The Early Development of a Canterbury Suburb Romano-British and Medieval Archaeology at Nos 19 and 45-47 Wincheap Helm

THE EARLY DEVELOPMENT OF A CANTERBURY SUBURB? ROMANO-BRITISH AND MEDIEVAL ARCHAEOLOGY AT NOS 19 AND 45-47 WINCHEAP RICHARD HELM AND JAKE WEEKES This report considers the findings of two new archaeological excavations in Wincheap, Canterbury, in light of other recent excavations and studies in the area, especially that at 10 Wincheap (Shand and Hicks 2013). The discussion here focuses on what the combined Romano-British and medieval evidence from these separate interventions is beginning to tell us about the development of a Canterbury suburb. 19 Wincheap ('Captains Cabin) In 2008, the Canterbury Archaeological Trust (CAT) was commissioned to undertake an archaeological excavation at 19 Wincheap Street (TR 14439 57250, centred). The work confirmed the presence of significant archaeological remains already diagnosed through a range of field evaluation and test pitting (Fig. 1) in 2007 (Boden 2007) and 2008 (CAT archives). 1 The natural substrata were seen as ex1ensive mid brown finn silty clay horizons (G41), overlying natural gravels at approximately +13.39m AOD near the road frontage and +13.5 l m AOD to the rear of the site. A lone residual Neolithic scraper and some residual Late Iron Age potsherds from the excavation phase of the work attested to earlier activity on the site, but no layers or features in the area could be securely dated to the pre-Roman period. Activity appears to have begun in earnest in the south-west of the site in the early Roman period, with the excavation oflarge intercutting pits (G 1; ?G8) probably as a means of gravel ex1raction (Fig. 2). These features were backfilled by tips containing pottery, ceramic building materials (CBM), animal bone, fish bone, oyster and mussel shell, and some evidence of local metal working (the significance of which will be seen below). Deposition seems to have concluded by AD 130 at the latest, but it is worth noting that earlier sherds were found amid the tipped material, including Southern Gaulish samian ware (e.g. form Dr.29, dated c. AD 43-85).2 The upper deposits of pits in group G 1 had been cut through by a partially seen shallow linear feature, a gulley or truncated ditch (G3) aligned at right angles to the projected Roman road, and potentially marking part of a plot boundary. Another large pit, seen cutting the natural gravels in the 2008 evaluation trench to north, was probably cognate; its fill was sealed by slumping silty layers which appear to 235 RICHARD HELM AND JAKE WEEKES Fig. l Location plan. tally with the ealiest deposits seen in the north-east area of the site during open area excavation, the remains of an e>..1ensive early soil horizon. The buried soil was up to O. l 2m thick and contained occasional flints and chalk, pottery, burnt flint, animal bone and flecks of carbon, and was cut by a small but intriguing feature (G6). This burial-shaped pit was aligned approximately west/east, 1.4111 by 0. 7m with near vertical sides 0.47m deep and a slightly angled base, and seemed to be lined 236 THE EARLY DEVELOPMENT OF A CANTERBURY SUBURB? NOS 19 AND 45-47 WINCHEAP I - -·- - - - - - I :, Tp4 90/105 Tp3: :, '·-·' Tp5 ',_- - - - - 2m. :, Tp1 ◄ GB ...... - - -, ' ' ' ' · · rp2 100/95 Fig. 2 No. 19 Wincheap. Early Roman large intercutting pits. with a linear band of slightly fire hardened clay. Within, the pit was filled with friable silt containing burnt flint, burnt animal bone (some very possibly human), a copper alloy hairpin with a conical head, broken into three joining pieces (SF l 02), and a ceramic loomweight (SF l 03 ), along with some small scraps of metalworking waste. The size and shape of the feature could suggest a perhaps juvenile inhumation burial where human remains had completely disintegrated, or some other funerary featme, but the pit might equally have been associated with industrial activity. Two undoubtedly residual fragments of human skull, recovered from the backfill of a later cellar just to the south-west, are more likely to reflect Romano-British use of the area for burial, but probably much later in the period (see below). A further soil layer, again extensive but ranging from just 0.0lm to 0.06m thick, sealed the backfill of the indetenninate feature, and yielded an assemblage of small pottery sherds dated between AD 25 and 70-100, along with occasional large and medium sized tile and brick fragments, bumt and worked flints, oyster shell, animal bone and a fragment of quern/lava stone. Silted or backfilled features in the south-west area and the upper soil horizons to the north-east formed the background for a significant change of use for the whole of the north-west area of the site from aound the second quarter of the 237 Tp3 ,- - - - - - - - \ :"i Tp4 ; 90/105 Tp5 :_· - - .. 