Additional Evidence of Roman (and later) Occupation Adjacent to the Marlowe Arcade Canterbury Excavations at Rose Lane 2002-4

235 Additional evidence of Roman (and later) occupation adjacent to the Mar lowe Arcade, Ca nterbury: Excavations at Rose La ne, 2002-4 jake weekes Between October 2002 and May 2004 various archaeological interventions were carried out by the Canterbury Archaeological Trust as part of the ongoing development of a commercial plot at Nos 6-8, Rose Lane, Canterbury, next to the Marlowe Arcade (Fig. 1). This area of land (approximately 0.17ha) has been near the heart of the city since its beginnings in the late Iron Age and early Roman period. The recent scheme of developer funded archaeology produced piecemeal new evidence in an area already much truncated, and indeed much excavated, yet still very incompletely understood. A new building was to be constructed using a raft and pile method, and so open area excavation (between November 2003 and January 2004) was limited to the ‘formation’ depth required by the developers, specified as a maximum of 0.9m below the existing surface (approximately 12.33m aod). Excavation was therefore curtailed from a purely archaeological perspective, affording only tantalising glimpses of the deeper stratigraphy. These results were slightly augmented, however, by a series of concurrent small-scale interventions forming part of the overall project, including an evaluation trench (Jarman 2002) and test pits (Jarman 2003), watching briefs on the laying of new services across the area (D. Boden, pers. comm.; Barrett and Westmacott 2005) and a geoarchaeological borehole survey (Branch et al. undated). Observations of deeper service trenching and removal of the existing building in particular revealed earlier archaeological sequences, especially within large service trenches at the north end and along the north-east of the site, and within a central modern basement. The net result of this general scheme of investigations is that a few new details (mostly gleaned from the evaluation and watching briefs) can be added to our understanding of the site in the Roman period, while the open area methodology to formation level at least allowed for more JAKE WEEKES 236 Fig. 1 Site location plan. systematic sampling of the medieval and post-medieval archaeology of this part of Canterbury. Group numbers are used in the following account to refer to associated archaeological features and deposits recorded in the open area excavation, and are those allotted during the post-excavation process (for more detail, see Barrett and Westmacott 2005). Depth of stratigraphy A borehole survey conducted around the proposed footprint of the new building (Fig. 2) produced the earliest evidence, indicating a sometimes surprising depth of stratigraphy even for a central Canterbury location. Natural Head Brickearth was encountered in all the positions except Borehole 4, which was abandoned when it struck a wall surrounding a 237 Fig. 2 Romano-British phase plan, also showing positions of boreholes (BH) 1-7. JAKE WEEKES 238 well. Interesting variations in the depths of the natural were also apparent, with the upper surface of the Brickearth (and therefore the depth of the archaeology) in Boreholes 2, 5, 6 and 7 proving to be particularly deep as compared with the evidence from Boreholes 1 and 3 (Table 1). The only sighting of natural Brickearth during the subsequent open area excavations (Group 212; group numbers are henceforth indicated by the prefix ‘G’) was in the section at the north-east corner of the site (i.e. nearest to Borehole 1). The surface of the natural here, chopped away on all sides by deeper features, was recorded at approximately 10.7m aod, which accords with the measurement from the Borehole 1 position, some 10m to the south. Early occupation The site lies less than 50m to the north-east of known late Iron Age round houses and within the projected line of triple enclosure ditches believed to have been the forerunner of Roman Canterbury (cf. Blockley et al. 1995, fig. 4). Archaeological deposits of interest are uniformly reported in the borehole logs above the deepest Brickearth horizons and beneath the excavation formation level, including (in summary and approximately from earliest to latest) silty clay, silt, and cess-rich deposits variously containing occasional charcoal fragments, marine mollusc shell, waterlogged wood and other organic matter, chalk pellets and gravel. A single fragment of late Iron Age pot is reported from 10.53m aod within Borehole 3, but this could easily have been residual, revealing nothing about the context from which it derived. A potentially early pit or ditch [G181], severely truncated, was observed within the base of a service trench in the eastern corner of the site during the excavation phase of the work (Fig. 2). The feature contained several sandy clay fills yielding occasional flint, carbon, oyster shell and Ta ble 1. Relative depths of the stra tigra phic sequence in boreholes Borehole