F osse Way a nd Akeman Street across the n ew c ourse o f t he r iver a nd i nto t he t own c ould b e connected w ith t he completion of t his canalisation perhaps marking t his a chievement w ith a more r ecognisable a nd traditional e xpression of c ivic d isplay. Better e vidence i s required f or t he r elationship b etween t he c analisation o f t he river and t he various p rimary defensive e lements i n s tone on this s ide o f t he t own. I f a c onnection b etween t he canalisation and t he construction of d efences can b e f ound, i t can h ardly b e t hought t hat t his e nterprise c ould h ave b een achieved w ithin t he c ircumstances o f t he A lbinian episode. Certainly t here a re i ndications t hat C irencester w as sufficiently developed and p rosperous i n t he 2 nd c entury f or the c onstruction o f d efences i n s tone t o h ave b een a f easible aspiration and one which would have b een a l ogical e xpression of c ivic p ride a nd e conomic a chievement ( Wacher 1 974, 3 01, 302-5) I n v iew o f t he e vidence f rom o ther major c entres f or t he initial construction of gateways i n s tone, i t i s worth pursuing t he p ossibility t hat s everal a spiring c entres may have begun t o i mplement l ong-term s chemes f or defences i n stone i n t he 2 nd c entury a s a n e xtension o f t heir municipal building p rogrammes a nd at a s tage i n t heir development w hen the c ompletion o f t he p ublic b uildings h ad b egun t o f ree manpower and r esources f rom o ther c ivic e nterprises. C ivic p ride c ould h ave b een a f actor i n t he a ddition o f stone walls t o e xisting earthwork c ircuits. Wacher ( 1966, 6 6) i nterprets t he r ebuilding o f t he ' emergency' e arthwork f ortifications i n more durable materials s imply i n t erms o f a return when t ime a llowed t o t he o riginal p lans b egun by Albinus for t he c onstruction of defences i n s tone. The desire to m ake permanent t he e xisting t imber a nd e arth c ircuits h as been put f orward as s ufficient e xplanation f or t he construction o f s tone walls ( Wacher 1 974, 7 5; F rere 1 978, 288) . But i t i s necessary t o e xplain why t he q uestion o f rebuilding o f t hese defences i n s tone s hould h ave b een considered at a ll o nce t he danger had passed. I nitially, i t

6 5

was t hought t hat S everus ( AD 1 96-211) may h ave b een r espons ible f or i mplementing a p rogramme o f r ebuilding in s tone ( Corder 1 955, 1 24; Wacher 1 966, 6 7; F rere 1 978, 2 86). B ut once t he emergency o f AD 1 96 was o ver, t he t roops restored, a nd H adrian's Wall r efurbished, t here w as n o l onger a ny d anger t o t he t owns and t heir defences w ere no longer needed. S everus i s a lso u nlikely t o h ave h ad e ither t he r esources o r t he i nclination t o c onsider t he i nterests of the t owns of an a nti-Severan p rovince ( pers. c omm. G . Webster). I t m ay s imilarly have been a l ack of specific i nterest i n t he u rban c entres o f t he p rovince w hich p revented S everus f rom i ssuing an o rder f or t imber and earth defences t o be levelled. T hese defences w ould c ertainly h ave p resented n o t hreat t o h is authority s ince, e ven i f t hey p roved t roublesome, which i s u nlikely, t hey c ould h ave o ffered n o s erious r esistance t o t he R oman a rmy. F or t he s ame r easons, i t i s unlikely, a s has b een s uggested more r ecently, t hat t hese e arthwork defences w ere made more permanent a s part of a p rogramme of r econstruction o n t he military s ites o f t he n orth u nder Caracalla ( AD 2 112 17) ( Frere 1 978, 2 88). I t s eems much more l ikely t hat, s ince t hese d efences w ere n o l onger n ecessary, t heir continued maintenance and r econstruction i n s tone can be e xplained i n t erms o f c ivic p ride. Under t hese c ircumstances, t he addition of a s tone wall was a matter n ot of p rovincial policy but o f t he i nclination a nd f inances o f e ach i ndividual c entre. N or i ndeed would t he i dea of a r eturn t o an A lbinian s cheme f or defences i n s tone e xplain t he c onstruction o f walled circuits at a l ater date e ither on s ites h itherto undefended o r on n ew a lignments w here e arlier d efences h ad p erhaps been a llowed t o decay. Only f or t hose centres which h ad a lready embarked u pon l ong-term s chemes f or d efence i n s tone p rior t o t he Albinian episode would t he r esumption of work i n s tone be a natural c onsequence o f e arlier a ctivity. One f urther p oint w hich must b e c onsidered i s t hat o f t he e xpense of walled c ircuits f or ' lesser' centres. S o far n o r ealistic a ssessment c an b e made o f t he f inancial c ompetence of i ndividual c entres t o f ulfill s uch a n undertaking without a ssistance. This w ould h ave a ffected t he a bility o f a c entre t o e rect defences as a matter of choice r ather t han as t he r esult o f c oercion by t he g overnment. The c ost o f b uilding a walled c ircuit must have b een c onsiderable even for t he smaller c ivitas c apitals s uch a s C aerwent ( 44 a cres/17.8 hectares enclosed) a nd Caistor-by-Norwich ( 35 a cres/14.1 h ectares) w here t he t otal l ength o f t he w all a mounted t o a lmost a mile ( Jarrett 1 965, 5 7-8). F or centres s uch as Water Newton, w here a c omparable a rea t o t hat o f Caerwent was e nc losed ( 44.5 a cres/18 hectares) , t he f inancial burden w ould p resumably b e much g reater f or r esources were p resumably r ather l ess, o r l ess e asy t o c ommand, t han i n t he t ribal capitals. f d efences i n s tone i nvolved T he c onstruction • o community i n a h eavy commitment o f r esources and, Jarrett(1965, 5 8) s uggested, one c an only s uppose t hat work t hat f lint

was

spread

over

a n umber

o f

years.

I t

h as

a a s t he

been c alculated

o ver 4 5,000 l oads o f bonding s tone a nd 1 05,000 l oads o f were u sed t o b uild t he 1 .5 mile ( 2.4km ) walled circuit

at S ilchester a nd t he Jurassic l imestone r equired f or t he levelling c ourses h ad t o b e t ransported a d istance o f 4 5-50 miles ( 70-80km ) ( Boon 1 974, 1 00-2) I n a ddition, t he c onstruction of the walls would often have r equired e xtensive roadworks. The c onstruction o f walled c ircuits i s u nlikely t o have b een borne w ithout d ifficulty. Wacher ( 1978, 1 02) suggests t hat t he decline i n public and p rivate b uilding ..

construction during t he 3 rd c entury may b e a ttributed t o s ome extent to t he n eed t o build masonry defences. B ut l ittle major building a ctivity may h ave b een n ecessary during t his period s ince the i nitial major period of public and p rivate building c onstruction i n s tone was l argely c omplete ( Rivet 1964, 9 3-4; F rere 1 978, 2 89-90). On t he o ther h and, t he construction o f s tone defences would n ot h ave b een a n unmitigated burden o n t he e conomy s ince i t would have p rovided a s timulus t o t rade a nd employment a nd t o l ocal s tone quarries, b rick and t ile works, and o ther t ypes o f p roduction. I n t he absence o f a ny e vidence a t a ll a s t o h ow c onstruction was f inanced, i t might b e f easible t o c onsider whether t he t ribe might n ot h ave b een a c ompetent b ody i n these matters; '. . . a lthough t he work must have b een sanctioned by t he c entral government, i t w as e vidently guided as w ell as f inanced by t he l ocal a uthorities; .. . . and i n each c ase t he c ivitates, t hrough t he decurions, would h ave h ad to f oot the b ill' ( Rivet 1 964, 9 2-3) . Corder ( 1955, 4 1) considered t hat a t ribal l evy would h ave b een n ecessary t o finance certain a spects a t l east o f defensive a ctivity, and Wacher ( pers. c omm.) s uggests t hat t he g overnment may h ave introduced a r emission of t axes f or t he l andowners who would be t he ones t o h ave t o f inance t he w ork. U nder t hese c ircumstances, t he possibility of a t ribal policy and t herefore of regional v ariations i n t he p rovision o f d efences may b e considered. I t i s d ifficult t o assess t he p otential o f s uch a line o f i nquiry. The c urrent i nsistence t hat a dministrative centres and government i nstallations were p rovided w ith defences w ould l eave l ittle s cope f or c hoice by t he t ribe i n the s election between one

of s ites r egion

worth defending. a nd a nother a re

H owever, d ifferences d iscernible. I n

Trinovantian t erritory only Colchester and G reat Chesterford received w alled c ircuits. A g ood d eal s till r emains t o b e discovered about t he defences of s ettlements i n t his a rea b ut, apparently, t he c olonia a t C olchester i s t he only s ettlement which has a l ong a nd a lmost c ontinuous defensive h istory. ' It would s eem t hat t here w as n o g eneral p rovision f or t he d efence of s mall t owns' a mongst t he T rinovantes and t his r equires explanation i n a n a rea vulnerable f rom t he l ater 2 nd c entury t o the a ctivities of s ea r aiders ( Rodwell 1 975, 9 3) . The existence o f a s eries o f c oastal f orts p erhaps a s e arly a s t he early 3 rd century would have o ffered p rotection t o t he settlements i n t his a rea a nd t his may h ave r estored c onfidence sufficiently f or any f urther d efensive a ctivity t o b e considered u nnecessary. There i s n evertheless a marked c ont rast with other t ribes, s uch a s t he Cornovii o r t he C oritani, i n t he concern f or d efence. I s t he e xplanation t o b e f ound i n their disinclination t o i nvest capital i n t his w ay a nd t heir lack o f c ivic i nterest i n p ublic works o f t his n ature?

6 7

Casey ( 1983) b elieves t hat i mperial p olicy h ad a c ons iderable i mpact on the ability o f civil settlements t o f inance defensive a ctivity i n t he 4 th . c entury at l east. H e has s uggested that Constantine's confiscation t o t he imperial t reasury o f t he r evenue n ormally a vailable t o t owns f or p ublic works meant t hat municipal f inances must have been at a l ow ebb i n t he e arly 4 th c entury u ntil A D 3 50-354 when V alentinian r emitted a t hird of t his c onfiscated municipal revenue for t he r epair o f p ublic works. Only f rom t his t ime onwards did t he i mperial government c oncern i tself w ith r estoring d efences a nd r emitting t axes f or t hat purpose. D efensive activity i n t he e arly 4 th c entury i s t herefore p ossibly a measure of t he wealth of i ndividual communities s ince i t would h ave been up t o t hem t o f ind t he money t o i mplement s uch w ork.

6 ) o r

The p rovision of s tone walls: maintenance a nd u pdating?

s eparate defensive

scheme

The p rovision o f walled c ircuits f or t he m ajority o f s ettlements which had been p reviously d efended by t imber a nd e arth, a nd a lso f or a n umber o f p reviously u ndefended centres, has b een e xplained as t he r esult o f an official d ecision t o make p ermanent t he e xisting d efences, b eginning i n t he e arly 3 rd 3 rd

c entury. Subsequently, a nd 4 th c enturies t o s peed

decisions were made i n the l ate u p a nd e xtend t his p rocess.

F rere's c hronology f or t he c onstruction of d efences i n s tone, w hich i nvolves a n emphasis o n t he l ate 3 rd c entury f or t he majority o f walled c ircuits, h as c onsiderable i mplications f or t he q uestion o f continued d efence o f t hose c entres p rovided w ith r amparts earlier ( Frere 1 978, 2 86-9; see a lso Wacher 1 975, 5 1). I f t he p rovision o f w alls i s a m ainly l ate 3 rd c entury development, t hen by t his t ime many o f t hese 2 nd c entury r amparts may h ave b een a llowed t o collapse w hile others may c onceivably have b een c ontinuously maintained. One o f t he i ssues w hich h as c oncerned s cholars, but w hich i s b edevilled by t he d ifficulties o f t he dating, is t he q uestion o f h ow much t ime e lapsed b etween t he building o f t he t imber and e arth c ircuit a nd t he construction of t he w all. The dating e vidence a llows c onsiderable f lexibility o f i nterp retation i n t his matter; J arrett a nd Wacher gave v ery d ifferent e stimates o f t he l ikely l apse o f t ime b etween t he e rection of bank and wall. Jarrett ( 1965) s uggested that a t t hat t ime t he e vidence f rom t hose c ircuits which had b een i nvestigated would a llow t he bank and wall t o have b een put up w ithin a s hort t ime a nd t hus t o r epresent p erhaps a s ingle s cheme which was p lanned t o span a n umber of years. Wacher ( 1965), o n t he other h and, u sed t he f lexibility o f t he d ating t o s tress t hat t he bank a nd wall were d istinct schemes s eparated by a c onsiderable n umber o f y ears. I n r eply t o Jarrett, both Wacher and F rere emphasised the d ifference b etween t he t erminus p ost q uem f or t he w all a nd t he terminus post

quem

f or

t he

p rimary bank

6 8

i n

s uch

e xamples

a s:-

B rough-on-Humber

mid 2 nd/early 3 rd AD 2 70 f or wall.

Caerwent

mid

C irencester

Hadrianic-Antonine period; early 3 rd c entury, a lthough v ery l ittle e vidence dating f or t he wall.

D orchester-on-Thames

bank c .AD 1 60; century l ater.

S ilchester

c .AD 1 60-170 f or bank; i n t he mid s equent walling.

2 nd

c entury

c entury

a nd

f or

f or

wall

f or

w all

r ampart;

AD

2 40.

a bout

a

I nner D efence p rimary 3 rd c entury f or s ub-

2 F rere ( 1965, 1 38) a nd Wacher ( 1965, 26) w ere p rimarily c oncerned t o make a c ase f or t he r ampart a nd t he walling b eing t wo s eparate s chemes. However, t heir a rgument w as b orne o ut n ot by any ' marked d ifference' b etween t he dating of t he r ampart a nd t he wall b ut by t he m ethods o f c onstruction f or each which made i t unlikely t hat b ank and wall were e nvisaged o riginally a s p art o f t he s ame s cheme. Jarrett's ( 1965, 5 8) p oint t hat on i ts own t he dating e vidence w ould a llow wall and b ank t o b e p art o f t he s ame s cheme was w eakened b y t he a bove e xamples, but t hey underline t hat t he t erminus post quem f or t he t wo b uilding o perations may n ot i ndicate a nything a bout t he l apse of t ime b etween t hem. This p roblem might be more r eliably e lucidated by a g reater emphasis on t he p hysical a spects of t he defences, s uch a s r enewal and r epair of the r ampart and i ts d itches, t he s tate o f t he r ampart a nd d itches a t e ach stage and particularly when t he wall was a dded. S o f ar t he i nvestigation o f defensive c ircuits h as r arely i nvolved a ttempts t o e xamine such a spects a s t he r ate of c ollapse and e rosion o f t he r ampart a nd t he r esulting a mount o f maintenance a nd r esources r equired t o k eep particular t ypes o f earthwork i n a defensible s tate. I n

t he

case

o f

t imber

a nd

e arth

defences,

e rosion

would

have been a major p roblem and i t would h ave r equired much e ffort t o maintain a r ampart i n a r easonable s tate o f r epair. T he a ddition of a wall a t t he earliest opportunity would have s aved much o f t his e ffort, a ssuming o f c ourse t hat t he e xpenditure could b e borne. I ndeed, Webster h as s uggested a p ractical e xplanation f or w alling w hich r elates t his t o t he c ontinued maintenance a nd r epair o f t he e xisting earthwork c ircuit. H e h as p ointed o ut t hat a n ormal e arthen b ank, e ven i f t urved, would r apidly b egin t o e rode a nd collapse i nto t he d itch, t hus r endering t he d efences u seless. The most c onv enient method of d ealing w ith a r ampart which was e roding o r would eventually e rode i nto t he d itch w as t o i nsert a s tone wall i nto t he f ront of t he r ampart. The a ddition of s tone d efences d id n ot t hen r epresent a s eparate p hase i n t he s ense t hat i t i s u sually a rgued; t he i mportant s tep c ame w ith t he i nitial

p rovision

I t p rovision

h as o f

b een n ew

o f

defences.

o bserved s tone

t hat,

f ronts

6 9

i n

i n t he

a m ilitary e arly

2 nd

c ontext, c entury

f or

t he t he

r amparts of l egionary f ortresses e rected i n t he early F lavian period may r eflect t he l ifespan o f e arth a nd t imber r amparts ( Jones 1 980, 5 1). At t he c oloniae o f L incoln and G loucester, t he n eed t o r eplace t he t imber r ampart r evetment o f t he l egionary f ortress d efences may h ave p rovided a very p ractical i mpetus i n t he e arly 2 nd century, o r possibly as e arly as t he l ate 1 st c entury a t L incoln, f or t he p rovision of a s tone wall ( Wacher 1 974, 9 7 & 1 22; Jones 1 980, 5 1) . Webster a lso a rgues t hat s ome e arthwork c ircuits would e rode more r apidly t han others, depending on s uch f actors a s t he composition of t he r ampart, a nd s o b e more q uickly i n n eed o f r epair. A difference i n date b etween various walled circuits could t herefore b e e xplained by t he f easibility, i n s ome c ases, o f maintaining a n earthwork c ircuit f or a l ong period, s o t hat i t may o nly h ave b een n ecessary t o i nsert a wall w hen t he p rovince was t hreatened i n t he l ate 3 rd/early 4 th c entury. At other c entres t he d ecision t o e rect a s tone wall m ight have been n ecessary much s ooner. This hypothesis extends t o civil s ettlements. The a rgument applied t o t he m ilitary s ites a nd t he c oloniae, where t he a ddition o f a wall i s s een as t he n atural r eplacement o f e arlier t imber work, was a ccepted. Webster ( pers. comm.) has s uggested t hat s eparate p ermission was n ot r equired f or w alling s ince i t c ould b e r egarded a s c onstituting r epair, and s uch an a rrangement would s uggest t hat i t was r ather a matter f or e ach s ettlement t o m ake i ts own decision a bout t he maintenance a nd r epair of t he c ircuit. This p ractical approach h as much t o r ecommend i t w here a n earthwork c ircuit a lready e xisted b ut Webster suggests t hat w here w alled c ircuits w ere e rected d e n ovo j e ither b ecause t he earthwork c ircuit h ad b een a llowed t o decay o r b ecause t he s ettlement h ad b een p reviously u ndefended, t he l ater a ddition of t he wall c ould r epresent a r esponse t o a r enewed t hreat. Three centres :

c hoices

w ere

t herefore

a vailable

t o

i ndividual

-

i ) The d efensive c ircuit c ould b e a llowed t o d ecay deliberately l evelled. I t may h ave b een n ecessary t o defences l ate 3 rd

o r b e e rect

i n s tone w hen t he i nsecurity o f t he p rovince i n century made t he p rovision o f d efences expedient.

i i) The e arthwork c ircuit c ould b e maintained a nd would i nvolve a nnual c leaning and r efurbishing. A w all h ave b een e rected e ventually.

t he

t his may

i ii) The d efences c ould b e maintained b y t he a ddition o f a s tone r evetment w all. This could be u ndertaken e ither i mmediately o r a s a l ong t erm a im t o b e i mplemented a t a ny t ime a s and w hen r esources were a vailable. Comparatively f ew e arthwork c ircuits a re k nown which w ere n ot s ubsequently p rovided w ith s tone d efences ( see Tables O ne and Two). Those s ites w hich w ere n ot m ay f all i nto c ategories i ) a nd i i) above. I n c ategory i ) Chelmsford i s t he only c lear e xample w here t he e arthwork d efences a re k nown t o h ave been s hort-lived and were a ctually l evelled e arly i n t he 3 rd c entury ( Drury 1 975, 1 70-2). The e arthwork defences a t Neatham

may

a lso

f all

i nto

t his

c ategory.

H ere

t he

d itches

a ccompanying

t he

r ampart

h ad

b een

f illed

i n

t he

m id

3 rd

century. I t i s n ot known i f t he s ettlement was s ubsequently i walled ( see Gazetteer).ExamPleS i n category i) a re l ess easily i dentified s ince s o l ittle work h as b een d one i n analysing t he maintenance of earthwork c ircuits. There a re two s ettlements w here t his s equence may b e o bserved a lthough more e vidence i s r equired. At B rampton ( Norfolk) a defensive c ircuit was p rovided b ut n o t race h as b een f ound o f d efences i n s tone. H owever, t he f act t hat there i s a lso n o t race o f a r ampart s uggests t hat t he e vidence h as b een r emoved by ploughing and n o f irm conclusions can be d rawn about t he p rovision o f a w all. The d itch appears t o h ave r emained open until the l ater 3 rd o r early 4 th c entury when i t was deliberately f illed ( Britannia ( 1975) 6 , 2 60; ( 1976) 7 , 3 41; ( 1977) 8 , 2 11-13). At F enny Stratford ( Mag i ovinium ) ‚ a n earthwork c ircuit a s y et u ndated h as b een f ound a nd, a lthough t here i s n o e vidence t hat a s tone wall was a dded, t he ditch appears t o h ave r emained o pen u ntil t he l ate 3 rd/early 4 th c entury a nd the r ampart r evealed e vidence f or two periods of c onstruction s uggesting t hat a ttempts w ere made t o maintain t he e arthwork circuit i n a defensible s tate a lbeit f or an unknown period o f t ime ( Britannia ( 1970) 1 , 2 88-9; ( 1971) 2 , 2 68). There a re many other centres where a t erminus post quem i n t he l ate 3 rd c entury f or t he c onstruction o f t he wall suggests t hat e ither c ategory 1 ) o r c ategory 2 ) applies. But very l ittle i s k nown a bout t he maintenance o r o therwise o f t he earlier earthwork c ircuits i n t hose c ases where t he wall was e rected s ome c onsiderable t ime a fter a nd w hether t he w alls may be s een as a s eparate phase o f defences e rected specifically as a r esponse t o t he i nsecurity o f t he l ate 3 rd c entury o r whether they r epresent part of a p rocess o f f airly c ontinuous defensive I n

a ctivity. a

l arge

n umber

o f

c ases

w here

t he

d efences

w ere

reconstructed i n s tone, t he walled c ircuit f ollowed t hat o f t he e arlier e arthwork. H ere t he r elevance o f Webster's ( pers. comm.) a rgument t hat t he a ddition of a s tone wall c an be s een s imply a s a q uestion o f t he maintenance o f t he e xisting c ircuit i s apparent. There a re t wo categories o f s ite where t he wall was n ot e rected a s a r evetment t o t he r ampart b ut h ad a n independent

f unction.

i ) S ites w here n o p re-existing known, t hat i s Ancaster, Caistor Canterbury, Catterick, Gatcombe, Chesterford,

H orncastle

a nd

e arthwork c ircuit i s ( Lincs.), C ambridge, Godmanchester, G reat

L eicester.

i) S ites w here a n earlier earthwork a lignment was i abandoned i n f avour o f a n ew a lignment, f or e xample, Exeter, Verulamium; o r where t he e xisting c ircuit was a dapted, such

a s

Mildenhall

a nd

Thorpe-by-Newark.

I n the f irst c ategory the d ecision t o build a s tone wall came i ndependently a nd w as n ot a s imple d ecision t o maintain an e xisting c ircuit. The l ate 3 rd and e arly 4 th century dates f or s ome o f t he s ites i n t his c ategory Ancaster, Canterbury, Cambridge, Catterick a nd G reat Chesterford may s upport t he -

-

7 1

i dea t hat t hese walled c ircuits m ay h ave r esponse t o t he i ncreased i nsecurity o f t he t ime ( FL 'ere 1 978, 2 89).

b een ' erected i n p rovince at t his

I n t he s econd c ategory, one p ossibility i s t hat t he earlier t imber and earth c ircuit ' had b een a llowed t o decay s o t hat t he s tone w all w as a n ew v enture. On t he other h and, t he earthwork c ircuit may have p roved t o b e e ither t oo r estricting o r t oo l arge i n w hich c ase t he d ecision t o e rect a w all m ay have p rovided a u seful opportunity t o a dopt a more c onvenient a lignment. I n e ach c ase t he q uestion o f t he continued maintenance of t he r ampart u ntil t he construction of t he walled c ircuit h as t o b e i nvestigated s ince i t i s i mportant t o establish whether a s ettlement a t a ny t ime e ffectively d ispensed w ith i ts defences e ither b y s imple neglect o r by deliberate demolition. i t h as b een s uggested by Wacher ( 1964, 13) t hat w here t he w all w as a v ery l ate c onstruction, a t 1 M ildenhall a nd possibly Thorpe-by-Newark, t hen t he a doption o f a n ew a lignment c ould b e a ccounted f or b y t he d isappearance o f t he earlier c ircuit by t he t ime of t he emergencies i n t he l ater R oman p eriod. However, partial r euse of t he earlier

b oth t hese e xamples i nvolve a lignment s uggesting t hat

t he t he

c ircuit c ould h ave b een e xtant i f n ot i n a d efensible s tate. E ven i n t hose c ases where t he s ettlement d ispensed w ith i ts defences, t he s ubsequent c onstruction o f a w alled c ircuit n eed not be a ttributed t o a s pecific t hreat o r to t he i mplementation o f a government p olicy. Webster's s uggestion t hat n o s pecific permission was r equired f or a walled circuit w here d efences a lready e xisted w ould a llow a s ettlement t o r econstruct i ts defences e ven a fter a l ong period o f neglect, a nd a t a convenient.

t ime

a nd

o n

a n

a lignment

w hich

i t

f ound

m ost

I t s eems t hat a h igh p roportion o f t hose t owns a nd v illages which were p rovided w ith earthwork c ircuits f all i nto t he c ategory w here a w alled c ircuit w as a dded s ooner o r l ater. I t i s however c rucial t o k now h ow quickly t his was u ndertaken a nd w hat h appened t o t he e arthwork d efences i n t he m eantime. I t i s i mportant t oo t o c onsider t he q uestion of whether t hose c ircuits w hich w ere n ever r econstructed i n s tone w ere maint ained f or a t ime o r abandoned i mmediately. Much m ore w ork n eeds t o b e done o n t he decay a nd maintenance of t imber a nd earth defences and t heir condition a t t he t ime when a walled c ircuit w as b uilt i n o rder t o a ppreciate t he a ttitude o f i ndividual communities t owards t he provision o f defences. T he n otion t hat t he d ecision t o maintain t he d efences w as a matter f or each s ettlement means t hat t he d efences h ave to b e considered much more a s p art o f t he h istory o f t hat s ettlement. H ere questions of a spirations, c ivic p ride a nd f inancial c ompetence h ave a v ery c onsiderable r elevance.

7 . i s t here a ny e vidence t o s uggest t hat t he construction o f a w alled c ircuit may h ave b een a p rolonged p rocess?

Considerations walled

c ircuit

of

would

f inance

s uggest

p robably h ave

7 2

t hat b een

t he

c onstruction

a c onsiderable

o f

a

under-

taking. At t he c oloniae o f L incoln a nd G loucester, a nd a t Exeter, the r etention o f t he l egionary defences s uggests a policy t o defend t hese s ettlements a nd a ttempts t o r evet t he military r ampart i n s tone s eem t o have begun w ithin a s hort period. Nevertheless, t here a re i ndications t hat t his t ook some t ime t o complete. A t L incoln t he t imber f ront o f t he l egionary f ortress rampart was r eplaced w ith a n arrow s tone wall a nd s tone f ronts were a dded t o t he gate t owers. S ome s ections a cross t hese upper colonia d efences have p roduced dating e vidence suggesting t hat t he wall w as b eing e rected a t t he e nd o f t he ist o r beginning o f t he 2 nd century, b ut t he e vidence f rom other s ections r aises t he p ossibility t hat t he b uilding o f t he wall lasted i nto t he Hadrianic period at l east ( Jones 1 980, 5 1-2). E ven t his r epresented o nly t he b eginning o f a l ong process of modifications which was t o l ast t hrough i nto the 4th c entury a nd w hich w as t o i nvolve t he a ddition o f s tone interval t owers t o t he r ear o f t he wall i n t he m id-late 2nd century, t he h eightening o f t he wall a nd r ampart a nd t he rebuilding of t he g ates i n t he l ate 2 nd/early 3 rd c entury, and the h eightening a nd t hickening o f t he wall i n t he 4 th c entury ( Wacher 1 974, 1 22; Colyer 1 975, 1 1; F rere 1 978, 2 83; Jones 1 980, 4 9-55). At G loucester t he i ndications t hat t he c onstruction o f the w alled c ircuit spanned a c onsiderable period o f t ime a re even more p ronounced ( Wacher 1 978, 9 7; F rere 1 978, 2 83). Stone gates a re k nown t o have b een i nserted i nto t he l egionary fortress r ampart o n t he n orth a nd e ast i n t he e arly 2 nd century; s hort s tretches o f wall were built on e ither s ide o f each gate. I t i s n ot c lear w hether a c omplete w all w as erected as much o f t he e vidence h as b een r emoved by s ubsequent activity i n t he l ater R oman o r medieval p eriods. Apart f rom the n orth and t he east gate a nd t he a djacent s tretches of walling, n o f urther a ctivity h as b een i dentified w hich c ould be a ttributed t o t he early 2 nd c entury a nd i t s eems quite possible t hat, a lthough t he r ampart w as e nlarged a t a date which has been p laced i n t he l ate 2 nd century, t he walled circuit w as n ot c ompleted u ntil s ome t ime i n t he 3 rd c entury ( Ant. J . ( 1972) 5 2, 2 9-32; ( 1974) 5 4, 9 -17; B ritannia ( 1970) I , 2 93; ( 1975) 6 , 2 73, f igs. 1 6 & 1 8). I t i s t he cutting back of t he r ear o f t he r ampart i n t he 2 nd c entury a long s everal p arts o f t he c ircuit f or t he i nsertion o f buildings which u nderlies t he s upposition t hat t he t runcated portion o f t he r ampart must t herefore h ave b een s upported b y some f orm of f ront r evetment; t here i s n o c lue a s t o whether this w as i n f act a w all ( Ant. J . ( 1972) 5 2, 2 9 & 3 2; Wacher 1974, 1 41) More e vidence i s n eeded f rom t he o ther c olonia a t Colchester where t he early 2 nd c entury f ree-standing wall found t o t he s outh o f t he B alkerne Gate i n 1 967 h as b een s een as a n i solated s tretch of a n uncompleted c ircuit ( Wacher 1 974, 1 13-14; 1 978, 9 7; F rere 1 978, 2 88). This c onclusion w as reached Balkern

on t he g rounds t hat t he G ate e xcavated i n 1 951

s tretch o f wall was f ound t o b e

n orth of t he o f one b uild

with

r ampart

t he

t his

t he

t o

i ts

r ear

and

7 3

material

i n

r ampart

p rovided a t erminus post q uem o f AD 1 40-170 f or t he c onstruction of t he wall ( Hull 1 958, 2 7) . However, e xcavations i n 1 971-1973 a long t he s outhern c ircuit at L ion Walk f ound a s imilar s equence o f f ree-standing w all of t he early 2 nd century t o which a r ampart h ad b een s ubsequently a dded, t hus u nderlining t he p ossibility o f a complete walled c ircuit o f t his date ( Crum my 1 977, 9 1-2, 1 01-2) There i s s ome c onfusion a bout t he d ating o f t his stretch of f ree-standing wall and i ndeed a bout t he dating of t he r ampart w hich was s ubsequently a dded b ehind t he w all. D ates i n t he early and mid 2 nd century h ave b een s uggested for t he w all a nd i n t he l ate 2 nd o r e arly 3 rd c entury f or t he rampart 1 ( Wacher 1 974, 13; 1 978, 9 7; F rere 1 978, 2 82), w hile b ased o n a n e xamination o f t he p ottery f rom a ll t hree s ites, Crummy p laces t he wall very e arly i n t he 2 nd c entury and t he rampart i n t he m iddle o f t hat c entury. The i nconsistencies i n t he e vidence f rom t hese s ections may a lso b e e xplained b y a p rot racted b uilding p rogramme s uch a s t hose i ndicated a t Lincoln a nd Gloucester. This u nderlines t he p oint t hat d ifferent parts of t he s ame c ircuit may have b een c onstructed a t d ifferent t imes a nd h ave u ndergone d ifferent s equences o f d evelopment. T herefore, t he dating e vidence f rom o ne s ection may n ot be r elevant t o t he c ircuit a s a w hole. As m entioned a bove i n t he case o f L incoln, t he c onstruction o f a defensive c ircuit r epresents only t he b eginning o f a p rocess w hich i nvolves maintenance a nd r epairs, modifications a nd a dditions t o t he e xisting circuit. There i s a s trong a rgument f or r egarding t he a ddition o f a s tone wall t o an e xisting c ircuit, a nd t he provision o f b astions a nd a ny o ther modifications s imply part of t he p rocess and n ot p rogrammes o f d efence.

8 .

I s

t he

e vidence

f or

w ith t he i nterpretation Theodosius a nd w ith a n ew Roman period?

t he

t o t he b asic s tructure a s a s d istinct and s eparate

p rovision o f

of r ole

a s ingle f or c ivil

b astions

c onsistent

s cheme i nitiated b y d efences i n t he l ate

The u se o f b astions a nd t heir f unction w ithin t he w ider context o f t he d efence o f Romano-British t owns and v illages r emains a matter f or f urther i nvestigation e specially i n v iew of t he apparent i nconsistencies i n t heir p rovision. By n o means a ll t he w alled c entres w ere p rovided w ith p rojecting bastions o r t owers. A lthough n ew s ites c ontinue t o appear, t he e xtent o f t he p rovision o f b astions i s s till u nknown. However, i t i s i nteresting, i n v iew o f ear l i er a rgumen ts t hat p reference was g iven t owards t he major t owns i n the m atters o f defence, t hat t his does n ot s eem t o h ave b een t he c ase w ith bastions ( see f ig. 3 ; Wacher 1 975, 5 1; F rere 1978, 2 91-2). Even w here b astions h ave b een f ound, t hey w ere n ot n ecessarily p rovided f or t he whole c ircuit. At Caerwent n o bastions h ave b een d iscovered a long t he w est a nd e ast w alls a nd t his may b e t hought o dd i n v iew of t he f act t hat t hese r epresent t he main approaches

t o

t he

t own.

One

7 4

o f

t he

p roblems

i n

attempting

t o

0

k nown

0

o possible

F IGURE 3

B astions

7 )

10K m

1 00 m i l e s

a ssess

t heir

l ikely

u se

i s

t he

f act

t hat

w e

d o

not

k now

whether t he absence of bastions on t he west and e ast parts o f t he c ircuit a t Caerwent was deliberate o r w hether t he c ircuit was n ever completed. S imilarly, a t Silchester, where t he cutting o f a n ew o uter d itch, w hich i s u sually associated w ith t he p rovision of bastions, may denote an i ntention t o p rovide t hem w hich was n ever i mplemented ( Boon 1 974, 1 06). At Wroxeter aerial photographs s how e vidence f or two p latforms on t he n orth-east p erimeter, b oth apparently n ot completed ( Baker 1 970, 1 97; Webster 1 975a, 1 07). There was a g reat deal o f variety i n t he s ize a nd s hape of b astions f rom one s ettlement t o another and a lso w ithin t he s ame c ircuit . Polygonal s hapes a re common ( Caerwent a nd Cirencester) while t hose at L ondon a re s emi-circular; s quare a nd r ectangular b astions a re k nown a nd t here a re t wo s ites where f an-shaped b astions have b een f ound ( Ancaster and Godmanchester) ( Todd 1 973, 1 26-7; 1 975, 2 17; F rere 1 978, 2 91). H ollow bastions appear t o b e most common b ut solid ones a re known, f or e xample a t Caistor ( Lincs.), a nd i t i s not c lear whether t his i s a f actor t o be considered w ith regard to t heir f unction. L incoln p rovides y et a nother v ariation f or h ere n o p rojecting t owers have b een f ound and t his may be e xplained by t he r estrictions o f t opography. I nstead t he e xisting i nternal t owers were demolished and apparently r eplaced by l arge s olid s tone p latforms s et a gainst t he i nternal f ace of t he wall a nd perhaps s erving t he same f unction as a rtillery e mplacements. This s ituation h as b een observed a t o ne p oint on t he western c ircuit of t he l ower colonia, where a t erminus post quem f or t he c onstruction o f t his s olid p latform was p rovided by a w orn coin o f Tetricus I ( AD 2 70-273) , and a t one point a long t he e astern c ircuit o f t he u pper c olonia ( Ant. J . ( 1975) 5 5, 2 556 ; Colyer 1 975, 1 6-17; Jones 1 980, 2 5 & 5 4). These d ifferences a nd i nconsistencies i n t he p rovision o f bastions cannot be easily e xplained a nd while we m ay feel t he n eed t o r ationalise t hese v ariations, i t i s c learly n ot r ealistic when s o l ittle i s k nown a bout t heir function and t he w ider r ole o f t own d efences i n t his p eriod. Wacher ( 1974, 7 7) s uggests t hat t he e xplanation f or t he l ack o f consistency m ay l ie i n careless s upervision and a n i ndifferent u nderstanding o f t he t actics i nvolved. I f t his i s t he c ase, t hen i t c asts s ome doubt upon t he competence of t he communities c oncerned t o make e ffective u se o f t hese b astions. An i nteresting suggestion h as b een made by Casey who c onsiders that the l ack o f c onsistency i n t he p rovision o f b astions and i n o ther defensive a ctivity i n t he 4 th c entury m ay b e explained by t he c onfiscation t o t he i mperial t reasury o f t he municipal r evenue which would n ormally have b een a vailable f or public works. This may h ave r esulted i n t he i nability o f communities t o p rovide more t han one o r t wo b astions i n s ome cases, o r t o c omplete t he c ircuit a s i n t he c ase o f Caerwent, o r i n t he n eed t o r euse masonry wherever possible. Work on d efences i n t his p eriod may t herefore b e a n i ndication o f t he w ealth o r poverty of i ndividual c ommunities ( Casey 1 983) Most t he

wall.

o f

t he The

b astions only c ase

e xamined where

7 6

s eem

t o

a b astion

h ave was

been

a dded

certainly

t o

bonded

i nto t he w all w as a t M ildenhall, where e xcavation o f t he south-east bastion r evealed t hat only t he l owest courses survived but t hat t hese were b onded i nto t he wall ( JRS ( 1959) 49, 1 31; WAM f or 1 958-1960 ( 1960) 5 7, 2 35). H owever, t here a re a number o f e xamples w here t he r elationship o f wall a nd bastion i s u ncertain. Caistor

( Lincs.)

-

Horncastle

-

Caistor-by-Norwich

-

Canterbury

-

a s olid b astion, s till e xtant, s eems t o have b een bonded i nto t he wall b ut t he e vidence i s n ot c onclusive ( Ant. J .(1960) 4 0, 1 83 & 1 85; Whitwell T h970, 6 9; Todd 1 973, 4 2). at t he n orth-west corner t hat t he bastion i s o f w ith t he wall ( Whitwell Todd 1 973, 4 2).

i t appears o ne b uild 1 970, 7 3;

t he e vidence consists of an early p hotograph which s hows o ne o f t he bastions apparently bonded i nto t he w all ( Wacher 1 974, 2 35-6). The photograph s eems t o i ndicate t he bastion w as c onstructed o ver a n e xisting wall p linth course ( Johnson 1 976, 9 8). e vidence

a gain

not good

b ut

suggests

t hat b astions a re e ither c ontemporary o r a dded at a l ater date b ut bonded i nto t he wall ( Wacher 1 974, 1 89) Verulamium

-

t he t wo k nown b astions a ppear t o have b een partly bonded w ith t he wall ( Frere 1 964a, 7 1-3; Wacher 2 Johnson h as observed 1 974, 20) . t hat t he f oundations o f t he b astion west of t he s outh-east gate a re c learly s eparate f rom t he w all ( pers. comm., F ebruary 1 980).

A lthough t here i s a t endency t o a ssume t hat f urther i nvestigation i s l ikely t o s how t hat t hese b astions were a dded t o w alls a nd n ot a ctually c ontemporary w ith t hem. Wacher ( 1974, 2 35) h as p ointed out t hat where t he material was f lint, as a t Caistor-by-Norwich a nd C anterbury ‚ o r b rick i t w ould b e easier than where r egular masonry b locks were u sed t o c reate a bond b etween b astion a nd a p re-existing w all. More work i s needed before we c an a ssume t hat a ll bastions r epresent l ater a dditioná. I ndeed, i t i s c onceivable t hat i n s ome c ases, where the wall may have been a v ery l ate construction, t he bastions may h ave f ollowed v ery s hortly a fter t he c onstruction of t he wall. There a re i ndications t hat t his may have b een t he c ase at Caerwent. R ecent r eassessment o f t he e vidence l ed Casey ( 1983) t o c onclude t hat t here was a t erminus post quer n of A D 3 30-335 a nd s ince h e was a lso a ble t o a rgue f or a t erminus

post

quem

o f

AD

3 48-349

7 7

f or

t he

a ddition

of

t he

n orth-east c orner b astion, t he l atter must have f ollowed shortly after t he building o f t he wall. The evidence f rom M ildenhall a lso p resents a n i nteresting s ituation. Here e xcavation of t he b astion c lose t o t he s outh-east c orner s howed t hat i t was b onded i nto t he w all a nd w as t herefore o f one b uild w ith the wall. This is n otable i n view o f the f act t hat a t erminus post q uem o f AD 3 60 e xists f or t he b uilding o f t he w all. This i s t he l atest date s o f ar a vailable f or a walled c ircuit ( Wacher 1 975, 5 1). A t erminus post q uer n o f AD 3 60 f or t he b astion would f it i n w ith t he t heory t hat t he p rov ision of bastions i s t o be dated t o t he period after AD 3 67. The only o ther c ircuits w hich a re t hought o n p resent evidence t o belong t o t he 4 th century r ather t han t he l ate 3 rd century a re Cambridge, Catterick, G reat C hesterford a nd possibly Thorpe-by-Newark but i n n one o f t hese e xamples a re bastions k nown ( Frere 1 966, 9 6-7; 1 978, 2 89; Wacher 1 971, 1 71-2; V CH Esseç ( 1963) 3, 7 5-6; JRS ( 1965) 5 5, 2 13; Alexander 1 975, 1 08). None o f t hese c ircuits h as b een w ell i nvestigated o r c losely dated and s o f ar dates early r ather t han l ater in t he 4 th c entury h ave b een s uggested. These 4 th c entury d ates h ave o ccasioned t he comment t hat s urprisingly none of t hese e xamples had b astions p rovided; w e m ight e xpect t hat t hey would r esemble the defences o f G aulish t owns of t he l ate 3 rd a nd e arly 4 th c enturies ( AD 2 84-305) w here b astions w ere p rov ided f rom t he f irst ( Wacher 1 969, 4 9; F rere 1 978, 2 86). G reat Chesterford, Cambridge a nd Catterick were p reviously undefended and one a ssumes t hat t he p rovision of a walled c ircuit dg r i ovo i ndicates t hat by t he e arly 4 th c entury t hese s ettlements h ad become c entres o f s ome i mportance. b e e xpected t hat t hey would h ave u ndergone t he f ull ment and be p rovided w ith b astions.

i t m ight develop-

The b astions a t v arious c entres a re r egarded a s contemporary and part of a s ingle u nified s cheme probably i nitiated by T heodosius. The most u sual r eason o ffered f or t he a cceptance of a Theodosian context f or t he provision o f b astions i s t he r ecord o f Ammianus t hat Theodosius r estored t he c ivitates o f t he p rovince ( Webster 1 975a, 1 06; F rere 1 978, 2 91). B ut w e c annot b e sure t hat t his n ecessarily r efers t o t he a ddition o f bastions; t here i s v ery l ittle s atisfactory a rchaeological e vidence. Many b astions a re k nown o nly f rom aerial photographs and, e ven where e xcavation has t aken place, dating e vidence i s v ery e lusive; m any o f t he e xcavated e xamples were i nvestigated s ome t ime a go. The latest date f or t he a ddition o f b astions t o a n e xisting c ircuit c omes f rom G reat Casterton w here t wo c oins o f A D 3 54 a nd AD 3 58 w ere f ound a ssociated w ith t he c onstruction of two d ifferent b astions ( Corder 1 955, 3; Wacher 1 969, 4 9; F rere 1 978, 2 91; 3 s ee a lso Casey 1 983) . At Aldborough a n e arly d itch, w hich w as p artly f illed t o make w ay f or t he b astions, c ontained pottery down t o t he mid 4 th century ( Corder 1 955, 2 7; Charlesworth 1 971, 1 62; F rere 1 978, 2 91). At L ondon t he e xcavation o f b astion n o. 6 ( Duke's P lace) a nd a s ite n earby p roduced s ome d ating e vidence i n t he f orm o f a c oin o f Constans o f AD 3 463 48 f rom t he upper f ill o f t he d itch p artially overlain by t he b astion. A l ayer f ace of t he b astion t erminating

i n

AD

w hich h ad contained

3 70.