2m. RICHARD HELM AND JAKE WEEKES , -􀀊 -, Tpl IIJ.II Structural features 100/95 Fig. 3 No. 19 Wincheap. Romano-British smithy (second-century). second century, which also retained clues of earlier activity nearby: a RomanoBritish smithy (Fig. 3). The area, undoubtedly near and possibly adjacent to the Roman road out of Wincheap, contained evidence of multiple phases of timber buildings, clay floors and occupation and dump deposits (G2, G4, G7) and large pits (G5, G9) associated with iron working. Thin spreads of silty clay (G2) above slumping fills of the backfilled gravel e>..'traction pits included charcoal rich lenses and various cultural materials, but most notably iron slag and harnmerscale, the latter typically indicating iron smithing in situ. The presence of large lumps of slag but no intermediate or small fragments demonstrated secondary deposition of material derived from a heap beyond the limits of excavation (McDonnell 2012). The constrnction and use of the timber structure or structures was represented in the excavation by nine post-holes and/or possible remnant post-settings (G4), some containing large lwnps of slag as packing (again pointing to an earlier phase of iron working off site: ibid.); some of the daub and charcoal recovered from the area indicated a traditional architecture of oak, hazel and willow wattles and cereal chaff tempered clay/manure. At least 3m by 2m of the south-west edge of the structure is suggested by the plan, probably an e>..'tension of the existing smithy which may have lain nearer the road frontage and/or to the south-west. The building 238 THE EARLY DEVELOPMENT OF A CANTERBURY SUBURB? NOS 19 AND 45-47 WINCHEAP seems to have been in broad axial alignment with the projected road. Several phases or modifications of this structure might be represented, and the structural elements were also associated with a complicated series of truncated levelling and floor layers and occupation deposits which sadly could not be attributed to overall phases of use. A partially seen feature (G5) at the edge of the excavation, was probably a pit associated with early metalworking activity· its upper fill overlain by later floors in the sequence. Finds from these levels also indicated that other activities were taking place, attested by small dumps of animal bone, oyster shell, CBM and pottery among occupation layers, as well as a copper alloy hairpin with a pine cone shaped head (SF59) from pit G5. To the north-east the metalworking area (G7) incorporated a truncated patch of gravel surface with overlying levelling and dump deposits of clay and burnt daub, clay floors and occupation layers, as well as three further post-holes, a beam slot and a shallow truncated pit. Once again, finds suggest more than just light industry here. The truncated area of metalling incorporated occasional small pottery sherds, CBM, burnt flint, animal bone and oyster shell, and a set of copper alloy tweezers (SF64), and was overlain by layers yielding occasional animal bone and some fish bone, potsherds, small fragments of glass, CBM and stone, copper alloy finds (one a pennanular ring, possibly a child's bracelet; SF6 9; Bevan 2012), several plain tesserae, an iron knife (SF84), nails and unidentified objects, along with further smithing residues. Quantities of earlier pot continued to be present, but it would appear that the structure and associated work surfaces came into use after c .130, with continuing occupation and undoubted use of the area for smithing during the latter half of the second century. The large pit (G9) to the north of the smithing area was also probably associated with this phase of activity. It may have started life as another gravel eA1raction pit but it was backfilled with various tips containing metal working residues along with reasonable quantities of potsherds, tile, brick and