no. Surface of natural (m aod) Depth beneath existing (m) 1 10.68 2.17 2 8.32 4.49 3 10.41 2.38 4 (unknown) (unknown) 5 7.32 5.55 6 7.94 4.86 7 7.45 5.30 EXC AVATIONS AT ROSE LANE , CANTE RBU RY, 2002-4: NE AR MARLOWE ARCADE 239 potsherds uncertainly dated to c.ad 25-75. A piece of slag also came from these contexts, along with an iron nail, occasional mammal bone, and some fish bone and scales. romano-british Previous excavations within and adjacent to the site had already produced a fairly detailed picture of the area during the Roman period (Fig. 2). Most pertinent to the recent work are the findings of the Canterbury Excavation Committee (CEC) excavation carried out in the late 1940s within the present site boundary by Sheppard Frere, and the late 1970s and early 1980s excavations in advance of the Marlowe Arcade development, immediately to the south-west (Blockley et al. 1995). The CEC work, albeit limited to what could be excavated within a number of narrow trenches within the old Rose Hotel Yard, uncovered part of a mid/late Romano-British building (see Blockley et al. 1995, figs 105 and 114), along with evidence of a metalled lane running along the north-west side of the building and remnants of further structures nearby. The Rose yard building itself was reconstructed (more or less tentatively) by Frere as comprising five rooms, forming a range aligned south-west/north-east. Rooms 2 and 4 had opus signinum floors, while Room 3 was tessellated. The most northerly room (Room 1) had also once been tessellated, but alterations, possibly for insertion of a stair well, had disturbed the tessarae (Blockley et al. 1995, 230-1). This range of rooms was supposedly backed on its south-east side by a corridor, although the evidence for this appears slight. Excavations in advance of the Marlowe Arcade further contextualised the Rose yard building as occupying the north-west corner of an insula to the south of the earlier public baths portico, and probably forming part of a courtyard arrangement with a building immediately to the south-west (building R26 of the Marlowe Arcade investigations, see Blockley et al. 1995, 174-5, 210, 243-4, fig. 2). This building project seems to represent some considerable modification of this part of the town in about ad 300, with the construction of the associated lane taking the place of the rear wall of the contracting public baths palaestra. The Rose Yard Building A number of features and deposits seen in the recent work produced corroborative evidence of the Rose Yard building and its associated lane. The north-west corner of the Rose Yard building was uncovered during the evaluation (Jarman 2002), which focussed on CEC Trench RYV . The sheer depth and restricted width of the trench made it impossible to fully record due to limited access and safety concerns but, in spite of the small JAKE WEEKES 240 area of wall exposed and difficulty of access, four phases of construction were tentatively identified [Walls 1-4, Fig. 2]. The earliest [Wall 1] was formed of neatly laid Roman brick courses with occasional flint nodules in a pale orange-brown mortar, 0.75m high. An offset [Wall 2] extended 0.26m from the south-west side of Wall 1 at right angles before neatly terminating. Walls 1 and 2 were superseded by Walls 3 and 4, entirely built from roughly coursed flint, and both abutting and utilising the earlier structure. No deposits, layers or finds were visible in association with the walls. The top of the Roman period masonry was recorded at 12.11m aod (Jarman 2002, 6). In one of the May 2004 watching briefs, new evidence from within the area previously designated Room 4 of the Rose Yard building came to light. A small sondage allowed partial recording of several contexts, including a clay floor overlain by occupational debris and a make up layer of demolition material. The relationship of these to other contexts assigned to this group [G174] was difficult to discern within the confined trench, but two compact opus signinum layers interleaved by a metalled surface, and a plaster, mortar and clay floor or bedding layer were also observed, along with a possible robber cut for a beam or wall. Roman tile, fragments of painted wall plaster and some animal bone were recovered from these contexts. Demolition deposits [G175] superseded them and were found to contain pottery dated no later than ad 300. Environmental samples from groups G174-5 produced small amounts of animal and fishbone and shellfish, along with crumbs of pot and ceramic building material (CBM) and a fragment of tessera. A metalled surface [G176], formed of 80% flint, overlay the demolition deposits, and was in turn sealed by further demolition material [G177] containing Roman brick and tile, occasional bone and slag and pottery dated c.ad 275-425. An accumulation of soil [G178], interpreted as post-Roman in date, completed the early sequence in this trench. A little over 5m to the north-west, a number of small trenches revealed the metalled surface [G205] of the lane seen here in