I t

a ccumulated a gainst t he outside c oins of t he H ouse of T heodosius

w as

7 3

s uggested

t hat

t he

b astion

m ay

have been e rected c . AD 3 41-375 b ut t he e vidence i s n ot v ery satisfactory ( Maloney 1 980, 5 9-60; 1 983; Marsden, 1 980, 1 72; Casey 1 983). At G loucester, a s ection o f t he wall on t he eastern circuit was r ebuilt a t s ome s tage i n the 4 th century a nd t he bastion f ound h ere may b e c ontemporary; u nfortunately, there was n o p recise dating ( Britannia ( 1970) 1 , 2 93). E ven i f t he i dea t hat b astions w ere a dded a s t he r esult of a single decision i s accepted ( Corder 1 955, 3; F rere 3 1 978, 2 91), e ven t he G reat Casterton e vidence, which i s t he latest in date s o f ar available, does n ot t ell u s how l ate t his development w as. Nor i ndeed n eed w e a ssume t hat t his p recludes t he p rovision o f b astions a t a much earlier date. Wacher ( 1975, 5 1) has cast doubt on a Theodosian context f or t he p rovision o f b astions a nd a s pecific e xample where h e f eels t he bastions m ight be much earlier i s B rough-on-Humber Here t here was n o d irect dating e vidence f or t he b astions a nd i t w as necessary t o base h is a rgument on t he evidence t hat t heir construction was l ikely t o b e c ontemporary w ith t he r ebuilding of t he n orth gate. On t his e vidence Wacher ( 1969, 4 9 & 5 4) o bserved t hat t he b astions would n ot f it i nto a Theodosian context, and i ndeed h e a rgues t hat a s B rough-onHumber was e ffectively abandoned i n t he mid 4 th c entury t he bastions must have been i n e xistence b efore t hat d ate. H is a ssessment o f t he e vidence i nvolved a much e arlier date a nd h e put f orward t he possibility t hat Carausius appreciated t he i mportance o f B rough-on-Humber a s a p ort a nd p ossible n aval base and t hat during h is n orthward e xtension of t he S axon Shore system ( AD 2 86-293) h e may h ave b een r esponsible f or t he p rovision o f bastions. Of course t he f act t hat a case has been made f or r egarding B rough-on-Humber a s p rimarily a n aval i nstallation f rom t he mid 2nd century t o i ts d esertion i n t he mid 4 th may u ndermine i ts r elevance t o t he d efences o f c ivilian s ites. H owever t he s tatus o f B rough-on-Humber i s open t o debate a nd i t i s p ossible t hat b astions may h ave b een provided o ver a much l onger period, beginning perhaps i n t he l ate 3 rd c entury, s imultaneously w ith t heir i ntroduction i nto t he m ilitary At L incoln,

a rchitecture o f t he r eplacement

t he p rovince ( Frere 1 978, 2 86). o f t he e xisting i nterval t ower a t

The P ark by a n ew w est gate f or t he l ower c olonia i s n ot s ecurely dated but a p re-Theodosian date i s p ossible. I f t he demolition a nd r eplacement o f t he o ther i nterval t owers w ere part of t he general operation o f r econstructing t he gate and t he t hickening o r r ebuilding o f t he c olonia w all, t hen a d ate earlier t han Theodosius may apply t o t he c onstruction of t he l arge s olid p latforms w hich may h ave b een e mplacements f or a rtillery ( Ant. J . ( 1975) 5 5, 2 40-1, 2 43, 2 56; Colyer 1 975, 1 6-18). U nfortunately t he e vidence f rom L incoln i s n ot r eally admissible i n t his c ontext s ince i t does n ot i nvolve t he p rovision o f b astions a s s uch. T he b astions a t C irencester a re u ndated b ut i t w as f elt t hat a hoard o f 4 80 c oins f rom t he s ilt o f t he i nner d itch o utside t he n orth-east gate m ight p rovide a n i ndirect d ate s ince the p resence o f b astions on t he e dge o f t his d itch s eems t o h ave n ecessitated t he d iversion o f t he main s tream t o t he outer ditch and t he consequent s ilting and s tagnation o f t he i nner. Dated material f rom t his d itch w as t herefore t hought

7 9

t o p rovide a t erminus a nte q uem f or t he c onstruction o f t he b astion a nd, s ince t he h oard r evealed o n p reliminary e xamination t hat t here w ere n o c oins l ater t han AD 3 46, i t w as s uggested t hatthe b astions w ere b uilt b efore t hat d ate and a t s ome t ime d uring t he f irst h alf o f t he 4 th c entury ( Ant. J . ( 1961) 4 1, 6 9). O f c ourse, t his c annot r eally b e a ccepted a s a t erminus a nte q uem f or t he b astion a nd b etter e vidence i s r equired b efore t hese b astions c an b e a ccepted a s p reT heodosian. I f t his e arlier 4 th c entury d ate s hould p rove t o b e c orrect, i t w ould c onfirm t he n eed t o r econsider t he v iew t hat b astions w ere b uilt a s a s ingle e vent i n t he T heodosian p eriod. C asey ( 1983) h as a rgued t hat b etter e vidence f or a d ate e arlier i n t he 4 th c entury i s a vailable f rom Caerwent f rom c oin e vidence f rom t he n orth-east a ngle b astion. T he f oundation t rench o f t he b astion w as s ealed by a c obbled f loor, i tself s ealed b y m ason's c hippings. O n t he f loor w as a c oin h oard w hich h ad p resumably b een d eposited d uring t he c onstruction o f t he b astion. C asey's a rgument t hat t his p rov ides a t erminus a nte q uem o f A D 3 50 f or t he c onstruction o f t he b astion i s b ased o n t he t hree c oin i ssues i nvolved, t he i nterrelationship b etween t hem a nd t he c ircumstances o f t he 4 th c entury c oin c irculation. T he d itches f illed f or t he c onstruction o f t he b astion a lso c ontained a s equence o f c oins o f a p articular i ssue o f C onstans ( Britannia ( 1972) 3 , 3 02). T he n umismatic e vidence i n g eneral f or t he c onstruction o f b astions i nclined C asey t o t he v iew t hat t he d ating o f t he i nception o f t his a ctivity t o, o r j ust b efore, t he a ccession o f M agnentius ( AD 3 50) h ad g reater v alidity t han t he c urrent a ttribution o f t hem t o t he a ctivities o f C ount Theodosius i n A D 3 67. H e p ointed t o t he f act t hat t he c oinage w hich s hould b e l inked w ith T heodosius' v isit h as n ot b een f ound i n a ssociation w ith t he c onstruction o f b astions. I n t his m atter, ' it w ould a ppear t hat B ritain a nticipated i mperial p articipation a nd c oncern f or u rban d efences b y s ome d ecades', f or C asey c onsiders t hat a t t he t ime w hen b astions w ere b eing p rovided i n B ritain, a ny d efensive a ctivity w ould h ave h ad t o b e f inanced f rom t he f unds o f i ndividual t owns b ecause t he i mperial g overnment w as n ot i n a p osition t o g ive i t f inancial b acking. A s w ell a s d oubts a bout t he d ating o f b astions t o a T heodosian c ontext, t here a re d ifferences o f o pinion c onc erning t heir f unction a nd t he i mplications o f t heir p rovision f or c ivil s ettlements. O n t he o ne h and, t he p rovision i s s een a s i nvolving a n a ltogether n ew c oncept i n d efence i n t hat t he e mphasis, i n a m ilitary c ontext a t l east, i s n ow p laced o n a d efensive r ather t han a n o ffensive r ole a nd o n e stablishing a s ystem o f w ell-defended s trong p oints r ather t han o n t he e arlier u se o f m ilitary i nstallations a s b ases f or t he o ffensive. I t h as b een a rgued t hat t he p rovision o f b astions f or m ilitary a nd c ivilian b ases a like t herefore r eflected t he d efensive r ole w hich t he R oman a rmy h ad b een f orced t o a dopt i n t he f ace o f i ncreasing p ressure o n t he f rontiers ( Frere 1 978, 2 91). T he p rovision o f b astions f or c ivil s ettlements i s r egarded a s a ' modernisation' o f c ivil d efences t o b ring t hem i nto l ine w ith m ilitary a rchitecture a nd a s a r eflection

8 0

o f

t he

r ole

w hich

t hese

d efences

might

n ow

b e

e xpected

t o

play. Webster's ( pers. comm.) view, not necessarily s hared by others, i s t hat, u ntil t his t ime, t own d efences w ere i ntended not t o withstand a s iege o r engage i n confrontation w ith a s erious enemy b ut t o operate r ather a t t he l evel o f a deterrent

a gainst marauding

bands.

O n t he other h and, t here a re u ncertainties a bout t he u se of b astions. They a re u sually r egarded as emplacements f or a rtillery m issile-firing c atapulta , a nd b allista ,, a nd i n t he b est p reserved e xamples i n Gaul, t he a rched c atapult p orts c an still b e seen ( Baker 1 970, 2 11-19; F rere 1 978, 2 90-1) . The catapulta was a s mall s iege machine f or d ischarging m issiles or b olts and t he b allista was a l arger v ersion. Both operated -

i n much t he s ame w ay a s t he c rossbow. Ballistarii, s pecially t rained a rtillery units, a re attested i n t he Notitia D ignitatum a nd t he a ccounts o f A nunianus i ndicate t hat by t he mid 4 th century s iege engines were widely u sed on campaign ( Webster 1 983). H owever, Boon ( 1974, 6 8) h as c ast d oubts on the u se of a rtillery i n c onnection w ith bastions i n t he l ate empire; ' there i s n o e vidence t hat t he t owers o f RomanoB ritish town walls were d esigned f or, o r c ontained, e ven f rom t ime t o t ime, b allistae ( though t his l atter p ossibility may b e admitted) . He s uggests t hat bastions and t owers were too small t o h ave b een s uitable a s emplacements f or ballistae a nd considers t hat t hey were p robably manned by a rchers who would p rovide f lanking f ire a long t he w all ( Boon 1 974, 6 8 & 1 3). Webster ( 1983) a lso n otes t hat t he only surviving e xamples o f openings i n a ny t ower i n B ritain a re a t t he S axon S hore f ort of P evensey and t he l egionary f ortress a t Y ork and i t i s questionable w hether a ny o f t he bastions i n B ritain w ere '

suitable

f or

carrying

s iege

engines.

The p rovision o f b astions f or t own a nd v illage d efences, i f t hey were i nter ited t o carry a rtillery, i mplies t he deploy7 ment o f p ersonnel t rained i n i ts u se ( Wacher 1 974, 7; 1 978, 1 00; F rere 1 978, 2 90-1). H owever, v irtually n othing i s k nown about t he manning o f c ivil defences a t a ny p eriod. I t i s assumed t hat t he d efences were p robably manned by t he t ownspeople t hemselves. I n t he 4 th c entury t his s ituation may h ave changed. F rere i nclines t o t he v iew t hat garrisons were n ow being p rovided t o man t he n ew ' urban' d efences a nd p oints t o the f act t hat t roops were c ertainly s tationed i n s ome of the t owns i n G aul ( Frere 1 978, 2 91, 3 97 & 3 99). There i s n o evidence t hat t he t owns and v illages had a nything i n t he way of t rained u rban u nits o r a l ocal m ilitia ( Todd 1 973, 1 33-4; 6 7 Boon 1 974, 6; Wacher 1 974, 7-8; Webster 1 983) Regulations a pparently s how t hat by t he m id 4 th c entury i t w as quite normal f or t roops t o be b illeted i n t owns ( Wacher 1 974, 7 8), but t his i s n ot n ecessarily t he s ame a s t he a ctual deployment of t roops f or t he purposes of defending c ertain c entres.

has

The b een

p resence o f o fficial-issue b uckles a nd b elt-fittings t hought t o i ndicate t he s tationing o f detached units

o f t he r egular a rmy, p articularly G ermanic centres i ncluding both major t owns and s ettlements

( Hawkes

& Dunning

1 961,

8 1

4 0-1;

l aeti, a t c ertain smaller r oadside Todd

1 973,

1 33-4;

B oon 1 974, 6 6-70; W acher 1 978, 1 00). B ut t he e vidence i s f ragmentary a nd d iverse a nd i ts i mplications d ifficult t o a ssess. B uckles a nd b elt-fittings w ere i ssued t o c ivilian o fficials a s w ell a s m ilitary p ersonnel a nd t hey w ere a m ark o f r ank r ather t han e thnic o rigin. W eaponry i s a m ore u seful g uide t o t he p resence o f m ilitary p ersonnel b ut e ven s o s uch f inds m ay r epresent t roops i n t ransit r ather t han t he u se o f a c ivilian s ite a s a p ermanent b ase. S imilarly, W ebster a rgues t hat t he d iscovery o f l ead-weighted j avelins ( plumbatae) i n t he B aths B asilica a t W roxeter i ndicates t he r equisitioning o f t he b uilding a s a t emporary b ase f or t he o peration o f a d etachment o f s pecial t roops a nd n ot t he s tationing o f a p ermanent g arrison f or t he d efence o f t he t own ( Webster 1 975a, 1 05-6; B arker 1 979, 9 7-9; s ee a lso S alway 1 980, 1 6). W ebster ( 1983) a rgues t hat s hortages o f m anpower i n t he l ater R oman p eriod w ould m ean t hat r egular a rmy p ersonnel a nd s pecially t rained u nits o f b allistarii w ould n ot b e a vailable t o d efend t he c ivil s ettlements a nd m an a ny s pecialist e quipm ent a ssociated w ith t he b astions. H e r egards t he p resence o f m ilitary u nits i n t he c ivil s ettlements a s s imply a n i ndication o f t heir u se a s b ases f or a rmy o perations a nd t heir w ider r ole a s p art o f t he m ilitary s ystem o f t he p rovince. H owever, h e d oes o ffer t he s uggestion t hat t he v eterans w ould f orm a n o bvious s ection o f t he p opulation w ho m ight b e e xpected t o o rganise d efence. T he i ntroduction o f b astions s eems t o m ark t he l ast m ajor w idespread p hase i n t he s tructural h istory o f t own d efences. B ut i t m ay b e a n i llusion t o r egard t heir c onstruction o r e ven t heir p eriod o f u se a s n ecessarily r elevant t o t he a bility o f t he t owns t o d efend t hemselves o nce t he o fficial g arrisons h ad b een w ithdrawn f rom B ritain. N othing i s k nown o f t he p eriod o f u se o f t hese b astions n or i ndeed p recisely h ow t hey f igure i n t he d efensive t actics o f t he s ettlements ' involved. T here i s n o i ndication o f t he s teps w hich n eeded t o b e t aken i n o rder f or t he t ownspeople t o b e a ble t o m ake e ffective u se o f t he b astions. I n c onclusion, t he m eagre a nd i nconclusive e vidence f or t he d ating a nd f or t he f unction o f b astions a llows c onsiderable s peculatation. I n p articular, w hile s ome a uthorities f avour a T heodosian c ontext ( after A D 3 67) f or t he p rovision o f b astions, o thers f avour t he v isit o f C onstans ( AD 3 43) a s t he m otivating e vent. B ut t here i s s till t he p ossibility t hat b astions w ere n ot p rovided a s t he r esult o f a s ingle s cheme a nd t hey m ay h ave b een p rovided a t d ifferent c entres a t d ifferent t imes e ither b ecause s everal i mperial i nitiatives w ere i nvolved o r b ecause t heir p rovision w as p urely ,a m atter o f l ocal i nitiative. F urthermore c onsidera tions o f f inance m ay h ave p rolonged t he p rocess a nd e ven a long t he s ame c ircuit d ifferent b astions m ay h ave b een p rov ided a t d ifferent t imes.

3 2

9 . Did Burgi a nd t he e ffect o f w ider m ilitary considerations i nfluence the p rovision o f walled c ircuits f or many c ivil s ettlements?

I n t he p revious s ection s ome c onsideration w as g iven t o t he a rgument t hat t he p rovision o f bastions was a r eflection o f t he f act t hat d efended c ivil s ettlements h ad a g reater r ole t o p lay i n t he defensive network o f t he p rovince i n t he l ater R oman period. This a rgument h as b een e xtended t o t he a ctual p rovision of defences at certain c ivil s ettlements, s ome o f w hich have b een g iven t he n ame b urgi t o emphasise t he contention t hat t he p rovision o f defences i n t hese i nstances i nvolved c onsiderations o ther t han t hose f or t he s ettlement. I n t he a rgument f or a s ystem o f b urgi, one c an s ee t he more specific application o f Webster's ( 1975, 5 3 & 5 9) general emphasis on t he s ecurity o f t he main r oads i n a ll p eriods a nd t he i mportance of t he communications n etwork i n maintaining o r e stablishing c ontrol o ver t he p rovince. According t o Webster ( 1971, 3 8 & 4 2) , i t i s t heir r elationship t o t he Watling Street t hat p rovides t he most c ompelling r eason f or b elieving t hat t hey w ere d irectly associated w ith t he p rotection of t his s trategic r oute, f or e ach o f t hese r ectangular e nclosures i s b isected by t he r oad and a t both L etocetum and T ripontium t he p resence o f a h illtop, w hich h as much g reater v alue f rom t he purely military point of view, was i gnored ( LSSAHS 1 963-1964 ( 1964) 5 , 3 ; T BAS 1 966 & 1 967 ( 1969) 8 3, 1 34). I n e ach case, the enclosure i s a ctually p laced at t he s ite of a n e xisting s ettlement w hich w as p reviously u ndefended, e xcept i n t he case of Bannaventa and possibly a lso Manudessedum where t here a re i ndications o f a n e arlier e nclosure w hich h ad b een f illed i n s ome t ime b efore ( TBAS 1 956 ( 1958) 7 4, 3 2, 3 4 & 3 9; 1 967-1970 ( 1971) 8 4, 2 5; Todd 1 973, 3 8; B ritannia ( 1973) 4 , 3 25). But i n e ach case t here can be n o ' question of any a ttempt t o e nclose t he s ettlement. O f c ourse, s uch a n a ttempt would not have b een s trictly f easible s ince t hese s ettlements i nvolved d istances

r ibbon a long t he

d evelopment s preading f or c onsiderable r oad, i n t he case o f L etocetum f or about

t wo m iles. But, i t i s n oticeable t hat a t b oth T ripontium a nd L etocetum t here was a ' focus' i n t he f orm of complexes which h ave been e xplained a s mansiones. I n n either c ase w as t his n ucleus enclosed w ithin t he burgus defences. I n f act, a t e ach o f t he burgi, t here i s a l ack o f e vidence f or a ny c ontemporary buildings o r o ccupation o f any k ind w ithin t he defences and t his c ould b e e xplained by t he r equisitioning o f t hat p art o f t he s ettlement by t he a rmy ( LSSAHS 1 963-1964 ( 1964) 5 , 1 6; Wilson 1 975, 1 4; Webster 1 975a 7 8). H owever, t his p oint h as perhaps been over-emphasised and t he d ismantling of b uildings a long , t he l ine o f t he d efences a t Manduessedum may h ave f urther ' p rejudiced t he p icture by s uggesting military a ction w ithout r egard t o t he i nterests o f p rivate p roperty ( TBAS 1 967-1970 ( 1971) 8 4, 2 1, 2 4-5; ( 1974) 8 6, 5 6; Webster 1 971, 4 2; 1 975b, 5 6) , a lthough t he l ate R oman walls o f Cambridge a nd G reat Chesterford w ere a lso l aid o ut o ver e xisting b uildings and i t i s d ifficult t o s ee how t his could h ave b een a voided w here t he d efences w ere e rected l ate i n t he l ife of t he s ettlement ( Alexander 1 975, 1 08) . F irst,

8 3

e xcavation h as t ended t o c oncentrate r ather more on t he defences t han on the i nterior and t here has b een l ittle work a t P ennocrucjum, U xacona a nd Bannaventa. The r elative s carcity of 4 th century pottery a nd coins i s c ertainly s ignificant but t hese s ites do s eem t o h ave s uffered f rom p loughing o r o ther e rosion o f t he l ate Roman l evels a nd t he r elevance of t his does n eed t o b e e valuated. Occupation i n t he 4 th c entury i s c ertainly n ot e ntirely a bsent a nd t ook p lace on part at l east o f t he s ite a t T ripontium ( TBAS 1 9711 973 ( 1973) 8 5, 9 3-144). At L etocetum, t he upper l evels w ithin t he enclosure have b een r emoved by t he p lough b ut 4 th c entury p ottery a nd c oins h ave b een r ecovered f rom i nside t he defences ( LSSAHS 1 963-1964 ( 1964) 5 , 8 & 1 6). The s ame i s t rue o f Manduessedum b ut l imited i nvestigation of t he i nterior did r eveal o ccupation down t o t he e nd of the 3 rd c entury a t l east a nd deposits c ontaining p ottery o f t he e arly a nd p robably mid 4 th century were f ound o verlying t he t ail o f t he r ampart; t hey may b e a ssociated w ith t races o f walling which may r epresent b uildings i nside t he enclosure ( TBAS 1 956 ( 1958) 7 4, 3-4; 1 967-1970 ( 1971) 8 4, 2 5). At a ll o f t hese 3 s ettlements, t he r emains a re i n any c ase ephemeral c onsisting o f t he f ragmentary r emains o f s uch s tructures a s t imber b uildings and yards. G iven a lso t he f act t hat t hese b urgi a re a ssumed t o h ave h ad a v ery s hort p eriod o f u se, i t w ould b e difficult t o p rove t hat a ny f eatures i n t he i nterior were c ontemporary. Any more positive i ndications o f t he f unction o f t hese enclosures must await f urther e xcavation. Certainly, t he a bsence o f a ny i ndications o f a m ilitary p resence i s n oticeable. E ven i f t hey were constructed s imply a s t ransit c amps, s ome e vidence i n t he f orm o f m ilitary e quipment might b e e xpected. A lso, i n connection w ith t heir u se i n k eeping open c ommunications, p rovision f or t he c ontrol of e ntry i nto t hese enclosures has yet t o b e i nvestigated. There i s n o e vidence f or massive masonry gates t o a ccompany t he s tone walls, a lthough, apart f rom t he w est g ate a t U xacona, n one o f t he e ntrances h as b een e xcavated. A p ossible gate t ower i n s tone was f ound a t U xacona, but i t w as s o heavily r obbed t hat v ery l ittle i nformation c ould b e r ecovered ( WMNS ( 1959) 2 , 7 ; SaNL ( 1960) 1 3, 4 ; ( 1961) 1 6, 2 ; ( 1962) 1 8, 3 ). T he e vidence i n g eneral f or t he e xtent o f t he s ystem a nd f or u se of i ndividual b urgi i s a t p resent very l imited and much o f t he a vailable e vidence c omes f rom a erial photographs. The dating e vidence i s very i nconclusive a nd t wo f actors i n p articular h ave b een r esponsible f or t he s uggested l ate 3 rd/early 4 th c entury date: t he l atest material i n t he ramp art a t Manduessedum b elonged t o t he l atter p art of t he 3 rd c entury ( TBAS 1 956 ( 1958) 7 4, 3-4; 1 967; 1 970 ( 1971) 8 4, 3 2 1), a nd t he p ottery a ssociated w ith t he f inal phase o f a b uilding demolished t o a ccommodate t he d efences was n ot l ater t han t he mid 3 rd c entury ( TBAS 1 967-1970 ( 1971) 8 4, 2 5; ( 1974) 8 6, 5 6) . At L etocetum, t here was l ittle o ccupation d ebris i n t he r ampart a nd t he dating r elied o n m aterial s ealed by t he d efences. B elow t he p resumed c onstruction l evel t he w all w ere s herds w hich s uggested a 3 rd c entury d ate l ater a nd a well s ealed by t he wall contained i n

8 4

f or o r i ts

deliberate f ill material which a gain s uggested a date n ot earlier than t he 3 rd century ( LSSAHS 1 963-1964 ( 1964) 5 , 3 , 4 & 7 ). Webster ( 1975b, 4 6) n otes t hat f our o f t he b urgi ( Letocetum, Manduessedum, T ripontium and U xacona) h ave b een dated t o t he e arly 4 th c entury only on t he a bsence o f l ater pottery which h e r egards as s ignificant ( TBAS 1 967-1970 ( 1971) 8 4, 2 5; ( 1974) 8 6, 5 6-7). H owever, i n o rder f or t he s ites t o f it t he limited context o f Constantius Chlorus' r estoration o f t he p rovince ( AD 2 96-306) r ather b etter e vidence i s n eeded, even f rom Manduessedum. T his p articular dating h orizon i n t he l ate 3 rd/early 4 th century does a llow t hese enclosures t o f all w ithin t he context o f t he more w idespread p attern o f t he p rovision o f s tone w alls which i s generally a ccepted a s t aking p lace a t t his t ime . I n . a rguing a gainst t his more general c ontext f or t hese p articular s ites, Webster ( 1975a, 4 6) has emphasised t hat t he defences were s hort-lived immediate t hreat

a nd h ad

d id been

n ot c ontinue dealt w ith.

i n At

u se a fter t he Manduessedum,

T ripontium a nd t J xacona t here a re f airly p ositive i ndications that e fforts were made t o l evel t he d itches a t an early d ate. At t J xacona, a c oin h oard b uried i n t he u ppermost f ill w as expected on detailed analysis t o show t hat t he enclosure had ceased t o f unction b y t he mid 4 th c entury ( WMNS ( 1973) 1 6, 2 1) I n the B roadclose a rea a t Manduessedum, where t he l arge pottery i ndustry w as l ocated, t he o uter d itch o f t he t riple ditch system had b een l evelled off w ith i ndustrial waste a round t he m iddle o f t he 4 th c entury, while t he u pper s ilt o f the d itches on t he e astern part o f t he c ircuit c ontained pottery of t he l ater 4 th c entury ( JRS ( 1964) 5 4, 1 64-5; T BAS 1 967-1970 ( 1971) 8 4, 2 5) . At T ripontium, t he e xcavator t hought t hat t he d itches h ad b een d eliberately a nd c arefully f illed after b eing open only a s hort t ime a s t his d itch i nfill contained e arly 4 th c entury p ottery a t a ll l evels ( TBAS 1 9711 973 ( 1973) 8 5, 9 6; ( 1974) 8 6, 5 7); while a t Bannaventa, t he outer d itch h ad b een a llowed t o s ilt u p n aturally u ntil i t became a r ubbish t ip i n o r after t he l ate 4 th century ( Britannia ( 1973) were s till open ( 1964) 5 , 3 ).

4 , i n

2 35 & 2 96). t he medieval

At L etocetum, t he d itches period ( L ISSAHS 1 963-1964

W ebster ( 1975a, 4 6) a rgues t hat t he d itches w ere deliberately f illed i n p resumably b ecause t hey were a danger , t o t raffic a nd h ere a gain t he i nvestigation o f t he e ntrances might p rove i nteresting i n t erms o f the l ater u se o f t hese enclosures. Within t he c ontext o f t heir s pecialised u se a s t ransit camps f or t roops e ngaged i n dealing w ith marauding bands, one would c ertainly e xpect t hat t heir p eriod o f u se would h ave b een s hort. H owever, Webster ( 1 975a, 4 6) h imself observed t hat a lthough t he t rouble may h ave b een q uickly r esolved, t he b urgi were n ot c onstructed a s s hort-term e xpedients, a s t here i s e vidence t hat s tone w alls w ere p rovided at l east a t L etocetum, Manduessedum a nd t i xacona and possibly a lso a t P ennocrucium. We c annot b e c ertain o f t he early abandonment o f t hese enclosures a s d efensive f eatures. There must h ave b een many s ubsequent o ccasions o n w hich t hey could have b een u sed by t he a rmy, i f only i n t he context o f t he w ork of Count T heodosius. Apart f rom t his, e ven i f t hey

8 5

p rove t o b e p urely m ilitary b ases,their s ubsequent h istory a nd u se c ould have been at l east partly a matter for t he i nhabitants o f t he i ndividual s ettlement c oncerned. After t he Roman w ithdrawal, one might have e xpected t hese a bandoned burgi t o h ave p rovided u seful s trong p oints f or t he l ocal population. I t s eems unlikely t hat t heir ditches w ere f illed i n b ecause t hey b ecame 'a danger t o t raffic s ince p art of t he o riginal f unction of t he burgi would h ave b een to p rotect a nd k eep open t he c ommunications n etwork. I f t hey w ere deliberately s lighted t hen i t s eems l ikely that t he l ocal c ommunity decided t hat t hey w ere detrimental t o t he a ctivities of t he s ettlement; perhaps t he c learest e xample of t his occurs at Manduessedum. Other c entres, s uch a s Letocetum a nd Bannaventa, may have f ound a u se f or t he defences. C ertainly, one m ight e xpect t hese f ortified r oad s tations t o h ave b een suitable s trong p oints f or Romano-British s ettlement c ontinuing o n i nto t he s ub- a nd p ost-Roman p eriods. I f t hese defended enclosures s hould prove t o b e a l ocalised p henomenon o f t he West M idlands a nd n ot p art o f a more e xtensive system, i t may b e n ecessary to c onsider a n a lternative e xplanation f or t heir c onstruction. I n t his c ase, perhaps more l ocal c ircumstances might b e considered, possibly w ithin t he c ontext o f t ribal r ather t han p rovincial policy. I t i s unfortunate t hat we know s o l ittle a bout t he r esponsibilities o f t he i ndividual t ribes i n t he matter o f t he policing of t he civitates and i n t he defence of t he settlements w ithin t heir t erritory. One f urther p roblem c oncerns h ow f ar t he s eries o f f ortified points a long t he Watling Street can b e s een a s s tanding a lone i n B ritain w ithout a ny apparent r elationship t o t he general pattern o f t own and v illage defences ( TBAS ( 1971) 8 4, 2 6; ( 1974) 8 6, 5 6; Webster 1 971, 4 2). Two p oints a re i nvolved. F irstly, R ivet ( 1975, 1 12) a nd Todd ( 1973, 3 9) a re both i nclined t o t he view t hat t hese b urgi o r f ortified centres a re s imply part o f t he l ate 3 rd century defensive a ctivity which i nvolved t he p rovision o f s tone w alls t o p rotect government i nstallations, particularly c ollection points f or t he a nnona ( taxes i n k ind) w hich w ere l ocated i n many c ases a t ' lesser' centres o r ' minor' s ettlements. T he multiplication o f s uch c entres i n t he 3 rd c entury w ould m ean t hat many more of t hese s ettlements would qualify f or defences a t t he t ime w hen s tone walls w ere b eing e rected ( Wacher 1 975, 5 1; 1 978, 9 9; Webster 1 975b, 5 9-64) . These stone walls w ere p rovided t herefore n ot f or t he d efence o f t he s ettlement b ut f or t he p rotection o f tax collectors and t heir goods f rom t he s ettlers t hemselves, f or i n t he 4 th c entury t he a nnona h ad become a c ause of c onsiderable r esentment ( Todd 1 973, 3 9). R ivet t herefore s uggests t hat t he b urgi o f t he West Midlands were constructed t o p rotect government p roperty not o nly f rom b arbarian marauders b ut f rom t he C ornovii t hemselves; t he apparent absence o f any s ettlement w ithin t hese e nclosures would b e e xplained by t he u se o f t he a rea f or b arns a nd s tockyards f or s torage of t axes ( Rivet 1 975, 12). Secondly, T odd 1 s uggests t hat many o f t he d efended ' minor' s ettlements m ay have b een part of a s ystem o f burgi o r defended r oad stations h ousing

small

garrisons

o f

l ine-of-communication

8 6

t roops

f or

t he p urpose o f p rotecting t he r oads i n t he l ater e mpire. This would i nvolve much t he s ame i nterpretation a s t hat o ffered by Webster f or t he b urgi a long t he Watling Street b ut i n a l ess specific historical and chronological context. T odd d rew attention t o t hree s ites on t he F osse Way s outh-west o f L incoln Brough-on-Fosse ( Crococalana), Thorpe-by-Newark ( Ad Pontem ) a nd Margidunum. B rough i s k nown o nly f rom f inds a nd a erial photographs w hich h ave r evealed i ntensive s ettlement extending a long t he F osse Way and f or a c onsiderable d istance e ither s ide o f i t. A d efensive s ystem w as e stablished w ithin the s ettlement and i s r epresented by a double d itch s ystem o f an a pproximately r ectangular e nclosure o f c . 7 a cres/2.8 hectares, o f which t he west and parts o f t he n orth and s outh s ides h ave b een r evealed by a erial p hotograph b ut t here i s n o t race of a wall o r r ampart. No c ontrolled e xcavation has t aken p lace a nd t he d efences r emain u ndated ( St. J oseph 1 966, 2 7-8; Todd 1 973, 3 8; Wilson 1 975, 1 4, p 1. X IIIb). This small e nclosure i s s lightly l arger t han L etocetum ( 6 a cres/2.4 hectares) and Manduessedum ( 5 a cres/2.1 hectares) and l ike t he burgi i t i s p robably r ectangular i n s hape, b isected b y t he Roman r oad and e nclosing only a small part of t he o ccupied -

a rea. A t Margidunum, e xcavations i n t he 1 940s, 1 950s a nd 1 9661968 e stablished s omething o f t he h istory o f t he s ettlement and i ts defences w hich f orm a n i rregular p olygon a stride t he Fosse Way. An e arthwork c ircuit ( terminus p ost q uem . c . AD 1 40) e nclosed a n a rea o f 7 a cres/2.8 h ectares a nd t he l ater, substantial, 8 f eet(2.4 m ) w ide wall ( undated) f ollowed the same a lignment. B ut i n c ontrast t o t he e arlier p eriods t here is v ery l ittle e vidence f or o ccupation e ither i nside o r outs ide t he defences i n t he 4 th c entury. The only b uildings which have p roduced e vidence o f u se i n t hat c entury a re t wo structures o f s ome s cale a nd p retension i n t he n orth-west corner of t he enclosure ( Todd 1 975, 2 11-15). The combination of m assive d efences w ith s ubstantial b uildings i n o ne c orner of a n o therwise d eserted s ite s uggested t o Todd ( 1970, 7 3, 5 45 ; 1 975, 2 15) t hat t he s ite m ight h ave b een t aken o ver a s a n official i nstallation s uch a s might have b een u sed by a beneficiarius c onsularis o r s ome o utstationed s oldier c harged with t he s ecurity o f t he main r oads c rossing t he t erritory o f the C oritani. B ut i t i s d ifficult t o s ee w hy t his s hould h ave involved t he absence o f s ettlement. This might h owever b e more r eadily e xplained i f t he s ite h ad b een r equisitioned a s a position o f s trengh where mobile u nits could camp and d raw supplies

( Webster

1 975b,

5 9).

A t Thorpe-by-Newark, a w all a nd d ouble d itch s ystem enclosing c . 4 .5 a cres/1.8 h ectares w as p rovided i n t he 3 rd o r possibly, t he 4 th c entury o n t he s ite o f a n e arlier e arthwork ( terminus p ost q uem AD 1 40-160) but w ithout r especting i ts a lignment ( Trans. Thoroton Soc. 1 965 ( 1966) 6 9, 3 1; JRS ( 1961) 5 1, 1 77; S t. Joseph 1 966, 2 8-9, f ig. 1 ; W ilson 1 975, 1 0). Aerial p hotographs s how b uildings w ithin t his rectangular enclosure i ncluding a r ectangular s tone b uilding just i nside a nd p arallel w ith t he n orth w all ( St. J oseph 1 966, 29). Both the earlier earthwork and t he l ater walled c ircuit enclosed s uch a n arrow s trip a long t he n orth-west f rontage o f

t he F osse Way t hat t here c an h ave b een l ittle r oom f or buildings on t hat s ide and i n f act t here may have b een l ittle o ccupation h ere f rom t he b eginning, f or t he g ravel t errace o n which t he s ettlement was l ocated ends a long t he l ine adopted by t he d efences on t his s ide o f t he r oad. The e nclosing o f such a narrow a rea i s t herefore easily e xplained by t he t opog raphy o f t he s ite a nd i s a r eminder t hat c aution s hould b e e xercised i n asserting t hat enclosures o f t his n ature o ften h ad l ittle r egard f or t he l ayout o f t he s ettlement w ith w hich t hey w ere associated ( Webster 1 975b, 5 6). The f act that t he earlier a s w ell a s t he l ater e nclosure coincides w ith t his natural f eature i s s ignificant s ince t hey d iverge e lsewhere a nd a erial p hotographs a nd t he d istribution o f f inds confirm t hat occupation l ay a lmost entirely s outh-east o f t he F osse Way ( Trans. Thorotor i Soc. 1 965 ( 1966) U ‚ 2 0 & f ig. 1 ). C ropmarks and surface f inds s how t hat the s outh-east s ide of t his e nclosure was d ensely built u p a nd t hat o ccupation c overed t he period f rom t he i st t o the end o f t he 4 th c entury b ut a gain t here i s n o i nformation a bout t he p recise nature of t his o ccupation at any period. I t might b e t hought unlikely t hat t his e nclosure r epresents a b urgus s ince i t would h ave i nvolved t he r equisitioning o f a n a rea which appears, i n t erms o f t he density a nd t ime s pan o f i ts o ccupation, t o r epresent perhaps t he ' nucleus' of the R oman s ettlement a t Thorpe ( Trans. Thoroton Soc. 1 938 ( 1939) 4 2, 1 -14; 1 965 ( 1966) 6 9, 1 9-22; Wilson 1 975, 1 0) Very s imilar t o Thorpe i n s ome r espects i s A ncaster o n t he E rmine Street s outh o f L incoln, where a r ectangular c irc uit e nclosed a n a rea on e ither s ide o f t he main r oad b ut apparently, l ike Thorpe, t he defences ran very c lose t o t he r oad f rontage on o ne s ide ( west) ( Whitwell 1 970, 6 5-7; T odd 1 981) . There i s n o certain i ndication a t Ancaster o f a n e arlier d efensive c ircuit a nd t he w all, t he construction o f which has b een p laced within the p eriod AD 2 50-280, r epresents a d efence d e n ovo ( Todd 1 981). The d efences enclose only a very small part ( 9.1 acres/3.5 h ectares) o f a m uch l arger s ettlement a rea o f s ome 6 0 — 7 0 a cres ( 24-28 h ectares) spreading a long t he E rmine Street f rontage and well away f rom i t t o t he e ast a nd w est. E xcavation h as b een l imited b ut t here i s no mention o f any e vidence w ithin t he walls to s ugg est t hat t he o ccupation o f t his a rea c hanged w hen t he defences were c onstructed. As a t Thorpe, t he a rea c hosen f or t he d efences i s o ne o f f airly d ense o ccupation ( Todd 1 975, 21) . 2 Ancaster does s eem t o b e t he o nly one of t he examples s o f ar d iscussed w here b astions w ere s ubsequently p rovided. I n t his case t hey h ave b een l ocated s o f ar a t the n orth-west a nd s outh-west angles a nd a re f an-shaped l ike t hose a t Godmanchester ( Todd 1 981) Todd ( 1973, 1 27-8) h as a lso d rawn a ttention t o a s mall r ectangular earthwork w ith t riple d itches a t Scaftworth n ear Bawtry, ( Notts.) n orth-west o f L incoln a nd s ited near a c rossing o f t he R . I dle. The a rea e nclosed i s l ess than o ne a cre ( 0.4 h ectares). Dating e vidence i s s canty but w hat t here i s s uggests a date i n t he s econd h alf of t he 4 th century, p robably a fter AD 3 70 a nd T odd c onsiders t hat t he e nclosure was

perhaps

constructed

a s

part

8 8

o f

t he

Theodosian

r ecovery

o f

t he p rovince. He s uggests t hat i ts p osition points t o a possible use i n t he policing o f t he i mportant r oute f rom t he T rent valley t owards D oncaster ( a p ermanent f ort) a nd t hence i nto t he Vale o f Y ork. Superficially, t his s ite would s eem t o r epresent a purely military post s ince t here i s apparently n o c ivil s ettlement h ere c omparable w ith a ny o f t he other c ases. T he characteristics which d istinguish a ll t hese enclosures t he b urgi o f t he West M idlands a nd s ites s uch a s Brough, Margidunum, Thorpe and possibly a lso Ancaster i n t he East M idlands a re t hat t hey l ie on major a rterial r outes a nd they i nvolve only a small defended r oadside nucleus w ithin a much l arger s ettlement. While t he b urgi o f t he West M idlands a re d istinguished by their small s ize, by t he b isecting o f t he defended a rea by t he R oman r oad a nd by t he r egular r ectangular shape o f the a rea enclosed, s trictly speaking t here i s n othing t o d istinguish t he s lightly l arger a nd l ess c onsistently regular enclosures o f t he East Midlands f rom t he very l arge number o f defended r oadside s ettlements s uch a s D orchester-onThames, G reat Casterton, and Water N ewton, e xcept perhaps i n the l imited s ize o f t he a rea e nclosed ( see Table F our f or comparative s izes) . The o ccasional very c lose p roximity o f one s ide o f t he d efences t o t he R oman r oad i s a p uzzling f eature which may underline the f act t hat n o attempt was being made t o p rotect t he s ettlement. H owever, t he f unction o f these enclosures and particularly t he e xistence o f any military o r o fficial p resence w ithin t hem r emains a matter o f speculation. I t i s d ifficult t o s ay t herefore h ow f ar i t i s t rue t o r egard t his t ype o f small r ectangular r oadside enclosure a s unusual, unrelated t o t he s ettlement, and r ep resenting s ome f acet o f military o r a dministrative a ctivity which took place w ithin the c ontext o f t he l ate 3 rd and 4 th centuries. The v arious s uggestions p ut f orward b y Webster, Todd a nd Rivet f or t he use o f t hese enclosures d o n ot c onflict t o t he extent t hat small p osts f or l ine-of communication t roops might have b een p laced where t hey could a lso conveniently p rotect c entres f or t he c ollection o f t he a nnona -

-

or s ome other official i nstallation, while t he possibility that t hese e nclosures may h ave o ffered r efuge i n t ime o f danger for s ettlers and t heir l ivestock does not s eem worthy of c onsideration ( LSSAHS 1 963-1964, ( 1964) 5 , 1 6). A long

w ith

t he

burgi

t here

were

other

defended

s ites

whose f unction i s u nclear. Caistor a nd Horncastle ( Lincs.) a re t wo such e xamples. Nothing i s k nown of t he s ettlement i nside t he defences b ut e xtensive e xtramural s ettlement s eems to b e indicated i n both cases ( Whitwell 1 970, 7 2-4; T odd 1 973, 4 3-4). I n e ach case, t he walled a rea a nd i s b astions a re u ndated. At Caistor, t he i rregular polygonal s haped c ircuit e ncloses c . 7 -8 a cres/2.8-3.2 h ectares a nd t he a lignment of t he defences s eems t o have b een d etermined by t he n atural contours o f t he p romontory on w hich i t w as s ituated J. ( 1960) 40, 1 75, 1 87 & f ig.4). H orncastle l ies i n a valley and t he t rapezoidal e nclosure d efines a n a rea o f c . 6 a cres/2.4 hectares ( Whitwell 1 970, 7 4 & f ig. 6 ). I t h as b een s uggested that t hese two s ites a re part o f t he system o f e ast coast defences a nd, i f t his i s t he c ase, t heir p osition i nland may indicate a system o f defence i n depth s uch a s t hat outlined by

8 9

Johnson i n h is a rgument f or t he i ntegration of t he defence s equence and s tyle o f civilian s ites a long the east c oast w ith t hat o f t he Saxon S hore f orts ( Whitwell 1 970, 6 9-74; T odd 1 973, 4 2-4; Webster 1 975b, 5 3; J ohnson 1 976, 1 23; 1 980, 1 01), a lthough b oth Caistor a nd H orncastle a re t oo f ar i nland f or t his a rgument t o be r ealy convincing. Todd ( 1973, 42-44) s uggests t hat t hese t wo s ites may p ossibly belong t o t he Theodosian r eorganisation o f t he defence system of t he p rovince a fter AD 3 69 a lthough h e f eels t hat a n earlier date i s possible and t his may point t o a context within Carausius' s ystem o f S axon S hore f orts e stablished i n t he l ate 3 rd century. I n n either c ase has i t b een possible t o decide whether t hese s ites s hould b e r egarded a s f orts r ather t han c ivil s ettlements. A f actor which Johnson ( 1976, 1 7-20, 1 24) a rgues c an b e s een i ncreasingly f rom t he 3 rd c entury onwards i s a military-inspired n etwork o f defences f or the province w hich i nvolved n ot only t he f orts t hemselves b ut s ites w hich were t owns, v illages, ports, etc., i n t heir own r ight. T his p roblem o f c ivilian o r m ilitary s tatus i s n ot confined t o t he l ater Roman period. L ittlechester and R ocester f or example may b e f orts r ather t han c ivil s ettlements b ut t he defences here s eem t o have o riginated much earlier with e arthwork c ircuits. The

p ossible

e stablishment

o f

garrisons

i n

defended

r oad

s tations and certain u rban centres and t he a cquisition o f c entres l ike Catterick a s b ases f or d etachments o f t he f ield a rmy i s t o b e s een i n t he wider context of the security o f t he p rovince a s a w hole. The p rovision o f bastions f or many t own and village defences i s s een by s ome authorities a s r epresenting a n a ttempt t o b ring c ivilian defences i nto l ine w ith military a rchitecture and t he g reater i mportance n ow p laced on t he manning o f w ell-defended s trong points s eems t o have b een e xtended t o c ivilian d efences ( Frere 1 978, 2 91; Webster 1 983). I t h as e ven b een s uggested t hat t he c loser i nteraction b etween c ivilian a nd m ilitary a nd t he i ncreased emphasis on c ivilian participation i n t he defensive s ystem o f t he p rovince e xtended t o t he movement o f c ivilians i nto t he f orts a long H adrian's Wall and t he c oast t o take o ver t he duties o f d efending t he f rontiers. All o f t hese d evelopments have b een attributed t o Theodosius' a ttempts t o reorganise t he defences o f t he p rovince t o meet t he i ncreased t hreat f rom t he barbarians a t a t ime of d iminishing military manpower ( Salway 1 965, 1 4-16, 1 98-200; 1 980, 1 6; F rere 1 978, 3 92-9 & f ig. 1 3; R amm 1 978, 4 2-3, 1 26-9) I t s hould b e e mphasised t hat t he a rchaeological e vidence i s a s y et i nsufficient t o s upport a d ate a fter AD 367 f or t hese d evelopments, o r i ndeed a ny o ther d ate w hich w ould p rove t hat t hey were i mplemented a s t he r esult o f a s ingle scheme. The w ider r ole a nd c ontexts p roposed f or defended c ivil s ettlements a re s urmised but not p roven.