daub, animal bone and shell. This phase of activity produced the most CBM, nearly all of which was of local (probably Canterbury) manufacture and all of which was heavily abraded and perhaps on its third use at least (Pringle 2012). Animal bone deposits were most common with this phase, probably representing small-scale domestic cooking, with sheep and goat predominating, including two neonates (Jones 2012). The pottery assemblage once more produced interesting early wares (in this case including fragments of a Durotrigan vessel from Dorset dated AD 0- 90); many of the sherds were again abraded, however, and later sherds indicated that at least some of this material was deposited during the last quarter of the second or early third century. Around AD 200 the first in a sequence of extensive metalled surfaces (G 10- 13) was laid down across much of the north-west corner of the site and beyond (Fig. 4). The first surface (Gl 0) survived only as scattered remnants, the largest area covered being approximately 16m2 in extent. If these disparate patches were indeed all remnants of the same surface, however, it could originally have been at least 50m2 square metres in eJ,.1:ent. Small patches of occupation deposits and repairs preceded the laying of construction and use of the second surface (G 11), again most clearly seen on the largest continuous metalled area the north-west corner of the site. A number of thin and truncated clay floor, occupation and ashy 23 9 RICHARD HELM AND JAKE WEEKES - ·, Tp4 Tpl I + 90/105 Tp2 Tp3' ; Tp5 :_- - - .. - - - _1 2m 100/'95 Fig. 4 No. 19 Wincheap. Sequence of ex1ensive metalled surfaces laid down c.200-300. deposits on this surface testified to new, albeit temporary, structures, perhaps indicating a continuation of metalworking activity. The occupation deposits, generally not more than 0. lm thick, incorporated carbon rich lenses and chalk as well as CBM, potsherds, burnt flint, animal, bird and fish bone, iron nails and metalworking residue and a copper alloy stud. The metallings themselves were typically composed of well sorted and compacted small and medium-sized flints in a silty clay matrix, although pottery, CBM daub, animal bone and oyster shell were also included.Two further re-surfacings of the same area followed (Gl2-13), with commensurate thin occupation deposits containing material suggesting that this phase of activity was concluding by c.AD 250 or at the latest 300. That some activity centred on or very near the site in the later Roman period is attested by Alice Holt potsherds (probably post- AD 250) from deposits sealing the uppermost metallings. Also, fairly fresh residual sherds from the earliest features to disturb this area proved to be from a convex-sided dish in Late Roman grogtempered ware, unlikely to be earlier than c.350 in date (Lyne 1994 fig. 52, 7B.10) and perhaps as late as the early 5th century. Centuries later, the first feature to disturb Roman levels was a narrow medieval 240 THE EARLY DEVELOPMENT OF A CANTERBURY SUBURB? NOS 19 AND 45-47 WINCHEAP :- - - - - - - - ' • Tp4 90/105 Tp3 ', :, 2m. 100/95 Fig. 5 No. 19 Wincheap. Medieval features. ditch (014), which had been recut at least once on a similar alignment (Fig. 5). The latest pottery from this feature dated to c.1150-1250. Some 3 .3m long and 0.5m wide, with steep sides to an angled base 0.55m deep, the ditch was also seen in testpits ne:,,,.1 to Wincheap Street at the north-west limit of the site, cutting metalling deposits and producing further pottery of c.1150-1250. Animal bone and oyster shell assigned to this context during excavation could just as easily have been dislodged from disturbance of the earlier surfaces. Interestingly, both the initial ditch and its re-cut appeared to be tenninating or fully tenninal within the area of excavation, and may represent one side of a causewayed entrance. The upper deposits of these features were overlain by a generalised build up of loamy soils (022) in this comer of the site, up to 0.35m thick and producing further pottery dated c.1125-1250, along with residual and probably residual CBM, burnt and worked flint, animal bone and oyster shell, and an iron nail.Cutting this horizon were two probable pits and a small post-hole (023; one pit with potte1y dated c.1175-1225/50) which were themselves cut by another ditch (024), similarly aligned