the 1940s and 1980s excavations and associated with the construction of the Rose Yard building. The surface, uncovered over a maximum area of 5.72m north-south by 2.79m west-east, was constructed of 90% flint in a sandy clay matrix. This metalled surface had been cut by a fence line or minor structural foundation [G206], aligned approximately north-south and 0.07m wide and 0.43m deep, at some during or after its use. Finds from the upper surface of the lane included occasional iron nails and CBM, animal bone (some burnt), pottery (dated c.ad 175-300), occasional oyster shell and some carbon. Environmental sampling of these deposits recovered iron nails (very common in this sample) along with fish bone, mussel and barnacle fragments and a coprolite (faecal concretion). An accumulation of thirteen dumped layers of clays and silts followed [G207], containing much chalk, flint, CBM, mortar, nails, glass, pottery (again dated no later than c.ad EXC AVATIONS AT ROSE LANE , CANTE RBU RY, 2002-4: NE AR MARLOWE ARCADE 241 175-300), animal bone and oyster shell. Significantly, the material included dumps of ash, deposits composed of 95% carbon, and burnt clay, suggesting dispersal of demolition material following a fire. Everyday personal items were also recovered from these contexts, including six bone pin fragments, an unidentified copper alloy object, and a number of beads and hobnails. Wall plaster fragments, tesserae (stone and ceramic), slag, hammerscale, fishbone and shellfish were present in small quantities. The animal bone from these deposits (fifty-one hand collected fragments) formed the bulk of the assemblage derived from securely dated Roman period contexts. The dumped deposits in this area were subsequently used as bedding for another, more lightly metalled surfacing, with associated structural features [G208]. This surface, which was mostly of flint but also contained CBM and occasional pottery, animal bone and oyster shell, appeared to be associated with the construction of a new, timber framed building. The structure included floors of clay and re-used tile, an east-west aligned beam slot and a possible post pit as well as occupation deposits. Most of the pottery derived from these contexts again ranged from the mid second to the third century, but some sherds were notably later (c.275-425; further pot of this date was recovered from later garden soils overlying this area). The thin occupation deposits associated with the structure also yielded small quantities of animal and fish bone and shellfish, fragments of opus signinum, painted wall plaster and glass, nails and unidentified iron fragments, hobnails, bone pin fragments and four beads. While probably not from primary contexts, the animal bone assemblage from these groups produced some interesting, if fleeting, details relating to diet and husbandry in the Romano-British town. A cattle bone from group G207 (demolition deposits on the lane, probably derived from the Rose Yard building) showed signs of having been removed from a draught animal (Bendrey 2006, 1-2). A pig mandible from the same area [G208] is also of interest, for though the animal had clearly suffered from a highly visible facial infection for some time, ‘(B)utchery evidence ... indicates that this suppurating lesion did not deter the local inhabitants from making use of this animal’ (ibid., 2). Further significant elements of the Roman period sequence in the area were seen in the deep north-west facing section revealed by a twentieth-century basement cut (Section 1, Fig. 3; hereafter referred to as ‘the basement section’), including groups G1-18; 23-27; 34; and 102. Two rooms of a building with opus signinum and plaster floors [G1-3] were identified here, approximately 15m to the south-east of the Rose Yard building and a little further to the north-east of Marlowe building R26. These rooms were replaced by a second set of rooms [G4] with a possible corridor between and another opus signinum floor. Significantly, this second phase seems to have concluded with the destruction of the building by fire [G5-7]. JAKE WEEKES 242 Fig. 3 The basement section. EXC AVATIONS AT ROSE LANE , CANTE RBU RY, 2002-4: NE AR MARLOWE ARCADE 243 Similar partial and tantalising views were offered by sections at the limits of excavation [G122; 187; 195; 202; 213] and across the site at formation level. In group G122, for example, layers that produced mostly second-century pot, the latest probably from the late second or early third century, were cut by a pit containing sherds dated to c.250-300. Further probable Roman period structures were also evidenced, albeit sparsely, by sporadic features and deposits that could not be investigated further, including flooring and demolition [G8, 10; 23-25; 27; 177; 187], pits for cess, rubbish or storage [G9; 26; 34; 213; 122; 195; 202] and soils [G178; 182; 212]. Two small areas of possible ‘dark earth’ [G28-29], traditionally interpreted as the accumulation of soil following abandonment of Roman period structures, were