9 0

1 0. What defences?

were

t he

other

developments

i n

l ate

R oman

c ivil

The 4 th c entury s eems t o mark a p eriod o f i ntense activity on t he d efensive c ircuits. The p rovision o f bastions r epresents o nly a s ingle e lement i n a n a pparently m ore g eneral phase of defensive a ctivity i n t he l ate R oman period. I n s ome cases t his s eems t o h ave i nvolved work on a v ery l arge s cale. For example, a t L incoln t he c onstruction o f a s ingle-arched gateway w ith f lanking g ate t owers a t t he w estern e ntrance o f the l ower colonia was undertaken apparently i n o r a fter t he mid 4th c entury; s hortly a fterwards, b ut s een a s p robably forming part o f t he s ame p rogramme, s ubstantial s tretches o f the walled c ircuit were r ebuilt a nd t he e xisting w all thickened i n o ther parts o f t he western and t he s outhern c ircuits a t l east. The r ampart a nd t he d itch w ere a lso r efurbished ( Ant. J . ( 1975) 5 5, 2 38-43, 2 56-7; C olyer 1 975, 1 8). I n t he u pper c olonia, t he n arrow w all ( 4 f eet/1.2 m wide) was a lso a dded t o o r r eplaced t o make a more s ubstantial wall c . 9 f eet/2.7 m t hick ( Wacher 1 974, 1 24; Colyer 1 975, 14; Jones 1 980, 5 3-4) . There were s ome i ndications f rom t he excavations o n t he w estern d efences o f t he l ower c olonia t hat this r emodelling and s trengthening o f t he defences, a lthough on a massive s cale, was e xecuted w ith a c ertain amount o f haste and t he e xcavator was t empted t o a ttribute s ome at l east of t he r ebuilding t o t he g eneral r eorganisation o f d efences under Theodosius. The dating e vidence w as n ot s ecure b ut indicated a date s oon after t he mid 4 th century f or t he c ons truction o f t he g ate ( Ant. J . ( 1975) 5 5, 2 40-3; C olyer 1 975, 1 8) A s imilar s ituation o ccurred a t G loucester . , w here stretches o f t he colonia wall, which i t appears may n ot have b een completed u ntil s ome t ime i n t he 3 rd c entury w ere r ebuilt on e ither s ide o f t he east gate a t s ome s tage i n t he f ollowing c entury. This m ay h ave t aken p lace a t l east p artly i n c onjunction w ith t he construction o f b astions, f or one n orth o f t he east gate a ppears t o b e c ontemporary w ith t he r ebuilt wall. E xcavations on t he n orth d efences s howed t hat t he wall west of t he n orth g ate w as a lso e rected a t s ome t ime i n t he 4 th c entury a nd p resumably r epresents a r ebuilding o f a n existing wall. A more p recise date w ithin t he 4 th c entury f or t hese developments c ould n ot b e e stablished ( Britannia ( 1970) 1 , 2 93; ( 1975) 6 , 2 73 & f ig. 1 6; Ant. J . ( 1972) 5 2, 3 1-5) The c onstruction o f t he r iverside w all a t L ondon h as b een dated t o AD 3 50-370 on t he b asis o f t he dendrochronology a nd r adio-carbon o f t he o ak p iles f rom t he f oundations o f t he wall. H ere a gain t he t emptation h as b een t o r egard t his a s part of t he T heodosian r eorganisation c ontemporary w ith t he p rovision of b astions ( Ant. J . ( 1977) 5 7, 5 0-1; Maloney 1 980, 6 0). A f urther p hase o f a ctivity i s r epresented by t he d iscovery, a t t he T ower o f L ondon, o f a nother r iverside w all of i ts

a later

date

p redecessor.

c onstructed T his

wall

parallel d id

9 1

n ot

t o

b ut

e xtend

4 m i nto

t o

t he

t he

n orth o f

s outh-east

c orner o f t he w alled c ircuit b ut 1 5m w est o f t his i t t urned s outh t o j oin t he e arlier w all, c reating a n a rea r esembling a s tronghold i n t he s outh-east c orner a pproached by m eans o f a n arrow c orridor o n t he west. Twenty-five c oins, t he l atest b eing of Valentinian I ( AD 3 75-392), were f ound in t he I material dumped during c onstruction a nd i t h as b een s uggested t hat a date a fter AD 3 90 i s i ndicated, s o t hat i t m ight h ave b een b uilt a s p art o f S tilicho's c ampaign o f c . A D 3 96-398 ( Britannia ( 1978) 9 , 4 53; Maloney 1 980, 6 0; Marsden 1 980, 1 78-9) Very l ittle i s k nown o f t he s trengthening o r r efurbishing of t he defences of o ther centres i n t he l ate R oman p eriod. A t Kenchester, modifications w ere made t o t he w est g ate apparently s ome t ime i n t he s econd h alf of t he 4 th century, a lthough t he d ating e vidence was a gain i nconclusive. T his a ctivity i nvolved t he abandonment of t he existing d ual c arriageway f or a much n arrower r oad a nd t he extension a nd r ecutting o f t he d itch once more ( TWNFC ( 1962) 37, 1 52, 1 56, 1 60 & f ig. 6 ). The t ime s pan i nvolved f or t hese e vents i s unknown. S ince t here i s a t erminus post quem of c . AD 3 50 f or t he c onstruction o f t his g ate i n s tone, t hese s ubsequent developments must have t aken p lace s ome t ime i n o r a fter t he s econd h alf o f t he 4 th c entury. The o ne major a nd s ubsequent minor phases may have spanned b etween t hem a f air p eriod o f t ime, b ut w hether t hese d evelopments c an b e a ssigned t o t he early 5 th c entury, a s was t entatively s uggested, i s i mpossible t o e stablish ( TWNFC ( 1962) 3 7, 1 52 & 1 67). This n arrowing of t he gateway i s a development which i s r epresented i n t he d efensive s equence o f many o ther t owns a nd v illages. G ates were s ometimes b locked c ompletely and t his h as b een s een a s a n i ndication o f t he i ncreased c oncern f or s ecurity which, i t h as b een s uggested, b ecomes apparent i n military a rchitecture f rom t he l ate 3 rd c entury onwards ( Boon 1 974, 1 06-7; F rere 1 978, 2 89). E xamples a re t he partial b locking o f t he n orth gate a t A ldtorough, a nd t he b locking o f t he n orth a nd s outh gates a t Caerwent, o f t he main w est g ate ( Balkerne G ate) a t C olchester, o f t he w estern f ootway of t he n orth g ate at G odmanchester, of t he n orth gate a t G reat Chesterford, o f t he w est g ate a t I rchester, of t he f oot passages o f t he n orth gate o f L incoln upper colonia, of t he n orth, w est, s outh-west a nd s outh-east g ates a t S ilchester a nd possibly a lso o f t he n orth-east carriageway o f t he V erulamium Gate a t C irencester a nd t he w est gate a t t i xacona. I n n one o f t hese e xamples i s a date a vailable f or t his e vent, a lthough C rum my ( 1977) s uggests t hat t he r adical s tep o f w idening t he d itch outside t he Balkerne G ate a t Colchester and c arrying i t r ight a cross t he c arriageway i s l ikely t o h ave t aken p lace s ome t ime i n t he 3 rd c entury.

9 2

H owever, t he b locking a nd n arrowing o f g ates r emain p roblematical. F ulford ( 1983) has observed t hat t he unmortared mass o f u ncut s tone u sed t o b lock t he s outh c arriageway o f the west gate at S ilchester may be t he r esult o f s ite c learance during t he medieval p eriod. At t he s outh-east g ate, where t he l atest r oad s urface, which s ealed c oins o f AD 3 503 60, s howed v ery l ittle w ear, F ulford s uggested t hat t he g ate may a ctually have gone o ut o f u se before r ather t han as a r esult

o f

b eing

b locked.

F airly l arge s cale r epairs t o t he n orth-east d efences a t Cirencester b ecame n ecessary a t s ome s tage i n t he 4 th c entury b ecause of t he u ndermining o f t he wall a nd t he n orthern g ate t ower of t he n orth-east gate by t he s tream which f lowed t hrough the i nner d itch o n t his s ide o f t he t own. A fter t his, haphazard a ttempts s eem t o have b een made t o p revent a r epetition o f t he p roblem by t ipping l ayers o f g ravel o n t o t he b erm and e recting a t imber g royne a cross t he vulnerable r e-entrant a ngle b etween wall a nd gate t ower. B ut t his w as p resumably i neffective f or some t ime l ater a s mall f loodgate was c ut t hrough t he w all i tself t o r elieve p ressure a t t he angle ( Ant. J . ( 1961) 4 1, 6 9). U nfortunately, t hese developments a re u ndated a nd t heir c hronological r elationship t o t he c onstruction of b astions a long t his part o f t he c ircuit, which i s r ather i nsecurely p laced i n t he f irst h alf o f t he 4 th c entury, i s n ot c lear. T here

c ould

h ave

b een

o ther

c ircuits

w hich

would

h ave

been i n n eed of r epair o r modification a t s ome s tage i n t he e arly mid 4 th c entury e ven i f t hey h ad o nly b een c onstructed at s ome stage i n t he l ate 3 rd c entury. Where t he wall had b een constructed v ery much e arlier, a s a t C olchester, G loucester, L incoln, L ondon, Verulamium a nd p ossibly C irencester, r epairs would h ave b een n ecessary -o n a c ontinuous b asis for a f airly l ong t ime. I t may b e t hat t he major r ebuilding o f t he w alled c ircuits a t G loucester a nd L incoln i n t he 4 th century s hould b e s een i n t erms o f s uch r epairs r ather t han i n t he c ontext o f a s pecific g overnment p olicy. I n general i nformation about t he c ondition o f walled c ircuits a t t he b eginning o f t he 4 th c entury, o r f or t hat matter a t a ny period, i s l acking. There a re o ccasional i ndications, a s a t C irencester, o f t he a ttention o r l ack o f i t g iven t o t he general maintenance of t he defences a t i ndividual c entres. -

T he d evelopments d iscussed a bove a t c entres s uch a s C irencester, G loucester, K enchester, L incoln and L ondon s eem t o i ndicate a p eriod o f i ntense a ctivity, t he m eaning i f w hich i s a s yet h ardly u nderstood. F urther work i s n eeded n ot only o n i ndividual e lements w ithin t he c ircuit, s uch a s b astions and g ates, but on t he d itches a nd t he walled c ircuit i tself s o t hat a chronological f ramework f or t he v arious modifications a nd r epairs can b e e stablished f or each c entre. The l ack o f s ecure dating makes i t v ery d ifficult t o e stablish a c hronology f or l ate R oman defensive a ctivity a nd a lso t o f it s pecific d evelopments s uch a s t he c onstruction o f b astions i nto a general s equence o f r ebuilding a nd r epair. Such

d evelopments

a s

t he

9 3

t hickening

a nd

s trengthening

o f

w alls, t he n arrowing o f gateways, a ddition o f bas tions a nd gatetowers a re l ikely t o have b een i mplemented i n r esponse t o c hanging c ircumstances o f warfare a nd a n eed f or i ncreased d efence. These developments n eed n ot b e s een solely a s t he r esult o f i ndividual government i nitiatives a t specific t imes; i nstead, t hey may b e s een a s part o f a continuous p rocess o f maintaining a nd u p-dating d efensive c ircuits once e rected. Casey's ( 1983) a rgument, t hat u ntil t he t ime of Constans ( AD 37-350) d efensive a ctivity was n ot p art o f i mperial p olicy, 3 would s uggest t hat any work undertaken on t he defences i n t he e arly 4 th c entury must h ave b een f inanced a nd i mplemented b y t he s ettlement c oncerned. E ven a fter t he r emission o f t axes f rom t he t ime o f Valentinian I ( AD 3 50-354), much o f t he work on d efences must have b een t he p roduct of the i nitiative a nd f inancial c ompetence o f i ndividual s ettlements. E ven i f t he p rovision of b astions o r o ther d evelopments a re s een a s a r esult o f o fficial a ssistance o r e ncouragement t o t he t owns, i ndividual centres must a lways have b een r esponsible f or t he g eneral maintenance o f t heir d efences. Some mention h as a lr eady b een made o f the n eed t o t reat a d efensive c ircuit as an o rganic e ntity u ndergoing c ontinual p rocesses of d ecay a nd modification. A g reat d eal o f emphasis h as b een placed o n t he ' restoration' of t he ' towns' by Theodosius, and F rere ( 1978, 2 92) s uggests t hat t his may h ave e nabled t owns t o h old o ut f ar i nto t he 5 th century. We cannot be c ertain what w as i nvolved i n t his r estoration, e specially s ince t here i s n o e vidence t hat t he t owns s uffered during t he b arbarian i nvasion of AD 3 67. We d o n ot k now w hether i t i nvolved t he p rovision o f b astions, and/or t own garrisons, a ssistance i n t he r ef urbishing o f n eglected c ircuits, o r f urther s trengthening o f o thers. Nor s hould we a ssume t hat l arge-scale work on t he d efences o f a ny t own o r v illage b elonged t o t he t ime o f Theodosius, s ince at n one of t he centres s o f ar d iscussed was t he d ating e vidence s ufficiently p recise. There may a lso h ave b een o ther o ccasions when a general p rogramme o f r ec onstruction may h ave b een i nitiated. The b rief v isit o f t he emperor C onstans i n AD 3 42-343 a nd t he p ossible e xpedition under Stilicho c . AD 3 96-398 a re s ometime c ited as o ccasions o n w hich d efensive r econstruction may h ave b een put i n h and. Both of t hese episodes a re p lausible i n t he s ense t hat t hey i nvolved a ttempts t o d eal w ith a c risis i n t he s ecurity o f t he p rovince ( Ant. J . ( 1977) 5 7, 5 0; Merrifield 1 978, 2 6; F rere 1 978,

3 87-9;

Marsden

1 980,

1 78-9;

C asey

1 983).

There may have b een many s ubsequent phases i n t he s tructural h istory o f t he d efences i n t he l ate R oman p eriod a nd possibly even a fterwards. I n t he l ast r esort, t he c ont inued maintenance o f a d efensive c ircuit must h ave b een a decision f or each i ndividual c entre. A longside t he s trategic a nd t actical r ole p layed b y e ach t own a nd v illage i n t he s ecurity o f t he p rovince a s a whole, t here i s the f act t hat i ndividual s ettlements a re s een a s b eing f orced t o r ely i nc reasingly upon t heir own r esources f or t heir defence a s t he m ilitary g arrisons were d epleted f ollowing t he R oman w ithd rawal

of

1 978,

1 02).

c .

AD

4 10

( Frere

Wacher

1 978,

( 1978,

2 92,

1 04)

9 4

3 97-9;

m entions

Wacher t hat

1 974,

t he

7 8;

a ctive

survival o f t hese s ettlements a fter AD 4 10 d epended o n t he bands of unreliable f ederate s oldiers a nd G ermanic l aeti which were r ecruited a s u rban garrisons, a nd o n t he a bility o f t he townspeople t o maintain d iscipline over t hem. The p roblems o f maintaining a nd d efending t he v ast c ircuits w hich e xisted a t some of t he major c entres, s uch a s C irencester a nd Wroxeter have f requently l ed s cholars t o s peculate t hat t he t ownspeople must h ave b een f orced t o abandon t hese centres s ooner o r l ater f or m ore e asily o r n aturally d efended s ites s uch a s h illforts, or s ites l ike Shrewsbury which were a lmost c ompletely s urr ounded by marsh a nd r iver, o r e ven f or t he amphitheatre a t Cirencester ( Wacher 1 974, 3 13-314; 1 978, 1 04). This i s a matter i nvolving l ocalised d evelopments w hich a re o utside t he scope of any overall f ramework f or t he p rovince a s a whole.

9 5

Conclusion

The a im o f t his p aper h as n ot b een t o e stablish a n a lternative f ramework t o r eplace t hat w hich i s n ow current, but t o b roaden t he s cope o f t he i nquiry t o i nclude m any n eglected a spects o f R omano-British d efences. This h as l ed t o t he i ntroduction o f a more f lexible a pproach t o t he r easons f or t he building of d efences, and t he f unctions which they m ay h ave p erformed. I n doing s o, a ttention h as b een d rawn t o t he n ature and t he l imitations of t he e vidence on w hich t he p resent f ramework i s b ased. A d etailed e xamination of t he evidence a vailable a nd of t he way i n which i t has b een i nterp reted was n ecessary i n o rder t o u nderline t he f act t hat w e a re n ot r eally i n a position t o construct a c hronological f ramework o r t o make a ssumptions a bout t he basis on w hich defences w ere p rovided. The e vidence i s i nsufficient t o d emonstrate a ny consistent pattern i n t he p rovision o f defences, whether e arthworks o r w alled c ircuits, a nd i t i s q uite c lear t hat i n n o i nstance i s t he s equence o f development of an i ndividual c ircuit p roperly u nderstood. Only t he c umulative evidence f rom a n umber o f s ections f rom s everal different s ites h as made i t p ossible t o b uild u p t he p resent p icture. But as t he evidence f rom t he i nvestigation of i ndividual c ircuits i ncreases, i t may b e p ossible t o d emonstrate m ore conclusively t hat t he p resent p icture i s t oo s implistic. T he a pparent a nomalies w hich e xist i n t he c urrent f ramework may b e r esolved b y f urther i nvestigation. O n t he other h and, t hey may p rove t o b elong t o o ther p hases o r a spects o f defensive a ctivity n ot y et r ecognised. ' Perhaps t he t ruth o f t he matter i s t hat w e h ave b een t oo a nxious t o r ead i nto o ur evidence n eat s chemes which would p rovide g eneral j ustification o f g rants o f murage a t a s pecific t ime .. . ‚ ( Hartley 1 966, 5 7). The a rchaeological e vidence i s u nlikely e ver t o b e s ufficient t o m ake watertight h istorical p oints a nd w e s hould a bandon t he s earch f or c onfirmation o f p recise d ates a nd h istorical c ont exts f or d efensive a ctivity a nd,concentrate o n a spects w hich t he a rchaeological e vidence can more p roperly r eveal. Once t he n ature o f t he e vidence a nd i ts l imitations a re appreciated, i t b ecomes e asier f or n ew e vidence t o b e assessed f rom a n objective v iewpoint r ather t han o ne w hich s eeks t o f it i t i nto t he c urrent f ramework. I h ave attempted i n Sections II a nd I V t o s uggest n ew a pproaches t o t he s tudy o f defences I by o ffering s ome a lternative i nterpretations of t he e vidence, a nd b y a sking q uestions a bout t heir r elationship w ith t he i ndividual s ettlement o f which t hey f ormed a part. The c ase h as b een m ade f or c onsidering d efensive c ircuits i n r elation t o t he c ircumstances a ffecting t he s ettlement and t he r egion i n w hich i t w as s ituated. The more w ork d one, t he more c omp licated t he p roblems become and t he more ' anomalies' t here a re i n t erms o f t he c urrent f ramework. The q uestions of ' why' and ' when' n eed t o b e a sked i n each c ase n ot f rom t he narrow v iewpoint o f a 2 nd c entury e arthwork a nd l ater wall, but f rom an approach which i s aware o f t he variety o f considerations t hat may h ave h ad a p art t o p lay i n t he c onstruction a nd subsequent h istory o f t he d efences. An appreciation of t he i ndividuality

a nd

c omplexity

9 6

o f

a

p articular

s ettlement's

defensive which has attempting

c ircuit c an o nly h elp r evive i nterest i n a s tudy been r educed t o t he r ather u nproductive a im o f t o date e arthwork d efences a nd w alled c ircuits.

F inally, i t must b e s tressed t hat t his paper h as b een prepared t o s timulate d iscussion a nd r evitalise i nterest i n the s tudy o f R omano-British c ivil defences. Many o f t he l ines of i nquiry s uggested c an o nly b e e xplored a s p art o f l ong-term r esearch p rojects b ut t hese a ims must b e k ept i n m ind i f t he nature and s ignificance o f R omano-British defences i s t o b e better understood. I t h as n ot b een p ossible t o e xamine a ll o f the e vidence f rom e very s ite a nd t he p resent a ssessment i s based only o n p ublished i nformation w hich i s o ften a vailable only i n summary o r i nterim f orm. Apology i s t herefore due t o those whose v iews may h ave b een m isrepresented. F urther i nvestigation w ill n ecessitate c onstant r evision o f t he e vidence u sed f rom i ndividual s ites, a nd t his i n n o w ay u ndermines t he validity o f e xamining a lternative i nterpretations and o f exploring n ew approaches.

9 7

Desiderata

f or

t he

Future Study

o f Civil

Settlements.

I n general, t he i nvestigation o f d efences has b een c onf ined t o t he p lacing o f t renches a t v arious p oints a round t he c ircuit. The i nformation p roduced h as b een l imited a nd t here a re many questions a bout t he h istory a nd development o f defences which a re l argely u nexplored. The s tudy o f defences h as f ocussed h eavily on e stablishing t he date a nd e xistence o f major phases of defensive a ctivity b ut, i n attempting t o a chieve a more c omprehensive u nderstanding o f t he h istory a nd u se of defences, t he f ollowing l ines of i nquiry n eed to b e pursued: i ) Which e arthwork maintenance and were w alled o nly a t a l ater i) i

F or

c ircuits w ere s ubject t o e ither n ot r eplaced by s tone date?

h ow l ong were

t hese

l ong t erm walls o r

c ircuits maintained?

ii) Which c entres h ad t o a bandon o r c hose t o dispense i t he t imber and earth c ircuit a nd r eturn t o an undefended a nd w hen d id t his o ccur?

w ith s tate

i v) Where t he e arthwork a nd w alled c ircuits f ollow a different a lignment ( see Tables O ne a nd Two) , had t he earthwork d efences d ecayed/disappeared b y t he t ime t he w alled c ircuit was e rected? V ) At w hat p oint d id i t b ecome more e conomical t o e rect a s tone wall r ather t han maintain t he earthwork d efences? D oes t his h elp t o e xplain t he r elatively e arly o r l ate p rovision o f a s tone wall at i ndividual c entres? v i) F urther i nvestigation i s n eeded o f t hose s ites w here p resent e vidence s uggests t hat walled c ircuits were erected de n ovo i n t he l ate 2 nd o r e arly 3 rd c entury. I s t he c ons truction o f walled c ircuits at a n e arly date f or s ites p rev iously u ndefended b y a n e arthwork c ircuit r eally a f actor t o be c onsidered i n t he h istory and development o f Romano-British defences? v ii) I s i t c orrect, a s F rere s uggests, t o a ttribute t he majority o f walled c ircuits t o t he l ate 3 rd c entury? If s o, w hich c ircuits w ere e rected a t a n e arlier date a nd h ow e arly a re t hey? A lternatively, i s t here a ny e vidence that c ircuits may h ave b een walled a t d ifferent t imes a s p art o f a p rocess of maintenance and r epair? v iii)

How

l ong

d id

i t

t ake

t o

i mplement

e ither a n earthwork c ircuit o r a walled n ecessarily c ompleted more q uickly w here smaller? This would b e d ifficult t o

t he

c onstruction

c ircuit? Was t he c ircuit establish but

o f i t w as a t

S ilchester i t h as b een c alculated t hat t he c onstruction of t he I nner D efence earthwork c ircuit c ould h ave b een a ccomplished w ithin a y ear o r 3 ,000 m an h ours ( Fulford 1 983). T here a re a lso i ndications t hat t he w alling p rogramme a t Gloucester m ay h ave b een p rolonged.

9 3

i x)

Was

t here

a n a cceleration

i n

t he

s peed

w ith

w hich

c ir-

cuits w ere e rected i n t he l ate 3 rd/early 4 th c entury w hen t he s ecurity o f t he p rovince b ecame a matter o f u rgent c oncern? For e xample, t here were apparently i ndications t hat t he massive r emodelling a nd s trengthening o f t he w estern d efences of L incoln l ower c olonia was e xecuted i n s ome h aste. x ) F or how l ong w ould a p articular w alled c ircuit r emain i n a defensible s tate g iven t hat i ndividual c ircuits would b e s ubject t o g reater o r l esser d egrees o f e rosion, by a r iver for e xample a t C irencester and L ondon? x i) Were s ome walled of t he Roman period?

c ircuits

a llowed

t o

d ecay

b efore

t he

e nd

T hese q uestions a re d ifficult t o a nswer a nd r equire a much g reater s cale of i nvestigation t han h as b een u ndertaken s o f ar. Such considerations a s t he s peed o f c onstruction o f a defensive c ircuit a re d ifficult t o a ssess a rchaeologically. Nevertheless, such assessments a re worth a ttempting a s i n t he case o f S ilchester. Certainly, much o f w hat w e may w ant t o know a böut defences can a rchaeological r ecord.

only b e

i nferred,

n ot

p roven,

f rom

t he

A b etter a ppreciation o f w hat t he a rchaeological e vidence can r eveal about d efences c ould b e a chieved by g iving more attention t o e xamining t he n ature o f t he e vidence p roduced b y this t ype o f f eature and by s tudying t he p hysical a spects a nd composition o f t he d efences. T he f ollowing l ines o f investigation a re s uggested: -

i nquiry

a nd

m ethods

o f

i ) s tudy o f t he a ctual p rocesses by w hich e arthworks particular were c onstructed a nd o f t he question o f whether bank i s automatically d efensive ( Bidwell 1 980, 6 0-62).

i n a

i) s tructural a nalysis of r ampart composition a nd o f i materials u sed i n w all c onstruction i n o rder t o e lucidate t he techniques a nd mechanics of c onstruction a nd t o i ncrease understanding o f t he h istory a nd d evelopment o f i ndividual circuits and of t he e conomic f actors i nvolved ( Lincoln upper colonia, Jones 1 980, 3 7-55). ii) a ttention t o a ssessing t he r ate o f c ollapse a nd e rosion i of e arthworks a nd walled c ircuits a nd t he amount o f maintenance r equired t o k eep v arious t ypes o f c ircuit i n a defensible s tate. i v) a ttention n ot s imply t o t he s ectioning o f t he w all a nd rampart at various points a long t he c ircuit t o e stablish t he major p hases b ut t he m inor p hases a lso, a nd t he more d etailed history of r eplacement, r enewal a nd r epair. V ) c onsideration o f t he s tate o f stages and, where r elevant, t o t he and

d itches

a t

t he

t ime

w hen

t he

9 9

t he d efences c ondition o f

w all

w as

b uilt.

a t t he

v arious r ampart

This

would

p rovide s ome means o f a ssessing h ow much t ime i s l ikely t o have e lapsed b etween t he building of t he r ampart a nd the w all a nd may p rove more r elevant t han a ttempts t o a ssess this o n t he basis of t he gap b etween t he t erminus post quem for t he r ampart a nd t he t erminus p ost q uem f or t he w all. I t w ould a lso t hrow s ome l ight on t he question of t he decay or c ont inued maintenance o f t he e arthwork c ircuit. vi) i nvestigation of g ates, t owers a nd bastions a s part o f a u nified c ircuit o f defences a nd n ot s imply t o e stablish t heir dates and s tructural h istory. v ii) g reater a ttention t o t he d itches, f or e nvironmental e vidence, f or i nformation on periods o f u se and disuse and f or i nformation on t he f unctioning o f t he d itch s ystem a s part o f t he defensive c ircuit. v iii) t he e xcavation o f l arger a reas l aterally a long t he defences and complete s ections t hrough t he defences, including t he d itch s ystem, wherever p ossible. Various l ines of i nquiry concerning t he r elationship b etween t he d efences a nd t he a ssociated s ettlement c ould a lso be pursued such a s t he i mpact o f t he d efences on t he settlement i n t erms o f: i ) i ts t opography. The r elationship o f t he defended c ircuit t o t he s ize a nd l ayout o f t he s ettlement n eeds t o be i nvestigated. Any c hanges i n t he t opography of t he s ettlement, density of b uildings, character o f certain a reas, p articularly s uburbs, n eed t o b e e xamined i n r elation t o t he construction of t he defences a nd t he a lignment adopted. S imilarly, a ny a lteration i n t he a lignment may be e xplained by changes i n t he s ettlement i) i ts g rowth a nd e conomic p rosperity. D id t he p rovision i of d efences contribute i n any w ay t o t he p rosperity, character o r d evelopment o f t he s ettlement e ither b y p roviding greater s ecurity o r by p roviding i ncreased s tatus o r by a ttracting f urther s ettlement? What w as i nvolved i n t erms o f manpower, r esources, t ransportation o f materials? D id t he p rovision o f d efences, particularly o f s tone d efences, a dversely a ffect t he public o r p rivate building p rogramme? ii) c ontinuity o f s ettlement i n t he l ate and s ub-Roman i periods. D id t he e xistence o f a defensive c ircuit increase t he a bility o r i nclination o f t he c ommunity t o maintain t heir i nterest i n t he s ite a fter t he R oman w ithdrawal? I n o rder t o u nderstand t he h istory a nd development o f i ndividual defensive c ircuits, t he i nvestigation of t he defences n eeds t o b e i ntegrated much m ore w ith t he overall s tudy of t he s ettlement w ith which i t w as c oncerned. Sections a cross d efences h ave t oo o ften b een v iewed i n c omparative i solation This i s compassed p erhaps

f rom what was h appening i n t he s ettlement c oncerned. t rue e ven i n t he v ery b asic s ense o f t he a rea e nby t he s ection. Too l ittle o f t he a rea , i nside a nd a lso

o utside

t he

a ctual

1 0 0

d efensive

c ircuit

i s

i ncluded

i n t he i nvestigation. T here a re p ractical p roblems i n t he e xtension o f s ections t hrough d efences t o i nclude a djacent a reas o f o ccupation w hich w ould n ot r eceive a dequate t reatment. Therefore, s ince t he d efences n eed t o b e s tudied a s p art o f t he s ettlement, e xcavations o n t he d efences s hould b e conducted n ot a s s mall s ections b ut a s p art o f a p rogramme t o e xamine a reas o f s ettlement a djacent t o t he d efences b oth i nside a nd o utside t he c ircuit. T hese a reas d eserve i nvestigation b ecause t hey p rovide s tratigraphic l inks b etween t he d efences a nd t he m ain o ccupation s equence.

1 01

0

C

w alls b uilt d e n ovo o r o n n ew a lignment w alls

s ubstantially r epaired o r added

t o e xisting e arthwork c ircuit

X defences

o f u ncertain n ature

F IGURE 4

4 th c entury d efensive c ircuits

1 02

1 50 1 (m

0

I

Gazetteer

' Lesser'

n ot b ut

Towns,

Villages

and Other

Civil

Settlements

T here a re a n umber of s ites i n t his gazetteer which a re o r may n ot b e c ivilian s ettlements i n t he s trictest s ense which have b een i ncluded f or completeness, e ither b ecause

t hey h ave b een r eferred t o i n t he t ext o r b ecause t he possibility o f a d efended c ivil s ettlement has b een s uggested b ut r emains u nproven. I n t his r espect, a ttention i s d rawn t o t he f ollowing s ites: B itterne, Caistor-by-Yarmouth, L eintwardine, L ittlechester a nd R ocester w hose s tatus a s c ivilian settlements i s uncertain. I t h as b een s uggested t hat t he l atter t wo s ites may b e f orts a nd t he f ormer t wo s ites S axon S hore f orts. T he defences a t D roitwich do n ot enclose t he c ivil s ettlement b ut a re p robably d efending a n o fficial o r a dministrative complex a nd a s imilar s tatus h as b een s uggested f or G atcombe. D efended v ici a ttached t o f orts h ave b een i ncluded and a n umber o f s ites i n E ssex L ong Melford, Maldon, Shoebury Camp, Upham Camp a nd W ixoe where R odwell h as s uggested enclosures possibly o f R oman date ( Rodwell 1 975, 9 2). -

-

1 03

Alcester

( Warks.)

( SP

Defences d iscovered s outh. E xcavations u npublished.

0 92576) i n on

1 965; t he

n ot east

l ocated o n n orth, i n 1 964-1966.

T he defended n ucleus l ies o ff t he k nown R oman r oads t he medieval and modern c entre; l arge e xtramural which r emained open l and i n t he m edieval p eriod. E arthwork

w est o r R esults

a nd b eneath s ettlement

defences

Turf a nd c lay r ampart ' dated t o t he 2 nd c entury o r l ater'. No d itch has been f ound on t he e ast possibly b ecause t he s lope o f t he g round down t o t he R iver A rrow made i t u nnecessary. N o i ndication of t he n ature o f t he dating e vidence, i f any. Walled

c ircuit

S tone wall i nserted i nto t he f ront o f wooden p iles i n t he f oundations. No dating WMNS ( 1961) 4 , ( 1978) 2 1, 7 9. 5 6, 2 06.

Alchester

3 ; JRS

( Oxon.)

( 1965) 8 , ( 1961) 5 1,

( SP

2 -3; ( 1966) 9 , 2 ; 1 72-3; ( 1965) 5 5,

defences

s treet p lan. f rom a major Walled

enclose

( 1967) 2 08-9;

u sing

1 Q, 2 ; ( 1966)

5 73203)

Aerial p hotographs a nd s urvival a s east. E xcavations i n 1 920s a nd 1 974 a long The

t he r ampart, e vidence.

a n early

e arthworks

o n

n orth-east

a nd

s quare

The s ite i s d eserted R oman r oad j unction.

r ectangle

a nd

l ies

o n

t he

w est

a nd

east.

w ith m arshy

a

r egular

g round

a way

c ircuit

E xcavations i n 1 974 o n t he e astern d efences f ound n o e vidence f or a p re-wall earthwork p hase. The mortar c onstruction l evels f or t he w all w ere i nterleaved w ith t he main b ulk of t he r ampart. The d efences on t his s ide f ollow t he l ine o f t wo b oundary a nd d rainage d itches l aid o ut i n t he 1 st c entury. The d efences c onsisted of a s and a nd g ravel r ampart f ronted by a s tone w all; s ingle s hallow d itch r ecut once. P ottery c ontained i n t he r ampart gave a t erminus post q uem i n t he l ate 2 nd c entury. B ritannia ( 1975) 6 , 2 56; Oxonier isia S ee a lso R owley 1 975, 1 22-3; Wilson Alfoldean E xcavations

( Sussex) i n

1 923

( TQ a nd

18330) 1 1 924.

1 04

1 975 ( 1976) 1 975, 1 1.

4 0,

1 36-70.

Small

r ectangular

E arthwork

e nclosure

S tane

S treet.

S ite

d eserted.

defences

I nformation enclosure. S ussex A rch. a lso C unliffe Ancaster

o n

i nadequate.

C oil. 1 973,

( Lincs.)

( 1923) 7 1-2.

( SK

Small

6 4,

r ectangular

8 1-104;

( 1924)

6 5

,

e arthwork

1 12-57.

S ee

9 82437)

D efences t raceable o n s outh, s outh-east and e ast. E xcavations o n n orth-east, s outh-west a nd e ast 1 962-1971. An i rregular L incoln. E arthwork

r ectangular

enclosure

on

t he

E rmine

Street

s outh

o f

defences

T here i s n o e vidence t hat t he wall w as a dded t o a p ree xisting earthwork, b ut Webster s uggests t hat an early d itch s ystem, on a s imilar c ircuit t o t hat o f t he w all, may b elong t o t he c ivil s ettlement and n ot t o a f ort a s h as b een s uggested by i ts e xcavators ( G. Webster, ' Fort a nd Town i n E arly B ritain' i n J . S . Wacher ( ed.), T he C ivitas C apitals o f R oman B ritain ( Leicester University P ress 1 966), 3 5). Walled

c ircuit

E xcavations i n 1 962-1971 s howed a n e arth, g ravel a nd s and r ampart f ronted b y c ontemporary s tone w all. A b uilding c onstructed o ver t he t ail o f t he r ampart o n t he e ast s ide was dated t o AD 2 80 o r l ater. T wo f an-shaped b astions ( cf. Godmanchester) a dded t o t he wall a t north-west a nd s outh-west angles. A 4 th c entury date s uggested. JRS

( 1957)

4 7,

2 10;

( 1961)

5 1,

1 71;

( 1962)

1 58 & F ig. 1; 1 B ritannia ( 1971)

( 1965) 5 5, 2 06 & F ig. 2 , 2 57 S ee a lso Whitwell

3 8,

3 4;

1 26-7

Whilton

& f ig.

L odge

1 975,

B annaventa

5 2,

1 2; 1 970,

1 67; ( 1968) 6 5-6;

( 1964)

5 4,

5 8, 1 84; Todd 1 973,

2 17-21.

( Northants.)

( SP

6 12646)

Aerial p hotographs, r esistivity s urvey a nd e xcavations 1 970-1972 on n orthern d efences. R ectangular e nclosure o n t he Watling S treet b etween T ripontium a nd T owcester. E arthwork

defences

E mergency e xcavations o n n orth-west c ircuit l ocated a d itch whose upper f illing s uggested i t w as d eliberately f illed, p robably i n t he e arly 4 th c entury, t o p rovide a b ase f or t he wall. E xistence o f a r ampart o f t urf a nd c lay i nferred f rom material i n t his u pper f ill.

1 05

Walled

c ircuit

S tone w all a nd t riple d itch s ystem l ocated i n e xcavations 1 970-1972 a nd f ill o f d itches i ncluded pottery of f irst h alf of 4 th c entury. Not c lear i f t his w as t he t rue d ate o f c onstruction o r i f a ll 3 d itches were contemporary. WMNS ( 1971) 1 4; S ee a lso Wilson Bath

Aquae

P ossible

B ritannia 1 975, 1 4.

S ulis

( Avon)

s urvival

plan. E xcavations west i n 1 970. I rregular

polygon

E arthwork

defences

( 1973)

( ST

1 963-1965

t he

2 96

& 3 25,

f ig.

9 .

7 49648)

i n medieval

o n

4 ,

walled

a nd

F osse

1 980

c ircuit o n

and m odern

n orthern

S treet

d efences

a nd

on

Way.

An a lmost c omplete s ection a cross t he n orthern s ector o f t he e arthwork defences was obtained by means o f s taggered t renches. Very l ittle dating e vidence b ut a f ew s craps o f s amian suggested a date i n t he s econd half of t he 2 nd c entury. O n t he w estern c ircuit i t w as observed t hat a wall e rected i n t he early 2 nd century may have r evetted t he t ail of t he r ampart. B ut possibly Walled

s ee B idwell 1 980, 6 2, t oo s mall t o r epresent a n

w ho s uggests t he r ampart i s i ndependent earthwork phase.

c ircuit

T he dated.

a ddition

Cunliffe 1 969, ( 1981) 1 -30. B itterne

( Hants.)

o f

t he

1 66-8;

( SU

wall

t o

t he

p re-existing

B ritannia

( 1971)

2 ,

r ampart

2 76;

w as

n ot

O 'Leary

4 35133)

E xcavations 1 938-1939 a cross n orthern c ircuit o f wall a nd t racing o f wall f or d istance o f 6 9 f t. ( 20.7m.) 1 951 complete s ection a cross I nner F osse a nd t wo f urther t renches a cross l ine o f palisade. 1 959 s ection n ear s outh e nd o f Outer Bank a nd Ditch. .

The s ite o f B itterne l ies o n a t riangular p iece of l and s urrounded on a ll s ides e xcept t he e ast by a bend i n t he R . I tchen. T he s ite was d eveloped a s a s uburb o f S outhampton i n t he 1 9th c entury. I ts d efences a re r epresented b y a n O uter Bank and D itch r unning n orth/south w ell t o t he e ast o f t he I nner F osse; I nner F oss a lso r uns n orth/south a cross p romontory; w alled c ircuit w hose c ourse h as b een s ecurely i dentified o nly o n t he n orth b ut i , t i s a ssumed t hat t he wall f ollows t he contours o f t he t riangular p romontory. On t he e ast, e arlier a ntiquarian r eferences s uggest i t r an j ust i nside t he l ine o f t he I nner F osse a nd s eem t o i ndicate t he p resence o f a b astion ( Cotton & G athercole 1 958, 4 1-2). H owever, a djacent t o t he s upposed b astion f ailed

1 ' )F

a n t o

e xcavation in l ocate t he wall

1 954 b ut

r ubble b locks f ound may b e d ebris f rom w all east ( Cotton & Gathercole 1 958, 1 48-9) Earthwork

l ying

f urther

t o

t he

defences

T he i nvestigations o f t he I nner F osse & 951) r evealed a d itch and i nner palisade. T he only dating e vidence c ame f rom o ne of t he s maller t renches a cross t he p alisade t o n orth o f main section. The dating c ame f rom a U-shaped d itch e ast of t he palisade and r egarded a s c ontemporary w ith i t. The r apid s ilt o f the d itch contained pottery of c .AD 1 20-150 a nd on t his e vidence t he p alisade w as d ated t o t he Antonine p eriod, ' perhaps c .AD 1 50 o r s omewhat l ater'. P ottery o f AD 1 50-170 i n s ucceeding s ilt l evels and a t hick l ayer s ealing t he d itch by c .AD 1 70-180 were t aken a s i ndications t hat t he palisade f ell i nto d isuse t owards t he e nd o f t he 2 nd c entury. Walled

c ircuit

E xcavations 1 938-1939 a nd 1 951 o n t he n orthern c ircuit f ound no d irect d ating e vidence f or w all. A l ate 4 th c entury date ( possibly Theodosian) was s uggested i n 1 938-1939 on t he b asis of material f rom t he R oman o ccupation l evel i nto w hich t he wall h ad been c ut. This c ontained much ' Late P eriod' N ew F orest pottery dated c . AD 3 30 o nwards t ogether Valens ( AD 3 64-379). I n 1 951, more p recise dating e vidence b ut date. P roc. JRS

H ants.

( 1952)

F .C. 4 2,

( 1920-1925)

1 00-2;

C otton

9 ,

w ith a much c orroded c oin o f t his l evel d id n ot p roduce any c onfirmed a l ate 4 th c entury

3 91-2;

& G athercole

1 958,

3 8-43.

D uring l evelling o f t he b ank i n 1 923, R oman c oins w ere f ound ' lying u pon t he s urface b eneath t he r ampart'. These i ncluded two coins o f Gallienus, o ne o f T etricus I a nd t wo a ttributed t o G ratian ( c.AD 3 68). 1 959 s ection n ear s outh end o f Outer Bank r evealed a b ank o f t he 4 th century.

t wo

p eriods,

t he

f irst

u ndated,

t he

Cotton & Gathercole 1 958, 3 3-4; JRS ' ( 1960) 5 0, 2 33. See a lso Webster 1 975b, 5 3; Johnson 1 976, 6 2, 1 42 & f ig. B rampton C rop r narks s ectioned

( Norfolk)

( TG

a nd a erial t he d itch i n

o f

7 7.

2 23237) p hotographs. 2 p laces.

E xcavations

1 974

Roman s ettlement/industrial c omplex a t c rossing o f R . east/west r oad f rom Water N ewton t o Caistor-by-Yarmouth. E arthwork

s econd

a nd

B ure

1 975

by

defences

N o t race w as f ound o f a w all o r b ank t o a ccompany t he d itch but t hese c ould h ave b een r emoved by p loughing. T he earliest d itch f illing c ontained p ottery o f t he l ate 2 nd/early 3 rd c entury and t his i s t he d ate s uggested by t he e xcavators f or i ts construction. T he s econd p hase o f d itch f illing c ontained d umped

1 07

s oil

a nd

B ritannia 2 11-13.

i ndustrial ( 1975)

d ebris 2 60



B rough-on-Fosse

d ated

&

p 1.

C rococalana

l ate

XVIIIa;

( Notts. ,)

Aerial photographs of c ontrolled e xcavation.

n orth-west

S ome modern building s ettlement u nknown.

b ut

E arthwork

3 rd/early ( 1976)

( SK a nd

l argely

4 th 7 ,

c entury. 3 41;

( 1977)

8 ,

8 36584) w estern

d efences.

d eserted.

C haracter

No

o f

defences

Two b road d efensive d itches l ocated b y a erial p hotographs t o north-west of F osse Way a nd apparently cutting earlier o ccupation. D efences u ndated. St. S ee

Joseph 1 966, 2 7-8 & p 1. a lso Todd 1 973, 3 8-9.

Caistor - on-the-Wolds P artial s urvival. s outh-west. I rregular Character Walled

V II;

( Lincs.) T he

only

W ilson

( c.

T A

1 975,

t he

& p 1.

X IIIb

0 15125)

e xcavation

p olygonal e nclosure o n of s ettlement u nknown.

1 4

t ook

Wolds

p lace

e ast

o f

i n 1 959

E rmine

o n

t he

Street.

c ircuit

N o e vidence h as b een f ound f or a p re-existing e arthwork c ircuit. The f ootings of t he w all w ere l aid on t he n atural s tratum o f c lay a nd r ocks o n t he s outh-west. D ating e vidence w as s canty and n o f inds i n t he l ayers a ssociated d irectly w ith t he wall. None o f t he p ottery w as e arlier t han t he 3 rd c entury a nd most o f i t was consistent w ith a d ate i n t he f irst h alf of t he 4 th c entury. Two appeared

o f t o

c onclusive Ant.

J .

t he be

s urviving b astions ( Cooper's a nd North b onded i nto t he w all b ut t he e vidence

b ecause

( 1960)

4 0,

o f

modern

a nd

o bstructions.

1 75-87.

S ee a lso Whitwell 1 970, Johnson 1 976, 1 23; 1 980, Caistor - by-Yarmouth

i ntrusions

6 9-72 1 01.

& f ig.

C aister-on-Sea

E xcavations i n 1 951-1 53, a nd n orth-west.

1 961-1963

6 ;

Todd

( Norfolk) on

t he

1 973,

( TG

s outh-east,

& 1 27;

n orth-east

o n t hat

c ircuit a nd p resumably e nclosing a l arger o f t he s ettlement h as n ot b een e stablished Shore f ort.