approximately with Wincheap Street and continuing beyond the north-eastern limit of excavation (heavily tnmcated to the south-west). As seen this was 3.78m long and 1.25m wide, with a less acute profile than the earlier 241 RICHARD HELM AND JAKE WEEKES ditches, but again filled by clayey silts containing much gravel, probably eroded from the metalled surfaces through which it cut. The primary silt in this feature produced occasional potsherds of Roman and again early medieval date (c.1150- 1250), CBM, burnt flint and animal bone and oyster shell, along with some fish bone, slag, window glass and an illegible copper alloy coin (probably late Roman; SF22), and unidentified iron and copper alloy objects. Further residual pottery dated to c .1050-1150, found sparsely distributed in a nwnber of later features across the site, also testified to an increase in activity at this time. Later medieval and early Tudor occupation was in evidence in the same northwest area of the site, but generally heavily disturbed by subsequent street frontage buildings. This earlier period had included masonry structures (G 19); a section of chalk and flint wall footing observed at the limit of excavation, incorporated into a later cellar, probably dates to the medieval period. To the north-east, it would appear that the the Romano-British metalled surface had prevented deeper cuts during the medieval period, and had in fact been utilised as a solid foundation. Nos 45-47 Wincheap Archaeological evaluation and excavation was carried out by the CAT at 45-47 Wincheap (TR 14363 57175, centred) between February and May 2012 (Tasker and Helm 2012) (Fig. 1). River Terrace Gravels (G l ) were identified in Area A at a depth of 1.54m (12.70m AOD) below the existing ground surface , in Trench 2 at 1.91m (11.43m AOD) below the existing ground surface, and in Trench 6 at 1.65m (11.36m AOD) below the existing ground surface level. While a residual prehistoric worked flint typically signified a likely earlier presence, the earliest features identified within the site were dated to the Roman period. In Area A, a sequence of nine deposits (G2), were interpreted as internal surfaces and features within a Romano-British building (Fig. 6). The deposits were only seen in section, and were not excavated. The earliest, a dark silty clay up to 0.04m thick containing flecks of chalk and carbon, was sealed by a layer of light yellow grey crushed lime plaster, 0.05m thick. l11is probable bedding deposit preceded an alternating sequence of compacted clay floors and dark silty clay occupation deposits. An area of intense burning above the final floor may have been a remnant hearth. No dating evidence was recovered from these contex1s, but an overlying soil horizon (G4) contained early Roman pottery giving a broad terminus ante quern of c.AD 150-300. Two cut features (G3), possibly pits, were located at the southern end of Area A. Both had been truncated by modem intrusions, so their shape and full ex1ent could not be determined; neither pit was excavated and no finds were retrieved. One of the pits cut the potential hearth that formed the latest activity \-vithin building structure (G2), and was sealed by the soil horizon (G4). None of these deposits were excavated, and their ex1ents were only seen in section to comprise mainly silty clays, between 0.03-0.20m thick, although a dark reddish-brown silty area, potentially representing in situ burning was noted. Dating evidence for the sequence was limited to three fragments of pottery, dated to c.AD 175-300. In Area B, a sequence of four deposits (G5) was interpreted as internal surfaces 242 THE EARLYDEVELOPl\.1ENT OF A CANTERBURY SUBURB? NOS 19 AND 45-47 WINCHEAP L--- - --- 2m. , , i, ' \ ! I -·-- - '-' - -·- ----l - -· Trench 2 I L - -·-·- __ 1 I .. ... - -- I I loo------ AREA A AREAS - -·- - ----- --- -·--- I -------·-· 1 1 III.Jm, '· g􀀍 ---·, ______ J ..:-i.. __ ,,..,.,...-:-:--:--􀀚---- - · G5 ----- -􀀌 ----- --- =:=ir =:= t-;- -- - - --- - - - - - -'􀀓-- ___ )_/ Fig. 6 Nos 45-47 Wincheap. Phase l features. 