<<

otesvnhcnuy(alnr20;Plie 00 atnl 06 2–0.Frmost For 325–50). 2006: dated Mattingly commonly 2000; Palliser now 2000; is (Faulkner century in seventh use the imperial proto-urban Roman rather the to of the of towns’ survive re-emergence not analysis the did in towns Subsequent and that ‘life support. view withdrawal, general some is, the had to that led has however, towns, has, life’, evidence ‘town Roman salient former economically of an concept than occupation claimed Winchester 408–21) non-urban with influential (1995: limited sites, Wacher’s an of of continuity. shifting number put Romano-British/English a mainly 103–12) demonstrating at (1976: opinion functions as place’ Biddle with ‘central century. continuing years, seventh for until case 50 the activity urban past in low-level takeover continuing the and Anglo-Saxon over decay gradual oscillated abandonment, continuation of speedy possible have discussions between the to towns about central Debates the been Roman society. has of post-Roman of towns of impacts Roman nature key of the fate the and empire the continuity of an Similarly, was 97). one Empire Roman (2013: as “[t]he cities” argues: regarded of Cleary Esmonde been as Britain, often of has conquest Roman urbanism of introduction The Introduction Britain Roman Lane Alan of end the and Wroxeter * eierna imports Mediterranean Keywords: ANTIQUITY  C niut ulctosLtd. Publications Antiquity eateto rhelg,CrifUiest,CrifC1 E,UK 3EU, CF10 Cardiff University, Cardiff Archaeology, of Department 8(04:5155http://antiquity.ac.uk/ant/088/ant0880501.htm 501–515 (2014): 88 oa rti,Woee,Cli et nl-ao,ubns,chronology, urbanism, Anglo-Saxon, West, Celtic Wroxeter, Britain, Roman ∗ 501 ntelt orho al fhcnuyAD. century fifth early or fourth largely late towns, been the Roman in centre have other with may urban along it abandoned, an that as suggests Wroxeter and of west survival the and challengestheevidenceforthepost-Roman in discoveries coupled new Britain Roman with Late evidence, on More work claims. recent the these on doubt of throws however, re-examination Careful chronology. years accepted 200 traditionally to the up beyond AD, century seventh or the sixth into continued that post-Roman activity urban of a phase suggesting evidence 1966–1990 produced from Roman by Wroxeter at conducted in Barker excavations Philip life The urban end? did Britain how and When

Research Wroxeter and the end of of the ‘Celtic West’ there is no urban life before the Viking period or, in many areas, before the Norman Conquest (Astill 2000: 44). A few sites—/St Albans, Canterbury, York—have continued to be debated as evidence of possible continuous low-level urban activity between the late fourth and mid/late seventh century. A key site for the discussion of urban life in the British west is the RomantownofWroxeterintheWestMidlands,thecivitas capital of the British tribe of the . Since the 1960s, Wroxeter has been cited as a classic excavation demonstrating major building activity post AD 400 in a Roman town and indeed the continuing existence of urban life well into the sixth or even seventh century (Barker et al. 1997; White & Barker 1998: 118–36). Philip Barker’s work at the Baths Basilica at Wroxeter was one of the most influential excavations of the later twentieth century in Britain. Excavated between 1966 and 1990, and published in 1997, it set the standard for several generations of archaeologists as well as being immensely influential for the early medieval period in western Britain. Barker’s 1977 book on Techniques of archaeological excavation drew heavily on the Wroxeter work as well as his other excavations and consequently had an impact much wider than specialists interested in Roman towns or post-Roman archaeology—“the bible for the new generation of excavators” (Collis 2011: 83). From the early 1970s, when the first interim statements of the scale of the Wroxeter discoveries were published, it provided a paradigm for urban excavations and a standard of work, and indeed expectation of discovery, for sites of many periods. In recent decades it has been central to a wider model of fifth-century and later western Britain. ‘Late Antique Britain’ as announced by Ken Dark in 1994 and Britannia Prima by Roger White (2007) depend on the Wroxeter Baths Basilica excavations as their central exemplar. As Ian Wood (2003: 429) says in the recent A companion to Roman Britain, “[t]he archaeology of sub-Roman Britain was effectively revolutionized by work carried out at Wroxeter from 1966 to 1994”. Most scholars accepted the Wroxeter model and indeed considerable effort has been expended trying to replicate it elsewhere (Carver 1987: 40–46). However, aspects of the Wroxeter claims have met with some criticism in the past, in particular the claim that the excavations demonstrated the continuation of town life in Britain into the sixth or seventh century (Ward-Perkins 1996: 9–10), and by implication that this evidence had been missed by previous excavators on other town sites in Britain and elsewhere in Europe; but the basic claims have not been directly challenged in print (Loseby 2000: 331–36). The failure of other excavators to replicate the scale of the Wroxeter evidence in subsequent urban excavations also led to it being regarded as a strange western English anomaly without any general implication for urbanism elsewhere in England (Faulkner & Reece 2002) or to the acceptance that it was the seat of a magnate or bishop, but not urban (Halsall 2007: 359). In 2002, however, a short but important review by Professor Michael Fulford questioned key aspects of the claimed Wroxeter sequence. Fulford noted the difficulties of the excavation with numerous robber trenches, earlier excavation trenches and other disturbances which isolated the stratigraphy into separate islands and suggested that much of the rubble of the major penultimate phase (Z) was the result of late Saxon stone robbing rather than being the foundation platform of a massive Romanised timber- framed building. While accepting that some structural evidence indicated timber buildings C Antiquity Publications Ltd.