1 08

4 2-4

5 17124)

C oastal s ite w ith a pproximately s quare w alled a rea d ifferent a lignment on t he n orth a t l east f rom e arthwork c haracter b e a S axon

bastion) was n ot

s lightly of t he

a rea. a nd i t

T he m ay

E arthwork

d efences

P alisaded e nclosure d ated i n t he e xcavations 1 951-1953 t o t he H adrianic period b ut dating evidence n ot s tated. Subsequent s ections p roduced n o f urther d ateable material. Walled

c ircuit

I n t he e xcavations o f 1 951-1953, t he w all a nd d itch ( possibly a lso t he o uter d itch) were p laced i n t he s econd h alf of t he 2 nd c entury. D ating e vidence n ot s tated. D ating e vidence i n s ubsequent s ections w as s canty b ut t he wall was a ttributed t o t he f irst h alf o f t he 3 rd c entury ' perhaps n earer 2 50 t han 2 00' o n t he b asis o f a'3rd' mortarium f ound i n t he r ampart a ccompanying t he w all i n t he n orth-east c orner i n 1 962. I n t he e xcavations o f 1 951, t he w estern j amb o f t he s outh gateway ( stone) was e xamined. No f urther i nformation a vailable. I n 1 961-1963, t he r ecutting o f t he i nner d itch a nd perhaps a lso t he c onstruction o f t he outer d itch w ere a ttributed t o a r eorganisation o f t he d efences p erhaps i n t he e arly 4 th c entury; dating b ased on pottery i n t he p rimary s ilt o f t he i nner d itch. The f oundations l ocated i nside t he n orth-east a ngle o f t he walled c ircuit ( possibly f or a s tair o r r amp ) w ere t hought t o b e c ontemporary w ith t his r eorganisation. JRS ( 1952) 42 , 9 6-7; ( 1962) 5 2, 1 76 & f ig. 2 3; ( 1963) Norfolk A rch. ( 1965) 3 3, 9 4-107; ( 1969) 3 4, 4 5-73. S ee a lso Johnson 1 976, 1 8, 9 7-8 & 1 24. C ambridge

D urolipons

E xcavations i nformation

( Cambs.)

a b asically

1 37;

4 43592)

i n 1 960s on t he s outh-west a nd w est. a vailable; r esults a re u npublished . .

I rregular e nclosure o f c rossing o f R . Cam. Walled

( TL

5 3,

r ectangular

N o

s hape

a dequate

s ituated

a t

c ircuit

N o e vidence f or p re-wall e arthwork c ircuit. Wall, r ampart and b road s hallow d itch a ttributed t o early 4 th c entury. I n 1 964, e xcavations o n s outh-west d efences ( Mount P leasant) f ound wall c ompletely r obbed; c lay, e arth and g ravel b ank c ontained 4 th c entury p ottery t hroughout. West g ate w ith g atehouses a nd p rojecting t owers e xcavated 1 967 a nd 1 969. N o f urther i nformation a vailable. JRS ( 1965) 5 , 2 13; ( 1968) ( 1978) Cambs. I I. 4 2. S ee a lso A lexander 1 975, 1 08 Catterick A erial

1 94;

B ritannia

& f ig.

a nd

e xcavations

R oughly L-shaped e nclosure s outh o ccupied b y R oman f ort.

1 ,

2 90;

VCH

2 .

C ataractonjum/Brompton-on-Swale

p hotographs

( 1970)

1 959,

( Yorks.)

1 972

a nd

( SE

2 25995)

1 974.

o f R . Swale and i ncluding a rea L argely deserted. D efended

1 39

v icus/planned Walled

early out t he

t own

o f

t he

l ate

3 rd/early

4 th

c ircuit

D efences were apparently f irst p rovided s ometime i n 4 th century f or a s ettlement which developed o r w as

a t t his early

t ime. F reestanding wall 4 th century s uggested by

a nd s ingle d itch. t ery c oins and po t

1 972 e xcavations. Eviden ce f or t wo periods t he s tretch of wall f acing t he r iver. B ritannia S ee

c entury.

a lso

( 1973) F rere

4 ,

2 79-80

1 966,

9 6-7

Chesterton -on Fosse

& f ig. & f ig.

( Warks.)

3 ;

( 1975)

2 1;

Wacher

( SP

enclosure

astride

t he

A d ate found

c onstruction

6 ,

i n i n

a long

2 35.

1 971,

1 70-2

& f ig.24

north

gate

3 41597)

Aerial photographs. E xcavation 1 961 a nd n orth west corner. R esults u npublished. Rectangular

o f

t he l aid

Fosse

1 967

Way.

at

I rregular

and

S treet

plan, Stone buildings, e xtramural s ettlement t o s outh-east. Known mostly f rom Aerial P hotographs which a lso i ndicate t he position of a gate on t he s outh-east. E arthwork

defences

Emergency e xcavations 1 961 i n a dvance o f r oad w idening a t t he north gate i ndicated p resence of c lay rampart and a t l east t wo d itches. 1 967 e xcavations on n orth-west r evealed t urf b ank and ditch o r possibly d itches. No dating evidence mentioned. Walled

c ircuit

1 967 e xcavations p laced t he w all i n t he f irst h alf o f t he 4 th c entury and evidence f or t wo ditches and a c ounterscarp r eplacing t he e arlier d itch s ystem. I n 1 961, ' the w all w as a lso attributed t o t he 4 th century b ut t he double d itch s ystem was r eplaced ( after i ndication given

other

l ate 3 rd c entury) by a s ingle of t he dating evidence.

North gate small i nformation. -

s imple

p lan

w ith

WMNS ( 1961) 4 , 3 ; ( 1967) 1 0, 1 8; JRS S ee a lso Wilson 1 975, 1 4 & p 1. X III6T o

d e

Military

Corstopitum o rdnance

d epot

No f urther 1 975, 1 4

i nformation.

Doncaster

D anum

Two

s ections

only

( W.

( Northumberl.) a nd S ee

defended Salway

R iding/Yorks.)

a cross

defences

1 ( 1

on

b road

2 p rojecting

( 1962)

( NY

v icus 1 965,

( SE west

5 2,

d itch.

No

t owers.

No

1 71-2

& f ig.

1 8

9 83648) a long

4 1,

S tanegate.

4 5-60,

5 74034) 1 976.

1 87;

W ilson

Defended v icus. Very l ittle k nown about t he c haracter o f t he defended a rea and t he s equence of military and c ivilian o ccupation h as s till t o be e stablished. E arthwork

defences

A p it cut by t he i nner d itch c ontained pottery o f AD 1 401 60. D itch backfil1ed, t he f ill containing l ate 2 nd o r early 3 rd c entury pottery, a nd t wo d itches dug f urther o ut, t he i nner o f which cut t hrough two pots. One o f t hese pots contained a coin h oard w ith a date r ange down t o AD 3 37 a t t he l atest. The d itch i tself contained pottery of t he l ate 4 th c entury. P ers. See J .

comm. D r. R . Magilton

P . C . 1 977,

Dorchester-on-Thames

Buckland 3 4-6.

( Oxon.)

1 979;

( SU

B ritannia

( 1977)

81

3 84.

5 87945)

Survival as a n earthwork on s outh and s outh-east. E xcavations 1 935-1936 a nd 1 962 o n western d efences, 1 963 on s outh a nd 1 972 north-west.

i n on

R ectangular e nclosure b eside R . Thames on Roman r oad f rom S ilchester t o Towcester. E astern a lignment o f defences u ncertain nd e xcavations 1 972 on e xpected l ine o f eastern defences f ound n o s ign of t hese ( Britannia ( 1973) 4 , 2 97). Early defensive

circuit

E xcavations ditch of t he l ate ( Britannia ( 1973) Earthwork

defences

E arth t erminus material Walled

o n t he n orth-west i n 1 972 l ocated ' a defensive i st C ' a ligned parallel t o t he western defences 4 , 2 97).

r ampart

a nd

V-shaped

post quem c . AD down t o c . AD 1 85.

d itch.

1 60

and

P ottery

t he

bank

b elow

b ank

i tself

g ave

a

contained

c ircuit

Wall c onstructed i n f ront o f r ampart a nd t he contemporary r ampart heightening c ontained material of c . AD 2 76-290. A n ew d itch s ucceeded t he o riginal d itch a t a date b elieved t o b e i n t he 4 th century. L ater two smaller V-shaped ditches were cut i nto t he u pper f illing o ver b oth l ips o f t he o riginal d itch. See F rere CBA G roup D orn Aerial

1 965, 1 38; R owley 9 N ewsletter 1 974,

( Glos.)

( SP

1 975, 3 -4

18 1

& f ig.

1 and

1 24

2 07338)

photographs.

R ectangular enclosure D eserted s ite.

Apparently w ith

n o

r egular

11 1

e xcavation. S treet

p lan

beside

F osse

Way.

Wilson

1 975,

D roitwich P artial

1 1

Salinae

s urvival

E xcavations 1 972 on t he

( He.

a s

a n

& Worcs.)

e arthwork

i n 1 955 on n orth-east.

t he west, Only t he

( SO

o n

8 98637)

n orth

a nd

e ast.

1 967 on t he n orth a nd 1 971 a nd 1 955 e xcavation i s p ublished.

R oughly r ectangular enclosure on n orth b ank o f R . Saiwarpe a nd o n opposite s ide o f r iver t o t he main s altworking a rea. D efences p rovided f or a v illa c omplex e stablished i n l ate 2 nd o r e arly 3 rd c entury a nd r egarded a s l ikely t o b e t he o fficial r esidence f or t he p rocurator of t he s alt i ndustry. T he e nclosure c onsists o f a double d itch s ystem o n t he n orth, w est a nd e ast a nd a r ampart o n t he west and east s ides. E arthwork

defences

C lay bank a nd d ouble d itch s ystem i dentified o n w est 1 955. No dating evidence. 1 967 s ection t hrough double d itch s ystem o n n orth. No d ating. On t he n orth-east, t he r ampart w as s ectioned i n 1 971 and 1 972 and a c oin o f Carausius ( AD 2 87-293) w as f ound i n t he r ampart material. The d itch w as n ot s ectioned o n t his s ide. The construction of d efences s eems t o b e r elated t o a r eplanning o f t he v illa s ite a fter c . AD 2 70 b ut b efore C . AD 2 90. The s ite was destroyed by f ire a t t he end o f t he 3 rd c entury. T BAS

f or

1 957

( 1959)

7 5,

2 -3,

f igs.

1 & 1 4;

WMNS

( 1967)

1 0

,

7 ;

( 1971)14, 1 7-19 ( 1975) 1 8, 4 8; B ritannia ( 1972) s ee a lso F reezer

& f ig. 8 ; ( 1972) j, 1 7-18; ( 1974) 1 7, 4 9-50; JRS ( 1956) 4 6, 1 30 & f ig. 2 8; ( 1968) 5 8, 1 8; 3 , 2 17; ( 1973) 4 , 2 87-8 1 977, 7 & f ig. 4

Fenny

Magiovinium

Stratford

E xcavations defences.

i n

1 969

E nclosure

a long

t he

E arthwork

defences

a nd

( Bucks.)

1 970

Watling

( SP

l ocated

S treet

8 90336)

s outh-east

s outh-east

o f

a nd

n orth-west

T owcester.

Rampart a nd d itch w ith a r ow o f s takes o n o uter l ip. R ampart of s outh-east d efences s howed t wo p eriods o f c onstruction. No d ating e vidence m entioned. F ill o f d itch contained material of l ate 3 rd early 4 th c entury. -

B ritannia

( 1970)

G atcombe

( Avon)

1 ,

2 88-9;

( ST

( 1971)

2 ,

2 68

5 26698)

Some s urface i ndications. E xcavations 1 965-1966, 1 967(?) a nd 1 973.

12 1

o n

t he

d efences

i n

1 954,

R oughly r ectangular a rea e nclosed a lthough t he p osition o f t he s outhern defences i s u nknown. The s ite does n ot l ie on any k nown R oman r oad a nd i ts c haracter a nd p urpose a re s till n ot f ully understood. S ettlement e stablished i n l ate 3 rd c entury and s eems l ikely that defences w ere p rovided f rom t he f irst. P ossibly a p rivate o r o fficial estate. Walled

c ircuit

T he walled c ircuit i s u nusual i n t he massive w idth o f t he wall and t he f act t hat no d itch h as yet b een l ocated. The d efences d id n ot p roduce v ery c lose d ating. No d irect e vidence f or t he c ontemporary e rection of t he d efences and o f t he b uildings i nside t hem b ut t his i s r egarded a s l ikely. On t he b asis of e vidence f or t he construction o f t hese b uildings, t he d efences h ave b een p laced w ithin t he p eriod AD 2 80-300. B ritannia

( 1974)

Godmar ichester



4 52;

Duroviqutum

B ranigan

( Cambs.)

1 975

& f ig.

1 & 1 81

( TL246705)

P ossibly s urvival i n modern s treet p lan and p roperty boundaries. E xcavations i n 1 955-1957, 1 959-1961, 1 972-1974 a nd 1 978. R oadside s ettlement with mansio s ituated at c rossroads on E rmine S treet beside R . O use. Walled c ircuit e ncloses h exagonal a rea a round the c rossroads. A long s outh west s ide t here i s a s tretch o f 1 16m where n o t race o f t he d efences h as b een f ound. E arly defensive

c ircuit

A c ase h as b een made f or a V-shaped dyke e nclosing t rapezoidal a rea a round t he mansio a nd c rossroads. No e vidence f or r ampart. E ntrances on s outh a nd e ast a nd p ossible c rossing points r epresented by wooden p iles on n orth a nd west. R egarded a s H adrianic a nd a ssociated w ith t he b uilding o f t he mansio. B ut s ome c onfusion e xists a bout t he e vidence f or t his date. I n 1 956, a d itch, which w as i nterpreted a s b elonging t o t his dyke s ystem, was f ound t o c ontain p ottery o f t he i st c entury AD and a l ater 1 st c entury date w as a ttributed t o t his d itch w here i t w as l ocated on t he e ast. H ere i t was f illed i n t he 2 nd c entury a nd b uilt over. D r. A . S . E smonde C leary p oints o ut t hat t he e vidence f or t his comes f rom t he e xcavations a t t he West g ate, where t he s o c alled dyke i s i n f act a l eat f or t he b aths n ear t he m ansio, a nd at t he south gate w here t he d itch was n ot d efensive a nd o ccurred only e ast o f t he g ate ( pers. c omm. 1 980). E vidence f or H adrianic date n ot s tated. P roc.Camb. Antiq. S oc. ( 1958) ( 1962) 52, 1 74, p lan; B ritannia f igs. 4 , 1 0, 1 4 Walled

5 1, 8 3; ( 1975) 6 ,

Circuit

1 13

( 1961) 4 , 2 51; G reen

7 5; 1 975,

JRS 1 96,

Wall w ith c ontemporary c lay r ampart. C oin o f T etricus ( AD 2 70-273) f ound i n r ampart a t s outh gate p rovides a t erminus p ost quem o f t his d ate. A C 14 d ate w as obtained f rom m aterial s ealed b eneath t he wall and was u sed i n s upport o f t his c oin e vidence. Subsequently, a t imber r evetment was a dded t o t he r ampart o n t he n orth s ide. A t rench a cross s outh-west d efences i n 1 979 l ocated t he w all and a d itch 1 9ft. ( 6m ) wide. The l atter p artially d isturbed e arly 3 rd c entury i nhumation b urials. The w all f oundations h ad b een r obbed e xtensively by t he e arly 4 th c entury. A s ubstantial s ingle — s pan gateway o n t he s outh w ith gatetowers a nd a covered a rchway was e xcavated i n 1 959-1961 and r egarded a s b eing c ontemporary w ith t he w all b uilding a lthough i t was a ctually b uilt i nto a gap l eft i n t he walled c ircuit. R ear o f n orth gate e xcavated 1 972-1973. S econdary a lteration t o n orth gate w ith b locking o f t he w estern f ootway. 1 973 e xcavation o f f an-shaped b astion o n s outh-east ( Cf. Ancaster) a nd e xcavations 1 957 and 1 974 on t he west d emonstrated t hat a n ew d itch ( 36ft/llm w ide) was d ug i n t he l ater R oman period. The d itch s ilt c ontained pottery p robably o f t he s econd h alf o f t he 4 th c entury. I n t he l ate 4 th c entury, a d itch ( 8ft./2.5m w ide) w as dug i n f ront o f a n p arallel t o t he s outhwest d efences b ut c utting a cross t heir l ine at s outh e nd. P roc. C amb. Antiq. S oc. ( 1958) 5 1, 6 8-82; Arch. NL ( 1961) v ol. 7 n o. 4 , 9 5; B ritannia ( 1973) 4 , 2 88-9; ( 1975) k 2 51; ( 1979) 1 0, 3 01; G reen 1 975, f igs. 5 , 1 0, 1 4 ,

G reat

C asterton

E xcavations

i n

( Lincs.) 1 950s

o n

0 02090)

( TF

v arious

p arts

o f

t he

I rregular polygonal e nclosure on t he E rmine R . Gwash b etween L incoln a nd Water N ewton. E arthwork

defen c

w alled Street

c ircuit. a t

c rossing

o f

s

The r ampart was r egarded b y t he e xcavator a s b eing contemporary w ith t he wall b ut t he early d ate ( late 2 nd/early 3 rd c entury) s uggested f or t he w alled c ircuit l ed t o d oubts a bout t his i nterpretation. Todd s uggests t hat t he p ossibility t hat a p re-wall r ampart c an b e i dentified f rom t he published s ection ( Todd 1 973, 3 7-8 & f ig. 7 ). Walled

c ircuit

Wall a nd r ampart r egarded a s c ontemporary by t he e xcavator a nd t he e vidence of pottery a ssociated w ith a h earth s ealed b eneath t he r ampart s üggested a l ate 2 nd c entury date o r t erminus post quem f or t he wall. On t he e ast a nd s outh-east, p ottery s ealed b eneath t he w all-footings w as d ated n ot l ater t han A D 1 70. The o riginal d itch was f illed i n a nd a s hallow o uter d itch dug; attributed t o t he 4 th c entury. B astions w ere b uilt o ver t he f illed i n i nner d itch. F our b astions i dentified; e xcavated ( north, e ast a nd s outh-east) i n 1 954 l ayer f or t he e ast b astion c ontained 14 1

a nd 1 955. a c oin o f

The c onstruction C onstans AD 373

3 50; a c oin o f Constans o r C onstantius I I ( AD 3 37-361) w as f ound i n t he s outh-east b astion f oundations. F rere n otes t hat t wo c oins m inted b etween AD 3 54 a nd AD 3 58 were a ssociated w ith t wo d ifferent bastions ( Frere 1 978, 2 91) b ut s ee C asey f orthcoming. o f AD 355-360 for t he T he e xcavators s uggested a d ate c onstruction o f t he b astions. C order 1 955, 3 2-3; 1 15; ( 1954) 4 4, 9 2 3 ; ( 1957) 4 7, 2 12;

JRS ( 1951) 4 1, & f ig. 4 ; ( 1955) ( 1959) 4 9, 1 12.

1 27 4 5,

& f ig. 1 32-3;

1 8; ( 1953) 4 3, ( 1956) 4 6, 1 31-

S ee a lso P . Corder ( ed.) The R oman T own and V illa a t G reat Casterton, Rutland. R eports 1 -111 ( Nottingham 1 951, 1 954 & 1 961) ( not c onsulted); Whitwell 1 970, 6 1-2 & f ig. 5 ; T odd 1 973, 3 7-8, 1 26, f igs. 7 & 3 4; Webster 1 977, 3 20-1.

G reat

C hesterford

( Essex)

( TL503430

Aerial p hotographs. E xcavations 1 948-1949 c ircuit o f t he n orthern and w estern parts o f

t raced t he t he wall.

c omplete

R oman s ettlement a long t he L ondon t o C ambridge r oad b eside t he R . C am. U nusual r egular polygon o r o val enclosure w ith r egularly s paced a ngles. D eserted s ite. Walled

c ircuit

F reestanding w all w ith b road s hallow d itch. D ating e vidence p rovided by pottery o f l ate 3 rd o r e arly 4 th c entury a nd o ne s herd o f early 4 th c entury i n a s haft c overed b y d ebris f rom a b uilding destroyed by f ire a nd s ealed b y t he wall. N orth gate e xcavated a nd d escribed a s h aving s ome p uzzling f eatures and a lterations. S imple gateway consisting o f a gap on e ither s ide o f w hich t he w all w as w idened e xternally; g ateway c ompletely b locked a t one period by t he c onstruction o f a t ower a t i ts opening; a djoining l ength o pening.

o f

n ew g ateway wall a nd t he

JRS ( 1949) 3 9, 1 05-6; ( 1963) 1 11, 7 2-6 & f ig.

( 1950) 1 9

s ee a lso Alexander 1 975, Wilson 1 975, 1 0 & p 1. V

H ardham

( Sussex)

( TQ

L imited

e xcavation

Small r ectangular S ite d eserted. E arthwork

4 0, '

1 06-7

R odwell

c ut t hrough t hrough t his

& f ig.

1 975,

8 8,

2 3;

VCH

f igs.

t he n ew

E ssex 2 & 4 ;

0 30175)

1 926. e nclosure

o n

S tane

S treet

s outh

o f

m id.

2 nd

R .

A run.

defences

E nclosure. S ussex

1 08;

s ubsequently r oad d iverted

A rch.

Abandoned

Coil.

( 1927)

a s 6 8,

s ettlement 8 9-132.

1 15

s ite

c entury.

s ee

a lso

C unliffe

L i e ybLigge

1 973,

s ee

-

Horncastle

& f ig.

2 1.

Maldon

( Lincs.)

Survival o f E xcavations 1 960s.

6 9-72

( TF

p arts o f on t he

2 58695)

t he w alled c ircuit o n s outhern d efences a nd

west, n orth a nd s outh. n orth-west corner i n

T rapezoidal e nclosure o n s outhern e dge o f Wolds east o f E rmine Street a t confluence o f r ivers B ain a nd Waring. Not o n any k nown Roman r oad. C haracter o f s ettlement u nknown. Walled

west w all.

c ircuit

Wall a nd b astion

Whitwell 1 970, ( 1976) 7 , 325;

l ichester

1 948

P ossible l ate w ith s uburbs.

e vidence. N orthbuild with t he

( Som. ,)

a nd

c ivitas

1 949

( ST on

Britannia

5 24226)

n orth-east,

c apital.

7 ;

O n

1 969

F osse Way;

o n

s outh

s izeable

and

1 974

s ettlement

d efences

C lay b ank I nitial c lay f ront. t han AD

N o d ating b e o f one

7 3 & f ig. 6 ; Todd 1 973, 4 2-4 & f ig. J ohnson 1 976, 1 23; 1 980, 1 01.

L indinis

E xcavations on w est.

E arthwork

c ontemporary r ampart. e xcavated a nd s hown t o

s hown t o b e o f t wo p eriods o n t he w estern c ircuit. r ampart l ater g iven a t imber f rame and palisaded

On n orth-east, t he d efences h ad b een l evelled n ot e arlier 1 50 a nd possibly n ot b efore AD 2 00 and a house w as b uilt

o ver t he d efences. On t he w est, t he r ampart h ad b een c ut b ack a nd partially l evelled t o f orm t he f oundation f or a l arge 3 rd c entury c ourtyard h ouse w hich c ontinued i n u se u ntil t he e arly 5 th c entury. The dating e vidence i s n ot g ood; t he i nitial c lay r ampart w as a ttributed t o t he l ate 2 nd o r e arly 3 rd c entury b ut t he e vidence i s n ot s tated. The s ubsequent palisading i s a lso u ndated. JRS ( 1949) 3 9, 1 08; ( 1975) 6 , 2 75-6. Walled

( 1950)

4 0,

1 10;

B ritannia

( 1971)

2 ,

2 78;

c ircuit

T he s outh w all a nd o ne t ower o f t he s outh g ate w ere e xcavated i n 1 969 a nd 4 th c entury c oins were f ound i n t he f oundation l evels o f t he w all. C lay b ank b ehind t he w all u ndated. O n t he w est a t l east t he wall h ad b een added t o t he 1 16

p re-existing B ritannia CRAAGS

I lkley

d efences.

( 1970)

I nterim

O licana

1 ,

Rpt.

2 96;

( 1971)

1 974.

( Yorks.)

2 ,

Casey

( SE

1 1

2 78

f orthcoming.

4 8)

I lkley s eems, on e vidence p roduced by B . H artley, walled i n t he f irst half of t he 4 th c entury ( Saiway No f urther i nformation.

t o h ave 1 965,

b een 4 1).

Defended vicus. S ite l ies s outh o f R . f rom r oute f rom Manchester t o Aldborough.

o n

r oad

Iping Aerial

( Sussex)

( SU

photographs.

Wharfe

a nd

t he

8 44261) No

excavation.

Small r ectangular e nclosure on t he C hichester t o S ilchester r oad c lose t o the c rossing of a t ributary of t he R . Rother. L argely deserted. Sussex A rch. Coil. ( 1953) 9 1; lso Cunliffe 1 973, 7 1-2. S ee a

I rchester

( Northants.)

Aerial photographs defences.

and

( SP

Wilson

1 975,

1 4

& p 1.

9 17666)

e xcavations

1 962-1963

on

Nearly r ectangular e nclosure s outh o f t he R . Nene Roman r oad r unning s outh-west f rom Water Newton. buildings and e xtensive s uburbs. D eserted s ite. Earthwork

of

X IVb.

t he

s outhern

a nd on t he Substantial

defences

E xcavations i n 1 962-1963 on t he t he south-west c orner f ound t hat

s outhern d efences t he r ampart w ith

j ust e ast a r ubble

core h ad been built u p o ver a n i nitial l ow marking-out b ank. Occupation material i ncorporated i n r ampart contained about haifa-dozen s herds dated a fter AD 1 20 a nd a n umber o f s herds paralleled at L eicester i n contexts dated AD 1 50-160. Pottery was n ot o f a t ype w hich a llowed p recise d ating. The a bsence o f colour coated ' Castor' ware, which f orms t he main b ulk o f Roman pottery f rom I rchester f rom AD 2 00 onwards, i s r egarded a s s ignificant. Rampart t herefore attributed t o t he period AD 1 502 00. Wide b erm s eparating r ampart f rom d itch w as t hought t o b e an o riginal f eature and i t was covered by i ntact 2 nd century o ccupation l evels. Walled

c ircuit

F ront of r ampart c ut away f or i nsertion o f w all. No evidence; construction t rench f or wall contained n o f inds.

17 1

dating

gate i nto

E xcavations i n 1 878 on t he w est g ate d emonstrated t hat t he had b een d ismantled a t some s tage a nd t his main e ntrance-way t he t own b locked. No dating e vidence.

JRS ( 1963) 5 3, St Joseph 1 966,

K elvedon

1 35 & f ig. 2 6-7 & p 1.

Canonium

Survival a s a n l ocated r eturn on

1 7; V ;

( Essex) e arthwork e ast s ide

Arch. Wilson

( TL

J , 1 967 ( 1968) 1 975, 1 1.

1 00-28;

8 64185)

o n t he n orth. Aerial p hotographs 1 976. L imited e xcavation 1 977.

R oman s ettlement w ith p ossible mansio s ituated B lackwater away f rom t he main L ondon-Colchester r oad. Earthwork

1 24,

b eside

R .

defences

A l ength o f s urviving b ank delimiting t he n orthern s ide o f t he s ettlement t hought t o be r emnants o f a defensive enclosure. Corner o f a d efensive e nclosure f ound o n w est a nd s outh i n excavation 1 977. Virtually n o dating e vidence f ound but d itch s tratigraphically p laced b etween AD 6 0-70 a nd AD 2 00-225. P ossible butt end of d itch f ound b eside R oman r oad on s outh-west. B ritannia & f ig. 2 .

( 1974)

Kenchester

5 ,

Magnis

f ig.

( He.

1 5;

( 1978)

& Worcs.)

( SO

9 ,

4 51;

Rodwell

1 975,

9 2

40428) 4

Aerial photographs and s urvival. T rench a cross defences at on n orth 1 925; e xcavations o n t he w est a nd t he west g ate 1 958, 1 960-1962; on t he n orth-west b astion 1 956-1958.

a ngle 1 956-

I rregular e nclosure a stride t he e ast/west r oad f rom W orcester t o C lyro a nd j ust e ast o f t he j unction w ith t he Watling S treet West and c lose t o a c rossing o f R . Wye. S ubstantial b uildings a nd e xtramural E arthwork

s ettlement

k nown.

D eserted

s ite.

defences

Rampart a nd d itch e xcavated 1 956. M uch p ottery f rom r ampart and l ayer o f o ccupation s ealed by i t. L arge quantity o f pottery o f m id-2nd c entury a nd f ragment o f s amian o f t he Antonine p eriod ( AD 1 40-180) P ottery of s imilar date i n t he r apid s ilt of t he d itch. General e vidence f or t he t ermination o f o ccupation i n t he f irst h alf o f t he 2 nd c entury i n t he a rea u nder t he r ampart a long t he w estern d efences. S outh o f t he w est gate i n 1 960-1962, a hearth c ontaining a c oin o f T rajan ( AD 1 01-103) i n g ood c ondition was s ealed by a l ayer w hich was c ut by t he i nner ditch. .

I n a llowed B ut possibly

t he 1 956 e xcavation, t o collapse, s ilt up s ee t oo

i t w as f ound t hat t he and b ecome f illed.

B idwell 1 980, 6 2, small t o r epresent

d itch

h ad

b een

w ho s uggests t hat t he r ampart a n i ndependent earthwork.

1 18

i s

Walled

c ircuit

I n 1 956 t he r emains o f t he c ore o f t he w all i n f ront o f t he b ank w ere f ound. The b ack of t he wall was f aced w ith m asonry a nd t he c onstruction t rench f or t his f acing c ould b e s een i n t he r ampart. No dating e vidence f or t he wall. S herds f rom t he w all w ere n ot l ater t han t he mid-2nd c entury. T he e xcavator w as i nclined to t he p ossibility t hat t he wall dated f rom t he m id-4th c entury and w as c ontemporary w ith t he c onstruction o f t he w est s tone g ate a lthough i t c ould b e earlier ( TWNFC ( 1963) 3 7, 1 64 & 1 67) West gateway, n orth h alf s tripped 1 956; 1 960-1962 e xcavation o f r est of gateway. Only t he f oundations o f t he double a rched s tone g ateway r emained w ith s quare g atehouses o n e ach s ide o f t he c arriageway. These g atehouses were o f d ifferent s ize and l ayout w ith a guard c hamber i n t he n orth t ower a nd a much smaller c hamber i n t he s outh t ower. E ach g atetower was f ronted by a s emi-circular b astion w hich p rojected o ut o ver t he f illed i n i nner d itch and t hat of t he s outh gatetower c ontained r eused s tone. I t i s possible t hat t hese p rojections, w hich w ere s olid a t the base at l east and of s imilar c onstruction t o t hat o f t he n orth-west b astion, were l ater a dditions t o t he f ront o f t he g atetowers. R elationship o f wall a nd gateway uncertain; , t he s outh t ower a nd t he w all b eing b onded i n a t u pper l evels o nly. F oundations of gate c ut t hrough o ccupation o f t he 1 st and 2 nd c entury; a c oin o f C rispus ( c. AD 3 20-324 ) c ame f rom t he t op o f t he f oundations on t he n orth s ide o f t he s outh t ower b ut was n ot s ufficiently well s tratified. F oundations o f g ateway c onstructed o ver a nd i nto t he f ill of t he c ulvert. P ottery i n t his f ill s howed t hat t his i nfilling ' occurred c . AD 3 50' ( TWNFC ( 1963) 3 7, 1 52; s ee a lso Wacher 1 969, 4 9, 5 3n3) S ubsequent a lteration t o gateway; r evised p lan f or m iddle o f gateway with c onstruction of a narrower r oad t hrough n orth part o f gateway. Some e vidence t hat f ront o f s outh c arriageway may have b een b locked b y a wall. D itch r ecut and e xtended n orth t o meet n ew narrower r oad t hrough g ateway. No d ating. S ubsequently d itch r ecut a nd f illed g radually. No d ating. 1 956 n orth-west b astion l ocated. S emi-octagonal f oundations c ontaining r eused b uilding b locks. E xcavation 1 957-1958 s howed t hat f oundations n ot i ntegrated w ith t hose o f t he w all s ince b astion f oundations were s et a t a l ower l evel t han t hose of t he w all. Bastion b uilt o ver i nner d itch w hich h ad b een h alf f illed w ith n atural s oil a nd g ravel; packed s tones l aid o ver f ill t o p rovide f irm b ase f or b astion. No d ating. A erial photographs s how i nner d itch a nd b road o uter d itch. 1 958 e xcavation o f o uter d itch. Two f ragments o f c olour c oated ware, p robably 4 th c entury, f rom s ilt o f o uter d itch . '

T WNFC ( 1958) 3 5, 1 38-45, 3 28; ( 1959) 3 6, 1 00-16, 3 61-3; ( 1962) 3 7, 1 49-78; ( 1966) 3 8, 1 92-5; JRS ( 1957) 4 7, 2 11; ( 1958) 4 8, 1 36-7; ( 1961) 5 1, 1 71; ( l962) 5 , 1 69; WMNS ( 1958) 1 , 3 ; ( 1960) 3 , 5 ; ( 1961) 4 , 5 ; ( 1962) 5 , 5 . S ee a lso A . R . W ilmott, K enchester A R econsideration ( MA d issertation, U niversity o f B irmingham, 1 979), 3 4-9 -

1 19

K irkby Thore P artial

B ravoniacum

s urvival

a s

a n

( West i norl.)

earthwork.

( NY63

A S ingle

2 5)

t rench

i n

1 961.

Defended v icus t o n orth a nd w est o f t he f ort. Civil s ettlements covers c . 3 0acres ( 12.1 h a.). S ite l ies e ast of R oman road f rom York t o Carlisle. Walled

c ircuit

S urface o bservations s uggested t he v icus w as defended. 1 961 t rench a cross l ine o f bank r unning n orth - e ast/south .-west l ocated t he f oundations o f a s ubstantial wall. N o d ating e vidence. CW ( 1964) 6 4, 6 3-4 & 6 9. S ee a lso S alway 1 965, 4 1

& 1 39-41.

L eintwardjne

( He.

B ravonium

& Worcs.)

( SO

4 04742)

P artial s urvival. E xcavations o n n orth-west d efences 1 958-1959 ( site A ) ; observations at s outh-west corner 1 964 ( site D ) ; e xcavation o n e ast d efences 1 971 a nd o bservations l ocated position of e ast gate 1 972. R oadside s ettlement w ith mansio a long Watling S treet W est b etween Wroxeter and K enchester. R ectangular e nclosure astride Watling S treet n orth o f R . T eme. D r. S . C . Stanford c onsiders t his enclosure t o b e a f ort; D r. G . Webster r egards i t a s a defended c ivil s ettlement ( Webster 1 975b, 5 3). E arthwork

defences

C lay r ampart w ith l og c orduroy b eneath r ampart a nd t imber l acing; i nternally f acetted c orners o n n orth-west and S outh-west. Two, p ossibly 3 , d itches. I ntervallum r oad on n orth-west. T erminus post quer n f or r ampart p rovided by a s herd of s amian dated C . AD 1 60-190, f rom t he p re-rampart c ircuit o n t he n orthw est and c onfirmed b y 2 nd c entury c oarse wares and s amian f rom t he p re-rampart c onstruction l evels h ere. An e xtensive deposit of c harcoal a nd wood a sh at t he e xternal f oot of the r ampart i n o ne t rench ( trench 3 ) s outh o f t he n orth-west c orner w as i nterpreted a s t he f iring o f t he p alisade o r a t ower S oon a fter t he d efences w ere e rected. S low degradation o f the r ampart,but i t i s p ostulated t hat t he c ollapse o f t he r ampart o n t he n orthwest may have r esulted f rom t he r emoval of a palisade post a nd t herefore perhaps s ome d eliberate d emolition. R ecutting of d itches a nd r epair o f r ampart. N ew t imber f ront, r ampart b uilt u p w ith b rushwood a nd c lay , a nd d itches r ecut. I n 1 971 e xcavation, t he t imber l aced r ampart w as located a nd t he s ubsequent e nlargement w as f ound t o c over t he o riginal i ntervallum s treet. 1 )

I n f urther s ilted up

p eriod o f n eglect i n and w as t hen f illed

w hich w ith

t he o uter d itch ( ditch c lay and s andstone.

Counterscarp t hrown i nto i nner d itch ( ditch 3 ). s ealed by a s tone spread which w as a ssociated w ith o f

t he

r ampart

o n

t he

n orth-west

c ame

1 20

a 3 rd

c entury

F rom a l ayer t he r ebuilding s amian

s herd.

This s tone s pread b ehind t he r ampart may h ave b een a ssociated w ith t he building o f a wall o n t op of t he r ampart a nd i t was s uggested t hat t he s tone f ill o f t he i nner d itch m ay h ave c ome f rom t his. R eplacement o f e arlier d itches by a s ingle d eep d itch. s ilt

F urther r ampart u p s lowly.

r epairs.

Not

d ated.

D itch

l eft

o pen

t o

E ast e ntrance o bserved 1 972. I n t he e xtreme n orth-west c orner o f t he e xcavation, c lay a nd t imber o f t he r ampart l ocated a nd d ownwash f rom t he r ampart, n orth s ide o f g ravel r oad a nd w all f oundations w hich w ere t hought t o r epresent a f ront r evetment wall p ossibly r elated t o a gateway. Walled

c ircuit

R e possibility defences, see a bove.

o f

a wall

o n

t op

o f

t he

r ampart

n orth-west

26; ( 1968) 3 9, 2 23, 2 58-66, 2 78TWNFC 1 958 ( 1959) 3 6, 8 7-99 & 1 ig. 9 ; 9 ; ( 1972) 4 0, 3 18-20; JRS ( 1959) 4 9, 11-12 1, & f 1 ( 1960) 5 0, 2 22; ( 1965) 5 5, 2 08; WMNS ( 1958) 1 , 4 ; ( 1959) 2 , 7 ; B ritannia ( 1972) 3 , 3 17.

L etocetum .

Wall

( Staffs.)

( SK

1 01066)

Aerial photographs. Antiquarian o bservations and s urvival o f p art o f t he w alled c ircuit i n 1 9th c entury. O ne t rench a cross d itches n ear s outh-east c orner 1 955, o ne s ection a cross s outhern r ampart 1 956; one t rench a cross n orthern d efences, 2 a cross western and one at s outh-west c orner 1 960-1961. Small r ectangular e nclosure a stride Watling S treet j ust w est o f j unction with Ryknield S treet a nd e xcluding much o f t he r ibbon d evelopment a nd t he s o-called mansio a nd b athhouse. S ite l argely d eserted. Walled

c ircuit

Wall w ith c ontemporary t urf r ampart a nd t riple d itch s ystem, c lay l ined and p robably wet i n R oman t imes. Wall e ntirely r obbed i n e ast a nd s outh s ections. I n 1 955 a nd 1 956, s herds f rom t he r ampart and a r oad o verlying t he t ail o f t he r ampart i ndicated t hat t he defences c ould n ot b e e arlier t han t he l ate 2 nd c entury. I n t he o ther s ections, t he r ampart c ontained l ittle o ccupation material and n o d ating e vidence. The r ampart h ad b een b uilt u p b ehind t he wall f ollowing t he c onstruction o f t he l atter. O n t he w est ( site L ) , t he w all h ad b een b uilt o ver a r ubbish p it w hich y ielded a c oin o f T rajan ( AD 9 8-117) a nd 2 nd c entury p ottery. T he g ravel b eneath ' the r ampart a lso c ontained 2 nd c entury p ottery; while on t he n orth ( site K ), s herds f ound b elow t he p resumed c onstruction l evel f or t he w all s uggested a 3 rd c entury d ate o r l ater f or t he c onstruction o f t he wall. O n t he w est ( site T ) , t he wall s ealed a w ell i n w hose d eliberate f ill w as p ottery suggesting a date n ot e arlier t han t he 3 rd t he w all. O n t he e ast, t here w as e vidence f or r ecutting and c leaning of d itches. O n t he n orth,

1 21

c entury f or c onsiderable e xcavation o f

t he

d itches

w as

d ifficult

b ecause

TBAS 1 957 ( 1959) 7 5, 2 4-29; f ig. 1 ; K eele A rch. G rp. NL S ee a lso Webster 1 971, 3 8-9, 5 5-7 & f ig. 2 .

L ittlechester

D erventjo

o f

t he

h igh

w ater

t able.

L SSAHS 1 963-1964 ( 1964) '5 , 3-7, 1 6 & ( 1961) 1 , 1 ; WMNS ( 1961) 4 ,6 4 2 f igs. 2 0 & 2 3; TBAS ( 1974) 8 6,

( Derbys.)

( SK

3 53375)

Survival of c ircuit e arly 1 8th c entury. E xcavation on s outhern d efences i n 1 960; on d itch o n e ast i n 1 966; t hree s ections on western defences i ncluding w est gate 1 968; t wo t renches on n orthe ast 1 968. R ectangular e nclosure L ittle i nvestigation c ertain. E arthwork

b etween t he R . of s ettlement

D erwent a nd Ryknield a nd c ivilian s tatus

S treet. i s n ot

d efences.

I n 1 968 e xcavations, t he r ampart h ad b een e rected o ver a t imber-framed b uilding which y ielded a s amian sherd no e arlier t han AD 1 25 . D efences much d isturbed o n s outh a nd n o t race o f a r ampart b ehind t he wall h ere. The s econdary d itch f ill contained n othing l ater t han t he 2 nd c entury. T he d itch on t he east w as c lay l ined. W est gate l ocated 1 968. Stone f oundations o f a s imple g ate which was n ot c ontemporary w ith t he s tone d efences w ere found a nd i t w as t hought t his m ight i mply a n e arlier e arthwork c ircuit o n a s lightly d ifferent a lignment. Walled

c ircuit

R ampart c ut b ack f or w all. Wall on t he west was ' dated t o t he mid-2nd p ottery i n i ts f oundtions b ut t his i s quem. D erbys. 5 7, 1 81; S ee a lso

L ong

A rch. J . ( 1961) 8 1, ( 196 ) 5 9, 2 11-13 Webster 1 975b, 5 3.

M elford

P artial

( Essex)

s urvival

a s

An

defences e arthwork

s ettlement may b een i dentified

8 5-110; JRS ( 1961) 5 1, 1 67; & f ig. 3 4 F Britannia ( 1970)

R .

( 1967) 1 , 2 83.

8 58453)

e arthwork.

R oman s ettlement b eside west f rom B raintree. E arthwork

( TL

r obbed o n s outh. The w all c entury' o n t he basis of p resumably a t erminus p ost

No

S tour

e xcavation. a long

R oman

r oad

r unning

n orth-

( ?) w hich

f orms

t he

s outhern

b oundary

of

t he

be o ne of s everal u ndated enclosures w hich h ave i n a ssociation w ith R oman s ettlements i n Essex.

1 22

R odwell

1 975,

M aldon

H eybridge

S urvival

a s

9 2.

a n

( Essex)

( TL

e arthwork.

N o

8 48072) e xcavation.

R oman s ettlement on n orth b ank o f R . Chelmer a t h ead o f B lackwater e stuary s preading a long a m inor R oman r oad f rom C helmsford apparently t owards C olchester. E arthwork

t he t he

defences

( ?)

I ron a ge a nd R omano-British r ectilinear enclosure which R . C helmer.

R odwell

1 975,

M alton

D erventio

E xcavations

9 2

& f ig.

( N.

material f ound l ies on h igher

i n 1 972 w ithin g round s outh o f

2

R iding/Yorks.)

1 949-1952 a nd

( SE

7 92716)

1 968-1969.

D efended v icus s outh o f t he p ermanent f ort a nd e xtending s outh o f t he R . D erwent. Major r oute c entre. R ampart a nd wall i dentified o nly o n t he e ast b ut a ssumed t o r un a long s outh a nd w est a lso. E arthwork

defences

E arth a ttributed W alled

r ampart to l ate

i dentified o n 2 nd/early 3 rd

e ast s ide c entury.

o f

v icus

1 968-1969

a nd

c ircuit

Wall l ocated o n e ast s ide 1 952 a nd 1 968-1969 where i t h ad b een a dded t o e xisting r ampart a t a date s aid t o b e i n t he l ate 3 rd c entury. T he 1 952 e xcavations i dentified a h eightening o f t he r ampart i n t he e arly 4 th c entury. A l ate 4 th o r e arly 5 th c entury d itch w ith a r ampart p alisade was a lso f ound on t he w est s ide o f t he v icus, a nd i t p resumed t o r epresent a r eduction i n t he d efended a rea. B ritannia

( 1970)

1 ,

M anchester

Mamucjum

E xcavation

1972

D efended D efences E arthwork

a long

a nd

2 80;

R obinson

( Lancs . .)

( SJ

1 978,

7 -8

a nd w as

& 1 0

8 35979)

1 978.

v icus ( ?) n orth o f i dentified o n w est

p ermanent only.

f ort.

Major

r oute

c entre.

defences

D itch w est

a nd t imber p alisade r unning n orth s ide o f v icus ( annexe?). D ating

1 23

f rom f ort s parse.

d efences T he d itch

h ad b een d amaged characterised t he Jones 1 974, 4 1, B ritannia ( 1979)

by workings s ite i n t he 4 5, 4 6-7, 1 0, 2 91.

Mancetter/Witherley

a ssociated w ith 2 nd c entury.

1 85;

J ones

t he

f urnace

& R eynolds

Manduessedum " ( Warwickshire)

1 979,

( SP

w hich

7 ;

3 28969)

Survival as a n e arthwork f or much of t he c ircuit. T rench ( A-B) 1 927 a cross b ank on n orth-east ( inconclusive); 1 954-1956 t rench on n orth-west j ust s outh of Watling S treet; 1 955 t rench ( C-D ) on n orth-east; 1 956 t rench ( E-F) a t s outh-east c orner; 1 964 t rench ( A) a cross e astern d efences j ust s outh o f Watling Street. Small r ectangular enclosure b i -sected b y Watling Street south of t he c rossing o f t he R . Anker. R ibbon d evelopment a long W atling Street on b oth s ides o f t he r iver w ith a l arge pottery production c omplex i n t he B roadclose a rea s outh-west o f t he r oad. P artial d esertion. E arly d efensive

c ircuit

?