243 RICHARD HELM AND JAKE WEEKES within another building. No evidence for structural features associated with the building was identified. The earliest deposit, an internal clay floor, contained one fragment of pottery dated to c.AD 50-250. The floor was sealed by a layer of offwhite lime plaster mixed with silt and occasional small flints, between 0.03-0.04m thick. This may have represented a plaster floor or collapsed walVceiling plaster associated with the abandonment of the strncture. A metalled surface sealed the plaster layer, made up of compacted flint in a silty clay matrix, with rare patches of mortar, and yielding three fragments of pottery (c.AD 150-300), one from a South Gaulish samian dish, and a fragment of imbrex. This was sealed by a dark grey brown silty clay deposit, up to 0.06m thick, interpreted as an occupation surface associated with the metalled floor. The upper surface of Roman archaeology lay at between 12.26-12.3 lm AOD. Notable residual Roman period finds from the site included coins (one a dupondius of Trajan, AD 98-117 (SF 1), another heavily corroded and illegible (SF 4), but dated to c.AD 260-380), three glass fragments and an amphora sherd. A dark grey silty clay loam (G6) containing small flecks of carbon and common chalk, up to 0. l l m thick, sealed G8, and likely fonned following abandonment. A small assemblage of six pottery sherds recovered from this deposit dated to c.250- 400, along with a fragment of imbrex. Sealing late Roman soils in Area A, a further series of deposits (G7; Fig. 7) formed a 0.53m thick horizon overlying the securely Roman phased material. Earliest in this sequence was a dark green grey silty clay layer, containing flint and carbon, and occasional chalk flecks and sandy clay, up to 0.28 m thick. This was sealed by a dark grey silty clay with common flint and chalk and occasional small carbon and shell fragments. The soils appear to represent a considerable period where activity in this vicinity was limited to agriculture or horticulture. Similar soil sequences, 0.87m and 0.5m thick respectively, were observed sealing natural deposits in outlying Evaluation Trenches 2 (G 16) and 6 (Gl7) to the southeast and north of Area A. Elements of a medieval building (G8) were exposed in a north-west facing section in Area A, though only partially seen. These consisted of a construction cut truncating the post-Roman soil horizon, and a compacted silty and flinty clay layer, up to 0.06m thick, which formed a bedding for a chalk and flint wall foundation. The foundation survived to a height of 0.19m, with a length of 3.06111 exposed, and appeared to be aligned approximately north-east to south-west. Remnants of a mid brown yellow clay sand deposit, up to 0.05m thick, partly overlay the foundation, perhaps representing the remnants of a floor extending beyond the limit of excavation. An area of metalling (G9) was located in the south-west comer of Area A, again overlying the post-Roman soil. The earliest metalling was fonned of a dark grey clay silt containing abundant small to medium flint nodules, up to 0 .2 l m thick. This was sealed by a dark compacted silty clay, containing occasional small flint and chalk flecks, up to 0.16111 thick, which was in tum superseded by a two further metallings, presumably resurfacing events, up to O. 07m and O. 9m thick respectively. The metalled surfaces survived in a roughly rectangular area (approximately 2.27 x 1.23m), and were abutted by a sequence of clay floors (G 10) on their north-east 244 THE EARLYDEVELOPl\.1ENT OF A CANTERBURY SUBURB? NOS 19 AND 45-47 WINCHEAP L.- - G11 2m. I/ I /i ·--'"t--􀀐-i"'􀀑-􀀒.:􀀓 --- --- I ! AREA A I , I , , I I 1_.. ..._􀀐·----- ___ j Trench 2 AREAB Fig. 7 Nos 45-47 Wincheap. Phase 2 features. 245 RICHARD HELM AND JAKE WEEKES side. These were formed of a mottled orange grey clay with patches of grey brown silty clay, chalk and carbon flecking, abutted the area of metalling. The deposits varied between 0.08-0.12m thick, and were 2.23 x 1.16111 in extent. In Area B, a north-west to south-east aligned wall foundation from another building (G 11) was identified, as evidence by compacted white chalk with patches of dark grey silty clay and rare flint within a foundation cut (as seen 1.7m by 0.65m). The south-west face of the wall was abutted by a clay, probably an internal floor, extending beyond the limit of excavation, in turn sealed by a metalled surface and probable demolition material, up to 0. 