502 tl oa reat otne oarv ntest ni 9–5.Tescn issue second site. The the 490–550. on structures until of site identification the the and fourth-century- sequence on the that arrive of claim the reliability to the from the and continued concerns objects 400–700 any artefacts period of Roman the of absence number to style complete a dating the are need least There not which 700. excavations claim, as sixth this late the as accepting in sixth dismantling in function Romano- deliberate and to difficulties continued its a fifth which until the town as centuries timber into function seventh a well by and to replaced buildings, was continued it masonry that it Roman and claim major centuries they with initially particular town, In British 1998). Barker & White “ 600– of burial a than earlier 650/700, after post-Z, (phase 790). effects and years. furniture 75 portable least at for lasted which 530/80–650/700) Z, (phase buildings two-storey 10 Building fac three-storey towered massive a the including with platforms, rubble on timber in reconstructed 500/550–530/580). Y, (phase market open an 367–480/530). as AD then W, and (phase century sixth early to fifth late the in time its evaluate Barker by to proposed order was in Wroxeter sequence at following 240–41): The sequence (1997: claims. claimed structural the and understand chronology to important is It sequence Basilica Baths The activity Wroxeter the of scale the disputed radically review Fulford’s 2002). 400, (Fulford AD post of nw hr olct h uidaha lcste eurdfrcuc ulig nhis In building. church for have required not would they robbers blocks stone ashlar Saxon buried late the rubble, basilica locate by the to believed, sealed he where and as known Y this if, that phase that aware in trench was suggested a Barker dismantled in colonnades. He was strap-end nave Trewhiddle-style structure. Saxon the ninth-century of late a basilica one of of out masonry find robbing result the the the the by dated of actually from be were remnants could evidence town) activity the timber isolated of the by phase robbing for the dated evidence stone that (the was be suggestion spreads could main rubble Fulford’s phases questioned. Z be different can identical. areas the not stratigraphic were that separate deposits the claim though even the 1) (Figure Thus stratigraphy of areas isolated these of uha ubesras.Bre asta lot5 e eto h iehdbe otto lost been had site the of cent per 50 almost that says (Barker contexts, disturbances Barker interpretation ‘matching’ previous correlating spreads”. spatial by rubble the stratigraphy as of that islands such result numerous the of linking with the disturbances involves other and trenches excavators’ ... ufr’ 20:634)sotrve oe h ifiute ae yBre’ team: Barker’s by faced difficulties the noted review short 643–44) (2002: Fulford’s (Barker Wroxeter about claims unique of number a make White and Barker r r r r h tairpyo h ah aiiaadisevrn a u yrbe rnhs early trenches, robber by cut was environs its and Basilica Baths the of stratigraphy the h niest a hncrflydmlse ihtermvlo agrtimbers, larger of removal the with demolished carefully then was site entire was The centre town entire the and demolished was structure basilica remaining 480/530–530/580) AD The X, (phase structures flimsy for used was but ruinous became It some until roofed and standing remained structure, masonry huge a Basilica, Baths The aeadaohr3 ibrfae ulig nldn te substantial other including buildings timber-framed 36 another and ¸ade tal et 97 ) h iespaepasrqie h correlation the required plans phase site’s The 7). 1997: . lnLane Alan 503  C niut ulctosLtd. Publications Antiquity tal et 1997; . tal. et

Research Wroxeter and the end of Roman Britain 1997: plan A1). et al. Figure 1. Plan showing the amount of destruction of the Baths Basilica site before Barker’s excavations (after Barker

C Antiquity Publications Ltd.