I n 1 954-1956 t rench o n t he n orth-west s outh o f Watling Street, i t was f ound t hat t he r ampart h ad b een constructed o ver a d itch w hich c ontained i st c entury p ottery. The d itch b ecame narrower and s hallower on t he n orth s ide o f t he cutting as i f f or a n e ntrance a t Watling S treet. No t race o f a r ampart b ut a l arge p ost-hole c ontaining a f ragment o f p ottery o f c . AD 1 00 may h ave b een a ssociated w ith t he d itch. P ottery f rom t he s ilt of t his d itch s uggested a c losing d ate o f c . A D 1 20. A V-shaped d itch f ound a t t he s outh-east c orner i n 1 956 was t hought t o r epresent t he s ame f eature. I t c ontained p ottery n ot l ater t han AD 9 0-100 and i t was s ealed by a l ayer o f g ravel a nd a l ayer o f s ilt , b oth o f w hich c ontained p ottery o f t he m id-2nd c entury. T he r ampart of t he walled c ircuit h ad b een e rected o ver t his. N o t race of t his d itch w as f ound i n t he 1 964 e xcavations j ust n orth o f t he 1 956 s ection a long t he e astern c ircuit. Walled

c ircuit

C lay r ampart, wall a nd t wo d itches. The e xcavators o f 1 9541 956 and 1 964 c oncluded t hat t hese f eatures were c ontemporary. E xcavations i n 1 954-1956 o n t he n orth-west l ocated a r ampart composed of d istinctive l ayers of c lay a nd earth c ontaining pottery ' mainly o f 2 nd c entury d ate, w ith a f ew s herds o f l ate 3 rd o r 4 th c entury ware'. I n 1 955 o n t he n orth-east, i t was t hought p ossible t hat t he b ank w as o f t wo p eriods b ecause o f t he c lear d istinction b etween t he c lay a nd t he o ccupation l ayers of t he b ank. T he p ottery f rom t hese l ayers c ould n ot b e s eparated i n date a lthough t hat f rom t he o ccupation l ayers was generally 2 nd c entury w hile t hat f rom t he c lay wa s no t ear l i er t han c . AD 2 50. T he b ank b ehind t he wall a t t he s outh-east corner i n 1 956 was r egarded a s b eing o f o ne p eriod o nly a nd i t c ontained a l arge quantity o f p ottery n ot e arlier t han t he l atter part o f t he 3 rd c entury 1 964 o n

a nd t he

p ossibly o f t he f irst e ast , t he s equence f or

1 24

p art o f t he 4 th t he c onstruction

c entury. I n o f t he w alled

c ircuit was

e stablished

a s

f ollows;

dismantling of a s ubstantial t imber building on t he l ine of t he defences, d igging o f r ear r evetment posts f or r ampart, b uilding of wall begins, dumping of material f rom i n f ront of t he wall and f rom t he digging o f t he d itches t ogether w ith d emolition material to f orm t he r ampart, s urfacing of berm a nd t ail o f r ampart w ith a l ayer of small c obbles. A t erminus post q uem f or t he construction of t hese • defences was p rovided by t he building destroyed f or t he c onstruction o f t he d efences; p ottery f rom t he upper l ayers of o ccupation a ssociated w ith t his b uilding contained a mortarium d ated AD 2 40-320. The a bsence o f 4 th century material f rom a ny o f t he l evels was t hought t o b e s ignificant. The e xcavator o f t he 1 964 s ection c oncluded t hat the d efences w ere e rected i n t he l ate 3 rd o r e arly 4 th c entury. I n 1 955 o n t he n orth-east a d eposit o f h eavy b lack s ilt containing pottery o f t he l atter part o f t he 4 th c entury o verlay t he t ail o f t he r ampart. This l ayer was f ound a lso o n t he s outh east i n 1956 where i t contained a c oin o f L icinius ( AD 3 07-324) i n g ood condition a nd a ssociated w ith mortaria p robably o f t he mid 4 th century. I n 1 964, 4 th c entury pottery was r ecovered f rom material overlying t he t ail o f t he r ampart. T he wall was l ocated i n a ll t he s ections e xcept t hat o f 1 927; evidence f or much r obbing i n t he medieval a nd l ater periods. I n 1 964 on t he e ast d efences, two l ayers of t umble i n f ront o f t he wall r epresented t he i nitial c ollapse o f t he w all before r obbing. T he i nner d itch was l ocated i n 1 954-1956, 1 955 a nd 1 964 a nd an o uter ditch was f ound i n 1 964. P ottery f rom t he b ottom o f t he i nner d itch i n t he 1 964 e xcavation c ontained s herds c overing t he period AD 2 60-370. The i nner d itch c ould n ot b e f ully e xcavated i n t he o ther t renches. I n 1 954-1956 a nd 1 955 o n t he n orth-west and s outh-east, t he upper f ill of t he i nner d itch c ontained pottery possibly o f t he l ate 4 th c entury; i n 1 964 t he s ame w as t rue o f t he upper f ill o f b oth d itches. TBAS 1 928 ( 1931) 5 3, 1 73-95; 1 956 ( 1958) 7 4, ( 1971) 8 4, 1 8-44; WMNS ( 1964) 7 , 8 . See a lso Webster 1 971, 3 9 & 4 2, f ig. 2 1; T BAS f ig. 2 .

Castle

H ill

Margidunum

( Notts.)

3 0-52; ( 1974)

1 967-1970 8 6,

5 5-7

&

( SK 7 00415)

Survival as an e arthwork on east s ide a nd much o f n orth and north-west s ides. One a lmost c omplete s ection a nd o ne t rench across i nner ( ]ate ) d itch a nd one a cross outer ( early ) d itch o f s outhern defences 1 966-1968 w ith a t rench n ear t he s ite o f t he south g ate e ast o f t he F osse Way. Small i rregular p olygonal and L eicester a ssociated Deserted s ite. Earthwork B ank

e nclosure o n F osse Way b etween L incoln w ith r ibbon development a long r oad.

defences a nd

p ossibly

t wo

d itches.

1 25

T he

d ifferent

a lignment

o f

t he d itches l ed t Q t he c onclusion t hat t hey w ere n ot c ontemporary , and t hat t he outer d itch r eplaced t he i nner d itch a t a n unknown date b ut b efore t he e arly 3 rd c entury Terminus pQ q uem C . AD 1 40 p rovided by a s amian s herd f rom t he r ampart; another sherd of c . AD 1 40 o r l ater c ame f rom b eneath t he r ampart; a lso f rom t he r ampart a s herd w ith a s uggested date r ange o f c . AD 1 30-175. Walled

c ircuit

S tone w all a dded t o f ront o f e xisting r ampart. i t w as suggested t hat t he outer d itch d iscussed above might a lso b elong t o t he w alled c ircuit. E vidence f or a l ater r ampart w as f ound i n f ront o f t he o riginal r ampart b ut b ecause o f wall r obbing i t w as not c lear w hether i t w as c ontemporary w ith t he wall o r whether t he wall had b een a dded l ater. No dating evidence. R emoval of w all f oundation a nd o f s tratified m aterial b eneath i t a lso p revented r ecovery of dating e vidence f or t he c onstruction o f t he wall. U pper f ill o f o riginal d itch, w hich w as f illed w hen t he w all was e rected, c ontained n othing e arlier t han c . AD 2 00. D ouble d itch s ystem a nd c ounterscarp u ndated b ut p robably a ccompanied wall b uilding and t he outer d itch r emained open until t he end of t he R oman p eriod; s herds o f l ate 4 th c entury f ound n ear t he bottom of i ts f illing. The l arger i nner d itch contained a d ump o f v egetable r efuse, dung a nd a nimal b one a pparently a ccumulating i n t he mid and l ate 4 th century. T rans.

T horoton

S oc.

1 969

( 1970)

S ee a lso B ritannia ( 1971) 2, 1 975, 2 11-15, f igs. 4 2 & 4 3.

Melandra

Castle

E xcavations

o n

A rdotalia s outhern

D efended v icus t o s outh Etherow a nd s tream j ust

7 3,

4 2-53 .

3 14-16;

( Derbys.)

d efences

1 966

T odd

( SK a nd

1 973,

3 6

& f ig.

7 ;

0 10950) e astern

and e ast o f f ort a nd e ast o f Manchester.

defences

1 969.

s ituated b etween R . B oth f ort a nd v icus

established c . AD 8 0 a nd a bandoned,c. AD 1 40. D efensive ditch o f a t l east t wo p eriods f orming a n a rc w hich e nclosed m ajority o f v icus. Western and n orthern l imits of t he defences n ot located. D eserted s ite. E arthwork

defences

I n e xcavations o n s outhern d efences, t wo p hases o f d efensive a ctivity i dentified. P hase I : c lay r ampart a nd ditch n ot d ated b ut v icus h ad a pparently g rown c onsiderably by t he t ime d efences c onstructed. S light e vidence t hat t his phase o f defences I l evelled a fter AD 1 20. P hase I: h asty r econstruction o f d efences apparently a fter c onstruction o f mansio which a ttributed I t o t he early Hadrianic p eriod ( early 1 20s). P hase I d itch p laced b etween AD 1 20 a nd AD 1 40 when s ite a bandoned and t he c ontext f avoured f or t his a bandonment s uggested t he d ate o f A D 1 38-139 f or t his. No d irect dating e vidence. I n

e xcavations

o n

t he

e astern

1 26

d efences,

a

more

complex

s equence of 4 phases was f ound. P hase I : o riginal V-shaped d itch; c lay b ank a ssumed. P hase I I: d itch r elined w ith c lay; c lay b ank assumed. P hase I II: a pparently t o c ounter s ubsidence o f the b ank i nto t he d itch, s takes were hammered i nto t he s ides o f the d itch t o r etain t he b ank r evetment w hich w as n ow p rovided; s takes a lso built i nto t he outer f ace o f t he d itch. P hase I V: d ecay o f r evetment i nto a lready s ilted d itch a nd c ollapse o f r ampart possibly a ssisted by deliberate d emolition. T erminus p ost q uem of c . AD 1 40 f r f inal a bandonment o f d efences p rovided by pottery evidence f or t he d estruction of t he mansio. P erhaps t he d ifference b etween t he t wo s equences may b e e xplained by t he p resence o f t he mansio which may h ave l ain w ithin a n a rea w hich b ecame a n a nnexe o f t he f ort. JRS ( 1967) 5 7, 1 81 9 6-8; 1 971, ( 1973) ig. 7 . 4 & f

& f ig. 6 ; D erbys. 9 1, 7 0-80 & f ig. 1 ;

Mildenhall

( Wilts.)

C unetio

( SU

A rch. J . 1 969, ( 1970) a9 , B ritannia ( 1970) 1, 2 83-

2 16695)

Aerial photographs s how s ytem o f two d itches on e ast, s outh a nd west; t hree r ectangular b astions o n s outhern p art o f e ast s ide a nd s ix on s outh s ide' r egularly spaced on e ither s ide o f monumental s outh g ate. Walled c ircuit e ncloses l arger a rea o n w est, r unning j ust outside early d itch s ystem on s outh and a l ittle i nside i t o n e ast. Walled c ircuit s hows l ittle r egard f or e xisting street l ayout a nd s outh gate o ccupies d ifferent position f rom e arlier gate. E xcavations o n s outhern d efences 1 956 ( 3 s ections) ; e astern d efences 1 958 ( 3 s ections) ; 1 959 o n n orth a nd w est; 1 961 on n orth-west; 1 960 and 1 964 on west g ate. R oute c entre. v isible i nside D eserted s ite. E arthwork

R oman s ettlement s outh o f R . and o utside t he d efences on

K ennet. B uildings a erial photographs.

defences

N one o f t he s ections s eems t o h ave i nvolved t he r ampart which a ccompanied t he two V-shaped d itches. S ome i ndications i n 1 961 e xcavations o n n orth-west d efences t hat t he p re-wall d itches e xtended to t he n orthernmost l imits o f B lack F ield a nd t herefore e nclosed a l arger a rea o n t he n orth t han t he w alled c ircuit. Walled

c ircuit

I nner f ace o f wall b uilt o n o uter l ip o f o riginal o uter d itch on southern part of c ircuit and i n o ne o f t he s ections on t he e astern c ircuit. I t i s p ossible t hat t he l ip o f o ne o f t he earlier ditches w as l ocated b eneath t he t ail o f a v ery c onsiderable r ampart w hich a ccompanied t he w all o n t he n orthwest. Pottery a nd c oins f rom t hese s ections s uggested a c onstruction d ate f or t he w all o f AD 2 80-350. B ut a c oin o f A D 3 60 f ound i n t he p rimary s ilt o f a d itch c ut by t he c onstruction o f t he wall o n t he w est i n 1 964 g ave a t erminus p ost q uem o f AD 3 60 f or t he wall h ereiand p resumably a lso f or t he b astions ' s ince

1 27

t he s olid b astion on t he s outh-east, was f ound t o b e bonded i nto t he wall evidence f or t he b astion. JRS ( 1953) ( 1965) 5 5, ( 1963) 5 8,

w hich f ace.

was No

e xcavated i n 1 958, i ndependent d ating

4 3, 9 0-1, p l.XIII; ( 1957) 4 7, 2 22-3; ( 1959) 4 9, 1 31; 2 17; WAM ( 1957) 5 6, 2 41-5; ( 1960) 57, 2 33-5 & 3 97; 3 5 & 2 45; ( 1965) 6 0, 1 37 .

See a lso S t. Joseph J . H aslam, Wiltshire ( 1976) , 7 , p lan

Neatham Vindomi

1 966, 2 4-6; Towns; t he

? ( Hants.)

( SU

Wilson 1 975, 12, p ls. VIIIa & b ; a rchaeological potential, D evizes

7 40412)

Aerial photographs 1 973 s uggestive o f t wo s ides of a double d itched enclosure t o west o f n orth/south Roman r oad f rom Chichester t o S ilchester. Aerial photographs s how a rc of another ditch much f urther s outh b etween t he c rossroads a nd t he R . Wey. F ield work l ocated t he west 1 974 on s outhern defences of

s ide o f t he enclosure.

e nclosure.

Excavation

Roman s ettlement a round c rossroads o f n orth/south Postulated e ast/west r oad f rom Winchester t o Ewell. e xtends as f ar s outh a s R . Wey. R ectangular e nclosure of c rossroads a stride Chichester t o S ilchester r oad. Boundary

road a nd Settlement l ies n orth

d itch

E xcavations n orth o f R . Wey 1 974 f ound a f ence a nd a s hallow ditch 1 30 f t.(40m.) west of t he n orth/south r oad and p arallel to i t. I t was s uggested t hat t his may h ave f ormed t he b oundary o f t he s ettlement until t he 4 th c entury when encroached u pon. E arly

d efensive A possible

c ircuit 1 st

c entury

d itch

west o f t he n orth/south r oad i n and t he R . Wey. D ating evidence E arthwork

f ound

s outhern l ocated a s

c auseway i n 1 974.

No d irect l ikely f or

e ntrance Bank n ot

dating evidence t he d itches had

Aerial , p hotographs s how t he a rc t he R . Wey and i t was s uggested t hat defensive c ircuit. B ritannia

( 1975)

6 ,

2 15,

( Sept. 1 975) 5 2, 1 34-6; Hants.F.C. f orthcoming .

P enkridqe

i n

1 974

a rea b etween s tated.

6 5

t he

f t.(20m.) c rossroads

d efences

T he enclosure d itches. r egarded c entury.

was t he n ot

P ennocrucium

2 78

o f t he f ound b ut

b ut l ate 2 nd c entury d ate b een f illed by t he mid-3rd o f a nother t his might

& f ig.

Wilson

( Staffs.) 1 28

double-ditched excavation o f

3 ;

d itch north o f i ndicate a l ater

Current

1 975,

1 4

& p 1.

( SJ

9 03106)

A rchaeoloqy X IVa;

P roc.

Aerial p hotographs. T rench a cross d itches d itches o f s outhern d efences n ear s outh-east

o n n orth a nd c orner 1 948.

Small r ectangular e nclosure a stride Watling S treet. known a bout r oadside s ettlement. D eserted s ite. E arthwork

defences

o r walled

c ircuit

V ery

JRS s ee

s o w ide

a s

would

b e

e xpected

A n arrow d itch a nd 5 o f a wall

( pers.comm.

( 1949) 3 9, 1 02; T BAS 1 956 ( 1958) 7 4, 2 -3 & f ig.3. a lso Webster 1 971, 3 8-9 & 4 2; TBAS ( 1974) 8 6, 5 5-7

R ocester

Veratinum

?

( Staffs.)

( SK

l ittle

( ?)

T riple-ditch s ystem. R ampart p loughed out. v ertical s ided t rench 3 f t.(0.9m.) f rom u p o f i nner f t. ( 1.5m.) wide may r epresent t he r obber t rench a lthough i t i s n ot Gould). No dating.

a cross

J .

& f ig.

2 .

1 11395)

Survival as e arthwork on east a nd n orth-west. at n orth-west c orner a nd small t rench o n e ast.

Two

t renches

1 961

R ectangular enclosure. L ine o f s outhern d efences n ot k nown. The k nown a rea o f t he e nclosure l ies n orth o f t he R oman r oad f rom L ittlechester t o M iddlewich and t o t he e ast o f t he R . Churnet. L ittle k nown a bout t he s ettlement a nd c ivilian s tatus i s n ot certain. E arthwork C lay consisted

defences bank. V ery l ittle s urvived of a small amount o f pottery

a nd t he d ating e vidence f rom t he r ampart; t he

l atest s amian s herds w ere T rajanic o r e arly H adrianic a nd a small piece o f colour c oated ware ' which c ould hardly b e e arlier t han AD 1 60'. Walled

c ircuit

C lay and c obble f oundations s urvived b ut t he w all i tself completely r obbed. R elationship b etween wall a nd b ank u nclear b ecause o f r obbing b ut i t was s uggested t hat t he w all w as a l ater addition, i ndications b eing a s and p it post-dated t he r ampart a nd i t w as t hought t hat i t may h ave b een dug t o s upply s and f or t he building of t he wall. The p it w as c ut i nto a l ayer which o verlay t he t ail of t he r ampart. This l ayer w as p erhaps d eposited a s a r esult of t he weathering o f t he b ank a nd t his l ayer p roduced a coin o f Tetricus L ( Ab 2 70-274). I t was t herefore s uggested t hat t his m ight p rovide a t erminus p ost q uem f or t he b uilding o f t he wall. The d eliberate f ill o f t he s and p it a lso c ontained a c oin of T etricus I a nd s ome 4 th c entury s herds. NSJFS

( 1961)

R ochester

1 ,

3 7-8,

Durobrivae

4 1-3

f igs.

( Kent)

1 & 2

( TQ

1 2

7 44694)

Survival of walled circuit to a considerable height in some places. Section across south-east corner 1961, south-west defences 1962, north-east defenc�s 1961, 1967 and 1975, various sections across east defences 1960-1971. Site at mouth of walled area may earthwork circuit medieval defences

R. Medway on London to Canterbury Road. The follow a different alignment from that of the on the north-east. Walled circuit followed by forming irregular penannular enclosure.

Earthwork defences Clay rampart set on a foundation layer of earth and flint; V-shaped ditch. Latest pottery from rampart and its foundation dated AD 170-190. The rampart was not located on the north-east, but in the two sections on the north-east in 1961 the possible remains of the flint foundation for the rampart and the ditch were found beneath the wall. Walled circuit Rampart cut back for insertion of wall and the resulting rampart debris deposited in the ditch which accompanied the initial rampart. Wall constructed on concrete foundation resting on a foundation trench filled with rammed flint and brick earth. The initial rampart encased in rampart constructed after the building of the wall. Material of AD 170-190 from various contexts associated with the wall's construction on the north­ east. In 1976, a section across the south-west defences produced a fairly well-worn coin of Julia Domna (AD 193-211) in a context apparently associated with the construction of the wall. In the excavations in the vicinity of the supposed east gate in i969, a layer of trampled subsoil some distance behind the wall and possibly associated with its const�uction, yielded pottery of the 3rd century. In this same excavation, there was evidence that the wall had collapsed and been rebuilt. The bank had been cut back in order to repair the wall and the repair trench contained material of the first half of the 3rd century. 1963 (1964) 78, liv; 1967 Arch. Cant. 1962 (1963) 77, xlix-1; (1968) 82, lviii-lix; 1968 (1969) 83, 55-104; 1972 (1973) JU, (1976) ?, 377. 121-57; Britannia (1970) l-, 304-5; See also Arch. Cant. -1961 (1962) 76, lxxiii-lxxiv; Wacher 1964, 103n9, 104; Johnson 1976, 18-20. Shoebury Camp

(Essex)

( c • TQ 9 5 6 8 6 2 )

Visible remnants of a small rectangular earthwork similar to No excavation Uphall Camp but now half destroyed by the sea. except in 19th century? Prehistoric and Roman finds from the site. unknown. 130

Nature of settlement

E arthwork

defences

A s ection of t he western d itch s eparate periods of c onstruction. R odwell

S lack

1 975,

9 2,

f igs.

C amulodunum

E xcavation

i n

1 890

s howed

t wo

n orth-east

f rom

1 & 4 .

( W.

R iding/Yorks.)

( SE

0 84176)

1 969.

Defended vicus ( ?) Manchester t o Y ork. E arthwork

published

n orth

of

t he

f ort

o n

t he

r oad

defences

O utline o f v icus defences t raced; b ank a nd s ingle d itch c onstructed n ot b efore c . AD 1 20 b ut p robably b efore c . AD 1 40 w hen f ort a nd v icus abandoned. No f urther i nformation r e t he n ature o f the dating e vidence. S ee YAJ f or 1 967-1970 ( 1971) 4 2, 8 3-7 & f ig. 1 , where defences a re r egarded a s b elonging t o an annexe t o t he r ather t han a v icus, a nd where t here a re i ndications o ccupation a fter c . AD B ritannia

may 1 40.

( 1970)

have

1 ,

T horpe by -Newark Aerial

continued

i n

t he

a rea

of

t he

t hese f ort t hat

v icus/annexe

2 81.

Ad

photographs

P ontem s how

( Notts.)

two

( SK

r ectangular

7 60504) enclosures.

On

t he

n orth-east and s outh-east s ides, t he w all e ncloses a l arger a rea t han t he earthwork c ircuit b ut cuts a cross i t o n t he s outh-west. E xcavations on n orth-west d efence 1 952; d itches o n s outh-east 1 955 a nd north-east 1 960. I rregular r ectangular a rea e nclosed a long t he F osse Way b etween L incoln and L eicester. R oman s ettlement s traddles t he R oman r oad b eside t he R . T rent. D eserted s ite. E arthwork

1 952;

defences

I nitial e arth r ampart c onsisting o f t erminus post quem of AD 1 40-160

b lack c lay p rovided by

i dentified two s herds

s ealed b eneath r ampart; h eightening o f b ank w ith g ravel a nd c onstruction of a r ear r evetment wall and a t imber t ower g ateway. The t ower o r gateway c ut t hrough t he b lack c lay a nd a ssociated post q uem 3 rd

t he o r w as

with t he g ravel r ampart. The pottery gave a t erminus o f AD 1 45-165 f or t his h eightening a lthough a n e arly

c entury

h eightening c onstruction

was

t hought

was r egarded of t he wall, b ut

r ampart had b een c ut b ack p hases o f p re-wall r ampart

possible. a s i t

O riginally

t his

g ravel

b eing c ontemporary w ith was s tated t hat t he g ravel o f

f or t he i nsertion o f t he a ctivity may t herefore b e

1 31

t he t he

w all. Two i dentified

but b etter r elationship

e vidence i s n eeded t o b etween t hese t wo f eatures.

e stablish

t he

p recise

A s ubsequent r efurbishing o f t he r ampart w ith f urther a dditions t o t he b ank a nd t he construction o fa n ew r evetment, a ttributed t o t he l ate 3 rd o r e arly 4 th c entury, may b e a ssociated w ith t he b uilding o f t he wall. Walled

been

c ircuit

I n c ut

1 952 b ack

i t w as f ound t hat t he r ampart o n t he n orth-west h ad f or t he i nsertion o f a wall b ut t he r elationship of

t he l atter t o t he various p hases o f r ampart a ctivity w as uncertain and n ot dating evidence w as r ecovered. E xcavations i n 1 955 a nd 1 960 e xamined t he d itch s ystem a nd s uggested a 4 th c entury date f or t he walled c ircuit. I nterim o nly a vailable a nd n o d etailed d iscussion o f t he d ating e vidence. T rans. Thoroton S oc. 1 965 ( 1966) 6 9, 1 9-39; JRS ( 1961) 5 1, 1 77. S ee a lso JRS ( 1964) 5A, 1 59 & f ig. 1 2; ( 1966) 5 6, 2 03 & f ig.10; Wilson 1 975, 1 2; F rere 1 978, 2 89.

T iddir igton

( Warks.)

E xcavation o f t he a cross t he d itch. S outh-eastern

( SP

2 16555)

s outh-east

c orner

o f

c orner

1 980-1981

p ossible

i nvolved

d efensive

4 s ections

e nclosure

o f

postulated r oughly r ectangular s hape i dentified during e xcavation. Roman s ettlement o n o r c lose t o a r oad r unning a long t he s outh b ank o f t he R . Avon t o e ast o f Fosse Way. T he r oad b isects t he p ostulated o utline o f t he e nclosure. Deserted s ite u ntil r edevelopment i n 1 920s a nd 1 930s r esuming a gain i n 1 980s. E arthwork

d efences?

L arge d itch w ith s lack V -shaped p rofile a nd i rregular d imensions v arying f rom 1 4.5ft.-26ft.(4.5m.-8m. ,) a nd 3 ft.6 ft.(l.lm.-1.7m.) d eep. No t race o f a n i nternal b ank b ut i t w as t hought t his would have b een r emoved b y l ater c ultivation. T he d itch c ut v irtually a ll o ther f eatures i t e ncountered i ncluding a P it c ontaining early/mid 4 th c entury p ottery. The f ill of t he d itch c ontained m id/late 4 th c entury p ottery. The s ettlement appears t o h ave u ndergone a major r earrangement at t his t ime. T he d itch c ut a cross a r oadway r unning o ut o f t he s ettlement t o t he e ast a nd w ithin t he e nclosure t he l ine o f t his r oad w as o bscured by n ew b uildings. WMNS pers.

( 1979) 2 2, 4 5; c omm. N . P almer

T owcester E xcavation

L actodorum 1 954

a t

( 1980) 1 982.

2 3,

( Northants.) n orth-east

1 32

1 21;

( SP c orner

( 1981)

2 4

f orthcoming;

6 97483) a nd

o bservation

o f

contractor's

t rench

on

s outh-west

1 968.

Roman s ettlement a long Watling Street Fenny S tratford a nd s outh o f R . Tove. Earthwork

R esults

u npublished.

b etween

Bannaventa

a nd

defences

R ampart a nd d itch c onstructed o ver a n umber o f 1 st a nd e arly 2 nd c entury b uildings a b n orth-east c orner. D ating e vidence n ot s tated but a l ate 2 nd c entury date s uggested. Walled

circuit

S tone wall r obbed o n s outh-west, d itch p rofile v isible 1 968. Stone wall a dded t o existing r ampart on n orth-east, b ut p recise r elationship n ot s tated. Wall a pparently f ell i nto d itch on north-east i n t he 4 th c entury o r l ater a nd was n ot r epaired , a lthough o ccupation o f t he s ite c ontinued. JRS ( 1955) 4 5, 1 35; ( 1969) 5 9, 2 19; B ulletin o f t he Northamptonshire F ederation o f A rchaeoloqical S ocieties ( November 1 967) 2 , 1 9-20; Northamptonshire A rchaeology f orthcoming. A published s ection i s t o b e f ound i n J . A lexander, T he D irectinq of

A rchaeoloqical

Cave's

E xcavations,

I nn/Churchover

L ondon

T ripontium

1 970,

f ig.

( Warks.)

( SP

6 2.

5 35795)

R escue excavations 1 961-1962 o n n orth-eastern d efences a nd northern entrance l ocated. 1 961-1962 and 1 966-1968 e xcavation of t he w estern d efences. 1 970 s outhern e ntrance l ocated during r oad widening. Roman

s ettlement

a long

Watling

Street

b etween

B annaventa

Venonae. enclosing ' mansio' .

Small r ectangular e nclosure b i sected by t he p art o f t he s ettlement t o t he n orth o f t he L argely d eserted.

Earthwork

defences

were

a nd

r oad a nd s o-called

N o t race o f w all. T races o f a c lay b ank i nside a w ide d itch f ound but i t appears t hat t he r ampart h as n ot s urvived and

e xcavation was t herefore c onfined t o t he d itch. O nly o ne d itch i dentified. The most u seful i nformation was obtained f rom t he e xcavations o n t he w est. H ere w ere i ndications t hat t he d itch had n ot r emained o pen f or l ong a nd h ad b een d eliberately f illed. Pottery a t a ll l evels o f t he f ill w as e arly 4 th c entury i n d ate and t he upper f ill c ontained i n a ddition a c oin o f V alentinian ( AD 3 64-375) a nd t wo c oins o f t he H ouse o f C onstantine ( AD 3 243 41). This u pper f ill w as t hought t o r epresent s ilting o ver t he deliberate f ill. A p it c ontaining 4 th c entury w ares h ad b een c ut by t he ditch. TBAS 1 966-1967 ( 1969) 8 3, 1 30-79; 1 971-1973 ( 1973) £.‚ 9 3-144 f igs. 1 , 2 , 4 & 8 ; WMNS ( 1962) . ‚ 7 ; ( 1965) E , 1 3-14; ( 1966) 9 , 1 2; JRS ( 1967) 5 j, 1 86; B ritannia ( 1971) 2 , 2 63 S ee a lso Webster 1 971, 3 9-47 & f ig. 2 2; TBAS ( 1974) a 6, 5 5-7 & f ig.2.

1 33

Uphall

C amp

I lford

( Essex)

( c.

TQ

4 4

8 5

? )

Survival of e arthwork enclosure o n e ast b ank o f R . t o t he T hames. An e ntrance i ndicated o n n orth a nd

R oding s outh.

R omano-British t he s ettlement

t he

R odwell

1 975,

f inds have i s u nknown. 9 2,

f igs.

R edhill/Oakenqates

come

f rom

t he

s ite

n ature

o f

1 & 4 .

U xacona

( Salop.)

Aerial photographs. L imited 1 972 d itch o n n orth e xamined.

( SJ

7 28109)

e xcavations 1 959-1961 R esults u npublished.

Small r ectangular e nclosure a stride Watling Wroxeter a nd P ennocrucium. Nothing k nown a bout D eserted s ite. Walled

b ut

c lose

o n west

a nd

Street b etween c ivil s ettlement.

c ircuit

I n 1 959-1961 e xcavations o n t he w est l ocated t he h eavily r obbed faundations of a wall a nd a gate t ower. The wall w as t raced f or s ome d istance, t he b utt e nd o f a s ingle d itch l ocated. The r emains o f a r oad, presumably Watling S treet, were a lso f ound t o t he west a nd n arrowing a s i t a pproached t he gateway. T here was no s tratified material associated w ith t he construction o f t he w all a nd g ate-tower a nd f inds w ere f ew. 2 nd c entury pottery was f ound amongst t he f oundation s tones a nd c oarse w ares i n t he f ill o ver t he f oundations. S tonework i n f ront o f t he g ateway s uggested t hat i t had p robably b een partly b locked at s ome s tage. T he d itch was e xamined o n t he n orth i n 1 972 a nd ' dated b y p ottery f inds t o t he f irst part o f t he 4 th century' . A c oin h oard i n t he u pper d itch f illing was e xpected t o i ndicate t hat t he enclosure had c eased t o f unction b y t he m id 4 th c entury. WMNS ( 1959) 2 , 4 ; ( 1961) 1 6,

7 ; 2 ;

( 1960) ( 1962)



1 8,

2; a lso Webster 1 971, 3 9 & 4 ( 1974) 8 6, 5 5- 7 & f ig. 2 .

Wanborouqh S outhern d iscovered

D urocornovium

T BAS

( 1 73) 1. 5 , B ritannia f or

( Wilts.)

d efences o f p reviously during e xcavation 1 976.

R oman s ettlement G loucester. E arthwork

9 ; 3 ;

at

r oad

j unction

2 1; Sa ( 1974)

1 962-1963

( SU



( 1960) 4 27-8

( 1964)

5 7

,

1 3, S ee

1 32-3;

1 95852)

u nknown

o n

N L

r oad

d efensive

f rom

c ircuit

S ilchester

t o

d efences

E xcavations 1 30ft.(40m.) w est a t r ight a ngles

o n of t o

t he w est s ide o f m ajor Roman r oad c . t he r oad j unction l ocated a defensive d itch t he r oad a nd d ated t o t he l ate 2 nd o r e arly

1 34

3 rd c entury. f ound n orth o f Walled

D ating e vidence t his d itch.

n ot

s tated.

A c lay

r ampart

w as

c ircuit

P resence o f w all i ndicated by r obber t rench b etween r ampart and a c obbled apron w hich had b een i nserted i nto t he d itch during t he 3 rd c entury. This a pron s uggested a g ate p osition o n t he s outh-east s ide of t he defences. B ritannia

Water

( 1977)

N ewton

8 ,

4 16

D urobrivae

& f ig.

2 7.

( Cambs.)

( TL

1 22968)

Aerial photographs i ndicate a h exagonal e nclosure, bastions and gates o n t he s outh a nd s outh-west. s ections near s outh-west g ate.

small s quare 1 956-1958, 4

R ibbon development a long E rmine S treet w ith s ettlement e xtending away f rom t he r oad t o t he n orth-west a nd s outh-west. S ettlement covers a v ast a rea. The e nclosure c rosses t he R oman r oad. D eserted s ite. Walled

c ircuit

I n s ection n ear c entre o f s outhern p art o f w alled c ircuit, wall a nd c lay r ampart f ound t o be c ontemporary. Rampart deposited a fter e rection o f wall b ase. P ottery u nder t he r ampart i s s aid t o have d ated t he construction t o t he 2 nd c entury o r after. D etails n ot g iven. R emains o f w alling a nd p ebbled l evels suggested a gate at a k ink i n t he s outh-west wall. JRS

( 1958)

4 8

,

1 39.

S ee a lso St J oseph 1 966,

Whitchurch

Mediolanum

2 5;

Wilson

( Salop.)

( SJ

1 975,

9 -19

& p 1.

h a.

5 41417)

Possible s urvival i n s treet p lan. E xcavations i n 1 977 north l ocated a h itherto u nsuspected defensive c ircuit.

o n

t he

Roman s ettlement a long R oman r oad t o Chester f rom W roxeter a nd P ennocrucium. E xtensive r ibbon development. Shape a nd e xtent o f defences u nknown, b ut t he t opography s uggests t he d efences may f ollow t he l ine o f t he small detached h ill on which t he medieval and m odern n uclei a re s ituated. Earthwork

defences

E xcavations i n 1 977 o n t he n orth ( Bargates) r evealed a c lay bank. A t erminus p ost quem of c . AD 1 70 f or t he c onstruction o f t his r ampart w as p rovided b y p ottery i n a d emolition d eposit which h ad b een u sed t o b ackfill d itches a nd gullies i n t his a rea. This d emolition s eemed t o buildings a long t he l ine c learance a nd l evelling w as

i ndicate of t he f ound o n

1 35

t he d estruction o f t imber d efences. S imilar d rastic a s ite t o t he s outh ( between

H igh Street and P epper Street), which was e xcavated i n 1 977, i mplying a s imilar s equence t o t hat o n t he n orth. T he defences t hemselves were not f ound but i t was t hought t hat t he s ealing of t he s ite w ith l ayers of c obbling a nd c lay might r epresent t he t ail t han

of t he

Walled

t he r ampart. P ottery Antonine p eriod .

s ealed

by

t his

c lay

t hat

a s tone

not

l ater

c ircuit i t

w as

f ound

had been a dded t o t he f ront r elationship of wall t o r ampart r obbed. No dating evidence.

On

of n ot

t he c lay i ndicated.

WMNS

w as

t he

( 1977)

Wickford Evidence

n orth

2 0,

1 977,

6 2-4;

( Essex) u nknown

Roman s ettlement o f a r ectilinear Earthwork

i n

B ritannia

( TQ but

( 1978 ) 9,

r evetment

rampart. The w all

P recise was m uch

4 37.

7 62938)

e xcavation

p resumably

on minor r oad s outh o f e nclosure i dentified.

i nvolved.

R . C rouch. T hree D eserted s ite(?).

s ides

defences

A small part o f t he s ettlement was e nclosed by a n earthwork which was begun i n t he l ate 2 nd century and never c ompleted. D etails of dating e vidence n ot k nown. Rodwell

Wixoe

1 975,

88,

( Essex)

9 2-3,

( TL

f igs.

2 & 4 .

7 10430)

A r ectilinear e nclosure was p lanned i n 1 801. evidence i s u nknown and apparently n o e xcavation. Roman s ettlement on b oth s ides Cambridge and Colchester. E arthwork

o f

R .

S tour

o n

Otherwise

R oman

r oad

b etween

defences

Undated s ettlement i n

r ectilinear e nclosure i n s outh-western a rea angle b etween R . Stour and a t ributary.

R odwell

9 2,

1 975,

Worcester The

t he

Vertis?

n orthern

north-west Copenhagen Street) i n

f igs.

c ircuit ( Little S treet) 1 965.

t he

a long

t he

2 & 4 .

( He. o f

o f

& Worcs.) a major

F ish 1 964, Only

( SO

R oman

8 50550)

d itch

w as

t raced

Street) 1 957 a nd ( Technical College, a nd o n t he n orth-east a nd e ast ( Lich t he d itch was l ocated; n o evidence

survived f or a r ampart o r w all. under very d ifficult conditions.

1 36

The i nformation T he e xcavation

was r ecovered of 1 964 remains

u npublished. Roman s ettlement w ith e xtensive s uburbs w hich w ere i nvolved i n i ronworking i n t he l ater R oman p eriod. Character o f t he s ettlement w ithin t he defences i s u nknown. Major r oute c entre and c rossing-point o f t he R . S evern. I t was p ostulated i n 1 965 t hat t he defences e nclosed a r oughly s emi-circular a rea, a lthough t he m ajor d itch w as n ot l ocated s outh o f t he j unction o f F riar Street and S idbury s o t hat t he l ine o f t he d efences o n t he s outh i s u ncertain. Early d efensive

c ircuit?

I n 1 965 o n t he L ich S treet s ite, a small V-shaped d itch w as f ound. I ts d imensions were n ot e stablished b ecause i t w as cut by another ditch. I t h ad b een c ut i nto t he s carp o f t he I ron A ge defences and apparently f ollowed t heir l ine. I ts c orresponding r ampart, which h ad p resumably b een p iled o n t op o f t he I ron A ge rampart, h ad d isappeared during t he l evelling of t he s ite i n t he post-Roman p eriod. This d itch h ad a d eep r apid s ilt o f v ery c lean s and a nd g ravel w hich c ontained a s amian s herd o f t he e arly F lavian period ( c. AD 6 0-70) n ear t he b ottom. The upper f ill o f this d itch c onsisted o f l ayers o f m ixed c lay a nd g ravel containing a f ew r ather characterless s herds o f pottery p robably of t he 2 nd c entury. The i nterpretation o ffered w as t hat t his probably r epresented t he r emodelling o f t he I ron A ge d efences i n t he s econd h alf o f t he 1 st c entury. E xcavations i n 1 977 i n S idbury j ust s outh o f t he F riar Street j unction d iscovered a f lat-bottomed d itch r unning north/south a nd b utt e nding o n t he n orth-west. L ater r oadways appear to b e a ligned on t his corner r einforcing t he i dea of a n early e ntrance h ere. The f ull d imensions o f t he d itch c ould n ot be e stablished b ut i t was 5 ft.(1.5m.) deep and at l east 6 ft. ( 1.8m.) w ide. F our s herds o f 1 st c entury d ate c ame f rom t he p rimary s ilt. I t w as t hought t his m ight b e part o f a 1 st c entury defence. I t might b e r epresent military E arthwork

t hought t hat a ctivity.

t hese

d itches

a re

m ore

l ikely

t o

defences,

T he only k nown e lement c onsists s imply o f a f lat b ottomed ditch. On t he L ich S treet s ite i n 1 965, t he dimensions o f t his ditch were c . 9 0ft.(27.4m.) w ide, 2 5ft.(7.6m.) d eep. I t w as observed i n 1 965 o ver a d istance o f 4 50ft.(137m.) f rom F riar Street i n t he s outh t o P ump S treet o n t he n orth-east a nd h ad b een water f illed. T his massive d itch h as b een e quated w ith t he d itch f ound i n 1 957 a nd 1 964. T he main s ilting o f t his d itch c ontained a l ate s herd, p robably 4 th c entury, b ut t he l evelling o f t he d itch h ad not t aken p lace i n o ne operation. O n t he century was

L ich S treet s ite p ottery o f t he f ound i n t he p rimary s ilting o f 1 37

l ate 2 nd o r t he d itch.

e arly

3 rd

I t was observed t hat t he d itch c urved of P ump Street a nd H igh Street b ecoming n arrower s uggesting a n e ntrance.

t owards t he junction a lso s hallower a nd

Outside t he p erimeter o f t his d itch o n t he L ich S treet t here was another d itch s hallow a nd V-shaped and w ith a s ilt c ontaining 2 nd-3rd c entury pottery. T his d itch r an s traight l ine a cross t he s ite a nd d id n ot c urve. s ignificance i s n ot k nown. Walled

s ite, r apid i n a I ts

c ircuit?

There i s n o e vidence t hat a walled c ircuit a ccompanied t he massive d itch i n a ny o f t he s ections b ut i t h as b een suggested t hat i t c an b e p aralleled w ith walled c ircuits i n 3 rd and 4 th c entury contexts. H owever, t he b road d itches of t he l ater,R oman p eriod t end t o b e v ery much s hallower. There i s n o definite evidence t o s uggest t hat t he d itch was a ccompanied a t any s tage by a wall. TWAS 1 958, ( 1959) n s 3 5, 6 7-70; JRS ( 1967) 5 7, 1 85; 1 970, 1 6-18, 4 8-50, f igs. 7 , 8 , 9 & 1 0, p is. 1 & 2 ; WMNS 2 0, 4 3-5- S ee a lso Webster 1 975b, 5 6.

Wycomb

( Glos.)

( SO

0 28202)

E xcavations i n 1 972 d efensive system. An e xtensive C irencester. E arthwork

B arker ( 1977)

l ocated

a d itch

w hich

may

b e

p art

R oman s ettlement b eside t he R . I sbourne No R oman r oad k nown i n t he v icinity.

d efences/walled

o f

north

a

o f

c ircuit?

E xcavations s outh-west o f t he main R oman s ettlement a nd o n t he l ine of t he Andoversford by-pass l ocated a b road d itch r unning n orth-east/south-west w ith l ate R oman p ottery a nd c oins i n t he f illing. I t was t hought t his m ight b e part o f a h itherto u nknown d efensive c ircuit. r itannia

( 1973)

4 ,

3 11

S ee

a lso

1 38

Webster

1 975b;

W ilson

1 975,13 -

GAZETTEER:

A ldborough

C IVITAS

CAPITALS

AND COLONIAE

I surium B rigantium ( Yorks.)

( SE404662)

E xcavations 1 924 a t t he n orth-east a ngle a nd n orth g ate, 1 9341 935 o n north-west a ngle bastion a nd bastion t o t he s outh o f i t; 1 937 e xcavations a t s uth-east c orner i ncluding w all, a ngle b astion and i nternal angle t ower; 1 938 n orth d efences n ear n orthe ast c orner a nd n orth g ate; 1 960s s ections t hrough t he d efences on t he e ast a nd west s ides; 1 973 a t s outh-east corner. C ivitas c apital. A lmost r ectangular e nclosure w ith f our g ates, a ll s lightly off-centre, i nternal t owers a nd b astions. E arthwork

defences

I t i s n ot c lear whether t he b ank o f s and b ehind t he w all r epresents an i ndependent earthwork phase. I t w as i nitially t hought t hat t here was n o p re-wall e arthwork, t he b ank h aving b een p iled up b ehind t he wall ( Wacher 1 964, 1 07; 1 966, 6 3; F rere 1 965, 1 37). During e xcavations i n t he 1 960s, t he observation t hat t ip l ines i n t wo s ections i ndicated a bank c entred behind t he wall a nd n ot p iled u p a gainst i t l ed t o t he s uggestion o f an i ndependent e arthwork ( Charlesworth 1 971, 1 569 ). T his i nterpretation h as b een c hallenged by B idwell o n t he g rounds t hat t he d imensions o f t he b ank a re t oo s light and t here was n o t urf l ine on i ts s urface. A lso, t he b ank s ealed t he e dge of t he wall f oundation i n t he t wo s ections r eferred t o above and a lthough Charlesworth a rgued t hat t his may h ave r esulted f rom t he collapse of t he f ront o f t he bank, B idwell considers t hat t he contemporaneity o f t he b ank a nd wall i n a s ection a cross t he n orthern d efences i s a s trong i ndication t hat t his was t he c ase e lsewhere a lso ( Bidwell 1 980, 6 0 & 6 2). D ateable a nd

l ater

material

2 nd

c entury

f rom

t he

b ank

( Charlesworth

of

s and is mainly

1 971,

o f

m id-

1 58-9).

Walled circuit T he bank o f r epresent s and b ank t he w all.

r elationship o f t he b ank t o t he w all i s u nclear. T he s and may b elong t o an e arlier earthwork phase o r t he f irst s tage i n t he c onstruction o f t he w all. This was

heightened

w ith

c lay

a s

part

o f

t he

c onstruction

o f

D ating e vidence: a v ery worn d enarius w hich c ould b e o f Julia D omna was f ound i n t he f ootings o f t he wall i n 1 9341and t wo s herds of a t ype d ated AD 1 90-280 were f ound i n t he s econdary bank c ontemporary w ith t he wall i n 1 965 ( Charlesworth 1 971, 1 602 ). T he and

f ound

i nternal t o

b e

a ngle

t ower

c ontemporary

B astions a dded n ear t he n orth-east o ccupation l evels.

a t

s outh-east

w ith

t he

t o t he wall. The c orner c ut t hrough P ottery f rom t hese

1 39

c orner

e xcavated

1 937

wall. f ootings o f t he b astion a f illed-in d itch a nd f eatures g ave a t erminus

p ost s ome

q uem i n e vidence

t he e arly 4 th c entury f or t he b astion. There w as t o suggest t hat a n ew d itch was dug f urther o ut.