also noted in the basement section, and succeeding stratigraphy in the same section could of course relate to disturbance in the later Roman or sub-Roman periods. Significant Romano-British finds found redeposited in later contexts should also be noted. This typically included pot and CBM , along with fragments of painted wall plaster (including some with what appear to be stucco details from within the CEC trench backfills, although these have yet to be properly identified), tessarae and eleven third- and fourth-century coins. There were also remnants of more specialized items such as a fragment of a pipe clay figurine from the area of the demolished Rose Yard building. anglo-saxon Previous findings of Anglo-Saxon occupation in the area include an early sunken featured building less than 5m to the south-west of the recent interventions (Marlowe excavations structure S15, dated c.550-650; Blockley et al. 1995, 306-9) and a later cellared structure (structure S29, dated c.925-75; ibid., 359-61) and probably associated pits approximately 10m to the west. No Anglo-Saxon features or deposits were noted in the recent excavations, nor any residual pottery securely dated as pre- Conquest or other finds from this period. It is quite possible that at least some of the post-Roman elements of the basement section (Fig. 3) relate to this period, but this remains speculative. medieval The medieval archaeology of the Marlowe Arcade excavations immediately to the south-west of the recent work was characterized by robbing out of Roman period masonry before 1100 (Blockley et al. 1995, 370) followed by refuse and cess pitting between 1100 and 1400 (ibid., 230), a picture commensurate with that seen in the CEC section drawings showing disturbance of the Rose Yard building (ibid, fig. 267). Between JAKE WEEKES 244 c.1400 and 1550 a lane, seen to the west, may have continued on a northeast/ south-west alignment some 20-25m to the south-west of the recent excavations. This is thought to be a ‘lost lane’ indicated by William Urry (1967, Map 1(B), sheet 5, cited in Blockley et al. 1995, 421), and was perhaps the focus for occupation features seen within the Marlowe Arcade excavation area in the form of a hearth, pits and a south-west/north-east aligned timber boundary line (i.e. at right-angles to the projected ‘lost lane’). Semi-basements (structures L9-11; ibid, 421ff) were seen beneath the building to the south of the Marlowe Arcade, currently British Home Stores. The area covered by the recent excavations could therefore be expected to lie to the rear of medieval tenements known or suspected in all directions, and to contain evidence of refuse dumping and other activities. In the recent work, medieval archaeology was encountered in several discrete areas of the site (Fig. 4), and in particular in the north, west and east corners and along the south-eastern edge of excavation, including the next phase of features and deposits seen in the deeper basement section (cf. Fig. 3). A further concentration, in a western central area of the site [G98-100], did not produce dating material but was suspected as medieval on the basis of feature type. These areas of medieval activity were only seen because they remained intact at the formation level (the limit of excavation), while other areas had been truncated beyond this level by later activity. Nevertheless, despite such accidental qualities of preservation and discovery, some chronological and functional variation between the different areas could be discerned. Further small ‘windows’ on the medieval archaeology were afforded sporadically across the excavated area and watching briefs occasionally provided the opportunity to collect additional (though limited) information at a greater depth. In the western corner of the site the area above the disused Romano- British lane was used for dumping refuse in layers and in pits [G209]. Along with much residual demolition material, deposits here produced occasional slag and medieval roof tile, animal bone, pottery dated to c.1050-1225 and a possibly residual copper alloy mount. These deposits were sealed by an accumulation of garden soils [G210]. Notably, a horse and a dog bone from the latter contexts showed signs of butchery (respectively a femur with a fresh fracture and a radius with transverse cut marks). While this may not represent human consumption of the animals, the evidence is suggestive (Bendrey 2006, 2). Complicated evidence of successive occupation was uncovered but only partially understood in the northern corner of the site. The earliest stratigraphy here (revealed in the base of a pipe trench) comprised a series of deposits interpreted as dumps of demolition and occupation material, apparently cut by post-holes and other features [G66]. A silty layer topped by a rammed chalk surface [G68] and a demolished hearth EXC AVATIONS AT ROSE LANE , CANTE RBU RY, 2002-4: NE AR MARLOWE ARCADE 245 base of fired orange clay [G69] were seen in adjacent sections at this level. The demolition layers