504 eulig fpaeZ snl uepafr frbl adoe costetpo the foundation the of as top interpreted structure the was basilica timber-framed across the hardcore’ vast of rubble a portico of north of and platform aisle huge north ‘single nave, A former Z. phase of Rebuilding’ Roskams late the represent indeed been may had and it dating envisaged. himself firm that Barker grave no which decided have the robbing subsequently spreads stone that rubble only notes the was report claims, it (Barker The Fulford platform basilica. and the the recognised through of initially cut end not western was the cut towards lay burial male period. Saxon quem late ante the terminus in visible been thus and Z phase in standing remained otoe n otpd a oe otenrh in of north, sequence the structural to noted convincing, rubble was more of pads perhaps recognition post and pattern and different, by postholes rather identified were a Wroxeter though at found spreads, buildings the of Most buildings The White though this, discuss not does report final The walls. find in 183) to (2007: order in site the across in report interim 1975 ubewsfudimdaeyudrtepog ol akrsae htakyfco that factor key a archaeological or that trenches emptied: robber states been any before Barker had observed trenches was soil. it that plough was This recognition the 114). its in under 1975: helped (Barker immediately century” found seventeenth inspired was the classically rubble before last Britain the in 2)—“among constructed (Figure buildings building towered three-storey rectangular utcs ob nisceiiiy h eotsget htte‘ltom a lae u to due clearer was ‘platform’ the robbers?) that stone suggests (by report order was The correct credibility. ‘structure’ the its in the on backfilled doubt of of been removal cast had much before must these visible how was that clear it and that disturbances immediately claim the The is excavation. Barker’s It before A131). interpretation destroyed the already these and from shows lifted of A130 3 structure together Figure the (A129, of evidence. piecing outline plans primary the the the with requires see and it to digitally and A125) together stitched is and rubble A124 evidence the (A123, The of the plans 175–76). drawings large depict and three 2011: actual which drawings the White drawings reconstruction in & A3 (Everill presented the loose-leaf only used at 177 presentation Barker the looked the method than recording have by rather primary readers difficult plans made most phase is suspect oft-repeated structure the I supposed evidence. this the of of nature the of Evaluation ufr’ te e bevto a oqeto h vdneta uilpoie a provided burial a that evidence the question to was observation key other Fulford’s h ubepafr fBidn 0teeiec eoe icniuu n fragmented and discontinuous becomes evidence the [ 10 Building more of looked platform were platforms rubble rubble really the the they so than order, complete right their and in homogeneous or immediately, roughly less trenches back or robber put more whether layers backfilled the most, and with because emptied been rubble had the trenches, of archaeological surface early the on visible were “Few ... ] ” es comm pers. (Barker rtni Prima Britannia tal. et o h ubepafrsfrtebidns hsnrhsuhorientated north/south This buildings. the for platforms rubble the for ) ti ulig1,hwvr hc stekysrcuefrte‘Great the for structure key the is which however, 10, Building is It .). Britannia 97 141) 1997: ossgetta h otenaseclnaemgthave might colonnade aisle southern the that suggest does ehpteie htte ih aedgtiltrenches trial dug have might they that hypothesised he tal et c . 33.5m . 97 6–8.I hsrltosi si ob,as doubt, in is relationship this If 167–68). 1997: . lnLane Alan 505 × 55 n eosrce ndaig sa as drawings in reconstructed and 15.5m [ ... insula ] hnteeaesbrce from subtracted are these When . 2(Barker  C niut ulctosLtd. Publications Antiquity tal et 97 23–24; 1997: .

Research Wroxeter and the end of Roman Britain

Figure 2. Reconstruction drawing of the key timber-framed Building 10 (Barker et al. 1997: fig. 252). its packing with coloured plaster and mortar from a demolished structure but unfortunately this evidence was not presented in the report (Barker et al. 1997: 141; cf. Everill & White 2011: 176–78). It is important to understand that the only evidence for structure 10 is the rubble platform. The reconstruction drawings are entirely hypothetical. Nothing remained of the supposedly elaborate superstructure because it was entirely removed with all the other timber structures when the whole site was demolished—“the strong impression is of systematic abandonment of the site, with the removal of all the larger timbers and the portable furniture and effects” (Barker et al. 1997: 241). None of these buildings had floors with the partial exception of the unique stone structure Building 31 which had a small area of internal floor surface (Barker et al. 1997: 146). Many were surmised to have had suspended timber floors (Barker et al. 1997: 169–91). Domestic hearths were absent too and so braziers were suggested for domestic heating. Nor was any evidence of roofing found; again it has to have either been removed or have been of perishable organic material. Structure 10 and the whole timber town hypothesis rely on the rubble platforms being convincing.

The dating of Wroxeter The other key problem with the claims made for Wroxeter is the total absence of any artefacts (apart from the late Saxon strap-end) which need to date later than c. AD 400. Whatever the nature of the Wroxeter buildings the dating of the late phases is clearly crucial to their C Antiquity Publications Ltd.

506 lnLane Alan 507  C niut ulctosLtd. Publications Antiquity

Figure 3. The rubble spreads of Building 10 with the outline of the supposed structure superimposed (after Barker et al. 1997: plans A123, A124, A125, A129, A130 & A131).