YAJ ( 1959) 4 0, ( 1966) 5 6, 2 00; ( 1974) 5 , 4 166 2.

1 -77; JRS ( 1961) 1 69; ( 1965) 5, 2 04; ( 1968) ‚ 1 80; Charlesworth 1 971; B ritannia S ee a lso C order 1 956, 2 7-8; Bidwell 1 980,60 &

B rouqh-on-Humber

P etuaria

( Yorks.)

( SE

9 40265)

E xcavations on t he east and n orth-east, i ncluding b astions and e ast g ate, i n 1 933-1937; o n t he n orth-west, s outh-east, s outhw est and west, i ncluding t he b astion a nd north gate 1 958-1961. C ivitas c apital o r n aval b ase? R oughly r ectangular. Northern l imit o f f irst earthwork c ircuit n ot k nown. Walled c ircuit k nown e xcept f or s outh-west c orner where Wacher s uggests i t w as p ossibly u ndermined by a r ise i n s ea l evel and n ot r ebuilt ( Wacher 1 969, 3 4). Wall h eavily r obbed, f our b astions i dentified o n e ast, one o n n orth a nd o ne o n west s ide; a ll oblong a nd h ollow . E arthwork

c ircuits

1 ) P eriod V . E xcavations i n 1 958 a cross t he w estern defences l ocated a r ampart i n f ront of t he f ort r ampart. I t w as s uggested t hat t his c ircuit e nclosed a l arger a rea t o t he n orth a nd e ast t han t he s ubsequent d efences b ut l ess on t he w est. T erminus p ost q uem o f AD by a s herd i n post-pit of f ort f rom r ampart. 2 ) P eriod V I. Turf e ast, s outh-east, n orth l ies b eneath H igh S treet. p robably t imber.

1 25-145 f or c onstruction g ate a nd e arly H adrianic

p rovided p ottery

r ampart o n s tone f oundations l ocated o n a nd n orth-west b ut w est s ide p robably E vidence f or c ontemporary n orth g ate,

D ating material s ealed by r ampart f oundations i ncluded c oarse pottery o f l ate 2 nd o r e arly 3 rd c entury and s amian d ated AD 1 50-180; c oarse p ottery o f l ate 2 nd/early 3 rd c entury f rom b ase a nd core o f r ampart; H adrianic a nd Antonine p ottery i n l ayers c onnected w ith r ampart. Wacher p referred a l ate 2 nd r ather t han an early 3 rd c entury date , f or t his c ircuit. Subsequently t he r ampart c ollapsed o ver t he d itch o f p eriod V c ircuit and t o r emedy t his t he i nterior of t he e xisting b ank was r emoved a nd p acked w ith r ed c lay o n t he n orth-west a nd e ast o f t he n orth gate. D ate unknown and may b e c ontemporary w ith t he c onstruction o f t he wall. Walled

c ircuit

P eriod V II. On t he n orth, t he f ront o f t he p eriod V I b ank w as c ut b ack f or t he construction o f t he w all and t he g ap b etween w all a nd b ank f illed w ith c lay. The r ampart w as e xtended t o t he r ear and p robably h eightened a t t he s ame t ime. At t he n orth g ate t he w all t urned a t r ight-angles t o e ither s ide o f t he g ate c arriageway t o b uttress t he r ampart a nd t he gate j ambs p rojected

10

f rom

t hese

i nturns.

I t

o verlay

t he

g ate

o f

p eriod V I.

T erminus post q uem o f AD 2 70 f or construction o f t he wall p rovided by a c oin o f V ictorinus ( AD 2 68-270) f rom t he r ampart h eightening. This a lso p roduced coarse pottery of c . AD 2 40-320. A fter north gate h ad b een built, b ut b efore t he r ampart h ad b een h eightened, a guardroom was b ui' .t i n t he e astern, a ngle b etween t he wall a nd gate a nd was p ossibly n ot c ompleted b efore b eing r ebuilt i n period V III. D ating e vidence f or guardroom c onstruction p rovided by s herds o f p robable l ate 3 rd o r e arly 4 th c entury d ate f rom b elow t he door s tep a nd i n t he west wall. P eriod V III. R ebuilding o f guardroom. S everal worn f loor l evels w ithin this guardroom and material f rom o ccupation l ayers i ncluded coin o f C laudius I I ( AD 2 68-270) a nd p ottery w hich ' can h ardly b e l ater t han mid 4 th C ' a nd w hich w as s ealed by t he d estruction d ebris o f t he guardroom. The p enultimate s treet s urface i n t he gate p roduced c oins o f G allienus ( AD 2 58-68) a nd T etricus I ( AD 2 70-273) a nd c oarse p ottery o f t he s econd half of t he 4 th century. A ddition o f massive e xternal t owers t o n orth g ate u ndated but t hought t o b e at t he s ame t ime as t he r ebuilding o f guardroom. C onstruction o f u nusual e xternal t ower a t • w est g ate u ndated; construction o f a s ingle t ower a round p rojecting a ngle a t east gate u ndated. Addition o f b astions a lso u ndated b ut i t has b een s uggested t hat a ll o f t hese c onstruction phases a re c ontemporary. T he a bsence o f T heodosian c oins a nd s carcity o f l ate 4 th c entury p ottery i n l evels o ther t han t he t opsoil a nd r obber t renches, i ncluding t hose f or t own wall, has b een i nterpreted a s i ndicating t hat t he d efences w ere n o l onger p roperly maintained a fter t he m id-4th c entury ( Wacher 1 969, 5 0, 5 3-54) C order

a nd

R omans

7 , 1-30 ( not Wacher 1 971,

C aerwent

1 934-1938

( not

c onsulted) ; Wacher 1 66-7; 1 974, 3 95;

Venta

S ilurum

c onsulted);

( 1942)

3 rd

s er.

1 960, 1 969, 3 -4, 2 7-55 R amm 1 978, 3 9-42, 6 1-2.

S ee

a lso

( Gwent)

( ' ST

JBAA

4 67907)

E xcavation o f n orth g ate a nd s tone i nternal t ower t o e ast 1 9011 903; s outhern d efences, gate, b astions a nd i nternal t owers 1 925; s everal s ections on n orth a nd a b astion 1 930; west g ate and w all t o s outh e xamined 1 930-1933; e ast gate a nd a djacent s tretches o f defences 1 949-1950; west g ate a nd a djacent s tretches of defences i n early 1 950s; n orth-east c orner b astion 1 971; o uter d itch o utside e ast gate 1 973. C ivitas capital. R ectangular walled a rea s urviving f or whole c ircuit and t o a c onsiderable h eight. I nternal s tone t owers. Bastions, s emi-octagonal a nd h ollow, on n orth a nd s outh. E arthwork ' defences F rom e xcavations on s outhern d efences i n 1 925 a nd o n e astern d efences i n 1 949-1950, an i nitial b ank was i dentified and d ated t o c . AD 7 5-100 b y p ottery f rom t he b ank. On t he e ast, ' three s uccessive h eightenings' o f t he b ank c ontained pottery o f e arly

1 41

2 nd c entury t o t he l ater 2 nd r e-examination 5 6 nn 3 3 & 3 4;

e arly Antonine date. B ut a t erminus p ost q uem i n c entury h as b een s uggested by H artley a nd F rere on o f t he p ottery a nd s ection d rawings ( Hartley 1 966, F rere 1 965, 1 38; 1 978, 2 99 n lO)

Double d itch s ystem w hose d ate a nd r elationship t o r ampart and wall i s n ot c lear. One d itch s ection only on s outh p roducing e arly a nd l ate 4 th c entury c oins f rom s ilt o f i nner d itch. More e xtensive i nvestigations on west b ut n o dating evidence. O uter d itch o utside e ast g ate y ielded n o material l ater t han l ate 4 th century. No e vidence f or l ater b road d itch r eplacing o riginal double-ditch s ystem. Wailed

c ircuit

Wall e rected i n f ront o f e arthwork o n e ast, s outh a nd w est a nd g ap f illed w ith c lean s and w hich w as a lso u sed t o h eighten o riginal bank. On t he n orth, e ast o f t he n orth gate, t he i nitial b ank h ad b een c leared b efore t he e rection o f t he wall a nd i t w as possibly built more over t he d itch, a s on t he east ( Wacher 1 974, 3 82) I nternal s tone t owers n ot b onded i nto t he wall a t l east o n t he north but r elationship t o r ampart u ncertain. West gate s hown t o b e s tructurally l ater t han t he w all; s outh g ate possibly earlier t han t he wall; north and east gates n ot bonded i nto t he w all b ut t heir c hronological r elationship t o i t u ncertain. B locking o f n orth and s outh g ates u ndated. None o f t he e xcavations p roduced a ny r eal dating evidence. T erminus post quer n of c . AD 1 80 f or e ast gate and a djacent w all on b asis o f p ottery o f e arly-mid 2 nd c entury i n s and f illing i n gap b etween e xisting r ampart and wall a nd gate. F rom s econdary r ampart on s outhern c ircuit, c oarse p ottery o f c . A D 7 5-200 b ut Hartley suggested a t erminus post quem o f AD 2 40 f or wall on b asis o f t his material ( Hartley 1 966, 5 7). Casey h as r ecently s tudied t he e vidence f rom various ' excavations a nd concluded t hat t he c oin o f t J rbs R oma ( AD 3 30-335) f ound by Nash-Williams i n h is e xcavation of t he s outh-east corner b astion and r egarded as p roviding a date f or t hat b astion, i n f act p rovides a t erminus , post quem f or t he wall s ince i t was s ealed by a sand a nd mortar l ayer d eposited during t he c onstruction o f t he w all. This l evel was r ecognised by C asey during h is e xcavations on t he n orth d efences a nd h e h as p ointed o ut t hat i t s eems t o h ave b een u niformly deposited a round t he c ircuit o f t he walls. T his u nworn c oin o f AD 3 30-335 t herefore p rovides a t erminus p ost q uem o f AD330 f or t he wall. Casey's a nalysis o f t he i nformation f rom various e xcavations l ed h im t o c onclude t hat t he entire c ircuit i ncluding t he gates w as of 4 th c entury d ate ( Arch. ( 1930) 8 0, 2 64, 2 76; Casey 1 983). Bastions n ot b onded i nto t he w all. Casey h as a rgued f or a t erminus post quer n o f AD 3 48-349 f or t he n orth-east c orner b astion o n t he b asis o f a c oin h oard s ealed by l evels a ssociated w ith i ts c onstruction. The f oundation t rench f or the b astion w as s ealed by a c obbled f loor o n w hich t he c oin h oard w as f ound. This

f loor

mason's

was

c hippings

a working

s urface

d erived

f rom w ork

1 42

a nd o n

i t was t he

i n

turn

s ealed

s uperstructure

a nd

by t he

h oard w as t herefore deposited during t he c onstruction o f t he b astion. This i s a very c losely dateable h oard, t he i ssues i nvolved t erminating i n l ate 3 48 o r 349. I f t his d ating i s c orrect, i t would p lace t he construction o f t he bastions s hortly a fter

t he

T he b astions

building

o f

t he

wall.

i nner d itch was f illed i n w ith and a b road o uter ditch p rovided.

t he

e rection

A rch. ( 1904) 5 9 pt.l, 8 7-97; ( 1930) Q , 2 51-88; ( 1931) 8 6, 2 10-15; ( 1933) 8 8, 1 14-19; ( 1954) 1 03, f or 1 952-54 ( 1954) 1 5, 2 31-6; B ritannia ( 1972) 3 , ‚

2 23;

Casey

o f

t he

A rch. Camb. BBCS 5 4-65; 3 02; ( 1975)

( 1983).

S ee a lso Corder 1 956, 1 967, 2 49 n2; Wacher

Caistor-by-Norwich

2 5-6; Jarrett 1 965, 5 7; F rere 1 965, 2 25-6; 1 974, 3 82; H artley

yenta

I cenorum

( Norfolk)

( TG

1 965, 1 38; 1 966,56-7.

2 30035)

S ection across n orth defences 1 930 and s outh d efences, e xcavation o f s outh gate 1 934 a nd b astion a t w est gate 1 923. Civitas capital. R ectangular a rea s urviving f or most o f c ircuit. One b astion i dentified o n west, r ectangular a lternating w ith s emi-circular E arthwork

bastions

on

s outh.

defences

E xistence

o f

p re-wall

e arthwork

p hase

h as

n ot

b een

p roved.

Aerial photograph o f double-ditch s ystem o n d ifferent a lignment to s tone wall c ircuit may r epresent p re-wall earthwork c ircuit but t here a re doubts ( Wacher 1 974, 2 34). D r. A .S. E smonde C leary suggests Walled

these may b e

a l ater

f ield

I n

t he

1 930s

f rom c .

N o

AD

1 980).

Johnson dating

( 1922)

1 2,

Canterbury

o f

1 976,

b astions

of

i ncluding

t he

was

s uggested

t hat

t he

w alls

t o

w all

u ncertain

( Wacher

1 974,

9 8)

e vidence. 2 59;

( 1931)

2 1,

2 33;

s ections gates

capital.

f rom

on

widely

t he

I rregular

( 1935)

2 5,

2 13;

Wacher 1 964, 1 06, 1 12 1 976, 1 5, 3 7, 9 8 & 1 00.

Duroverr ium Cantiacorum

n umber

Civitas

i t

2 00.

1 32. See a lso F rere 1 965, 1 37; 6 2; 1 974, 2 34-6; J ohnson

A

c omm.

and a nd s outh d efences t he wall bank p iled up b ehind t he wall.

e xcavations,

R elationship 2 35-6;

JRS

( pers.

c ircuit

I n s ections a cross n orth bank w ere o f one b uild w ith t he

dated

b oundary

( Kent) s eparated

1 930s o ctagon

1 43

( TR

t o

t he

r eused

( 1961)

& 1 13;

5 1,

1 966,

1 49578)

parts

o f

t he

c ircuit

1 970s. f or

medieval

c ity

wall

a nd s urviving o n s outh, east a nd n orth-east. Near t he n orth g ate t he wall s urvives possibly t o parapet l evel. At l east o ne i nternal s tone t ower. Bastions on n orth-east a nd s outh-east. Walled

c ircuit

No e vidence a lignment.

f or

p re-wall

e arthwork

o n

e xisting

o r

o ther

Bank p iled u p b ehind constructed i n two s tages,

t he wall. O n t he n orth t he b ank b oth contemporary with the w all.

Some e vidence i n t he square i nternal t ower t o n ot k nown.

e xcavations o n t he s outh-east, f or a r ear of wall b ut r elationship to w all

D itch

s ystem d estroyed

by medieval

w as

r ecutting.

Terminus post quem f or wall construction p rovided b y coin o f AD 2 70-290 a nd p ottery f rom b eneath t he r ampart o n t he e ast. A small hoard of c oins o f t he e arly 4 th c entury was found in a p it cut t hrough t he t ail o f t he r ampart o n t he s outh-east. Bastions on n orth-east a nd s outh-east. Undated b ut possibly c ontemporary o r b onded l ater. Only o ne e xcavated; JRS ( 1961) 5 1, 1 91. JRS ( 1937) 2 7, 2 45; ( 1945) 3 5, 8 8; ( 1949) 3 9, 1 11; ( 1950) 4 0, 1 14; ( 1953) 4 3, 1 27; ( 1954) 4 4, 1 02; ( 1955) 45, 1 43; ( 1956) 90; B ritannia ( 1970) 1 , 4 6, 1 44; ( 1961) 5 1, 1 91; ( 1962) U , 1 3 04; ( 1972) 3 , 3 51; ( 1978) 2, 4 68. S ee a lso Wacher 1 974, 1 889 1.

Carlisle

Luguvalium

E xcavation e xpected o n

( Cumbria)

i n 1 973 f ailed t he n orth-west

( NY

3 956)

t o l ocate a ny d efences s ide o f t own.

where

t hey w ere

L ate c ivitas c apital? A walled c ircuit i s anticipated b ecause o f a documentary r eference t o a v isit by St. C uthbert i n A D 6 85 w hen he was s hown t he walls and various o ther f eatures suggestive o f s urvivals f rom a R oman t own. B ritannia

( 1974)

Carmarthen 1 968 and s outhern l ocated;

5 ,

4 11;

Wacher

1 974,

Moridunum Demetarum

4 06.

( Dyfed)

( SN

4 15203)

1 969 t wo s ections a cross n orth d efences; d efences i n 1 976 a nd p robable d itch o f 1 977

s ection

C ivitas c apital. i n modern s treet

a cross

Roughly p lan.

w est

s quare

1 44

section a cross east defences

d efences. e nclosure

p artially

p reserved

E arthwork

defences

On t he n orth, s outh, o nly the t ail T erminus

post

e arth a nd c lay r ampart o f t he bank s urvived.

q uem mid-late

2 nd

a nd

c entury

d itch.

f rom

n orth

On

t he

d efences.

Walled c ircuit O n t he n orth, d itch s ilted a nd was f illed p artly o ver t he b erm and partly o ver t he h eightened. Wall n ot s urviving o n s outh.

i n a nd w all b uilt d itch. Rampart

U ndated. B ritannia ( 1970) 1 , 2 70; ( 1977) 1 978, 6 4-5, 6 9-70. S ee a lso Wacher

Chelmsford CaesaromaguS E xcavation 1 972 C ivitas capital? and

e astern

o bservations E arthwork

( Essex)

8 , 3 60; ( 1978) 1 974, 3 92-3.

( TL



4 08;

James

7 08063)

o f s outhern d efences j ust west o f s outh gate. I rregular r ectangle. Northern l imit u ncertain and

western

a lignments

established

during

only.

defences

D ump r ampart t erminating w ithin t he e xcavation a rea a nd i ndications o f a s tructure, p ossibly a guardhouse a ssociated w ith the s outh gate. Two d itches l inked by a s hort s ection o f r ounded p rofile. The outer d itch was s lighter a nd i t w as s uggested t hat t his o uter ditch may have b een s econdary r epre ting an e ffort t o s trengthen t he p re-existing s ingle d itched e nclosure w hich w as never completed. E lsewhere only a s ingle d itch has b een observed. The d itch s ilt was consistent w ith o nly a s hort l ife and t he defences were deliberately l evelled and b uilt o ver. ' Material s ealed u nder t he r ampart s uggests a c onstruction date c . 1 60-175 1 . The f illing of t he d itches contained s ome burnt d ebris i ncluding much l ate Antonine s amian s uggesting t hat t he d efences were l evelled a fter a f ire ( AD 1 70-200) w hich i s observed on o ther s ites i n t he t own. The l evelling of t he d itches

i s

s uggested

Essex A rchaeology 3 02; D rury 1 975, 1 974,

a s

h aving

a nd H istory 1 65, 1 70-2;

t aken

p lace

c .

( 1972) 4 , 2 9; R odwell 1 975,

AD

2 00-220.

B ritannia ( 1973) 4 , 9 3. S ee a lso Wacher

1 99-200.

Chichester

NoviomagUS

RegnenSium

( Sussex)

( SU

8 60049)

Small s cale e xcavations and observations a t various p oints a ll a round t he c ircuit, i ncluding t he b astions, f rom 1 930s t o 1 970s. For g azetteer and s ummaries, s ee D own a nd Rule 1 971, 7 -14; D own 3 1 974, 3-5; 1 978, 2 0. C ivitas c apital. I rregular p olygonal c ircuit.

S ix

b astions

on

n orth-west,

1 45

s outh,

s outh-west

a nd

e ast.

Position defences

o f gates n ot and s urvival

E arthwork

c ertainly i dentified. Re-use of a l arge part o f c iruit.

f or

medieval

defences

E xistence o f s eparate e arthwork p hase n ot e stablished u ntil 1 959 s ection on s outh-east. H ere and i n s ubsequent e xcavations i t was n oted t hat t he b ank h ad b een c ut b ack w hen t he wall w as built.

l ate

T erminus post q uem i n l ate 2 nd century i n r ampart.

Double-ditch uncertain. Walled

s ystem

b ut

2 nd

c entury.

r elationship

Material

t o

bank

of

and

1 st -

w all

c ircuit

Until

1 959

t he

wall

a nd

b ank

were

r egarded

a s

c ontemporary.

T erminus post q uem i n l ate 2 nd/early 3 rd c entury f or b asis o f pottery of t his date f rom wall t rench on various

wall o n s ites.

S emi-circular b astions a dded t o t he wall. I nner d itch f illed f or bastions and wide f lat bottomed d itch dug p artly i nto e xisting o uter d itch. Bastions ' dated o n t he e vidence o f p revious e xcavations t o c . mid 4 th century'. S ussex Arch. Coll. ( 1962) 1 00, 8 0-92; ( 1934) 7 5, 1 07-27; 9 0, 1 64-220; ( 1957) 9 5, 1 16-45; JRS ( 1960) 5 . Q , 233-4; 5 1, 1 89; ( 1962) 1 00, 7 5-9, 8 0-92; B ritannia ( 1973) ( 1974) ‚ 4 58; ( 1978) 9 , 4 67; D own a nd Rule 1 971, Down 1 974, 5 9-71. S ee a lso Wacher 1 974, 2 49-52.

C irencester

Corinium Dobunnorum

( Gloucestershire)

( SP

( 1952) ( 1961) 4 , 3 21; 1 49-52;

0 25015)

E xcavations i n t he 1 950s-1970s a t various points a round t he c ircuit i ncluding t wo b astions o n t he n orth-east, i nternal t owers a nd gates. C ivitas c apital. R oughly o val. The c ircuit i s u ncertain o n t he n orth-west. I nternal t owers. Bastions i dentified on n orth-east, e ast and possibly t he south-west a nd s outh-east. E arthwork

defences

Rampart o f t urf, c lay a nd g ravel. D ouble-ditch l east on e ast, e lsewhere d itch s equence u ncertain.

s ystem a t

R ampart s tratigraphically l ater t han t he s tone-built n orthe ast gate and a s tone i nternal t ower on t he n orth-east circuit. Observations a t t he s outh-west gate ( Bath g ate) s howed that i t p receded t he wall b ut i t was n ot p ossible t o establish t he e xistence o f a r ampart. I t i s n ot c lear w hether a s tone t ower o n t he n orth i s a lso contemporary. On t he s outh-west t here a re i ndications o f a t imber i nternal t ower a ssociated w ith t he r ampart.

1 46

Terminus post q uem p ottery. But s ee B idwell small

t o

r epresent

a n

p rovided by 1 980, 6 2, who

i ndependent

Hadrianic a nd Antonine s uggests t he b ank i s t oo

e arthwork

p hase.

Walled c ircuit R ampart c ut back f or i nsertion o f wall a nd h eightened. S ome e vidence f or e rosion a nd a c ertain amount o f l evelling o f r ampart b efore t he wall was b uilt o n t he s outh-west. The wall t hickness varies t hroughout t he c ircuit and t he e vidence s uggests t hat t he n arrow ( 1.2m) wall was t he o riginal w all a nd t hat t he w ider s tretches r epresent s ubsequent r ebuilding a fter t he c omplete r emoval of the earlier wall. T hese p hases u ndated and v ery l ittle dating t erminus post quem f or wall c onstruction p rovided e arly

3 rd

e vidence b ut by pottery of

c entury date.

Bastions a dded t o wall. On t he n orth-east, e stablished t hat t wo b astions were built out o ver t he

e xcavation i nner d itch

a nd t he l atter p acked b eneath t he a ctual b astion. No d irect d ating. Hoard of 4 80 c oins f rom s ilt of i nner d itch outside n orth-east gate a nd i t i s a rgued t hat t his g ives a t erminus post quem

c .

gate

R epairs t o wall a nd g ate t ower t o r emedy and p revent f urther

i nner

AD

3 46

f or

d itch.

t he

bastions.

Undated

but

o n n orth s ide o f damage f rom f lood

t hought

t o

b elong

t o

a

n orth-east water i n l ate

R oman

( 1963) 4 3, 4 7, 1 88-92;

1 5-26; ( 1973)

period. Ant.J. ( 1964) 5 3,

( 1957) 3 7, 9 -18;

4 4 ,

2 00;

( 1978)

3 , 3 39; ( 1975) S ee a lso Wacher J arrett

1 965,

Colchester

2 06-15; ( 1966) 5 8,

( 1961) 4 1, 2 40-54;

4 .,

7 1-2;

6 , 2 73. 1 964, 1 09;

6 3-71; ( 1967)

B ritannia 1 965,

2 26-7;

( 1970) 1 966,

1 ,

2 27-39;

6 0;

( 1972)

1 974,

3 01-3;

5 7.

Camulodunum

( Essex)

( TL

9 96251)

E xcavations at various points a ll a round t he c ircuit 1 908-1951, i ncluding Balkerne ( west) g ate, a nd gates on n orth, s ummarised i n Hull 1 958. E xcavations 1 967 s outh o f Balkerne G ate ( St. Mary's 1 971-1973 a long t he s outhern s ection a t L ion Walk, a nd R ectory) , 1 973-1976 a t Balkerne L ane a long t he west d efences a nd t he B alkerne G ate, s ummarised i n C rummy 1 977. The i nterpretation i s b ased o n t he 1 967-1976 e xcavations. C olonia. R ectangular a rea w ith i nternal s tone i nterval b onded

t o

t he

t owers

wall.

E arly d efensive a ) of

current

c ircuits

A n orth/south d itch on t he w est s ide o f t he c olonia t he s ite of t he colonia wall was i dentified i n t he

j ust w est B alkerne

L ane e xcavations o f 1 973-1976. No t race o f a r ampart. C rummy s uggested t his may have f ormed part of an early defensive c ircuit. An e ast/west d itch w as f ound w ithin t he n orth-west p art o f

t he

s ettlement

n orth

o f

i nsula

1 47

9 6

a nd

was

t hought

t o

f orm p art

o f t his early d efensive c ircuit. The b ackfill of t he ditch was s ealed by material o f AD 6 5-120. At Balkerne Lane, t erminus post q uem o f AD 6 0 f or t he d itch a nd b ackfilled c . AD 7 5. b )

An east/west d itch on

s outhern . line o f

l ater

colonia

e xtending f urther t o w est , o f walled c ircuit a nd f illed i n c . AD 1 50. C rummy s uggested t his r epresented another e rected

c .

E arthwork

AD

wall

but

b efore c ircuit

7 5-100.

defence

S ome a ttempts h ave b een made t o r einterpret t he r esults o f earlier e xcavations t o demonstrate t hat t he earthwork b acking the wall may have e xisted, f or p art o f t he c ircuit a t l east,, a s an i ndependent earthwork p rior t o t he construction of t he wall. But t he r elationship o f t he w all a nd r ampart i n t he e xcavations o f 1 908-1951 r emains u ncertain ( Hull 1 958, 6 3; Wacher 1 964, 1 02 & 1 06; c f. B idwell 1 980, 6 0). Walled

c ircuit

I n t he s ections on t he w est ( St. Mary's R ectory) i n 1 967 and on t he s outh ( Lion Walk) 1 971-1973, t he wall was i nterposted as a f reestanding s tructure, p rovisionally a ttributed t o t he early 2 nd c entury, t o which t he r ampart was s ubsequently added. Terminus post q uem i n e arly 3 rd c entury f or a ddition o f r ampart a t St Mary's R ectory and of c . AD 1 50-175 i n s ection a t L ion Walk. B ut i n p revious e xcavations, n orth o f t he Balkern G ate, t he wall c ontemporary. Terminus p ost q uem o f i n t he r ampart i n 1 951 s ection.

i ncluding o ne 1 951 s ection and r ampart were r egarded as AD 1 40-170 f rom t he material

Balkerne Gate e xcavations i n 1 913, 1 917 a nd 1 973 i dentified t hree p eriods. C rummy s uggested t he g ate may have o riginated as a f reestanding monumental a rch w hich w as t hen a dapted f or t he west gate w ith t he building o f t he' wall i n t he early 2 nd c entury. I n t he e xcavations a t Balkerne L ane 1 973-1976, t he d itch o utside t he Balkerne G ate gateway possibly i n b ridge Hull 3 04;

s o

t hat

t he

was widened a nd t he 3 rd c entury. gateway may have

e xtended r ight No e vidence was c eased

1 958; B ritannia ( 1970) 1 , 2 90; ( 1974) 5, 4 39-42; ( 1975) ‚

( 1978) 9 , 4 49-51; C rummy S ee a lso Wacher 1 964, 1 02

Dorchester

Durnovaria

1 977. a nd 1 06;

( Dorset)

( SY

t o

( 1971) 2 63;

F rere

b e

a cross t he f ound f or a

used.

2 , 2 72; ( 1973) 4 , ( 1976) J , 3 43-44;

1 974,

2 88.

6 93904)

E xcavations a long west defences n orth o f w est gate 1 938, 1 951, a nd 1 972; s ections a cross s upposed l ine o n n orth 1 938; a long s outh d efences w est o f s outh gate i n 1 955, 1 969, 1 970 and 1 972. C ivitas c apital. I rregular p olygon. P osition o f gates a nd l arge part o f l ine on n orth and n orth-east s ide uncertain. D emolition o f w all v ery t horough but r ampart s urvives f or much o f t he c ircuit,although damaged by 1 8th c entury l evelling f or The Walks.

1 48

E arthwork

defences

1 938 r ampart l ocated o n n orth; n o d etails but e xcavators r egarded it as t he s ame as t hat f ound on t he west where t he r ampart c onsisted o f a small n ucleus o f c lay o verlain by a lternate layers of c halk , and c lay. I n t he e xcavations a cross t he s outh defences i n 1 969 t he c halk r ampart w as o f o ne p eriod o f c onstruction but a c ontinuation o f t his t rench i n 1 970 and a nother s ection i n 1 972 r evealed t he c halk r ampart w as s econdary t o a p rimary bank 4 9 f t.(15m.) wide at whose t ail was a c halk and f lint f oundation 1 0ft.(3m.) w ide s ealed by t he s ame t urf l ine a s t he b ank i tself. The chalk r ampart was r egarded a s a l ater a ddition. The 1 972 e xcavation d emonstrated t hat t he t ail o f t he r ampart had b een extended over t his f oundation at t he e nd o f t he 3 rd c entury. Wacher h as s een t he c halk a nd f lint f oundation a s a possible attempt t o c onstruct a walled c ircuit ( Wacher 1 974, 3 21) . I n 1 969, t he r emains o f a t imber s tructure, p ossibly a t ower, c ontemporary w ith t he r ampart were f ound n ear t he s outhw est

c orner.

D ating e vidence s carce. The 1 955 s ection t hrough t he s outh d efences produced a t erminus p ost quem of c . AD 1 40. I n 1 969, t he one-period r ampart o n t he s outhern d efences p roduced p ottery g iving a t erminus post quem c . AD 1 50; t he t ail o f t he r ampart w as o ut by t he s tone f oundations o f a b uilding w ith a c oin o f Commodus ( AD 1 77-192) on t he f loor. Walled

c ircuit

Wall extensively r obbed a nd w as n ot l ocated o n t he n orth i n 1 938 o r the s outh i n 1 969. The r elationship b etween wall and r ampart h ad b een d estroyed i n t he e xcavation on t he w est i n 1 938. I n t he 1 951 s ection, t he o riginal r ampart on t he w est had b een cut b ack f or t he a ddition o f t he wall. Wacher n otes t hat t he wall w as built on a r evised a lignment ( Wacher 1 974, 3 21). N o dating

e vidence.

E xcavations o n system; the i nner y ielding undated. w idening

t he w est i n 1 972 r evealed a a nd s maller was o f two phases,

d ouble d itch t he earlier

pottery o f t he 2 nd o r 3 rd c entury a nd t he l ater p hase The outer d itch a lso s howed a n e arly phase a nd a l ater associated w ith p ottery o f t he 3 rd o r 4 th c entury.

P roc. o f the Dorset Nat. l ist. & A rch. Soc. 1 938 ( 1939) 6 5; 1 969 ( 1970) 9 1, 1 81-3; JRS ( 1939) 2 9, 2 19; ( 1952) ( 1956) 4 6, 1 42; B ritannia ( 1970) 1 , 2 98; ( 1971) 2 , 2 79;

4 ,

See

9 ,

3 15. a lso Proc. o f t he Dorset Nat. l ist. & Arch. Soc. 1 967 1 83; • RCHM Dorset Vol.11 ( 1970) S outh-East, 1 04-5;

1 974,

Exeter

( 1968) Wacher

3 21.

I sca Dumnoniorum

E xcavations e ast

6 0, 5 14 2, 9 9; ( 1973)

p art

E xcavations

f rom 1 930s o f

t he

a cross

( Devon) t o

w alled t he

( SX

p resent

on

c ircuit

s outh-west

1 49

9 20925) t he

s outh,

i ncluding

corner

o f

t he

s outh-west t he

s outh

r e-used

and g ate.

l egionary

f ortress d efences i n 1 975, 1 977 a nd 1 978. Civitas c apital. The r e-used l egionary f ortress defences enclose a r ectangular a rea e ntirely w ithin t he l ater walled c ircuit. I rregular r ectangular walled a rea t wo and a half t imes l arger t han t hat o f t he l egionary f ortress. M uch o f t he wall survives a nd was r e-used f or medieval defences. E arly d efensive

c ircuit

R etention o f l egionary f ortress d efences a fter a bandonment o f f ortress c . AD 7 5. Outer l egionary f ortress ditch f illed i n c . AD 8 0. A l ater d itch c utting t hrough t he f illing o f t he t wo l egionary f ortress d itches s uggested t he d efences of t he fortress were r etained by t he t own. D ate o f t his d itch u ncertain b ut r emained open until Antonine period. Neglect o f Antonine period o f Antonine date E arthwork

t hese d efences p rior t o t heir d emolition i n t he i ndicated by domestic r efuse c ontaining material i n d itch a nd i n p it dug t hrough i ts s ide.

defences

A n ew c ircuit w as e rected o n a n e ntirely d ifferent a lignment f rom t hat of t he l egionary f ortress d efences. The bank w as c onstructed f irst a nd w as c ut f or t he i nsertion of t he wall a nd t hen h eightened, but i t i s uncertain w hether t he earth bank t o t he r ear o f t he wall r epresents a n i ndependent e arthwork p hase. I n t he e xcavations up t o and i ncluding 1 950, i t was concluded t hat t he r ampart r epresented a n i ndependent e arthwork p hase w ith a c onstruction date i n t he H adrianic o r Antonine period on t he b asis o f material o f AD 8 0-120 f rom t he r ampart. B idwell h as s uggested t hat t he bank may merely r epresent upcast f rom t he d igging o f a d efensive d itch a nd f rom t he f oundation t rench o f t he wall i n v iew o f t he s light d imension of t he bank and t he a bsence o f a t urf-line o n i ts s urface ( Bidwell 1 980, 6 0). The s outh-west t ower o f t he s oüth g ate was built b efore t he wall w hich a butted i ts n orth-east s ide. I f t he bank b elongs t o a n i ndependent earthwork, t he s outh gate may b e contemporary w ith i t. Walled

c ircuit

U ntil r ecently t he w all was r egarded a s s econdary, i ts f oundation t rench cutting t he wall and a t erminus p ost quem i n t he l ate 2 nd o r e arly 3 rd c entury b eing p rovided by w orn s herds o f t he 2 nd c entury a nd a s herd o f s amian o f l ate 2nd o r e arly 3 rd c entury date f rom a ' thin s eam o f d isturbed s oil between the c ore o f t he w all a nd t he l owest l ayer o f t he b ank w hich i s n ow r egarded as . marking t he e dge o f t he f oundation t rench ' ( Fox 1 952, 2 0-1). T he p ottery f rom r ecent a nd e arly e xcavations h as b een e xamined by B idwell who n oted t hat p ottery f rom the o ccupation l evels b elow t he r ampart a nd w all, f rom t he f illing o f t he d itch a ssociated w ith t he early defensive c ircuit, f rom t he r ampart a nd t he w all i s l argely Antonine w ith s ome H adrianic material a nd g ives

a l ate

Antonine

t erminus

post

1 50

quem

f or

both

bank

a nd

wall.

Two s eparate d itch s ytems; o ne t hought t o b e c ontemporary with t he construction o f t he wall a nd r ampart, was r ecut once a nd deliberately f illed. The other d itch i s thought t o be 4 th century

and

s ilted

u p g radually.

Fox 1 952, 1 9-21, 5 2-9; JRS ( 1962) 5 2, 1 84; B ritannia ( 1975) 6 , 2 76; ( 1976) 7 3 60; ( 1977) 8 , 4 15; ( 1978) 4 59; ( 1979) 1 0, 3 3 2-46. See a lso Wacher 1 974, 31-2; Bidwell 1 980, 2 3, 4 6-7, .



59-66.

Gloucester

Glevum

( Glos.)

( SO

8 35183)

Excavation 1 931-1932 of s outh-east d efences i ncluding angle tower; 1 934, 1 958, 1 960, 1 970-1971 o n n orth-east s ector; 1 961 t rench across defences east of s outh gate; 1 969 e xcavation o f 3 28 f t. ( lOOm.) s tretch e xtending n orth f rom e ast g ate ( 47-51 Eastgate St.) i ncluding b astion a nd a n i nternal t ower a nd observations on e ast and n orth d efences; 1 970 e xcavation o f west defences ( 13-17 B erkeley St.) s outh o f w est gate i ncluding internal t ower; small-scale e xcavation and observations 1 9711 973 on s outh-east, s outh, e ast a nd n orth defences; 1 974 excavation on n orth gate; 1 974 and 1 978 on east gate. Colonia. R ectangular defended a rea p artially r eused f or medieval defences. H eavily r obbed on w est. Stone i nternal t owers. Only one b astion c ertainly k nown n orth o f e ast gate but s ome indication o f

another

Early defensive

at

n orth-east corner.

c ircuit

P ossible r etention o f l egionary f ortress d efences. Rear o f legionary f ortress on e ast s ide c ut a way f or c onstruction o f buildings a lthough r est of r ampart t o east was l eft s tanding suggesting t hat t he o uter f ace was p ossibly s till r evetted. Early 2nd century p ottery i n t he ' o ccupation l ayer o f o ne of t he buildings. S imilar s equence on s outhern d efences w here t he l evels associated w ith a b uilding between t he l egionary f ortress r ampart and t he s ubsequent r ampart c ontained l ate 1 st t o e arly 2 nd century

pottery.

Earthwork

d efences

There i s s ome e vidence t hat t he n orth a nd e ast g ates a nd part of the walled c ircuit, on t he e ast at l east, was built ' in the early 2 nd c entury'. I f t his d ate i s c orrect t hen t his precedes t he construction of t he s econdary r ampart. On t he e astern c ircuit, t he l egionary f ortress r ampart a nd the buildings c ut i nto t he r ear o f t his r ampart were s ealed by a s econd r ampart w hich c ontained p redominantly m id 2 nd c entury b ut also some A

l ate

s tone

certainly

2 nd century pottery. i nternal

a ssociated

t ower

with

t his

n orth

o f

r ampart.

1 51

t he

e ast

I ts

g ate

w as

c onstruction

a lmost p it

cut

t hrough t he l egionary f ortress r amparts a nd Another t hree s tone i nternal t owers, one t o t he above, one a t t he s outh-east a ngle a nd a nother on may be contemporary w ith t his r ampart.

b uildings. n orth of t hat t he west s ide

A s imilar s equence o f a s econdary r ampart l egionary f ortress r ampart was r evealed i n t he 1 970s on t he w est, e ast, n orth a nd s outh d efences. Walled

s ealing t he e xcavations

c ircuit

The e ast gate a nd 1 48ft(45m.) o f w all o n e ither s ide o f t he gate were apparently built b efore t he r est of t he walled circuit. The l egionary f ortress d itch was d eliberately f illed i n t he e arly 2 nd c entury t o s upport t his wall. The n orth gate w as inserted i nto t he l egionary f ortress r ampart i n t he early 2nd c entury a nd was s imilar i n c onstruction t o t he east gate which w as regarded a s c ontemporary w ith i t. On t he e ast c ircuit, t he r ampart h ad b een cut b ack f or t he i nsertion of t he colonia wall, t he east e nd of t he stone internal t ower n orth o f t he e ast g ate was d emolished a nd r ebuilt t o j oin the colonia wall. The completion of t he colonia wall has b een a ttributed t o t he 3 rd c entury a lthough t here s eems t o b e n o dating e vidence. There i s e vidence f or t he s ubsequent r eplacement o f sections of t he colonia wall by a w ider wall. A gain t here s eems to b e no dating e vidence a lthough t he 4 th c entury h as been s uggested. Structural e vidence f rom t he e xcavations on t he east s uggests t he r ebuilding o f s ections o f t he w all m ay b e l inked w ith t he construction of bastions. The o riginal t he l ower date

i n

b astion n orth o f t he e ast g ate w as s econdary t o t he wall and c onstructed o f l arge r eused b locks s imilar t o c ourses o f t he r ebuilt w all. No d ating evidence b ut a

t he

The period'

4 th

n orth

century has g ate

was

been

' s uggested.

d emolished

' at

t he

e nd

o f

t he

R oman

TBGAS 1 931 ( 1932) 5 3, 2 67-75; 1 933 ( 1934) 5 ., 2 77-91; 1 962 ( 1963) 8 1, 1 0-40; Ant.J. ( 1972) 5 2, 2 4-37; ( 1974) 54, 9 -17; 3 3 , 39; ( 1975) B ritannia ( 1970) 1 , 2 93; ( 1971) 2 , 2 75; ( 1972) . ‚ 2 73; ( 1979) 1 0, 3 21; ( 1980) ) J . , 384. See a lso Wacher 1 974, 1 41, 1 47-9.

Leicester

Ratae

Coritanorum

( Leics.

) ( SK 5 85045)

1 952, 1 955 a nd 1 958 s ections a cross t he n orth defences; 1 966 s ection t hrough t he s outh defences; 1 950, 1 968 and 1 974 sections t hrough t he e ast d efences. C ivitas c apital. R oughly r ectangular. L ine o f defences on w est uncertain and t he gates have n ot been l ocated. The c ircuit w as f ollowed by t he m edieval w all a nd h as b een e xtensively r obbed. E arthwork

defences

1 52

T here i s n o e vidence f or a n i ndependent e arthwork p hase. I t h as b een suggested t hat an undated t imber palisade i n t he e xcavation a cross t he s outh d efences may b e c onnected w ith a p rewall d efence but i t i s n ot at a ll certain t hat i t f orms part of a d efensive c ircuit, a nd t here a re i ndications t hat t he w alled c ircuit cut a cross a reas of o ccupation on t he n orth, e ast and s outh. Walled

c ircuit

S uch e vidence a s i s a vailable s uggests t hat t he w all, t urf and g ravel r ampart a nd double-ditch system a re contemporary. I n t he e xcavations on t he n orth d efences i n 1 952 a nd 1 958 a nd o n t he east d efences i n 1 968 n o e vidence was f ound f or t he i nsertion of t he w all i nto a p re-existing r ampart. T erminus post quem i n l ate 2 nd/early 3 rd c entury on t he basis o f pottery f rom t he r ampart. I n t he 1 952 e xcavations o n t he n orth, p its under t he wall and b ank c ontained coarse ware o f t he

T rajanic

p eriod

( AD

9 8-117).

T he i nner e dge o f excavation on the n orth was

t he l ater

e xternal d itch i n r ecut f urther n orth.

t he

1 958

TLAS ( 1952) 2 8, 1 9-22; ( 1953) 2 9, 1 5-29; ( 1956) 3 2, 8 9-90; ( 1959) ‚ 7 8-80; f or 1 968-1969 ( 1970) 4 4, 3 -5; JRS ( 1959) 4 91 1 13; ( 1967) 5 7, 1 83; ( 1969) 5 9, 2 15; B ritannia ( 1975) 6 , 2 46. See a lso Todd 1 973, 3 6; Wacher 1 974, 3 51-3.

Lincoln Upper

Colonia

L indum

( Lincs.)

( SK

9 75716)

1906-1979 e xcavations at various points a long t he c ircuit especially o n n orth a nd n orth-east s ides: North d efences 1906, 1 975-1979 ( Cecil Street); 1 946 North R ow; 1 969 Temperance Place; 1 964-1966, 1 970-1975 E ast ' B ight. North g ate a nd a djacent wall 1937, 1 954 a nd 1 972. S outh defences 1 955-1958 O ld Bishop's Palace; West d efences 1 938-1946, 1 973; E ast d efences 1953, 1 971, 1 978-1979. E ast gate 1 959-1966. A ll o f t hese excavations a re r eported i n Jones 1 980 a nd t he s ummary b elow i s -

-

-

-

-

-

taken

f rom

Colonia.

t hat

R oughly

Early defensive R etention founded Walled

s ource.

c .

AD

s quare

e nclosure.

c ircuit o f

l egionary

f ortress

d efences

w hen

c olonia

8 5-95.

c ircuit

1 ) Timber f ront o f l egionary f ortress r ampart r eplaced w ith a narrow stone wall which was built s ome d istance i n f ront o f t he legionary r ampart a nd o ver a n e arlier l egionary d itch, w hich h ad been f illed i n when t he l egionary d efences were modified c . AD 71-78.

Stone

f ronts

a dded

t o

t he

1 53

gate

t owers.

Dating e vidence f rom t he i nfill b etween t he legionary r ampart and t he wall f rom various s ites on t he north defences p rovides a t erminus p ost q uem i n t he e arly 2 nd c entury but o n t he s outh and east defences a t erminus post quem i n t he Hadrianic p eriod was obtained. I t was t herefore s uggested t hat the w all b uilding may h ave b egun i n t he l ate 1 st/early 2 nd c entury a nd l asted i nto t he H adrianic period. 2 ) of wall e ast.

Stone i nternal t owers i n mid-late 2 nd c entury

3 )

R ebuilding

o f

e ast

f ound t o a t t hree

gate

i n

h ave b een p laces on

e arly

3 rd

a dded t o t he north

r ear and

c entury.