in the base of the trench were overlain by a probable make up layer topped by vestiges of clay floors [G67], these being cut by a small post-hole and burnt in one area, probably by a cooking fire. A build-up of soil above these contexts [G70] contained (along with occasional slag, glass vessel shards and oyster shells) pottery, Fig. 4 Medieval phase plan (scale 1:250). JAKE WEEKES 246 a ceramic lamp and part of a crucible all dated not later than c.1175-1300. A shallow pit cut into this soil [G71] produced pottery dated c.1175-1225, a small quantity of animal bone, worked and quern stone fragments and occasional fish bone and shellfish. Two consecutive flint and clay hearth bases [G72], overlying the soil, yielded pot dated no later than c.1225. Next to the north-western section, consecutive phases of partially seen surfaces with beam slots [G73-4] succeeded these earlier deposits, probably representing a number of phases of rooms within a building. The earlier of these rooms [G73] included remnants of an oyster shell surface and produced a medieval roof tile, and the later phase [G74] yielded potsherds dated at the latest to c.1250-1300, as well as occasional animal bone, slag and slate. This building may have been demolished following a fire: an extensive carbon, flint and tile rich clayey deposit [G75] sealed all. This deposit included animal and occasional fish bone along with pottery dated c.1200-1250. A new and extensive rammed chalk surface followed, overlain by occupation deposits (including animal bone) and evidence of later patching of the floor, and also cut by post- and stakeholes and a possible beam slot [G76; G79]. The latter produced pottery dating to c.1225-1300. To the south a possibly contemporary oyster shell surface was cut by a post-hole, and overlain by a patch of burnt clay floor, again possibly representing separate rooms [G77]. A deposit of dark soil overlying the clay floor seems to represent abandonment, dated between c.1250-1300 on the basis of pot evidence. In the adjacent northern section further contemporary surfaces were recorded, including make-up layers, rammed chalk and metalling [G95]. A remnant of a clay floor or make up layer [G86] some 5m to the east (cut away by a large modern drain) might also be related, while areas adjacent and to the east (revealed by the deeper cut of the same drainage trench) were characterized by probable garden soils and domestic dumping [G188; 193] immediately superseded by a clay floor [G189]. This was sealed by deposits signalling demolition and abandonment [G190] and the construction and use of a new structure [G191]; medieval tile and occasional glass fragments were present, and traces of animal and fish bone and shell were recovered from the floor level. These partially seen features and deposits represent a strikingly similar sequence to that seen in the north corner of the site, with pottery from the early demolition deposits in the sequence [G188; 193] providing a terminus post quem for the structures in about the late thirteenth century. More evidence of probably contemporary occupation was observed on the southern side of the modern drainage cut, where a small stake hole overlain by a clay floor [G196] was sealed by garden soils [G197], 0.3m thick and apparently containing pot dated no later than c.1225-1325. The soils were cut by a number of intercutting refuse and cess pits [G198], and these were in turn superseded by six clay floors, a large post pit and EXC AVATIONS AT ROSE LANE , CANTE RBU RY, 2002-4: NE AR MARLOWE ARCADE 247 two occupation layers [G199]. Further refuse pitting [G200] completed the sequence in this area, while more probable medieval archaeology was revealed further along the modern drain at the eastern corner of the site. This presented another typical sequence seen only in section, including (from earliest to latest) pitting, dumped deposits, a structure and occupation, further dumping, a build up of garden soils and another structure [G214-9; none datable]. Medieval occupation was also apparent near the south-east boundary of the site. The basement section (Fig. 3) revealed a typical sequence of robber pits and trenches focused on the Roman period buildings, followed by refuse pits and isolated structural features [G11-18; 22; 33; 35-7; 40- 3; 56; 59; 102]. Intact archaeology was investigated to formation level adjacent and to the south-east of this section. Early deposits [G52] were cut by a pit either containing the structural elements of a furnace, or dug for disposal of the remnants of one [G54], including a section of flint and chalk walling/foundation and remnant burnt clay surfaces, localized areas of 100% carbon and rake out materials. A pit with similar fills was seen in the basement section. Two small iron strip fragments, much corroded but possibly indicating knife manufacture were recovered from one of the carbon rich deposits along with an iron fragment. A small area of metalled surface overlain by a further (dumped) carbon rich deposit [G55] sealed