Research Wroxeter and the end of Roman Britain wider significance in discussions of early post-Roman Britain. The single archaeomagnetic date for phase X is central to the absolute dates presented by Barker et al. (1997). This date was revised at a late stage in the writing up from AD 400−+100 to AD 500–550 (Barker et al. 1997: 103). This is clear from the published text and implicit in the discrepancies between the general plans for phases X, Y and Z which cite dates of c. 410–550, whereas the text cites dates of c. 480–700 (Barker et al. 1997: cf. figs. A9, A10 and A11 with pages 240–41). The date of each of the ‘post-Roman phases’ in the text has been moved later by 50 to 100 years (or more) after the revision of the archaeomagnetic date, whereas the phase plans contain the original dating. Unfortunately there are reasons for rejecting the archaeomagnetic date. Mid to late first millennium AD dates are notoriously imprecise due to the scarcity of dates and a plateau effect (equivalent to the flat parts of the radiocarbon calibration curve) and an optimistic view of the date would span the period AD 470–980 (Cathy Batt pers. comm.). However, as the report states, the “material was only moderately stable magnetically, and only five of the 12 samples taken were sufficiently well grouped to provide a reasonable result” (Clark in Barker et al. 1997: 103). Paul Linford (pers. comm.) of English Heritage has stated that the date cannot be relied on as it is “based on only five samples and the statistical calculation is only strictly valid provided eight or more are used”. Both he and Cathy Batt (pers. comm.) have indicated that this would not now be seen as sufficient for a reliable date. Its alpha-95 reading of 4.9 degrees is outside the range of modern acceptable error. In consequence, the archaeomagnetic date for the oven attributed to phase X needs to be set aside in any consideration of the site chronology. The other absolute dates for the Baths Basilica sequence are provided by 10 radiocarbon dates. These consist of one date for a human burial, three for fragments of human skulls and six charcoal dates, including one split sample. These samples were all excavated in the 1970s and processed in the early 1980s and have various issues with their collection and processing which may limit their value. The skull dates span the period AD 70–530 at 2 σ and occur in Barker phases W, Y and pre-Z. The excavators regarded these as re-deposited cult objects and therefore of little dating value. A pooled mean for the skulls would be AD 210–410 at 2 σ and support the view that they are Roman. One split charcoal sample from phase W calibrates at 170 BC to AD 130 and is dismissed as old wood possibly from major roofing timbers of an earlier phase (Barker et al. 1997: 89). Two dates come from phase Z. These are described in the report as charred beams and have dates of 190 BC–AD 330 and AD 20–410, which are again dismissed as old wood from residual structural timbers (Barker et al. 1997: 168). The Harwell laboratory identifications, however, suggest these timbers are hazel and poplar, which are relatively short-lived trees. The hazel charcoal was adjacent to the phase X oven and may represent rake-out from the oven (Alex Bayliss pers. comm.) and hence provides a useful date for that structure (AD 20–410). Three further dates come from wood charcoal from a hearth, a dump and a surface. These all have dates at 2 σ between AD 390 and 660. Unfortunately, no wood identifications are available, which might indicate whether there is any old wood in the samples. Two of the radiocarbon certificates indicate they were hand sorted from bulk samples. Although these three dates are attributed by the excavators to three different phases, W, X and Y, they do indicate the possibility of some activity in the late fifth to early sixth century. The last date is for a human burial which C Antiquity Publications Ltd.

508 07 6.Salqatte of quantities Small 26). 2007: emt aerahdBiani arynro iezn from zone time amphorae narrow Roman fairly (Campbell Late a established and in ware firmly Britain Slip been reached Red have chronology African to (PRS), its seem ware has detail Slip increasing recently Red in only studied Phocaean been 2007). but has material 1930s This sites. the of claim since number confidence a with ceramics at his Mediterranean can identified began imported we been of Barker have where examples West’ since because ‘Celtic activity decades the sixth-century the and of in later-fifth- of areas archaeology transformed are the There been Wroxeter. of has at knowledge Britain work Our western areas. of nearby excavation of its parts and archaeology some site the the from within also from merely but not report examined be to need claims Wroxeter The west the from viewed Wroxeter 2 a has This Z. phase their post-dated believed excavators the ugt 00.TeodrecvtosrcvrdRmnmtra n al medieval early and material Roman recovered excavations Arnold 97–98; older 1988: Lane The iron & 2000). structures, (Edwards the of timber activity Huggett in site industrial of excavations & of evidence enclosed Limited indications found small ditch. and (1996–2000) a clay and 1990s fired is bank the slag, a It in by again Breiddin. then defended the and acres), of 1930s (1.2 hillfort hectare Age a Iron half and about Age Bronze Late the great in River identified the been up penetration have their sites 4). to two (Figure Wroxeter However, barrier of towns Glamorgan. political Roman vicinity a and the in be Somerset found may denunciations beyond been Gildas’s there have Severn in imports and these glimpsed England of western aristocracy none that of that military sites striking British defended is the It and 2007). enclosed to (Lane identify belong to to us likely allow northern are they and western exceptions of Ireland rest some Wales, the from With northern absent western England. be in Cornwall, to on seem outliers imports centred Such occasional zone Scotland. with southern a and Wales, in that southern densely most sites and lie Somerset identify These Devon, 500. imports late AD to Mediterranean around sixth The use mid in perhaps 54–73). were period, also 2007: same date, glass, the (Campbell in imported late-seventh-century century occur of to to seventh quantities seem late-sixth- Substantial origin, French of 46). Subsequently western 2007: France, of 32). largely (Campbell western 2007: Ireland from and (Campbell again Britain France in ware, western ‘E’ from find arrived we century, sixth mid the oa n osbeps-oa ciiy u ontcnr h ie hsn lie by claimed phasing wider the indicate confirm of not dates excavators. phase do the other the but The to activity, strap-end. belong post-Roman Trewhiddle-style possible may the and and by Roman century dated seventh/eighth burial, robbing the the stone about than Saxon correct later is late Fulford is If rubble sequence. Z proposed phase the confirm the to used be can them of burial the that decided rubble. and excavation the after relationship through only stratigraphic cut was was it the as the reinterpreted burial that excavators the suggested sealed the robbing and that stone interpretation from this rubble Z disputed phase has supposed 645) (2002: Fulford already, noted e icsi oae utoe 0mt h eto rxtro ilsu eiethe beside hill-spur a on Wroxeter of west the to 30km over just located is Pieces New h euto hsrcnieaino h bouedtsfrteBtsBslc sta none that is Basilica Baths the for dates absolute the of reconsideration this of result The d ´ eriv ´ e sigill ees lnLane Alan ´ e pal ees 509 ´ eochr ´ etiennes DP) rbbydtn to dating probably (DSPA), σ c aeo D6070 As 600–780. AD of date  7 o50(Campbell 550 to 475 . C niut ulctosLtd. Publications Antiquity