4 ) H eightening o f t he r ampart l ate 2 nd/early 3 rd century. North o f t he e ast g ate t his s tratigraphically post-dated t he r ebuilding of t he gate and p roduced material o f the l ate 2 nd/early 3 rd c entury. Four o ther s ites p roduced r ampart material of s imilar type but i t i s n ot c lear whether t his was a ll p art o f a s ingle s cheme. I t w as c oncluded t hat t he wall a nd r ampart were p robably h eightened i n t he l ate 2 nd/early 3 rd c entury. 5 ) Thickening a nd h eightening o f t he w all a t s everal p oints e ither by a dditions o r r ebuilding b ut l ittle evidence for d ate a lthough s ome p reference s hown f or l ate 3 rd/4th century. N orth wall of i nternal t ower n orth o f t he east gate demolished late 3 rd o r e arly 4 th c entury t o make w ay f or a s olid p latform extending t o n orth at r ear of colonia wall. Colyer

1 975,

1 1

a nd

1 4;

Jones

1 980. /

L incoln Lower 1 948-1950

Colonia

s ection

L indum

t hrough

( Lincs.)

r ampart

o n

( SK

w est

9 75712)

( Beaumont

F ee) ;

1 968

e xamination o f wall c lose t o 1 948-1950 s ection; 1970-1972 complete, b ut c omposite, s ection t hrough w est d efences i ncluding west gate and i nternal t ower ( The P ark); 1 971 excavation o f a s trip c . 1 6ft.(5m.) a long l ine o f w est d efences t o t he north o f The P ark ( West a reas a long t he e xcavation on

P arade/Motherby H ill) ; 1 973 e xcavation of t hree e ast d efences n ear s outh-east c orner; 1973-1974 t he e ast part o f s outhern defences i ncluding a

gate. L ower s outh f orm

Colonia/Peregrine Community. R ectangular o f upper c olor iia and a djoining i ts s outhern a n e longated r ectangle o verall. R e-used

e nclosure t o d efences t o f or medieval

defences. E arthwork

d efences

On t he west d efences a t Motherby H ill, t he wall a nd r ampart were f ound t o be contemporary. The r ampart was heaped up d uring t he b uilding o f t he wall a nd s ealed b uilding debris f rom t he c onstruction o f t he wall ( Ant. J . ( 1975), 5 5, 2 32, 2 34, 2 48 50). At

The

P ark,

i t

was

n ot

1 54

p ossible

t o

e xamine

t he

stratigraphical r elationship b etween t he w all a nd t he r ampart s ince t his had been d estroyed by l ater d efensive a ctivity but t he surviving f ront p ortion o f t he r ampart s outh o f t he w est gate s ealed a deposit o f mortar and l imestone chippingS u sed i n t he construction

o f

t he

w all.

H owever, e xcavation o f t he e ast d efences i n 1 973 i ndependent earthwork phase w ith a n earth a nd t urf

r evealed a n b ank which

appeared t o s urvive t o i ts o riginal h eight w ith e vidence f or a substantial f ence o n i ts c rest. The r ampart was e rected o ver a demolished b uilding w hich h ad b een b uilt i n t he e arly 2 nd c entury and w as of s everal phases. The r ampart was attributed t o ' the l ater 2 nd c entury ' ( Britannia ( 1974) 5 , 4 21-4). Walled circuit T he earliest f orm o f with c ontemporary r ampart,

t he d efences a t The P ark berm a nd d itch ( Ant. J .

w as a ( 1975)

wall 5, 5

2 35). bank

H owever, on t he e ast d efences i n 1 973 i t w as f ound had b een cut back f or t he i nsertion o f t he wall i n

c entury

o r

l ater

a nd

t he

r ampart

w as

t hat t he

t he 3 rd

h eightened.

I n the e xcavations on t he west d efences i n 1 970-1972, i t was c oncluded t hat t he wall a nd r ampart w ere p robably c onstructed i n t he l ate 2 nd o r early 3 rd c entury. At The P ark, a t erminus post q uer n w as p rovided b y p ottery s ealed b eneath t he r ampart. Most o f t he p ottery had a t erminal date i n t he 2 nd century but t hree s herds o f s amian c ould p ut t he f irst p hases o f t he d efences i nto t he b eginning of t he 3 rd c entury. At Motherby H ill, a t erminus p ost q uem o f AD 1 60-190 was p rovided b y a s amian s herd, s upported by c oarse p ottery, f rom l owest r ampart deposit ( Ar )t.J. ( 1975) 5 5, 2 34-5, 2 50). I n 1 948-50 e xcavation o f t he w est r ampart r evealed two p hases o f construction. The f irst s ealed p ottery down t o t he middle o f t he 2 nd c entury, t he s econd w as ' dated t o t he f irst quarter

of

t he

Various walled

The

d efences. 2 )

a ddition

a fter

f ollowed

t he

c onstruction

o f

t he

The

T erminus

t owers.

l east

t wo

i nternal

t owers

t o

t he

w est

D isuse

w est

gate

t ower

a t

o f

i nternal

t ower

p latform n earby.

p ost

t hickening

t he

o f

quem,AD

R efurbishing and

g ate

o n

t he

s ite

o f

o ne

o f

t he

s outh

g ate

w as

b uilt

U ndated.

R ebuilding

6 )

o f

Undated.

i nternal

w all.

i nternal 5 )

a t

c onstruction

The

the 4 )

o f

Undated.

i nternal 3 )

o f

modifications

circuit.

1 )

t he

3 rd century.'

t he

a t

Motherby

Terminus

w est

g ate

H ill;

p ç t g ui

w ith

AD

p rojecting

c onstruction 2 70. g ate

t owers.

3 64-375. o f

h eightening

t he o f

w all

o n

s ections

1 55

a massive o f

t he

s cale

wall.

w ith Undated

t he b ut

l ater

t han

t he

r ebuilding

o f

t he

w est

g ate.

7 ) Other u ndated defensi .ve a ctivity, a ) T races o f a doubleo r t riple-ditch s ystem d estroyed b y a w ide d itch o f t he l ater R oman period. b ) The a ddition o f e xternal gate t owers t o t he s outh g ate. P hases 4 -6 a nd t he w ide d itch o f t he l ate Roman p eriod a re r egarded a s part of t he s ame operation. There was e vidence of h aste i n t he r ebuilding o f t he w est g ate a nd r efurbishing o f t he wall. The r e-use o f masonry s eems t o h ave b een mainly f or t he w est gate a nd t he a ddition o f masonry t o t he e xternal w all f ace n ear t he s outh gate and at t he o ther p ossible w est g ate. On t he w est d efences, i nternal t owers a ppear t o h ave b een a dded t o t he r ear o f t he wall. At The Park, an i nternal t ower e xisted o n t he s ite s ubsequently o ccupied b y t he w est g ate. Another i nternal t ower was a dded t o t he wall a t Motherby H ill. No dating e vidence b ut p ossibly i n t he f irst half o f t he 3 rd c entury and n ot l ong a fter t he construction o f t he wall ( Ant. J . ( 1975) 5 5, 2 37, 2 47. 2 50-4). T he i nternal t ower a t T he Park w as d emolished a nd a s tone gateway w ith e xternal gatetowers was e rected. This w est gate w as r ebuilt w ith i nternal a nd e xternal g ate t owers. No d irect d ating e vidence but a date a fter t he mid-4th c entury w as s uggested b ecaused o f t he s tratigraphical r elationship w ith t he s trengthening of t he wall and t he n umismatic evidence f rom t he gateway. The l atest c oins i n t he e arliest r oad s urface were o f t he H ouse o f Constantine ( early m id 4 th c entury) a nd l ater s urfaces contained c oins o f t he H ouse o f Valentinian ( AD 3 643 78). However, t he r elationship b etween t hese r oad s urfaces a nd t he gateway had b een destroyed. A coin o f Valentin an I ( AD 3 643 75) e mbedded i n t he c ore o f t he n orth w all o f t he n orth g ate t ower was apparently deposited during t he c onstruction of t he gate, b ut i t may h ave b een r edeposited d uring t he 1 9th c entury. -

At

The

P ark

t he

berm

was

r esurfaced

soon

a fter

t he

c onstruction o f t he g ate t owers a nd o verlying i t w as a t hick r ubbish deposit containing c oins o f t he H ouses o f Valentinian ( AD 3 64-378) a nd Theodosius ( AD 3 78-402) a nd a l arge g roup o f l ate R oman p ottery. Along t he s outh d efences t he wall w as r efaced a nd w idened a t t he r ear i n t he l ater R oman p eriod ( Britannia ( 1974) 4 24). North o f t he w est gate. t he wall was demolished and r eplaced b y a w ider w all; t o t he s outh o f t he gate, i t w as t hickened a t t he r ear f or s ome d istance. At Motherby H ill, t he wall w as t hickened t o t he n orth o f t he i nternal t owere w here i t w as built u p a gainst t he f ace of t he t ower. A lso h eightening of t he r ampart at The P ark a s p art o f t his p rocess. No d ating e vidence b ut p ost-dated t he r ebuilding o f t he west gate ( Ant. J . ( 1975) 2 39, 2 41-3). ‚



There was s ome e vidence P arade r epresented by masonry o f t he w all. At

Motherby

H ill,

t he

f or a nother w est gateway at a dded t o t he i nner and o uter

i nternal

1 56

t ower

w as

West f ace

partly d emolished

during t he R oman p eriod a nd was p ossibly r eplaced i n f unction by a s olid i nternal s tone p latform perhaps f or a rtillery, built a gainst t he r ear f ace o f t he wall f urther s outh. The chronological r elationship b etween t his p latform and t he r efurbishing o f t he wall was n ot . established. I t w as t hought t hat t he demolition o f . t he i nternal t ower and e rection of t he p latform were p robably r elated t o t he 4 th c entury r emodelling o f t he d efences. A t erminus post quem of AD 2 70 f or t he c onstruction o f t he p latform w as p rovided b y a worn c oin o f T etricus i n t he c onstruction t rench. 9 JRS ( 1950) 4 0, 9; B ritannia ( 1971) ( 1973) 4 , 2 86; ( 1974) 5 , 4 21-4; Ant. S ee a lso Whitwell 1 970, 3 5-6; Wacher

2 , 2 57-8; ( 1972) a, 3 15; 2 J . ( 1975) ‚ 27-66. 1 974, 1 32-3; Col .yer 1 975,

1 5-26.

London

Londinium

( TQ

3 25813)

Excavations i n t he 1 950s, 1 960s a nd 1 970s o n t he w est i ncluding an i nternal t ower and b astion; i n t he 1 970s on t he east and n orth-east i ncluding b astions a nd o n t he r iverside w all o n t he s outh. P rovincial capital. I rregular polygonal c ircuit r e-used i n m edieval t imes. Survival o f i solated s tretches o f t he w all. Four i nternal t owers k nown. Bastions on west and e ast, none known on t he n orth e xcept b astion 1 1 a nd n one h as b een f ound a long t he r iverside wall. Those on t he east a nd n orth-east a re s olid and c ontain r e-used material f rom d emolished Roman buildings. Those o f t he west a re hollow a nd do n ot c ontain r eu sed m aterial apart f rom b astion 1 7 which i s a lso s olid. Bastion 1 o n the n orth b elongs geographically t o t he e astern g roup and 1 contains r e-used R oman material b ut i s h ollow a nd b uilt o n a plinth unlike t he o ther b astions o n t his s ide. / Early d efence? I n t he a bsence o f a n i ndependent e arthwork p hase o n t he l ine of t he wall, i t has b een suggested t hat t he p eculiarities i n t he s treet plan may b e t he r esult o f a n e arlier e arthwork d efence o n a different a lignment f rom t hat o f as y et, no a rchaeological e vidence

t he walled t o s upport

c ircuit. t his.

There

i s,

I n 1 977 e xcavation on t he n orth-east r evealed a d itch a t least 1 64ft.(50m.) l ong w hose f ill c ontained material o f c . AD 1 20. The wall f ollowed t he s ame a lignment l ess t han 6 .5ft.(2m.) t o t he s outh. The d itch w as n ot d efensive b ut w as t hought t o belong Wacher Maloney Walled

perhaps 1 964, 1 980,

t o

a n

earlier

13; 1 5 5.

1 974,

b oundary. 9 4;

F rere

1 965,

1 39;

1 978,

2 83;

c ircuit

T he w all w as b uilt o f K entish r agstone w ith f oundations o f c lay and f lint o r c lay a nd r agstone s et i n a f oundation t rench. Offsets on t he i nner f ace r educed t he w idth o f t he w all a bove t he f oundations. ditch

A b ank

constructed

i n

w as

built

f ront

o f

up behind t he

1 57

w all.

t he

wall

The

and

l atter

a v-shaped h as

b een

obliterated

by

t he

medieval

d itch

a t many p oints.

The C ripplegate f ort was i ncorporated i n the n orth-west c orner o f t he c ircuit by i nserting a nother w all a gainst t he i nside f ace of the f ort wall t o i ncrease i ts t hickness. T erminus p ost q uer n o f AD 1 83-184 f or t he w all p rovided b y a worn c oin o f Commodus f ound i n e xcavations i n 1 956, i n material deposited p rior t o t he t hickening o f t he w all. P ottery f rom within and b eneath t he bank i n various e xcavations r anged down to t he l ate 2 nd c entury a nd i n 1 977, o n t he n orth-east, p ottery f rom various contexts p rovided a t erminus post quem o f c . 1 80 f or t he wall.

1 967

On t he f loor o f a n was a l ayer of

i nternal r ubbish

t ower o n t he w est c ontaining early

e xcavated i n 3 rd century

material, c oin moulds a nd f our c oins r epresenting t he d eposit o f a f orger t hought t o have b een a t w ork i n t he period AD 2 15-225. I t h as b een a rgued t hat t his p rovides a t erminus post q uem o f c . AD 2 25 f or t he wall ( Merrifield 1 969, 1 19; Marsden 1 980, 126). The r iverside wall, p ostdating t he l andward w all, w as constructed on a r aft o f chalk and oak p iles and supported b y a bank t o t he r ear. Radio-carbon a nd d endrochronological a nalysis of t he oak p iles i ndicate a construction date of AD 3 50-370. H owever, t he w est e nd e xcavation i n 1 975 ( Blackfriars) showed t hat t he wall a t t he west e nd o f t he s ite d iffered f rom t hat s een e lsewhere i n i ts c onstruction a nd i n t he f act t hat i t c ontained r e-used s culptured s tone. N o d ating evidence but i t was s uggested t hat t his s ection was perhaps b uilt at a d ifferent t ime. ( At t he T ower o f L ondon i n t he s outh-east corner o f t he walled c ircuit, e xcavations i n 1 977, d iscovered a wall c ontaining r e-used s tone a nd c onstructed 1 3ft.(4m.) t o t he n orth of a nd parallel t o t he r iverside wall. 4 9ft.(15m.) w est o f t he s outheast c orner, i t t urned s outh t o j oin t he e arlier r iverside w all which r emained and o verlapped t he l ater wall c reating a possible s tronghold i n t he s outh-east c orner o f t he c ity a nd a pproached v ia a narrow corridor, on t he west. The bank associated w ith t his w all p roduced c oins, t he l atest o f V alentinian I I ( AD 3 753 92) a nd a t erminus post quer n o f AD 3 90 w as s uggested. Bastions a dded t o t he w all. L ittle d ating e vidence. H ollow bastion ( h A) w est o f C ripplegate w as e xcavated i n 1 966 and i ts c onstruction c ut t hrough a d eposit c ontaining m edieval p ottery. I n Aidgate)



t he 1 971 e xcavation on t he n orth-east t he f oundation o f b astion 6 h ad b een

( Duke's P lace, cut t hrough a

r ubbish deposit c ontaining 4 th c entury m aterial a nd o verlying t he f oundation c ontained 4 th c entury p ottery of t he H ouse o f Theodosius t erminating i n AD 3 70.

a a nd

I n 1 977, e xcavation of

t he t hat

o riginal

e xcavation bastion 4A

V-shaped

construction

o f

t he

d itch

n ear B astion i t ( Craswall) ‚ h ad

b astion.

b een Material

1 58

6 a nd i n was f ound

b ackfilled d umped

i n

p rior t he

l ayer c oins

t o

d itch

1 980 t he t he

i n

t he

stretch near bastion 6 contained coins of Constans AD 346-348 (Casey 1983). It was suggested that bastion 6 at least may have been constructed in the period c. AD" 346-375.

JRS (1957) 4 7, 220; Britan nia (19 70) .!. , 2 92; (1972) J , 33 5; (1977) §., 408; (1975) .§, 265; (1976) 1,- 347; (1978) 2., 452-3; (1979) 10, 311-13, 317; (1980) 11, 379; Ant. J. (1977) 57, 44-66; Grimes 1968, 47-56, 64-78; Merrifield 1965, 70-2, 101-Ll; 1969, 117-32; Maloney 1980 and 1983; Marsden 1980, 119-30, 16979. See also Corder 1955, 28-30; Frere 1965, 137; Hartley 1966, 57; Wacher 1974, 94�5, 99; Casey 1983. Silchester

Calleva Atrebatum

(Hants.)

(SV 640625)

Excavations on the defensive circuits in 1909, 1910 and 1938-1939 and more recently in the 1970s-1980s. Civitas capital. Two roughly polygonal enclosures - The Inner Earthwork and the Inner Defence. A third earthwork - the Outer Earthwork - is less certainly of Roman date and its precise 1imits unknown. The significance of a possible defensive ditch and later palisade 72ft. (22m.) inside the east gate of the Inner Defence is The ditch was aligned north-north-west/south-south­ uncertain. east and it contained early Flavian material in its lowest silts. In the late 1st or early 2nd century, a palisade or revetment was constructed along the southern tip but neither ditch nor revetment was in use by the mid-late 2nd century (Britannia (1981) .U, 362). Outer Earthwork 1909 and 1939 three sections across defences on south-west (Rampier Copse); 1956 four s�ctions across Primary Outer Earthwork on the west including west entrance; 1938-1939 two sections across Secondary Outer Earthwork on the west; 1952 observation of Secondary Outer Earthwork in area of its junction with Primary Outer Earthwork on north; 19?6 section south of· west entrance through Secondary Outer Earthwork; 1970s fieldwork and six sections across supposed alignments on east and south. Originally regarded as a polygonal circuit enclosing an area larger on all sides than that of the Inner Earthwork which it was regarded as superseding. On the west, this circuit involved a primary or outer alignment and a secondary reduced alignment. The circuit is marked by standing earthworks on the south-west, north-west and north-east. However, the investigations in the 1970s along the east and south parts of the supposed circuit failed to find any trace of the defensive circuit. As a result of the excavations on the south-west and western part of the circuit, it was suggested that the Salient Dyke in Rampier Copse was probably pre-Roman and subsequently enlarged for the Primary Outer Earthwork (Arch. (1947) 92, 138-40; (1969) Ant.J. (1958) 38, 114). A few sherds of Claudio­ 102, 16-18; Neronian pottery from beneath the bank represent the only dating 159

evidence f or t he P rimary Outer E arthwork. The d itch a t t he w est entrance appeared t o have s ilted b efore t wo s tages of i nfilling, t he f irst c ontaining material o f l ater 2 nd-late 3/4th c entury a nd t he upper f ill containing l ate 4 th c entury material ( Proc. Hants. F .C. ( 1958) 2 1 pt.l, 1 4 & 1 7; A rch. ( 1969) 1 02, 1 8-19, 3 9; B oon 1 974, 4 6) The S econdary Outer E arthwork a lso p roduced v ery l ittle evidence f or i ts dating and phasing but i t was r egarded a s having t wo p hases o f c onstruction i nvolving a n i nitial bank a nd d itch and s ubsequent h eightening o f t he b ank. N either phase was d ated b ut B oon t hought t hat t he P rimary Outer E arthwork was a bandoned s oon after i ts c onstruction and r eplaced a lmost immediately by t he S econdary Outer E arthwork p ossibly i n t he context o f t he Boudiccan r ebellion. The h eightening o f t he o riginal S econdary Outer E arthwork b ank was s een a s f ollowing s oon a fter i ts construction. A possible t hird phase of c onstruction w as s uggested by Boon on t he b asis o f a r e-examination o f t he Rampier Copse e xcavations o f 1 909 and 1 939. T his was r epresented by a r e-cutting o f t he d itch a nd t he a ddition o f a f lint r evetment t o t he bank. No dating e vidence but Boon s uggested i t w as possibly r elated t o t he l aying o ut o f t he s treet p lan i n t he F lavian period. R ecent work on t he d efences o f S ilchester by F ulford h as s uggested: 1 ) The S econdary Outer E arthwork on t he west may b e a p re-Roman p romontory e arthwork. The d efences i n Rampier C opse and t he two a rms of t he S econdary Outer E arthwork on t he w est f orm a p romontory e arthwork. M uch e arly R oman p ottery w as r ecovered f rom t his a rea during f ieldwalking but s ince n one h as b een f ound i n a ssociation w ith t his e arthwork o r b eneath i t, i t may b e p re-Roman ( Fulford 1 983) 2 ) The Outer E arthwork h as b een d emonstrated t o e xist o nly o n t he wes . No f irm evidence f or e xistence o f Outer E arthwork on the e ast. A ll p ossible a lignments h ave b een s uggested b y f ieldwork a nd a s eries o f s ections h ave b een c ut b ut a ll ?roved n egative apart f rom o ne on t he n orth-east where t here i s a s light earthwork. Charcoal f rom t he d itch p roduced a d ate o f AD 1 0-20. The e xtent t o w hich t he s tanding earthwork on t he n orth-east c ontinued east towards 3 t he amphitheatre was n ot r esolved ( Britannia ( 1979) Q , 31; F ulford 1 983) .

A rch. ( 1910) 6 2 p t.l, 3 17-19; 1 02, 1 -82; P roc.Hants F .C. ( 1976) 7 , 3 68-71; ( 1979) 1 0, S ee a lso Boon 1 974, 4 6; 1 983. I nner

( 1947) 9 2, 1 21-68; ( 1969) 3 26-7; 1 p t.l, 9 -21; B ritannia ( 1958) 2 3 31. Wacher 1 974, 2 57-60; Fulford

E arthwork

T renches 1 954-1955, i ncluding s outh entrance, d efences.

1 957-1958 a cross a cross n orth-west,

s outhern d efences, n orth-east and e ast

I rregular p olygon, n ot s ecurely d ated b ut a pparently p redating t he s treet g rid o f t he C ivitas c apital which overlies most o f t he e nclosure a nd a cross t he c ircuit o n t he e ast s ide. The

c onstruction

of

t he

s treet

g rid

1 60

has

b een

dated

t o

A D

9 0-120.

T he c ame

bank

f rom

the

l argely d estroyed

a nd

e xcavation o f

ditches.

t he

t he

majority

o f

t he

e vidence

D irect d ating f or t he I nner . E arthwork i s s carce a nd t he p roblems of dating h ave b een i ncreased by t he a lmost complete l evelling o f t he bank. The t renches a cross t he I nner E arthwork i n t he 1 950s p roduced p ottery of ' unromanised appearance' i n and b eneath c ircuit

t he s urviving t ip l ines o f ( Trench B , P roc. Hants F .C.

t he b ank o n t he j , pt-1, ( 1958)

n orth-east 1 3; A rch.

( 1969) 1 02, 9 ). F rom w ithin t he d itch f ill i n t his a rea, a l ayer which i s regarded a s r epresenting t he pushing of t he counterscarp b ank i nto t he d itch c ontained s amian a ssociated w ith a mass o f coarse pottery o f w hich about 2 /3 was of native character and i ncluded various G allo-Belgic p roducts. The o ccupation l evels b eneath the bank c ontained no Roman material, only native p ottery. The c onstruction o f t his c ircuit w as a ttributed b y t he e xcavator, B oon, t o b e s oon a fter AD 4 3-44 a nd i n t he c ontext o f Cogidubnus' c ontrol and p rotection o f much o f t he c entral s outh a s a c lient k ingdom ( Boon 1 974, 4 3-44; Wacher 1 974, 2 56). H e a rgued t hat t he earthwork did n ot s tand f or l ong b efore t he l evelling p rocess b egan a nd h e s uggested a date f or t his a s early a s AD 6 0 and certainly n ot l ater on t he basis of s herds which h e dated t o t he C laudio-NerOnian p eriod a t t he l atest f rom the l ower f illing o f t he ditch. But t he r elevant f illing layer i s one w hich B oon c onsidered w as f ormed by p ushing t he c ounterscarp bank i nto t he d itch t ogether w ith t he s uperficial s oil s craped up i n t he vicinity o f t his b ank. While t he l ayers which h ad a ccumulated i n t he d itch p rior t o t he deposition of t he bank material c ontained a s ubstantial amount o f p re-Flavian p ottery, there was a lso s ome material w ith a possible F lavian horizon. Although he n oted p ottery w hich c ould b e d ated u p t o r c . AD 6 0 i n t he s oil t hrown i nto t he d itch with t he c ounterscarp bank, h e decided i n f avour o f a date the d efences and he r elates the

F osse

Way

f rontier

i n

I t i s i mportant selection o f a date c . direct bearing construction o f T he

b ank

t he

AD

o f o f

4 7.

t o r ealise AD 5 0 f or

o n h is t he Outer a nd

s oon a fter AD 5 0 f or t he l evelling i ts r edundancy t o t he c onstruction

t he t he

r easoning b ehind B oon's l evelling b ecause i t has a '

a rgumen t r egarding Earthwork.

c ounterscarp w ere

t he

p ushed

d ate

i nto

f or

t he

t he

b ank

prior to t he main f illing up o f t he d itch. The l atter s eems t o have t aken p lace i n o r a fter t he T rajaniC o r e arly Hadrianic period. The dating ' is based on c oins a nd pottery f rom t he d itch f ill i n t he t renches a cross t he n orth — west a nd n orth — e ast circuit. O n t he e astern part o f ' t he c ircuit, where t he I nner Earthwork ±S o verlain by t he c entral c ore o f t he s treet g rid, one might expect t he e ffective f illing t o b e e arlier i n view o f t he F lavian date a ttributed t o t he l aying o ut o f t hese c entral streets. B ut here t oo t he upper f ill o f t he d itch contained, besides F lavian p ottery, l ater w orn p ottery a nd c oins o f Hadrianic a nd Antonine period w ith a l ater coin, a lso worn, o f Marcus Aurelius ( AD 1 71-172), s o t hat e ven on t he e ast t he circuit was f illed i n only g radually and n ot e ffectively obliterated

u ntil

t he

l ate

2 nd

c entury.

1 61

P roc. Hants F .C. ( 1958) 2 1 S ee a lso Boon 1 974, 4 4-6; I nner

pt.l, 9 -21; A rch. Wacher 1 974, 2 56.

( 1969)

1 02,

1 -82.

D efence

1 909 e xcavation a t e ast g ate; 1 909 s ections o n n orth-west, n orth, west, s outh and i n f ront of t he n orth, south and w est gates; 1 938-1939 s even s ections o n n orth; 1 974 o n s outh and e ast; 1 975 at s outh gate; 1 976 s outh-east p ostern gate; 1 978 a cross d itch s ystem o n s outh-east. R oughly p olygonal e nclosure d efining a n a rea much smaller t han t hat o f t he Outer E arthwork b ut e ncompassing v irtually t he whole of t he a rea of t he I nner E arthwork t ogether with an a dditional a rea o n t he e ast. The wall s till s tands f or much o f t he c ircuit. Earthwork

defences

Dump

r ampart

o f

c lay,

s and

a nd

g ravel

a nd

a

d ouble-ditch

system. A ll t he main gates and t he p ostern on t he s outh-east may b e c ontemporary w ith t he r ampart s ince t he f ormer a re p eculiarly p laced i n r elation t o t he wall b eing deeply r ecessed b ehind i t. The e xcavation o f t he s outh g ate i n 1 975 s howed t hat t he w all overrode t he f ootings o f t he gate a nd t he r ampart was a lso s tratigraphically l ater t han t he gate b ut p ossibly o f t he s ame phase. The n orth gate i s of s imilar p lan t o t he south gate, t he east gate s imilar t o t he w est g ate a nd t he s outh-east postern gate butts up t o t he wall. Dating e vidence c onsists o f Antonine p ottery a nd c oins Hadrian and Antoninus P ius f rom t he r ampart, a nd pott ry of c . 1 30-150 f rom b eneath t he r ampart i n t he 1 938-1939 excavations. Walled

o f AD

c ircuit

The r ampart w as c ut b ack f or t he i nsertion o f t he w all. T he r ampart was h eightened. The i nner d itch was f illed a nd one o r t wo d itches dug f urther o ut. Dating e vidence f or t he w all c omes f rom t he f illing of t he construction t rench. A date i n t he l ate 2 nd/early 3 rd century w as s uggested o n t he b asis o f p ottery f rom t he 1 938-1939 e xcavations but r e-examining t his material Hartley s uggests t hat t he wall i s u nlikely t o b e e arlier t han t he m id 3 rd c entury ( Hartley 1 966, 5 7 & - n 39) . I n 1 974, t he e xcavation on t he s outhern defences p roduced material i ndicating a c onstruction date i n t he mid-late 3 rd c entury. South-east p ostern g ate b locked a t u nknown date b ut l atest s urface w ithin t he gate s ealed c oins o f AD 3 50-360 a nd s howed v ery l ittle s ign o f w ear. The w est c arriageway o f t he s outh g ate was a lso b locked but t his may have been t he r esult o f medieval c learance ( Fulford 1 983). Arch.

( 1909)

1 23-34;

6 1

p t.2,

B ritannia

4 74-6;

( 1976)

7 ,

( 1910) 3 70-1;

1 62

6 2

p t.1, ( 1977)

3 17-32; 8 ,

( 1917)

4 18-19;

9 2,

( 1979)

1 0, S ee

3 31. a lso

F ulford

Boon

1 974,

5 3,

6 5-6,

1 00-7;

Wacher

1 974,

2 64-6;

1 983.

Verulamium

( Herts.)

E xcavations

i n

t he

( TQ

early

1 35075) 1 930s

and

i n

1 955-1961.

M unicipium. Three i rregularly s haped c ircuits o n d ifferent a lignments. No i ndication o f r iverside defences until walled c ircuit. 11 955' d itch l ies e ntirely w ithin walled c ircuit. F osse E arthwork e xtends b eyond t he walled c ircuit on t he w est w here i t s urvives a bove g round. I t i s r egarded a s u nfinished. Much o f t he walled c ircuit s urvives. Two i nternal t owers and two b astions, one o f e ach a t s outhern c orner, a nd w est o f t he s outhe ast

g ate.

E arly d efensive 1 955,

and

d efences. a nd

c ircuit

1 959

s outh

(1 1 955'

1 960 a nd

D itch)

i nvestigation

1 960

by magnetometer

d itch

o f

t raced

s outhern

o n

and

n orth-west,

eastern

s outh-east

survey.

D itch c . 9 .5 f t.(2.8m ) d eep a nd c . 1 8 f t.(5.7m ) L ittle r emained o f t he bank which on t he s outh h ad b een i nto t he ditch f orming b ands o f g ravel i n t he u pper f ill. I n

the

1 955

s ection on

t he

s outh,

t he

w ide. pushed

deliberate

f ill

b etween t he p rimary s ilt a nd t he b ank l evelling c ontained p ottery o f AD 8 5-100. I n t he two s ections o n t he e ast and n ear t he s outhern corner i n 1 960, pottery i n t he l ower l ayers o f t he d itch s uggested a s lightly l ater date f or t he abandonment c. AD 1 25. P ottery of f irst q uarter o f 2 nd c entury f rom f illing l evels a nd d itch l evelled b y d emolition o f r ampart n ear s outhern c orner. C onstruction had

g one

o ut

o f

u ndated, u se

by

but c .

i t

AD

was

1 25

concluded

a nd

t he

t hat

bank

t he

w as

d itch

l evelled

s hortly after. Ant.

J .

( 1956)

3 6,

4 -5;

( 1960)

, S ee a lso F rere 1 964a, 6 1- 3 1 974, 2 04-5; F rere 1 978, 2 99 Fosse

4 0,

, n 10. 6 5

6 9

2 -4, ;

2 1-4;

1 964b,

( 1961) 1 04

4 1,

& 1 07;

8 0-4. Wacher

E arthwork

1 932 s even s ections c ut a cross c ircuit b etween s outh c orner and n orth-west ( Chester) gate. 1 960 a t rench f ailed t o l ocate c ircuit where e xpected o n t he s Outh-east a nd a magnetometer survey f ailed t o f ind a ny t race o f i t on t he s outh-east and e ast and

i t I n

( 15m )

i s

t herefore

1 932, t he wide a nd c .

r egarded

a s

d efence w as 1 8ft(5.5 r n )

i ncomplete. f ound t o deep a nd

c onsist o f a d itch a g ravel bank w ith

c .49ft a t urf

r evetment considered a dditions.

t o t he r ear f our l ayers I t h as b een

a nd a c ounterscarp bank. Wheeler t o t he r ear o f t he bank t o be l ater s uggested t hat t he s tone g ates o n t he

north-west

and

( Chester

s outh-east

1 63

and

L ondon

gates)

a re

p robably

contemporary with the Fosse earthwork since they are of early plan and the city wall abuts them (Frere 1964a, 69-71). The Fosse Earthwork has not be traced as far as the London gate. Frere has suggested that the hollow into .which the internal tower west of the London gate subsided may be the ditch of the Fosse Earthwork but it has not been traced beyond insula XXV on the south-west where it crosses the walled· circuit (Frere 1964a, 73). The 'original bank' contained pottery of 'native or imported Belgic fabric' which was ascribed to AD 20-60 but a fragment of samian and of a 1st century Roman bowl were found beneath the rampart. The 'additions' contained pottery of AD 70-140 and the ditch silt produced late Belgic wares and Roman pottery of the mid to late 1st century. Wheeler attributed the Fosse Earthwork to the Flavian period. However, Frere argues that the 'additions' to the bank are part of the original construction. These 'additions' contain pottery down to the Antonine period provided a terminus post quern Most of this date for the construction of the Fosse Earthwork. recently Frere has suggested that this circuit was under construction at the time the Antonine fire occurred since one· of the 'additions' consisted of burnt material containing mid 2nd He concluded that the Antonine fire caused the century pottery. work on the defences to be interrupted for a while, the 'additions' representing the resumption of work (Frere 1981). Whee1er 193 6, 1, 49-5 6; Ant • J • ( 1961 ) 41, 8 2 • See also Frere 1964a, 69-71; 1964b, 108-9; 1981; Wacher 1974, 213-15.

197 8,

285-6;

Walled circuit Early 1930s three sections through the defences and excavation of two internal towefS, the London (south-east) Chester (north-west) gates and the Silchester (south-west) gate. 195 5 section across circuit on south-west; 1959 section across north-east defences; 1961 section across circuit on north. 1930s excavations discovered a chalk, clay and earth rampart contemporary with but structurally secondary to the wall, with the rampart sealing an offset in the wall and mortar from the wall construction beneath the bank (Wheeler 1936, 56-9; Ant. J. (1956) 36, 6). 1959 and 1961 sections on north-east and north, wall and rampart contemporary although structurally the wall post-dates the bank. .on the north-east the wall footings had cut through the bottom of the rampart but this was explained in terms of the initial construction of the ·core of the bank with the wall footings being inserted into layers formed by the erosion of the bank in the meantime, and the main body of the bank being deposited between the primary core and the wall and over the primary core. Again, in 1961 on the north, the lower part of the bank had been deposited before the footings were laid and in 1955 on the south-west there was a small primary mound below the main body of the rampart. In 1955, 1959 and 1961 robbing of the wall had destroyed the relationship between the wall and the main body of the rampart but in the 1961 section on the north the bank 164

o versailed

t he

f oundation

o ffset

o f

t he

wall.

I n a ll t hese e xcavations the l atest material f rom and b eneath t he bank was o f Hadrianic-Antonine date. I n 1 955 o n t he s outh-west, s ome pottery f rom the main body of t he r ampart b etween t he w all a nd p rimary mound w as n ot b eing manufactured b efore AD 170-180. F rere has suggested t hat a hoard o f c oins e nding w ith a c oin o f AD 2 27-229 f ound i n t he o riginal f loor o f t he t ower west o f t he s outh-east gate p rovided a t erminus p ost q uem o f c . AD 2 40 f or t he c onstruction o f t he walled c ircuit ( Ant. J . ( 1956) 3 6, 6 ; F rere 1 964a, 7 1; 1 978, 2 87). But more r ecently i t was r eported t hat t he b uilding f ound b eneath t he n orth d efences i n 1 961 c ontained t wo b owls dated n ot e arlier t han AD 2 65-270 which would make n onsense o f t he c . AD 2 40 date . ( Frere 1 98 1 ) T he i nternal t ower w est o f t he L ondon g ate w as o f one b uild w ith t he wall. I t s ubsided and f ell i nto decay as d id t he t ower a t t he s outhern c orner. A r epair t o t he o riginal f loor i n t he t ower n ear the L ondon gate contained a h oard o f coins ending w ith a coin o f AD 2 27-229. I n t he d ebris o f t his t ower w as a h oard o f 5 2 c oins with a t erminal date o f AD 2 73 ( Wheeler 1 936, 5 9-63) F rere h as r einterpreted t he s ection s uggesting t hat t he h oard w as deposited during t he c ollapse o f t he t ower f rom t he upper part o f t he t ower. destruction

H e o f

o f

f or

t he

c onstruction

London

and

Chester

c .

AD

2 73

T he t owers

a nd

T he

c oncluded t hat t he h oard d id n ot t he t ower but m ight p rovide a t erminus

r esembled London gate

o ne

gates

a nother

s ealed

o f

i n

t he

had

t he quem

wall. p rojecting

d esign

p ottery o f

date post

a nd

AD

s emi-circular

c onstruction.

1 25-150

a nd

a c oin

o f

T rajan ( AD 1 03); pottery o f e arly 2 nd c entury f rom t he o ccupation l ayers s ealed by t he gate and a c oin o f Hadrian ( AD 1 18-119) f rom b eneath t he o riginal r oad m etal o f t he Chester gate. Wheeler r egarded these gates a s con tem porar> T w ith t he walled c ircuit, t he b ank a t t he Chester g ate o verlapping t he i nner e nds o f t he g ate t owers ( Wheeler 1 936, 7 0). H owever, F rere a rgues t hat t hese gates p redate t he w all a nd a re p robably c ontemporary w ith t he F osse E arthwork pointing t o t he f act t hat t he gate p lans a re o f a n e arly type, a re n ot bonded w ith t he w all a nd w ere c onstructed at a d ifferent l evel f rom t he wall ( Frere 1 964a, 6 9-71) T he built a t o verrode 2 nd

Silchester

g ate

a s light angle t o a p it c ontaining

c entury

( Wheeler

1 936,

w ith

p rojecting

s quare

t he wall. The p ottery d own t o

t owers

n orth part o f t he b eginning

w as t he o f

a lso gate t he

7 1-3)

T here i s a c ertain a mount of doubt a s t o w hether t he bastions, which f ront t he two i nternal t owers, a re bonded i nto t he w all ( Wheeler 1 936, 5 9-61, f ig. 4 , p l.XXI; F rere 1 964a, 7 1-3; Wacher

1 974,

Wheeler

1 936,

( 1962) 4 2, See a lso 1 981;

2 20) 5 4-75;

1 50-3. F rere

Wacher

1 974,

Ant.

1 964a, 2 13-15,

J .

6 9-73; 2 20.

1 65

( 1956)

3 6,

1 964b,

5 -6; 1 09;

( 1960)

4 . 0 . ,

1 978,

4 -6; 2 87;

Winchester

yenta Belgarum

( Hants.)

( SV

4 82295)

E xcavations i n t he 1 950s-1980s;- C olebrook St 1 950 ( east); County Council Offices, 1 951, 1 955-1957 west); North walls 1 955 a nd 1 959; T ower St 1 960 and 1 964 ( west); Wolvesey Castle 1 960 ( southe ast); Castle Y ard 1 969-1971 ( west); 1 964 St. T homas' H all ( south); S outh gate 1 971; s outh o f s upposed east gate 1 980. C ivitas c apital. I rregular t rapezoid w ith s alients o n s outh-west and s outh-east. Re-used f or medieval d efences. One b astion o nly i dentified e ast o f s outh g ate. E arly defence

t he

Evidence f or a n e arly r ampart t urf r evetted w est a nd s outh s ides o f t he walled c ircuit.

at

t he

r ear,

o n

I n t he Council Office e xcavations 1 951, 1 955-1957, t he s tump o f a n e arly R oman bank was f ound t o r un t he f ull l ength o f t he s ite; a t Tower St i n 1 960 t here was e vidence t hat t he w all h ad c ut i nto a b ank o f c layey g ravel. I n 1 964, a t Tower S t., 3 s ections o ver a l ength o f 1 30ft. ( 40m ) p roduced a consistent s equence o f an i nitial d ump o ver t he I ron A ge l evel. D ating e vidence s carce but a f ew s herds i n t he r ampart p rovided a t erminus post quem o f AD 7 0. I n 1 969-1971, a t C astle Y ard, a l ow b ank o f c lean c lay w ith f lints and bands of chalk r ubble l ay partly o ver t he o ld g round s urface a nd p artly d irectly o ver t he l atest f ill of a p re-Roman d itch. A t urf r evetted earth wall, s een a t t he other s ites a lso, w as c onstructed o n t op. The b ank w as d evoid o f a ll f inds b ut p ottery f rom t he f inal f ill o f t he p re-Roman d itch s uggested t he b ank w as c onstructed t owards t he e nd o f t he 1 st c entury. At

t he

s outh

gate

i n

1 971,

i

was

t hought

c lear

t hat

t he

r ampart l ayers o f c lean g ravel a nd l oam h ad b een deposited f rom t he digging of a n adjacent d itch. The c onstruction w as s imilar t o t hat s een a t Castle Y ard a nd T ower S t. a nd t here w as a c lear contrast b etween t he materials o f t he f irst . r ampart a nd of t he s econd

p eriod

r ampart.

Contemporary

t imber

gate.

D ating e vidence consisted o f an a s o f Nero AD 6 4-68 i n t he f illing o f one o f t he p ost p its o f t he t imber s outh g ate. I t w as c oncluded t hat t he defences were o f l ate Neronian o r F lavian d ate ( c. AD 7 0 o r s hortly a fter). Some e vidence f or d ecay o f t his c ircuit a t Castle Y ard w here a b rown s oil developed on t he r ear s lope o f t he rampart and t he t urf s tack w as h eavily e roded. This w as i nterpreted a s a p eriod o f s oil g rowth and general decay b efore t he r econstruction i n p eriod 2 . At t he s outh g ate, t he t imber g ate was d eliberately r emoved a fter a n u nknown l ength o f t ime a nd a n ew S treet l aid a cross t he u nbarred g ap i n t he r ampart. I t w as concluded t hat t he f irst defences were s hort l ived. The

f ront

o f

t he

r ampart

defences.

1 66

h ad

b een

c ut

a way

by

t he

l ater

The

s alient

P roc.Hants ( 1970) 5 0,

on

t he

w est

F .C. ( 1962) 2 2, 2 82-3; ( 1975)

was

s hown

t o

5 7-8 & 6 0; 1 01, . 1 10

5,

b e

a p rimary

Ant.J. & 1 12;

( 1965) Wacher

f eature. 4 5, 2 35-6; 1 974, 2 77-

8 0. E arthwork

defences

O n t he e vidence f rom t he e xcavations i n t he 1 950s a nd 1 960s, i t was concluded t hat t he wall a nd rampart were contemporary. S ubsequent e xcavations o n t he w est a nd s outh s ide o f t he w alled c ircuit have s hown t hat t he i nitial r ampart was overlain by a s econd r ampart r epresenting a n i ndependent e arthwork p hase a nd i t i s t hought t hat t his may be r epresented i n t he Colebrook S t. s ection on t he east ( 1950) a lso. I n 1 980, t rial t renches i n t his s ame a rea attributed t he r ampart t o t he e arly 3 rd century. T he 1 964 e xcavation 1 00ft.(30m ) e ast o f t he s ite o f t he p resumed south gate r evealed a c onstruction t rench b ehind t he w all c utting t hrough t he l oam a nd g ravel b ank. No dating e vidence. A t Tower St. ( 1964) a nd Castle Y ard ( 1969-1971) t he b ank c onsisted of l ayers of c lay and l oam a nd o ccupation material. P ottery f rom t he r ampart a t Castle Yard p roduced a t erminus p ost quem

i n t he

Antonine

period.

A t

t he s outh g ate ( 1971), t his r ampart e ntirely c overed t he early r ampart and cut i nto s ite o f e arlier gate. The f ront o f t he b ank was s et o n a massive f oundation o f c halk a nd g ravel s et i n an e arly r ampart. This s econdary r ampart c onsisted o f l oamy s oil w ith a f ront o f t urf a nd g ravel. No g ate s tructure w as f ound b ut t he b ehaviour o f t he l ater w all s uggested t hat t he n ew r ampart may h ave b een r elated ' to a detached s tone s tructure n ow s et b ack f rom t he l ine o f t he r ampart'. Gate s tructure n ot s een ( Ant.J.