16m thick, apparently derived from the wall footing. An equivalent sequence was noted in section to the south-east. Finds were limited to a small assemblage of brick and peg tile. Remnant cellars (G 12 and G 13) were located at the south-east boundary of Area B. Cellar G 12 was a rectangular structure of roughly coursed flint and chalk wall within a construction cut, bonded with a very friable off-white lime mortar. The cellar wall was 0.78m from north-west to south-east, and 1.15m from north-east to south-west, e:,,.1:ending beyond the limit of excavation to the south-east. Only two courses survived, to a height of O. l l m, filled by a dumped deposit of mortar flecked silty clay. Cellar G 13 was located 2.52m to the south-west of G 12, and somewhat better preserved, with three sides represented by roughly coursed chalk and flint walls within a construction cut. This wall, e:,,.1:ant to a height of O .5 l m, was bonded with a friable, light yellow lime mortar, and had an internal render of the same mortar. As seen, the structure was 1.69m from north-west to south-east, and 1.78m from north-east to south-west, e:,,.1:ending beyond the limit of excavation. The cellar was backfilled by a chalk-flecked sandy clay silt with occasional flint. A series of relatively homogenous deposits (G l 4) of silty clay with inclusions of flint, chalk, carbon and shell e:,,.1ended across much of Area B, varying between 0.28m to 0.60111 thick. Sparse finds from this material included one sherd ofpotte1y dated to c.1200-1500, and fragments of Tyler Hill type roof tile (dated 1175-1400). Two sub-rectangular post-holes (G 15) truncating the soil horizon are likely to have been contemporary, and produced fragments of Tyler Hill roof tile. The development of Roman and medieval Win cheap By Jake Weekes The early Romano-British phases at 19 and 45-7 Wincheap clearly focus on what was probably already a significant route in the Late Iron Age (Margary 1955, 42), linking the settlement and probable cult focus (see Bennett et al. 2003) at Canterbury to settlements near Ashford (e.g. Booth et al. 2008). From there, an onward journey would lead towards the iron rich Weald, which we may suspect as an early economic driver towards proto-urban and infrastructure development in Kent. Gravel quarries such as those at 19 Wincheap are not untypical in early suburban locations around Canterbury (cf. Helm forthcoming), no doubt resulting from local infrastructure development. Even a smithy on the road leading into Canterbury, while regionally important because it adds another ironworking centre to the early Roman Kent and Sussex map, is to some e:,,.1:ent an expected find at this location, and while this second-century phase at 19 Wincheap produced tantalising evidence of life and work there 1800 years ago, the most striking finds at the site were in fact the extensive metalled surfaces. The site at 45-7 also contained an area of metalling 246 THE EARLY DEVELOPMENT OF A CANTERBURY SUBURB? NOS 19 AND 45-47 WINCHEAP in Area B, and such findings are all the more noteworthy because they are in keeping with a developing pattern along the line of the early road. Similar smfaces have now been discovered at 84 Wincheap (Lane 2013), on the opposite side of the road to Nos 45-7, and were already recorded at the junction of 5 Wi.ncheap and Gordon Road (Rady 1999; see also CCUAD 1147; although the latter at least may be evidence of the other road in the area heading south to link with Stone Street), and at 10-16 Wincheap, nearer the city defences (Shand 2004; Shand and Hicks 2013). Of key relevance here is the fact that the Roman defensive circuit was only built in the late third century, clearly changing the conceptual and physical landscape in terms of what lay 'within' and 'without' the town (Fig. 8) at this time. An earlier topographical shift in this area is evidenced by the fact that the new defences bisected a first- to second-century cemetery area (area 'South-West A', see Weekes 2011, 30-32), which by law and tradition would have lain beyond the perimeter of the early Romano-British settlement (apart from a probable conquest period fort: Bennett et al. 1982, 30ft). It may have been only after the early cemetery went - or was put - out of use, in the early second century, that Durovernum saw linear, mercantile and associated domestic ribbon development along the Ashford road, as is suggested by the cumulative dating evidence from Nos 5 (Rady 1999),10-16 (see Shand 2004, 16), 19, 28 (the Maiden '.s' Head; Bennett and Sweetinburgh 2007, 9), 45-7 and 