Research Wroxeter and the end of Roman Britain

Figure 4. Finds of fifth–seventh-century imported pottery and glass in the Celtic West (Ewan Campbell pers. comm.). glass, while the more recent excavation recovered more early medieval glass and pottery. The import assemblage now has nine glass vessels including seven cone beakers and a glass bowl. Seven vessels have white trailed decoration showing them to be Continental imports of C Antiquity Publications Ltd.

510 ntecmoiino h nmlbn sebaebtentedfeetBre phases. Barker different the between assemblage bone Wroxeter animal changes at the slight assemblage noted of pottery has composition (2011) the of Hammon the in Similarly coin in use. changes bronze pottery that stray continuing suggested a some indicate of has find 231–35) (Sims-Williams old (2006: century The Cool sixth 538–39). or 2007: fifth ( Lewis III the Valentinian & in Redknap site the 25–26; on 2002: in connections, road British western the with of north elaborate the less the to for postholes evidence and the (Fulford pads that post suggest to with happy insula those was particular He in 400. buildings, AD after Wroxeter at as maintain to state”. desire Roman and Late culture the material of province their possible the of as of terms much half in eastern Roman Romanised, defiantly the remained in who “living Romans and imports Mediterranean and west” fifth or the east in further of wearable found fashions was dried British Roman or trade anything Germanic (Barker and the if to as dwindled preference almost in stocks is centuries as sixth “it century that sixth suggested They the up”. deposition through the see material we that Wroxeter after fourth-century which outside; from at of material period new 490–550—“the of quantities AD large receive period to presumably the ceased and in use, sometime in continued until material production, Romano-British in that suggest Hitchcock to (e.g. was 400 Scheme AD Antiquities post Barker Portable 1988: 114). the Lane no. by 81, & 2006: in (Edwards Anglo-Saxon reported seventh stray being and the are 60–64), finds in the 2007: Wales, Campbell in 221–23; Somerset, northern 1991: Congresbury, Graham-Campbell in to Cadbury 64–66; is Dinorben, and it Wales and harder western southern century the elite in activity, reaching sixth Powys, the a were Dinas later finds as The be Anglo-Saxon such use. artefacts. significant sites to Anglo-Saxon in of been thought absence have the to is credit said Wenlock be 1989: can Much (Pretty site the at site British site earlier (Campbell The possible vessel French 48). Anglo-Saxon a western & an on 175–78). a of 45 monastery of sherd medieval Anglo-Saxon figs. sherd one from seventh-century 117, recovered and one 54–73, were beaker respectively glass cone 2007: as white-trailed medieval these C early identified group of has sherds Campbell two Wroxeter, deposits. of east the site, the post-Roman of early the exploration in modern and importance importance limited considerable century the of been sixth was reinforce has it centuries. early century, there suggests to sixth imports Although of fifth the site. quantity later the of the the of probably to phase ware), this dated (‘D’ of plate DSPA PRS of (Campbell a sherds centuries of two seventh sherds and two sixth later addition, the In to dated C, Group Campbell ufr’ imsa ftepaeZtme onde o enteewsn activity no was there mean not does town timber Z phase the of dismissal Fulford’s late how about questions serious raises Wroxeter to near imports these of presence The to 11km Priory, Wenlock At vicinity. the in imports with site only the Pieces New is Nor rtni Prima Britannia Cunorix ,myb eun n oprbet h aesrcue epsuae tSilchester at postulates he structures late the to comparable and genuine be may 2, tal et tal et 97 0) hsve a enrsae yWie(07 5)i i account his in 151) (2007: White by restated been has view This 203). 1997: . tn,woeLtnisrpinsest niaeahg-akn rs figure, Irish high-ranking a indicate to seems inscription Latin whose stone, 06.Sm otRmnatvt lehr tWoee sdmntae by demonstrated is Wroxeter at elsewhere activity post-Roman Some 2006). . c D403)hsrcnl encnre Ad ilas20:31). 2006: Williams & (Abdy confirmed been recently has 430–35) AD . hr eevsgsadvso ewe rtoi lt ntews with west the in elite Brittonic a between division a envisages he where tal et s(97 1)rsos oteasneo aebeartefacts dateable of absence the to response 218) (1997: ’s lnLane Alan 511  C niut ulctosLtd. Publications Antiquity es comm. pers. ).