5 5,

( 1975)

1 12-13)

A t Castle Yard a nd T ower S t. o n t he w est, t here was t he possibility t hat t his rampart w as constructed i n two s tages w ith a s ubsequent a ddition o f c lay a nd l oam t o t he o riginal construction. This ' addition' c ontained Antonine material and o ne s herd dated a fter c . AD 1 60. Walled c ircuit Wall west

a nd

o f

mortared

south may

A t Colebrook

f lint;

gully

i n

f ront

o f

t he

w all

o n

t he

c lay

b ank

r epresent marking-out d itch. S t.

o n

t he

e ast

i n

1 950,

t he

c halky

h ad b een f aced w ith a masonry w all a nd t he i ntervening s pace f illed w ith b rown e arth. Samian s herd o f c . AD 1 45-175 f rom t his i nfill. I n 1 980, t rial t renches i n t his a rea d ated t he i nsertion of t he wall i nto t he f ront o f t he r ampart t o c . AD 2 30. A t cut a s

i nto

Castle the

Yard

chalk

a f oundation

o n

t he

r ampart

r ather

t han

w est i n

i n

s uch

f orming

1 67

1 969-1971 a way

t hat

t he i t

a r evetment

w all u sed

t o

i t.

h ad t he

b een

r ampart

At t he s outh g ate ( 1971), t he f ront a nd butt e nd o f t he r ampart were cut b ack f or t he wall and t here was s ome evidence t hat t his may h ave o ccurred s hortly a fter t he c onstruction o f t he r ampart. The wall t urned i nwards , h ere, p resumably t o l ink up w ith a g atehouse b ehind t he w all l ine a s i n t he r ecessed east a nd west gates a t S ilchester but t he gate s tructure w as not s een ( Ant.J. ( 1975) 5, 13-15). Two s treet s urfaces w ere l aid i n 5 1 f ront of t he gate during t he l ate 4 th century, t he earlier s ealing 1 3 c oins o f w hich t he l atest w as a c oin of G ratian ( AD 3 67-378). At s ometime t hereafter, t he s outh g ate collapsed o nto t he s treet. A g roup c onstruction 5, 1 15). 5

o f e arly 3 rd c entury p ottery was a ssociated w ith t he l evels o f t he wall a t t he s outh gate ( Ant.J. ( 1975)

The r emains o f a p robable b astion w ere i dentified d uring observations e ast of t he s outh gate i n 1 971. The s tructure a butted t he w all a nd i ts f oundations h ad b een c ut i nto t he b erm. I t was a ssigned on s tratigraphic g rounds t o t he Roman p eriod b ut i t

w as

otherwise

u ndated

( Ant.J.

( 1975)

5, 5

15-16). 1

P roc. Hants. F .C. ( 1962) 2 2 pt.II, 5 1-81; Ant.J. ( 1965) 45 , 2 358 ; ( 1970) 5 0, 2 81-5; ( 1975) 5 5, 1 00-17; B ritannia ( 1981) 1 2, 3 63. S ee a lso Wacher 1 974, 2 83-4, 2 86.

Wroxeter

Viroconium

( Salop )

( SJ

5 65087)

1 923-1927 t renches a cross w alled c ircuit a round n orth-east a ngle. 1 936-1937 t renches a cross e ast defences; 1 975 s ectio i c a cross e ast d efences; 1 960 s ection a cross s outh d efences; 1 965, 1 967, 1 9691 971 s ections a t s outh-east a ngle; 1 963 s ection on n orth-west; 1 975 a nd 1 976 observations o n w est , ; 1 936-1937, 1 963 and 1 975 1 976 s ections a cross n orthern l ine o f p ostulated e arly defence s outh o f Bell B rook. C ivitas c apital. I rregular p olygon. No s atisfactory evidence r egarding t he position of g ates. Aerial photographs h ave r evealed t he p ossibility o f b astions i n t he f orm o f p latforms p rojecting out i nto t he d itch a nd d iscontinuity i n t he w all a lignment a t t hree a ngles o n t he n orth s ide ( TSAS ( 1967-1968) 5 8, 1 97-8, 2 00, f ig. 3 9 ( 1969-1970) 5 9 p t.l, 2 8). T he wall i s e xtensively r obbed a nd t he b ank s everely p loughed b ut s urface i ndications o f t he defences r emain f or m ost o f t he c ircuit a part f rom t he w est. E arly d efensive

c ircuit?

As a r esult o f K enyon's e xcavations o n t he east d efences i n 1 936-1937, i t was s uggested t hat Wroxeter may h ave b een enclosed b y a n e arthwork c ircuit w hich l ay b eneath l ater r ampart on t he east but enclosed a smaller a rea o n t he north. I nner s outhernmost s ection ( A) K enyon l ocated a t urf-revetted b ank a nd a double-ditch system s ucceeded by a w all, which cut i nto t he e arl y b ank, a nd a much d eeper s ingle d itch. To t he n orth ( section

B )

,

t he

e arly

d itch

l ay

1 66

b eneath

t he

l ater

w all

and

h ad

b een f illed w ith material f rom t he e arly r ampart w hich h ad b een pushed i nto t he d itch. F urther n orth ( section C ), only t he s econd e arly d itch w as f ound s uggesting t hat t he l ine o f t he e arly b ank and ditches swung away f rom t he l ater wall l ine. The l atest s herds f rom t he bank w ere c . AD 9 0-120. ( Arch. ( 1938) 8 8, 1 76-7) I n 1 975 a s ection a cross t he e ast d efences l ocated t he i nnermost ditch a s i n 1 936-1937 and two t urves o f the f ront o f t he a ssociated r ampart t he r est o f w hich h ad b een l evelled ( WMNS ( 1975) 1 8, 4 2-3; White et a l. 1 975, 6 -7; B ritannia ( 1976) 7 , 3 29

) .

S ubsequently, i t h as b een s uggested t hat t he e arly r ampart and i nner d itch b eneath t he l ater c ivil d efences on t his east s ide i s p robably m ilitary a nd p erhaps p art o f t he l egionary f ortress defences ( TSAS ( 1969-1970) 5 9, 2 10; WMNS ( 1975) 1 8,

4 2-3;

White

S imilarly,

e t

a l.

t he

1 975,

5 ).

n orthern

l ine

o f

K enyon's

p ostulated

e arly

c ircuit has b een re-examined. H ere, K enyon cut a t rench a t t he junction of t he t wo modern l anes s outh o f t he B ell B rook a nd l ocated a f lat-bottomed d itch w ith a l ow bank c ontaining a l ittle 1 st c entury material a nd o verlain by a good d eal o f 2 nd c entury material. The d itch had b een r ecut and i ts f ill c ontained l ate R oman m aterial ( Arch. ( 1938) 8 8, 1 78). A further s ection i n 1 963 l ocated at l east one d itch; i n 1 975-1976 a c lay a nd c obble f oundation w as d iscovered b ehind which w as a s ump o r t rench p ossibly connected w ith t he aqueduct, and a d efensive d itch a nd t wo small d itches p arallel w ith i t. T he defensive

ditch

was

r egarded

a s military.

c

( Sa N L ( Sept.1963) 2 3, 1 ; WMNS ( 1963) 6 , 8 ; B ritannia ( 1976) 7 , 3 29-30; ( 1977) 8 , 3 96; White et a l. 1 975, 6 ). See a lso TSAS ( 1957-1960) 5 6, 1 37; ' ( 1961-1964) 5 7, 1 23; ( 19671 968) 5 8 pt.3, 2 06-8, f ig.42; Webster 1 962, 2 9-31; Wacher 1 974, -

3 67-8. Earthwork

defence

1 936-1937 a nd 1 975 s ections r evealed e arthen r ampart a long east d efences overlying and s et f urther eastwards t han the e arly t urf r evetted ( military) r ampart. P ottery down t o c . AD 1 50 f rom the b ank i n 1 936-1937 but Webster has n oted t hree s herds o f c . AD 1 50-195 ( Webster 1 962, 2 9; ( Arch. ( 1938) 8 8, 1 76-9; T SAS ( 19671 968) 5 8, f ig. 4 2; WMNS ( 1975) 18 , 42-3; White et a l. 1 975, 6 ). 1 963 s ection o n n orth w est r evealed a c lay b ank c ontaining late 1 st á entury p ottery and a ditch partially o verlaid by another bank o r c ounterScarp c ontaining a n a s o f D o r nitian ( Sa N L Jan

( 1964)

2 . 5 v

2 ;

WMNS

( 1963)

6 . ,

8 ).

1 960 s ection a cross s outh d efences r evealed a g lacis bank w ith e vidence f or t urf work a t t he f ront and r ear. l atest pottery f rom t he r ampart w as Antonine ( WMNS ( 1960) Webster

1 962,

3 2).

1 6

s tyle The ‚ 1 0;

1 969-1970 a t t he s outh-east a ngle a nd a djacent t o 1 960 s ection, t he f irst Roman r oad s urface ( located i n 1 965) w as o verlain i n p art by t he r ampart c onstructed f rom d itch u p-cast and a ttributed t o t he l ate 2nd c entury. The rampart was o f g lacis s tyle a nd t he c lay-lined d itch f ormed a b utt e nd n ear t he r oad e dge ( WMNS ( 1969) 12 , 2 0-1; ( 1970) 1 3, 3 4-5; B ritannia ( 1971) 2 , 2 61). T he w estern s ide o f t he c ircuit i s u ncertain b ut i n 1 975 a s ection t hrough the r iver bank o n t he n orth-west r evealed t he b ase o f t he r ampart a nd i n 1 976 a l arge d itch c ontaining m id-2nd century pottery was a lso f ound on t he w est ( Britannia ( 1976) 7 , 30; ( 1977) 8 , 3 96). 3 S ee a lso T SAS ( 1967-1968) 5 8, 2 02-4; Webster 1 975a, 4-5; 4 Wacher 1 974, 3 69. Walled

c ircuit

The s ections

wall h as b een e xtensively r obbed a nd n one o f t he various a cross t he c ircuit has p roduced s atisfactory r esults.

I t would appear f rom t he s ections o n t he e ast ( 1936-1937 a nd 1 975) and n orth-west ( 1963) that t he wall was i nserted i nto t he e arlier r ampart. There i s a lso e vidence t hat t he o riginal double-ditch s ystem was r eplaced by a s ingle ' middle' d itch. I n 1 923-1927 t he c lay a nd c obble f oundation f or t he wall w as t raced f or s ome d istance a round t he n orth-east angle b etween t he Watling Street a nd t he modern l ane t o t he e ast. The d ouble-ditch system was r educed t o a s ingle ditch t owards t he east a nd d ecreased i n d epth a s i t a pproached t he l ane p ossibly i ndicating a gateway on t he n orth-east ( Atkinson 1 942, 3 28-32) I n 1 960, on t he s outh, v ery l ittle t race of t he w all w as f ound, ' only t he most careful s crutiny r evealed a s lightly darker f illing i n f ront o f t he b ank w hich must r epresent t he r obber t rench o f t he f ootings' ( Webster 1 962, 3 1). The position o f t he w all i n r elation t o t he f ront o f t he b ank w as r ather p recarious and i t was s uggested t hat the bank had e roded and t he d itch b ecome p artially f illed b efore t he w all w as b uilt ( WMNS ( 1960) ‚ 1 0) H ere r eplaced c ontained 3 1)

t he double-ditch s ystem o f by a s ingle d itch which a 4 th c entury s herd i n i ts

t he e arthwork c ircuit w as c ut t he i nner d itch a nd l ower , s ilt ( Webster 1 962,

I n 1 967 a t t he s outh-east a ngle, t he o uter d itch o f t he early system had b een deliberately f illed and covered b y a metalled s urface w hich s ealed l ate 4 th c entury p ottery a nd w as i n t urn c overed by o ccupation material o f a s imilar date ( JRS ( 1968) 5 8,

1 86-7).

I n 1 969 t he l ater t o h ave a c ounterscarp ( 1969) 1 2, 2 0-1) The

p ossible

s ingle d itch on w hich c ontained

e ntrance

on

t he

1 70

t he 4 th

s outh-east w as c entury s herds

s outh-east

w ould

s eem

t o

f ound ( WMNS

h ave

b een a bandoned w hen t he wall w as b uilt. I n 1 965 a t t he p oint w here t he p resent r oad t o I ronbridge c rosses t he r ampart, a R oman r oad w as f ound a djoined by f ragmentary f oundations s uggesting a p ossible stone-built g ate. I n 1 970 t he t riple-ditch s ystem was f ound b utting u p t o t his r oad b ut t he t own wall w as c arried a cross i t ( WMNS ( 1965) § , 15; ( 1969) 1 2, 2 0-1; ( 1970) 1 3, 3 4-5; B ritannia

( 1971

2 ,

2 61).

On t he west, t he 1 963 s ection r evealed t he w all-footings s et i nto a bank, a d itch and c ounterscarp. I n 1 975, a s ection t hrough t he r iver b ank l ocated t he r ampart base, t he r emains o f a c ut at t he c rest which may have b een a ssociated w ith t he r obbing o f the wall. This was s ealed by a ' later r efortification' a ssociated with a c oin o f T etricus I I ( AD 2 68-273) ( WMNS ( 1963) 6 ,

8 ;

B ritannia

( 1976)

7 ,

30). 3

N o dating e vidence f or t he w all. Webster i s i nclined t o p lace t his later i n t he Roman period and p ossibly e ven a s l ate a s c . AD 3 80 a nd h e m entions t hat ' the l evel a t w hich t he w all f oundations have b een i nserted s uggests t hat i t must h ave b een a l ong t ime a fter t he o riginal work' ( Webster 1 975a, 1 06-7, 1 18). S ee a lso TSAS 2 11-19; f or

( 1961-1964) 5 7, 1 969-1970 ( 1972)

1 21-7; 5 9, 2 8;

( 1967-1968) 5 8, 1 97-202, Wacher 1 964, 1 05; 1 974,

3 69.

York

Eboracum

( Yorks.)

N o

a rchaeological

a nd

early

2 0th

( SE

6 00515)

i nvestigation,

o nly o bservations

i n

t he

1 9th

century.

C olonia. Some i ndication o f defences e nclosing t he C olonia s outh o f t he R iver Ouse. C ircuit uncertain e xcept on n orth-west but t he medieval d efences, f orming a r oughly r ectangular e nclosure, may h ave f ollowed t he Roman c ircuit. No dating but i t has b een s uggested t hat t he s ettlement may h ave b een walled i n t he e arly 3 rd

c entury

when p romoted

t o

colonia.

1 839: a wall w as f ound w ithin t he r ampart o f t he medieval defences on t he west; 1 840: a massive wall f ound east o f above a nd r unning s outh-east; 1 874: f oundations o f a w all f ound during construction of r oad a rchway t hrough medieval d efences on w est; 1 939: r ubble w all e ncountered w hen c utting c able t rench t hrough medieval

defences

R CHM

n os..

( 1962)

Wacher

1 974,

4 9,

on west.

3 4a-34g

a nd

1 61-3.

1 71

n o.

5 1,

f igs.

3 7,

3 8

& 3 9;

Bibliography Alexander,

J. (1970) The Excavations. London":-

Directing

of

Archaeological

Alexander, J. (1975) The Development of Urban Communities: the Evidence from Cambridge and Great Chesterford. In W.Rodwell and T.Rowley (eds.) Small Towns of Roman Britain, 103-109. British Archaeological Reports 15, Oxford. Atkinson,

o. (1931) Caistor Excavations. �, �, 93-139.

Norfolk Archaeo-

Atkinson, D. (1942) Report on Excavations at Wroxeter (the Roman City of Viroconium) in the County of Salop 1921=° 1927. Oxford University Press. Baker,

A. (1970) Viroconium: A Study of the Defences from Aerial Reconnaissance. Transactions of the Shropshire Archaeological Society, �, 197-219.

Barker,

P. A. (1970) The Origins of Worcester. Transactions of the Worcestershire Archaeological Society, 3rd ser.,2, T=T24.

Barker,

P. A. (1979) The Plumbatae from Wroxeter, in M.W.C.Hassall and R. Ireland (eds.) De Rebus Bellicus, British Archaeological Report Supplementary series 63. Oxford.

Bidwell, P. T. (1980) Roman Exeter: Fortress and Town, Exeter City Council. Birley, A. (1977) Vindolanda. London. Birley, Boon, Boon,

E.

(1953) Roman Britain and the Roman Army. f

Kendal.

G. C. (1969) Belgic and Roman Silchester: the excava­ tions of 1954-58 with an excursus.on the early history of Calleva. Archaeologia, 102, 1-82.

G. c. (1974) Silchester. Newton Abbot.

The Roman Town of

Calleva.

Branigan, K. (197S) Gatcombe. In W.Rodwell and T.Rowley (eds.) Small Towns .of Roman Britain, 175-182. British Archaeo­ logical Reports 15. Oxford. Bushe-F6x r J. P. (1916) Third Report on the Excavations on the site of the Roman Town at Wroxeter Shropshire 1914. Report-of--the ResearchComrnittee of the Society of Antiquaries of London 4. London. Butler,

R. M. (1959) Late Roman Town Walls in Gaul. Archaeo­ logical Journal, 116, 25-50.

173

Butler,

Casey,

R . M . ( 1971) The d efences o f t he r ess at Y ork. I n R .M. B utler C ivilian i n Roman Yorkshire, U niversity P ress. P . J . ( 1983) i n B ritain, U rban R eport

Imperial c ampaigns a nd i n J. Maloney a nd B .

Defences i n 5 1. L ondon.

t he

West,

f ourth c entury f ort( ed.), S oldier a nd 9 7-105. Le11c ester

4 th century defences H obley ( eds.) R oman

1 21-124.

C BA

Research

Charlesworth, D . ( 1971) The d efences o f I surium B rigantum i n R . M . B utler ( ed.) Soldier a nd C ivilian i n Roman Yorkshire 1 55-164, L eicester U niveF Tty P ress. Chevallier,

R .

( 1976)

Roman Roads.

Collingwood, R . G . and Myres, J . t he English Settlements.

L ondon.

N . L . ( 1937) Roman B ritain a nd ( 2nd e dition) Oxford.

Collingwood, R . G . and R ichmond, I . Roman B ritain. ( 2nd e dition)

( 1969) The L ondon.

Archaeology o f

Colyer,

C . ( 1975) L incoln. The Archaeology o f an L incoln A rchaeological T rust, L incoln.

Corder,

P . ( 1956) The R eorganisation o f t he D efences o f R omanoB ritish T owns i n t he F ourth Century. Archaeological Journal, 1 12, 2 0-42.

Corder,

P . ( ed.) 1 951, 1 954, 1 961. T he Roman town a nd villa G reat Casterton, Rutland. Reports I -Ill. Nottingham.

Corder,

Journal o f P . a nd R ichmond, I .A. ( 1942) P etuaria , 1t he B ritish Archaeological Association, 3 rd s er., 7 3 0.

Corder,

P . Fort

a nd R omans,

T .

( 1934'1938)

at B rough-on-Humber

h istoric

Excavations

1 934-1937,

parts

1 938-9.

t he

a t

R oman

Hull.

Cotton,

M . A . (1947) E xcavations a t l ogia, 9 2, 1 21-167.

Cotton,

M . A . a nd Gathercole, P . W . ( 1958) Excavations a t Clausentum, Southampton, 1 951-1954, Ministry of Works A rchaeological R eports n o.2, H MSO, L ondon.

C raster,

. ( 1954) The east gate and a djoining t own w all o f t he R oman t own a t Caerwent, Monmouthshire. A rchaeologia Cambrensis, 1 03, 5 4-65.

C rummy,

P . ( 1977) D evelopment

Cunliffe, B . W . Committee L ondon.

S ilchester

a t 1 -5.

c ity,

Archaeo-

Colchester: The R oman F ortress and o f t he Colonia, B ritannia, 8 , 6 5-106.

t he

( 1969) Roman Bath. R eport of t he Research of t he S ociety o f A ntiquaries of L ondon, 2 4.

Cunliffe, B . W . ( 1973) series. L ondon. Down,

A .

Chichester

Excavations Down,

The

Regni.

Excavations

P eoples

of

R oman

B ritain

I I,

Chichester

C ivic

S ociety

I II,

Chichester

C ivic

S ociety

Committee.

A . Chichester Excavations Excavations Committee.

Down,

A . and R ule, M . ( 1971) Chichester Excavations Chichester C ivic S ociety E xcavations Committee.

D rury,

P . J . ( 1975) Roman Chelmsford Caesaromagus i n W . Rodwell and T . Rowley ( eds.) Small Towns of Roman Britain, 1 59-173. B ritish A rchaeological R eports 15. Oxford.

Fox,

I ,

-

A .

( 1952) R oman E xeter Excavations i n t he War-damaged Areas 1 945-1947. Manchester University P ress.

Freezer,

D . F . ( 1977) F rom Saltings t o Spa Town. The Archaeology of D roitwich. D roitwich A rchaeological Committee and H ereford a nd Worcester County Council, D roitwich.

Frere,

S . S . i nterim

( 1956) Excavations a t Verulamium, 1 955: r eport. Antiquaries Journal, 3 6, 1 -10.

f irst

Frere,

S . S . i nterim

( 1960) r eport.

f ifth

Frere,

S .

S .

( 1961)

Frere,

S . S . ( 1964a) Institute o f

E xcavations a t Verulamium 1 959: Antiquaries Journal, 4 0, 1 -26.

Civitas

-

a myth?

Verulamium Archaeology,

S . S . ( 1964b) 3 8, 1 03-112.

Verulamium,

Frere,

S .

Town

S .

3 9, Frere,

Frere,

( 1965)

Three

D efences

i n

Roman

R oman

S . S . ( 1966) The e nd o f t owns i n Wacher ( ed.) The Civitas Capitals 1 00. Leicester University P ress. S .

S .

S .

S .

( 1967)

S .

o f

Bu kletin

t he

C ities.

Antiquity

B ritain.

Antiquity

B ritannia.

A History

o f

S .

( 1978)

The

o rigins

B ritannia.

L ondon.

o f

small

A History

( Cardinal

S . S . ( 1981) Verulamium and M . H enig ( eds.) Century,

Roman B ritain.

i st

L ondon.

( 1975)

edition. Frere,

2 9-36.

R oman B ritain. I n J . o f Roman B ritain 8 7-

t owns.

and T . R owley ( eds.) Small Towns o f British A rchaeological R eports 1 5. Frere,

3 5,

1 37-139.

edition. Frere,

Then and Now. , 4 , 6 1-82.

-

Frere,

Antiquity,

3 83-392.

R odweli

Roman B ritain.

4 -7.

2 nd

1 974).

i nthe t hird century. The Roman West i n

B ritish

1 75

o f

p aperback

I n W .

Roman B ritain, Oxford.

Archaeological

I n A . K ing t he Third R eports,

I nternational

S eries

1 09.

Oxford.

Fulford,

M . ( 1983) The Defensive S equence at Silchester. J .Maloney a nd B .Hobley ( ed.) Roman U rban D efences t he West, 8 5-89. CBA R esearch Report l, L ondon.

G reen,

H .

J .

M .

( 1975)

Roman

Godmanchester.

I n

W .

R odwell

T . R owley ( eds.), Small Towns of Roman B ritain, 2 10. B ritish A rchaeological R eports 1 5. O xford. G rimes,

W . F . London.

Hartley,

B .

R .

( 1968) London. ( 1966)

The

Dating

J . Wacher ( ed.) 5 2-59. L eicester H artley,

Rome and her G loucester.

J .

Wiltshire

( 1976)

by

M . J .

I .

R .

Hull,

Roman

buildings

t owns:

t he

a nd

The

Spatial

I . R. and Hassall, M . R omano-British walled

Medieval

R oman

P rovinces,

a rchaeological

D istribution

( 1971) t owns.

1 83-

s tructures.

of

W . C . ( 1977) Review of The Towns Wacher, Antiquaries Journal, 5 7, ( 1975)

a nd

a nd

I n

B ritain

Romano-British Towns i n Hartley a nd J . Wacher

Northern

Small Towns, i n W . R odwell a nd Towns of Roman Britain 6 7-74. Reports 1 5. Oxford. Hodder,

o f

The Civitas Capitals University P ress.

A .Sutton,

H assall,

H odder,

t own

B . R . ( 1983) The Enclosure o f t he S econd Century A . D. I n B . ( e d s.)

Haslam,

Excavation

I n i n

8 4-95.

evidence.

of R oman 1 26-128. of

Britain

Romano-British

mall T . R owley ( eds.) , S B ritish Archaeological

The Man,

( n on-random s pacing 6 , 3 91-407.

M . R . ( 1958) Roman Colchester. Report Committee of t he Society of Antiquaries

of of

o f

t he Research L ondon, 2 0.

Oxford. Hawkes,

James,

S . C . and Dunning, G . C . ( 1961) Soldiers a nd Settlers i n B ritain, F ourth t o F ifth C entury. With a Catalogue o f Animal-Ornamented B uckles a nd R elated Belt-Fittings. Medieval Archaeology. 5 , 1-70. H . ( 1978) Excavations i n Church S treet, Carmarthen 1 976, i n G . C . Boon ( ed.) Monographs and Collections.I. Roman S ites, 6 3-106. Cardiff.

Jarrett,

M . 3,

G . ( 1965) 5 7-59.

Johnson,

J . S . ( 1973) Gallica Belgica.

Johnson,

S .

( 1976) The

Johnson,

S .

( 1980)

Town

D efences

R oman

B ritain.

A G roup o f L ate R oman Britannia, 4 , 2 10-223.

Roman Forts

Later

i n

Roman

o f

t he

Britain.

1 76

Saxon

L ondon.

Antiquity

C ity Walls

Shore,

London

i n

Johnson, S. (1983) Late Roman Urban defences in Eµrope. In J. Maloney and B. Hobley (eds.) Roman Urban Defences in the --West, 69-77. CBA Research Report 51. London. -,

,

Jones, G. D. B. (1974) Roman Manchester, s. Grealey (ed.) Manchester Excavation Committee, Altrincham. Jones, G. D. B. and Reynolds, P. (1979) Roman Manchester: The Deansgate Excavations 1978. An Interim Report, Man­ chester. Jones, M. J. (1980) The Defences of the Upper Roman Enclosure. Archaeology of Lincoln, 7/1. C. B. A., London. M agilton, J. R. (1977) The Doncaster District: an Archaeo­ logical Survey. Doncaster Museum and Art Gallery. Maloney, J. (1980) The Roman Defences of London. Archaeology (August, 1980), No.73, 55-60.

Current

Maloney, J. (1983) Ree ent Work on London's Defences. In J. Maloney and B.Hobley (eds) Roman Urban Defences in the CBA Research Report, 51. London-.West, 96-117. Marsden, P. (1980) Roman London. Thames and Hudson, London. Merrifield, R. (1965) The Roman City of London. London. Merrifield, R. (1969) Roman London. London. Merrifield, R. (1978) A Handbook to Roman London. London. Nash-Williams, V.E. (1930) Further excavations at Caerwent, Monmouthshire, 1923-5. Archaeologia, .!!.9_,229-88. Ordnance

Survey (1956) Chessington.

Ordnance

Survey (1978) Southampton.

Map Map

of of

Roman Roman

Britain

(3rd

ed.)

Britain (4th

ed.)

von Petrikovits, H. (1971) Fortifications in the North Western Roman Empire from the Third to the Fifth Centuries AD. J ournal of Roman Studies,_§_!, 178-218. Ramm, H. (1978) The Parisi. London. Reece, R. (1973) Roman Coinage in the Western Empire. Britannia !, 227-251. Rivet, A. L. F. (1964) Town and Country in Roman Britain (2nd. ed.) London. Rivet, A. L. F. (1975) Summing Up; The Classification of Minor Towns and Related Settlements. In W. Rodwell and T. Rowley (eds.), Small Towns of Roman Britain, 111-114, British Archaeological Reports 15. Oxford. 177

Richmond, I. (1969) Roman Archaeology and Art. ·Essays and Studies� Sir.� Richmond. P. Salway (ed.), London. Robinson, J. F. (1978) The Archaeology of M alton and Norton. Yorkshire ArchaeologTcal Society, Leeds. Rodwell, W. (1975) Trino�antian Towns and Their Setting: A Case Study. In w. Rodwell and T. Rowley (eds.), Small Towns 85-101, British Archaeological of Roman Britain, Reports 15. Oxford. Rodwell,

w. and Rowley, T. (1975) (eds.) Small Towns of Roman Britain. British Archaeological Reports 15. Oxford

Rowley, T. (1975) The Roman Towns of Oxfordshire. In w. · Rodwell and T. Rowley (eds.), Small Towns of Roman Britain, 115-124, British Archaeological Reports 15. Oxford. St.

Joseph, J. K. (1966) The contribution of aerial photo­ In J. Wacher (ed.), The Civitas Capitals of graphy. Roman Britain, 21-30. Leicester University Press.

Salway, · P. (1965) The Frontier People Cambridge University Press.

of

Roman

Britain.

Salway,

P. (198 0) The Vici: Urbanisation in the North. in K. Branigan (ed.) Rome and the Brigantes. The impact of Rome on Northern England, 8-17, Universit y of Sheffield.

Todd,

Roman The (1969) M. the of Transactions Nottinghamshire, 73.

Settlement Thoroton

at

Margidunum. Society of

Todd,

M.

Todd,

M. (1973) The Coritani, Peoples of Roman Britain series, London.

.!.,

(1970) The small towns of Roman Britain. 114-130.

Britannia,

Todd,

M. (1975) Margidunum and Ancaster. ·In w. Rodwell and T. Rowley (eds,), Small Towns of Roman Britain, 211-223. British Archaeological Repqrts 15, Oxford.

Todd,

M. (198 1) The Roman Town at Ancaster; Lincolnshire: the excavationsof 1955-?T:- Exeter.

Wacher, J. (1960) Petuaria. New evidence for the Roman town and its earlier fort. Antiquaries Journal, !Q_, 58-64. Wacher,

J. (1961) Cirencester, 1960. Antiquaries Journal, 41, 63-67.

Wacher,

J. s. (1964) A Survey of Romano-British Town Defences of the Early and Middle Second Century, 103-113. Archaeological Journal, 119 , 103-113.

178

First Interim Report.

Wacher,

J.

11,

(1965) Towns Defences in Roman Britain. 225-227.

Antiquity

J. (1966) Earthwork Defences of the Second Century. In J. Wacher (ed.), The Civitas Capitals of Roman Britain

Wacher,

60-69.

Leicester University Press.

Wacher,

J. s. (1969) Excavations at Brough-on-Humber, 19581961. Report of the Research Committee of the Society of Antiquaries of London 25. London.

Wacher,

J. S. · (1971) Yorkshire towns in the fourth century. In R. M. Butler (ed.), Soldier and Civilian in Roman Yorkshire, 165-177. Leicester University Press"-:-

Wacher, J. (1974) Towns of Roman Britain.

London.

J. (1975) Village Fortifications. In W. Rodwell and T. Rowley (eds.), Small Towns of Roman Britain, 51-52, British Archaeological Reports 15, Oxford.

Wacher,

Wacher, J. (1978) Roman Britain. London. Webster, G. (1962) The Defences of Viroconium (Wroxeter) • Transactions of the Birmingham and Warwickshire Archaeological Society, ..I!!_, 27-39. Webster,

G. (1971) A Roman System of Fortified Posts Along the Watling Street, Britain. In S. Applebaum (ed.), Roman Frontier Studies, 1967. Proceedings of the 7th Inter­ national Congress, 38-45. Tel Aviv.

Webster,

G. (1975a) The Cornovii. Peoples of Roman Britain series, London. Webster, G. (1975b) Small Towns Without Defences. In w. Rodwell and T. Rowley (eds.), Smail Towns of Roman Britain, 5366. British Archaeological Reports-15, Oxford. Webster,

G. (1977) Reflections on Romano-British Pottery Studies, past, present and future. In J. Dore and K. (eds.), Roman Pottery Studies in Britain and Greene Beyond, 317-333, British A�chaeological Reports supple­ mentary ser. 30, Oxford.

Webster,

G. (1983) The Function and brganisatiQn of Late Defences. In J. Maloney and B. Hobley (eds.) Roman Urban Defe�ces in the West ', 118-120. CBA Research Report 51, Londori":-

Webster, G. Fort and town in early Britain. In J. (ed.) Civitas Capitals of Roman Britain, Leicester University Press. Wheeler,

Wacher 31-46.

R. E. M. and Wheeler, T. V. (1936) Verulamium. A Belgic and Two Roman Cities, Report of the Research Committee of the Society of Antiquaries of London,11 Oxford. 179

White,

P., Webster, G., Johnson,° s., Barker, P. and Brown, P. (1975) Wroxeter Roman City. Work in Progress 1975.

Whitwell, B. J. (1970) Roman Lincolnshire. History of Lincolnshire, 2 Lincolnshire Local History Society, Lincoln.

Wilmott,

A. R. (1979) Kenchester - a reconsideration. Un­ p ublished MA dissertation, University of Birmingham.

Wilson,

D. R. (1975) The "Small Towns" of Roman Britaii From the Air. In W. Rodwell and T. Rowley (eds.), Small Towns of Roman Britain, 9-49, British Archaeological Reports-15, Oxford.

Wright,

T. (1872) Ur iconium: A Historical Account of the Ancient Roman City, And of the Excavations Made Upon � Site at Wroxeter, in Shropshire, Forming a Sk·etch of the Condition and History of the Welsh Border During the7foman Period.---Y:ondon andShrewsbury.

180

1 !

o

.

�l



-

9

cc c1

1

I n ,I �

4

-

-

� -M� '

c

L i U 0

‚ 4

U

1

9

c j

T i

J H

0

O

1

1 !

i t e t e S s w i h p r l a t e 2 n d C e a r t h d f n c e s e e

wo r k

w a i l

e d e e d S i t s f n e d b y r e a t h w o r k n l

o s a l i n e n n e

a n d

l a t e r l a t e r w

e d e d S i t s f n e d b y r e a t h w o r k a l

b u t n o

L

I e a r t h o d r t l i n e n i f e n

e d e d S i t s f n e d b y w or k i t h w a w l

n d e e n e n t N o i p d e n r t h o i l n t f e d i d e i u i r c t e n v c i r c t e e e d o a n d w

i l

0

Ii

0

0 C IO

-

o

'

C

-S '

c ' .

c r

( ..

6

c .

c n

)

c ' .

( I ) -

0

C f l .-

‚ -

C )

'-

U )

F

'-

4 -

( 4 )

1 82

S a l E e v i d e n c e

d e f e n c e s

r c u i t c i

-

l e a s t ,

l a t e

2 n d C ' e a r t h w o r k d e f n c e s

r c t e d e

2 n d C ?

f r o m t h w e d e e l

l a t e

2 n d

0 U

o0

1 83

j

0 U

O L

( r c t d e v o ? i n e a r l y e e e d n o

W a e d l e a r l y r p l a c e e

W a e d u i r c t e d l c i r c t e e e a r l y 2 n d u n t l H d r a n i c i a i x s t l e g i o n a r y e i i n g a m p e r t r

l c i u W a e d r c i t b e g u n n t e o c a n p i t e d A e d t o d a i n p e r t r e s r

l c i r c t e e a e d u i r c t e d e a r l y

C , ç x i b l y s

e x s t i n g i

u i r c l a t e 2 n d ! c i r c t e e t e d 3 d C i t b r d e n v r t o r e e o o o e x s t i n g i

-

d m i 2 n d C .

n o e w l i r E

e a t h w o k r c u i t r r c i

l a t e c o t u i n g n i n

e a r l y e a r l y l e g i n a y f o r t o r

p r i o . A e d t o e d d f o t e r s

u n t i l

U

e a t h w o r k r C ? P i b l y o n a d r t l i n e , o s i f e n o n h w e s a t t e t r c u i t . c i

r i i g o o r r s e f u r b s h n f l e g i n a y f o t e r a m p e r t

' L a t e U

e G l o u c s t e r

I Earlyearthwork 4 J

I

3 r d C )

i s t !

2 n d C ; 3 r d C ?

( 2 n d l a t e r )

C l )

4 ) r H

-

I

1 84

1 5 0

A D e i c i u n d f n s v e r c i t s u k o w n

e a r r w a l D i t c h d u n a t e d ;

o f

w al a n d ? o r c n t e m p o a r y r a m p a r t

H

O u t e r B k a n d a n s 2 p h a e ; f i r s t t i o h t o r e l a n s i p o t h e r

p o i b l e s r v e t e t n

r e l a n s h i p t o t i o p o s i b l e a r l i e r d e f n c e

a n d a l c o n t em p o a r y r m p a r t r a

u n k n o w n

4 J

p l i s a d e a d a n d i t c h ; t p q L .

o f

0

a l w a n d o r c n t e m p o a r y r a m p a r t

i n e t ' s r e d i n t o f r o r a n t o f m p a r t ' c i )

9 .l a / 3 . t h a

a i g r i T r i t

o d i r e n t n f e

o n d i f e r e n t a l i g r n E n t / a r e a e n c l o s e d

a s b f o r e e

L 3 I

( 2 n d h a l f 2 n d C )

f e a t u r e s / p h a s e s -

t p r o v e n / n o d a t e u k o w n n

C o r

o t h e r

r

6

r r o

L i

e -

I D

. 1 4

r

-

-

U

T h

n 1 ) i t e r n a l t i o l a n s h i p t o 2 ) s t o n e w l n n t o a u k o w n .

a s o c i a t e d t o w a l , e t c .

a g t e w a y

t r n r e r u e t i S E c o n ; r e l w a

f e a t u r e s / r e l a t i o n s h i p

t e s n o t a b l i s h e d . i o n S ; r e l a t o n s h i p

t o

f a n a t N W w a l

t o w a n

h p e d b a t i o n s x c v t e d s a s e a a n d S 1

b a s t i o n s / r e l a t i o n s h i p

f i

U

1 85 o r o l w e p r t i o n s s i s v o b a t o n u r i e ; 2 n s o u t h , n r a p o n e o o t h ; a r e n t l y d t l I n e : i n o t h e w a .

a t l e a s t o f 3

c o r n e r s / a d e d

t i t s y e 1 ) r e c u n g o f d i c h t m e a r l y 4 t b C . d i o o r a r 2 ) . a t n f m p o s t a i r r a o l o n r n e d e r f w a N E c o r ; u n a t e d .

s e oa e n c n d p h a s o f O u t e r B k a n d i t t t t D c h a r i b u e d o 4 t h C

I

I w ay t o

l s t C e e d f n s i v e r c u i t c i

r j

' o f 1 s t

d i c r

4 -3

Ii

h

1 86

d j

d i

n o d a t i n g

c o

c . A D 2 9 0 m p a r t s h 2 p e r i o d s r a o e d f s r u c t i o n S E o c o n t o n

w i t h t l a e s t

f l l c t w i o n s r u c t e d n f o i r n t o f r a m p a r t

n o i n o f r m a t i o n r e i r e l a t o n s h i p r t o e a t h w o r k

h

p c . A D 2 7 0

d u g

L

d i t c h l a t e

i c e l d t h b c k f i e d ; 2 d i t c h e s 3 A l ) 7 a t

I

f l j

wa a n d l c o n t em p o a r y r m p a r t r a

t o e a r t h

a l i h / r e l a t o n s h i p w or k

44

o n l y

I z r

d i t c h l o c a t e d ; A D l 4 0 1 f J

f t u r / p h a e e a e s s

J i I

m at i o n

o t h e r

o d i r n t n f e a l i g n m 3 .6 4 . c h a

n f e o d i r n t a l i g t / n r a n c l o s e d a e e

e a s b f o r e

s a e e m a s b f o r e

L

n o i n f o r

n o d a t i n g kno w n

n c l o s e d e

a r e a

e n t ;

9 1 h /

-

d

C

4 4

4 4

' . 'z f

d i

4)

x c v a t e d / n o e a

f u / r e l a t o n s h i p e a t r e s i b a / r e l a t o n s h i p s t i o n s i t o w a l

c c e d i i t h s u e d o r g i n a l n e w d d i t c h i n 4 t h C ? t c 2 d i h e s c u t .

v i d n c e e e s r u c t i n o r t h c o n t o n o n f o r 2 p

I

n o r h g a t e t f r m a t i o n i n o

s g x c v t e d / n o e t a t e e a i n f o r m at i o n

o c i a t e d a s o w a , e t c . t l

u e S b s q u e n t l y

o l er i d s o f w a

I

j jHj

1 87

i s t C e e d f n s i v e r c u i t c i

4

w or k -

-

1 88 m i 2 n d C d

W r a m

p ar t

per i o d s

o n W ;

-

w a e d t o a l d r a m p ar t

a n d a l c o n t em p o a r y r m p a r t r a

C.

i n s r t e d e r n t o f f o r a m p ar t

i n t o

k a l

w or k

O

l o w r k i g o u t h a n k r n p c e d i g r a n t r e n s r u c t i o n c o n t

1 ) m a r t o f 2 r a 2 n d p h s u n t e d a l s o a e d a 2 ) o n N E a n d a t r v e , ( 3 r d C ? ) l e l e l d

I b a o n e d a n d

a n d c n t e m p o a r y o r r a m p ar t

m a i v f r s e e s t a d i n g n

/ r e l a t o n s h i p a l i o e a r t h t

c J

f r m a t i o n n o i n o

9 4 J

( l a f e 2 n d / e a l y 3 r d C ) r

p o b l i t y o f s i e a r t h u s t e d s g e

o t h er f t u r / p h a e s e a e s s

o n d e r t i f i a l i g m e n t / n r a n c l o s e d a e e

-

L

U

C

o n e t o i o n r e l a t s h i p n o t

w e r o f S g a t e

t s a t e d

a l e a ar t i y x c v a t e d ; t s a t e d .

f u / r e l a t o n s h i p e a t r e s i

e a x c v a t e d ;

l w i t h u a

c o n t

r e mp o a r y w i t h a f n s h a p d o n S E ; a e d t o w

b a / r e l a t o n s h i p s t i o n s i t o % l

ed b a s t i o n

a n d N a n g l e s ;

I i

H a s t b i o n s a t N W o n d b e d i n t o u a L l

l a

e a x c v t e d

1

t p q 3 6 1

A D

I

t e o r N g a ; c n t e m p o a r y

t e n 1 ) S g a i s t o n e ; l t 2 ) n o r h g a t e p i o n r e l a t s h i p n o t

o c i a t e d a s o w a , e t c . t l

3 7 -

-

d is m a n t l i n g d u n a t e d

s 2s u b se q u en t p h a e s t a t e d o N g ; u n a t e d

l o t d i s h a w o u e r t c h d i ; 4 t h C ?

t o f s g a t e ;

e e t n e t d 1 ) t i m b e r r v n a e d t o n t t d r am p a r t o n or h 2 ) o u e r i t c h d u g l o n u e s t 3 ) o n S W , w a b e y r o d b e a r l y 4 t h C .

I

1 89

a t o f l e r a t i o n

o f u n d a t i o

n s

0

o n l y ;

4 J r

c l )

1 90

4 . J

r H

• ( l a t e 2 n d / r e a l y 3 r d C )

s po i b l y c o n t t o n e s g a t e o n W

o f r e r a m p

s u e 2 b s q u e n t

r r p ai & e f u r b i s h m t t e d ar t & f i c h s ; u n a t e d .

i o pos t i n o f a t l e a s t g a t e

r e mp o a r y

e a s t en t o f

l o c a t e d s p h a e s

1 L / r e l a t o n s h i p a l i

UH d a e d t o

r a m p

t o p

w or k n o t

r a m p ar t ?

e pos i b l s t o n e l w a o n

e a r t h e s t a b l i s h e d

r e l a t i o n s h i p t o

e a t h , o r k t o r

ar t

o f

.

i n o f r m a t i o n

-

o t h er f t u r / p h a e e a e s s

. f

n o

C-4 C6

d i t c h n o d a t i n g

i s t C d f n s i v e e r c u i t c i

f l •

h

.

c i kr c c o

H i

r

I

OH

o n o n bas t i N W ; n o r e t i o l r e l a n s h i p t o w a

a t e i n n W g ; r e l a t o s h i p u c e r t a i n ; t p q c . A D 3 Q

‚ a l

t o

l s t i o n s / r e l a t i o n s h i p

a s o c i a t e d f e a t u r e s / r e l a t i o n s h i p t o w a l , e t c .

e e v i d n c e

1 ) b l o c k i n g

dd

1 91 e e r c t e d o n a t e l 4 t h / e a r l y

4 t h C ? r a a a d a 2 ) d i t c h , m p r t n P l i s a d e

h e i g n 1 ) h t e i n g o f

a f t e r m i d

d u n a t e d a t 2 ) l e r a t i o n s

5 t h

C

r a m p

a r i c a g e i y ; t t s o d i c h y s t e m C .

f e e n t n d i r a l i g m e n t

4 t h

o f S

I ar t ; ( e a r l y

i s t C e e d f n s i v e c i r c u i t

z 1-

H

4 J

• N

4 i

1

n o d a t i n g

bI

1 92 2p e r i o d s c o n t s r u c t i o n

m i d

o f d e f n c e s ; 1 4 0 ) S i t e c . A D 1 4 0 ?

o f

l d i t c h e s f i e d b y t e r n e n 3 r d C ; s o u h t r a n c e i c a u s

r e c o n s t r u c t i o n ( b e f o r e c . A l ) a b o n e d ? e n d

o f b y o u t e r 3 r d C )

d i t c h o n

c u t b a c k

c .

r a m p e r t l f o r w a

s l p o i b y l a t e r d i t o n t a b u n o t c l e a r

e e l x i s t n c e o f u a n i s u c e r t a n , p o i b l y o r o s t h r k o n l y

i l e i l b u t o v r f e d i n n i e r e w n o t h r i s e u k o w n

w a e d f r o n t a l d t o f a l i p o r t o r

S ;

7 1 / . b a

r a l i g n t ;

n f e r o d i r n t

o d i r n t n f e a l i g r u Q n t / a e e r a n c l o s e d

-.

r a m p e r t a n d d i t c h ; ( b f o r e e A D 1 2 0 )

p o i b r p l a c i e n t s l e e e o r i g n a l i n e r d i t c h d i t c h ; ( b e f o r e e a r l y

t e r e a e s s o h f t u r / p h a e s

. -. - - - ‚

3

( ' _ )

U I

4 J

I C

K

C l.

U )

h

'°w""""'

Si.ACX

SfDEHJRY CAMP

RCXRESirR

IU:ESTER

P.ENtU:RlrnM

NFATHAM

MARGIDJNLM

MANO-IESIER

site

rrajor gate on S; W gate in stone excavated and wall abutted gate on both sides; no dating

8S'30Ciated features/relationship to ¼011, etc.

9 1:astions identified; one on SE excavated and bonded into ¼011

1:astions/relationship to ¼011

no independent dating

.-Jating

double ditch systan with COilllterscarp; undated

other phases

.i:,-

f--'



WIO