84 Wincheap. This pattern of development is for now exemplified by the sequence at No. 19, with a probable late first-century date for an early soil horizon, gravel ex-traction pits backfilled by AD 130 in tandem with the development of a roadside smithy, and ex'tensive metalled surfacing, which was begun by c.200 and apparently out of use by c.300. Does the confirmed late second- to early third-century dating of cemetery features nearby, at the Invicta Service Station (CCUAD 1874; 116-190 Wincheap) and Simmonds Road (Shand 2005), to the rear of llO Wincheap, respectively, indicate the furthest south-westward ex-tent of this ribbon development? Admittedly, a burial found in 1969 (CCUAD 1926) about 40m south-west of 45-7 Wincheap, was dated to the conquest period, but this would merely represent an earlier phase of funereal use. To the north-east, the roadside development seen in Wincheap conceivably began from well within the location of the later defensive boundary. Excavators at Worthgate (Frere et al. 1982, 53) noted a clear 'cut-off' point, and clearance of earlier occupation to make way for the rampart and wall at the end of the third century. The two-phase building that pre-dated the defences here (ibid., 52-3) was very similar to the structures, and indeed sequences, seen at both 19 Wincheap and Nos 10-16, and produced no pottery post-dating AD 220. Later Roman pot and other finds, like the coins, are rare and residual at 19 and 45-7, and apparently absent altogether at 10-16 Wincheap, and this suggests considerable change and possibly abandonment of the Wincheap development around AD 300; it seems unlikely that the building of the defensive circuit at the same time, severing the straightforn1ard connection of Wincheap with the rest of the town, is a coincidence. The probably late Roman inhumation cemetery at Pin Hill and Station Road East, closer to the wall (see Weekes 2011, 30-31), was surely founded at least partly on the basis of its extramural location. We can thus suggest an overall phasing of Roman Wincheap, beginning with 247 RICHARD HELM AND JAKE WEEKES 248 Fig. 8 Topographical development of the Wincheap suburb during the Roman period. a) First to early second century. b) Sites indicating secondto third-century ribbon development. c) Contraction following establishment of walled circuit: fourth century. THE EARLY DEVELOPMENT OF A CANTERBURY SUBURB? NOS 19 AND 45-47 WINCHEAP delineation well to the north-east in the first century, marked by early cemeteries, followed by ribbon development e;\.l)ansion along the road towards the south-west during the second and third centuries, and finally contraction within defences in the fourth (Fig. 8). This sequence would be an interesting hypothesis to apply to other areas outside the walls, and could make more sense of the complicated mixture of industrial, domestic and funeral activity in the Romano-British suburbs at Canterbury as currently defined. A lack of Anglo-Saxon features or finds from 19 and 45-7 Wincheap is also part of a wider pattern (see Shand and Hicks 2013, 40; Sweetinburgh 2013, 44). The only evidence that (so far) seems to survive of an Anglo-Saxon 'waegn ceap' or 'wagon market' is its name. Instead we have archaeological evidence of a buildup of soils from agriculture and/or horticulture in the vicinity of 45-7, and only later signs of Norman land divisions at No. 19. These ditches would seem to be part of the delineation of rental plots investigated by William Urry (1967), and, remarkably, found to be still extant as some of the boundaries in the layout oftltis part ofWincheap. The excavation at 7-9 Gordon Road picked up the rear boundary ditch of the plots at the north-east end of the street (Rady 1999). With the building of Norman castles and renovation of walls and gates, the Wincheap area was once again underlined as 'outside' the town, and seems to have retained a mostly rural character. Urry points out that Wincheap land plots were named in relation to old field systems, and that the area was slower to develop than other Canterbury suburbs in the medieval period, only becoming built up in the later twelfth and earlier thirteenth centuries (Urry 1967, 189; see Sweetinburgh 2013, 45). The suburb was clearly well established in the thirteenth and fourteenth centuries (although details are hard to discern from the rentals: Sweetinburgh 2013, 44), and it is perhaps from this point that Wincheap emerges as the forerunner of what we see today. In the recent excavations, this medieval Wincheap was evidenced by remnants of buildings at Nos 19 and 45-7, and the construction of a new metalled surface at the latter site. Interestingly, later medieval and Tudor builders apparently rediscovered and re-used the ancient Romano-British metalled surface at No. 19. ACKNOWLEDGEMENTS The authors thank both Bignold Elevations Ltd and Cardy Constmction Ltd, who respectively funded excavation, post-excavation analyses and publication of the archaeology atNos 19 and 45-7 Wincheap. BIBLIOGRAPHY Allison, E. and Vokes, A., 2013 'Captain's Cabin, Wincheap, Canterbury. Brief report on remains recovered from bulk sample residues' CAT Archive report. Barber, L. 2012, 'Canterbury Captain's Cabin Wincheap. The post-Roman Pottery', CAT Archive report. Bem1ett, P., Frere. S.S. and Stow, S., 1982, Et:cavations at Canterbury Castle. The Archaeology of Canterbury Volume 1, KAS. Bennett, P., Hicks, A. and Houliston, M., 2003, 'Canterbury Whitefriars: 'The Big Dig', in Current Archaeology 185, I 90-196. 249 RICHARD HELM AND JAKE WEEKES Bennett, P. and Sweetinburgh S., 2007, 'An Archaeological and Historical Desk Assessment of the Wincheap Industrial Estate'. unpubl. CAT report. Bevan, L., 2012, 'Captain s Cabin Wincheap. The Small Finds CAT Archive report. Boden, D., 2007, 'Preliminary results of a programme of archaeological investigation and assessment at 19 Wincheap, Canterbury', unpubl. CAT report. Booth, P., Bingham, A. and Lawrence, S., 2008, The Roman Roadside Settlement at Westhm11k Fam1, Ashford, Kent. Ex.cavations 1998-9, Oxford Archaeology Monograph 2, Oxford. Frere, S.S., Stow, S. and Bennett, P., 1982, Excavations on the Roman and Medieval Defences at Canterbury, T he Archaeology of Canterbury Volume 2, KAS. Helm, R., forthcoming, Outside the Town: Roman Industry, Burial and Religion at St Augustine House, Canterbury, Kent, CAT Occas. Paper. Jones, S., 2012, 'Assessment report on the fauna! remains from excavations at Captains Cabin, Wincheap', CAT Archive report. Lane, R., 2013, 'Land to the rear of 84 Wincheap, Canterbury, evaluation report', unpubl. CAT report. Lyne, M.A.B .. 1994, Late Roman Handmade Wares in South-East Britain, unpubl. PH.D. thesis, University of Reading. Lyne, M.A.B., 2012, 'The Roman Pottery from Evaluations and Excavations at the Captain's Cabin Site in Wincheap, Canterbury', CAT Archive report. McDonnell, G., 2012, 'Report on the metalworking debris recovered from Captain's Cabin, 19 Wincheap Street, Canterbury, Kent', CAT Archive report. Margary, I., 1955, Roman Roads in Britain 1, Phoenix., London. Pringle, S., 2012, 'Assessment of the Ceramic Building Materials from Captain's Cabin, Wincheap, Kent', CAT Archive report. Rady, J. 1999, 'Nos. 7-8 Gordon Road , inCanterbury'sArchaeology 1996-97, 9. Savage, A., 2012, 'Captain's Cabin Wincheap, 2008 (CCWEX08). Pottery from environmental samples', CAT Archive report. Shand, G., 2004, 'Land adjacent to Nos 10-16 Wincheap', in Canterbury's Archaeology 2002-03, 16-17. Shand G., 2005, 'Simmonds Road'. Canterbury's Archaeology 2003-04 12. Shand, G. and Hicks, A., 2013, 'Roman and medieval development of a Canterbury Suburban Area: Excavations at Land Adjoining No. 10 Wincheap', Archaeologia Cantiana, cxxxm, 33-65. Smith, W., 2012, 'Charred Plant Remain from Romano-British and Medieval deposits at Captain's Cabin, Wincheap, Canterbul)', Kent', CAT Archive report. Sweetinburgh, S. 2013, 'The Documentary Evidence', in Shand and Hicks, 44-9. Tasker, P. and Helm. R. 2012, 'Archaeological evaluation and excavation at 45-47 Wincheap. Canterbury, Kent', unpubl. CAT report. Weekes, J., 2011, 'A review of Canterbury s Roman cemeteries', Archaeologia Cantiana, CXXXI, 23-42. Urry, W., 1967, Canterbury under the Angevin kings, The Athlone Press. ENDNOTES 1 Throughout the site narrative, excavation group numbers, prefixed with a 'G', are used, while evidence from the evaluations, test pits and various watching briefs are identified by site code and relevant context numbers. 2 For more detail on pottery, see Lyne 2012, Savage 2012 and Barber 2012. Details of ceramic building materials and enviromnental analyses cited here are derived from Pringle 2012 and Allison and Vokes 2013 respectively. 250