Research Wroxeter and the end of Roman Britain

However, neither of these studies demonstrated any longevity of the site nor showed any serious critique of the stratigraphic claims of the excavators. The absence of diagnostic artefacts is the key problem with the Wroxeter claims. This is not an upland site where poverty and acid soils can be used to explain the absence of finds—3400 Roman coins, a quarter of a million pottery sherds and over 100 000 animal bones were recovered in the Barker excavation (Armour-Chelu 1997; White 2000: 106; Hammon 2011). There is also evidence of bronze- and lead-working and many examples of hearths, although not associated with floors, in the hypothetical buildings. A comparison with the assemblages of bone, fine metal-working and other small finds from sites such as Dinas Powys (Alcock 1963; Campbell 2007: 83–101) shows what is missing from Wroxeter. It is important to understand the intellectual context of the Baths Basilica excavation between 1966 and 1980. Frere had argued for continued fifth-century activity at Verulamium and in Britannia, his history of Roman Britain, he referred to towns being defended into the late fifth century (Frere 1967: 375–77). Biddle was arguing for Roman towns as important central places in the early Anglo-Saxon period. Anglo-Saxon specialists were trying to close the gap between late Saxon burhs and Roman towns and a significant interpretative lobby was minimising the cultural break between Roman Britain and Anglo-Saxon England (Biddle 1976). Few Anglo-Saxonists would now argue for town life before the mid/late seventh century. There is no evidence for towns in most of the Celtic West before the Viking Age and none in Wales before the Normans (Davies 1982: 57–58), so to claim urbanism on the Welsh Marches in the sixth and seventh centuries is to fly in the face of most of the evidence. Barker’s assertion that the Baths Basilica was re-roofed and continued in use until the late fifth to early sixth century (Barker et al. 1997: 224–25, 240) is in stark contrast to the near universal evidence in Britain and elsewhere on the Continent for the earlier abandonment of such massive masonry structures with the decline of competitive aristocratic public building. Nor does the Continental evidence support the idea that large towns survived in adjacent areas of Gaul. Contraction of urban space into smaller defended areas and substantial reduction in complexity and economic activity can be seen on many sites. Moreover, activity on Continental sites is indicated by artefact and structural sequences dateable to the fifth and sixth centuries by continued glass and ceramic production (Wickham 2005: 674–81; Esmonde Cleary 2013: 431–41). Barker’s work at Wroxeter and other sites was both beneficial and extremely influential (Roskams 2001: 13; Collis 2011). It fitted the (now uncontroversial) view that structures of many periods could only be recovered by meticulous excavation and recording. Many scholars thought that medieval deposits had often been destroyed by excavators interested in clearing upper layers to get at solid Roman structures (Collis 2011: 75). Barker was one of the excavators of the 1960s and 1970s who helped make field archaeology professional. He was a charismatic figure who had a profound effect on those who knew and worked with him (Everill & White 2011). He was also an artist and he knew that reconstruction drawings could mislead the unwary: “[o]ne problem which arises from the publication of such a drawing is that it is seized on by authors who want to illustrate their books on Late Roman Britain, often without the very strong reservations which ought to accompany a piece of kite-flying like this” (Barker 1986: 172). The interpretation C Antiquity Publications Ltd.

512 eoe osbet icr opeeya nepeainwihde o tteemerging the fit it not that does so which interpretation, interpretation its be an to and completely evidence”. of is evidence discard city’ thing the to important ‘ghost between “the possible the wrote, difference becomes that himself the probable 147) time of (1986: and is aware Barker late-fourth- As it continually the rest. Perhaps claim to than be. to put other reason to is no Wroxeter anything appear seems are they There chronology. coins material or and earlier-fifth-century phasing pottery site Roman the bolster the to W,that little phases through do spread Y contexts and stratigraphic their X idea but the activity support raise century may must fifth/sixth dates vicinity some radiocarbon of the The sequence. in Wroxeter activity the about fifth/sixth-century questions dateable serious of evidence emerging the Roman and White’s Late Roger of to life fundamental the also is extend Prima It to 2010). Speed trying sixth 2010; are and Rogers fifth scholars (e.g. the when sequences in quoted society frequently Antique is Late It flourishing a centuries. was Britain western that claim 2000) 324). fig. akto otfcstme on(iue5 c.Bre 96 g 0;Barker 90b; open fig. 1986: busy Barker presents, its (cf. excavation buildings, 5) the timber (Figure town elaborate evidence timber of the soft-focus drawings on or reconstruction market now seductive judged its be on to not needs however, Wroxeter, of (Barker Wroxeter at town timber Z phase The 5. Figure rxtrsubnsau n t asv eta uligaecuilt e aks(1994, Dark’s Ken to crucial are building central massive its and status urban Wroxeter’s oe.Hwvr h alr orpiaetecam fteBtsBslc elsewhere Basilica Baths the of claims the replicate to failure the However, model. lnLane Alan tal. et 513 97 g 324). fig. 1997:  C niut ulctosLtd. Publications Antiquity tal et Britannia 1997: .

Research Wroxeter and the end of Roman Britain Note The full outline of Building 10 is only shown in the overall plan of phase Z (Barker et al. 1997: plan A11). The building outline in Figure 3 is taken from the dotted interpretation in Barker et al. (1997) plans A129, A130 and A131.

Acknowledgements This paper has been in gestation for several decades. I would like to thank Cathy Batt, Paul Linford, Alex Bayliss, Ewan Campbell, Steve Roskams, Andy Hammon and Richard Reece for their comments. Roger White very generously discussed the evidence with me, though he would still radically disagree with the views cited here. The opinions expressed here are my own and I realise they may be seen as heretical by those who knew and worked with Philip Barker. They are not intended to diminish his many significant contributions to British archaeology.

References CARVER, M. 1987. Underneath English towns, interpreting English towns. London: Batsford. ABDY,R.&G.WILLIAMS. 2006. A catalogue of hoards and single finds from the British Isles, c.AD COLLIS, J. 2011. The urban revolution: Martin Biddle’s 410–675, in B. Cook & G. Williams (ed.) Coinage excavations in Winchester, 1961–1971, in J. and history in the North Sea world, c. AD 500–1200: Schofield (ed.) Great excavations: shaping the essays in honour of Marion Archibald: 11–73. archaeological profession: 74–86. Oxford: Oxbow. Leiden: Brill. COOL, H.E.M. 2006. Eating and drinking in Roman Britain. Cambridge: Cambridge University Press. ALCOCK, L. 1963. Dinas Powys. Cardiff: University of Wales Press. DARK, K.R. 1994. Civitas to kingdom: British political continuity, 300–800. Leicester: Leicester University ARMOUR-CHELU, M. 1997. Appendix 8: faunal remains, in P. Barker, R. White, K. Pretty, H. Bird & M. Press. Corbishley, The Baths Basilica Wroxeter: excavations – 2000. Britain and the end of the Roman Empire. 1966–90: 350–64. London: English Heritage. Stroud: Tempus. ARNOLD,C.J.&J.W.HUGGETT. 2000. New Pieces, DAVIES, W. 1982. WalesintheearlyMiddleAges. Criggion, Powys. Interim report on the archaeological Leicester: Leicester University Press. investigations 2000. Glasgow: Glasgow University. EDWARDS,N.&A.LANE (ed.). 1988. Early medieval ASTILL, G. 2000. General survey 600–1300, in D.M. settlements in Wales: a critical reassessment and Palliser (ed.) The Cambridge urban history of gazetteer of the archaeological evidence for secular Britain. Volume 1: 600–1540: 27–49. Cambridge: settlements in Wales. Bangor: Research Centre Wales, Cambridge University Press. University College of North Wales; Cardiff: Department of Archaeology, University College BARKER, P. 1975. Excavations on the site of the Baths Basilica at Wroxeter 1966–1974: an interim report. Cardiff. Britannia 6: 106–17. http://dx.doi.org/10. ESMONDE CLEARY, A.S. 2013. The Roman West AD 2307/525992 200–500: an archaeological study. Cambridge: – 1977. Techniques of archaeological excavation. London: Cambridge University Press. Batsford. EVERILL,P.&R.WHITE. 2011. Philip Barker’s – 1986. Understanding archaeological excavation. Wroxeter, in J. Schofield (ed.) Great excavations: London: Batsford. shaping the archaeological profession: 167–80. Oxford: Oxbow. BARKER,P.,R.WHITE,K.PRETTY,H.BIRD &M. FAULKNER, N. 2000. The decline and fall of Roman CORBISHLEY. 1997. The Baths Basilica Wroxeter: excavations 1966–90. London: English Heritage. Britain. Stroud: Tempus. FAULKNER,N.&R.REECE. 2002. The debate about the BIDDLE, M. 1976. Towns, in D.M. Wilson (ed.) The archaeology of Anglo-Saxon England: 99–150. end: a review of evidence and methods. The London: Methuen. Archaeological Journal 159: 59–76. FRERE, S.S. 1967. Britannia: a history of Roman Britain. CAMPBELL, E. 2007. Continental and Mediterranean imports to Atlantic Britain and Ireland, AD 400–800 London: Routledge & Kegan Paul. (CBA Research Reports 157). York: Council for FULFORD, M. 2002. Wroxeter: legionary fortress, baths, British Archaeology. and the ‘great rebuilding’ of c. AD 450–550. Journal of Roman Archaeology 15: 639–45.

C Antiquity Publications Ltd.

514 L L F P P H H M G H eevd eebr21;Acpe:1 eray21;Rvsd 1Otbr2013 October 11 Revised: 2013; February 12 Accepted: 2012; December 3 Received: ALLISER ULFORD OSEBY RETTY ANE RAHAM ITCHCOCK AMMON ALSALL ATTINGLY Sport. 2004 xod Wiley-Blackwell. Oxford: literature Arthurian (ed.) Fulton for H. context in archaeological ‘Arthur’?, an Saxons: the of coming abig nvriyPress. University Cambridge rti.Vlm :600–1540 1: Volume Britain. (ed.) Palliser D.M. ae&Penguin. & 409. Lane BC–AD 54 Empire, Roman the Studies brooches. zoomorphic penannular enamelled of dating the and Press. 376–568. west oe Lletres. Bones 400–800) (ed.) Gurt J.M. G. & in Ripoll England, Anglo-Saxon early and Britain Press. kingdoms (ed.) Bassett ( Wroxeter from perspective zooarchaeological a transition: medieval Romano-British–early rmto fRmnStudies. Roman of Promotion in excavations Silchester: insula Roman Late in labour and http://dx.doi.org/10.1017/S0068113X11000055 iooimCornoviorum Viroconium .20.Teedo oa rti n the and Britain Roman of end The 2007. A. , ..20.Pwradtwsi aeRoman Late in towns and Power 2000. S.T. , ..18.DfiigteMgnat,i S. in Magonsaete, the Defining 1989. K.B. , .2007. G. , ..20.Teoiiso rts on,in towns, British of origins The 2000. D.M. , -C . .C A. M., , odn eatetfrClue ei and Media Culture, for Department London: . .21.Udrtnigthe Understanding 2011. A. , Xsne1997. since IX 8 220–32. 38: .2006. D. , AMPBELL .(d) 2006. (ed.). F. , 7–3 ecse:LietrUniversity Leicester Leicester: 170–83. : 1–0 acln:RilAcad Reial Barcelona: 319–70. : h rgn fteAnglo-Saxon the of origins The abig:CmrdeUniversity Cambridge Cambridge: LARKE abra irtosadteRoman the and migrations Barbarian .19.DnsPwsmetalwork Powys Dinas 1991. J. , nipra ossin rti in Britain possession: imperial An 52.Mlo M)& (MA) Maldon 15–29. : h abig ra itr of history urban Cambridge The odn oit o the for Society London: &H.E ultno h or fCeltic of Board the of Bulletin raueAna Report Annual Treasure ee eie(ann. Regiae Sedes ). Britannia opno to companion A 72.Cambridge: 17–24. : CKARDT odn Allen London: 2006. . 2 275–305. 42: made emia ` Life lnLane Alan 515 R W W R W W S 2007. – R W W S PEED IMS OSKAMS OGERS EDKNAP ARD ACHER OOD ICKHAM HITE HITE Blackwell. Britain Roman to companion eln,A 100–1600 (ed.) AD Slater decline, T.R. in urbanism, Roman Late tod Tempus. Stroud: Batsford. eaigubns ihnadbyn h al A.D. walls the 300–700 (ed.) beyond Speed and G. within & urbanism Sami Debating D. in Britain, of Roman towns Late the in decline re-conceptualising places: Press. ra vlto nLt niut n h early the Ages and Middle Antiquity Late in evolution urban (ed.) Loseby S.T. & Christie ecse nvriyPress. University Leicester 300–700 A.D. walls the G. & Sami D. (ed.) in Speed towns, post-Roman early in life Studies period. pre-Norman the adf:Uiest fWlsPress. Wales of border. University English Cardiff: the and Wales Wales. south-east in 1: sculpture Volume stone and stones inscribed medieval xodUiest Press. University Oxford 400–800 Mediterranean, the and Europe et faRmncity. Roman a of death nvriyPress. University -W .21.Mn h acaooia)gp tracing gap: (archaeological) the Mind 2010. G. , -P .20.Tefia hs,i .Td (ed.) Todd M. in phase, final The 2003. I. , ,R.H.&P.B of transformation the and Wroxeter 2000. R.H. , .21.Lt oa on smeaningful as towns Roman Late 2010. A. , ILLIAMS rtni rm:Biansls oa province Roman last Britain’s Prima: Britannia ERKINS .1995. J. , ,M.&J.L .2001. S. , .2005. C. , 4 1–36. 44: 78.Lietr ecse University Leicester Leicester: 57–81. : .20.Tefielnugso ae in Wales of languages five The 2002. P. , .19.Ubncniut,i N. in continuity, Urban 1996. B. , –7 leso:Scolar. Aldershot: 4–17. : eaigubns:wti n beyond and within urbanism: Debating on fRmnBritain. Roman of Towns Excavation EWIS ARKER rmn h al ideAges: Middle early the Framing  C 2007. . tod Tempus. Stroud: 619 leso:Ashgate. Aldershot: 96–119. : niut ulctosLtd. Publications Antiquity 1998. . 319 Leicester: 83–109. : abinMdea Celtic Medieval Cambrian abig:Cambridge Cambridge: . 2–2 Oxford: 428–42. : opso early of corpus A on ntransition: in Towns rxtr h ieand life the Wroxeter: London: .Oxford: on in Towns A .

Research