at least some of the latter deposits to the north, while dumped deposits formed the make-up layer for a sequence of remnant clay floors to the south [G57]. The silty clay make up deposit, formed of re-used demolition material containing CBM , mortar, clay, carbon and chalk inclusions, also produced a ceramic cauldron rim with handle and other sherds with an approximate date range of c.1250-1450. Approximately 5m to the south-west was a truncated furnace structure and working area [G50]. The furnace was constructed of flint and tile, with a shallow circular area for firing floored with burnt red clay, and a flue. A slumped clay floor and an overlying dump of burnt material probably originating from the furnace lay immediately to the south-east. These surfaces were associated with another wall, constructed mostly of chalk with a stacked tile foundation. Ferrous slag and an animal bone was recovered from the furnace structure, while the clay floor itself yielded a cooking pot base dated c.1225-1350. Overlying furnace-related materials and surfaces to the north-east [G52; 55; 57] was a structure of unknown function [G58]. It was made of well-built flint walls, some forming a rectangular tank or trough. The walls continued to the north and south, with small areas of clay floors abutting. At least part of this building appears to have been adapted to use continuing into the post-medieval period (see below). Two phases of another probable late medieval building of flint construction were seen to the south-west of group G50. The truncated remnant of a flint wall JAKE WEEKES 248 bonded in a light brownish yellow mortar and abutted by a clay floor [G110] was succeeded by three sides of another flint structure covering just 1.85m by 1.33m and enclosing a small space. This building continued beyond the section and probably therefore formed a bay within a larger structure to the south-east. Unfortunately the building had been scoured of any associated deposits by later disturbance and it could not be dated or further characterised in terms of function. Probable medieval features, partially seen at the limit of excavation in the western central area of site, also defied adequate characterisation. Cut into an area of ‘garden’ soils, these certainly included another furnacelike structure [G98] and possibly related pits [G99], partially overlain by remnants of a structure with an internal clay floor, post-settings and a probably external yard formed by a metalled surface over a crushed chalk make up layer [G100]. These contexts were again undated, as were isolated remnants, possibly of the same structure, nearby [G115; 158]. late medieval and post-medieval Given the location of the site, continuing activities to the rear of buildings lying along streets to north, south, east and west, could be expected in the post-medieval period, coupled perhaps with the development of more permanent structures. To the south-west, the Marlowe excavations had shown that a medieval semi-basement [structure L9] continued in use into the Tudor period, while after c.1680 a substantial flint and chalk wall replaced an earlier boundary (noted above) and pits for refuse were dug in the area adjacent to the south-west of the Rose Lane site (Blockley et al. 1995, 446-7). Documentary evidence informs us that it was during this period that Rose Lane received its current name (although only much later its current alignment), having previously been called respectively ‘Dungeon Lane’ and ‘Pillory Lane’ due to nearby correction facilities in the medieval and early post-medieval periods (Sweetinburgh 2002, 9-10). The Rose Tavern, recorded from 1674, apparently gave its name to the thoroughfare in the later eighteenth century, a time when there were also slaughterhouses in the area, succeeded, from c.1800, by the construction of large houses fronting Rose Lane with associated coach and store houses, stables, granaries, yards and gardens (ibid., 10). During the late eighteenth or early nineteenth century the Rose Tap, with a coach yard and stables, was added to the rear of The Rose Tavern for coach drivers and the less well to do. This, which would have backed on to the eastern area of the site, and early twentieth-century residential and commercial developments in the area, were severely damaged by wartime attacks on the city. The remnants were finally demolished to make way for post-war development and the re-alignment of Rose Lane (ibid.). Once again, though the limit of excavation was governed by formation EXC AVATIONS AT ROSE LANE , CANTE RBU RY, 2002-4: NE AR MARLOWE ARCADE 249 level rather than characterisation of the archaeology, general zones relating to chronology and land-use could be posited. An extensive loamy soil horizon was the earliest deposit encountered in the southern area of the site [G20; Fig. 5]. Not surprisingly, this produced mixed finds including oyster shell, CBM, peg tile, animal bone, charcoal, Fig. 5 Late Medieval to early post Medieval (c.1475-1700) phase plan (scale 1:250). JAKE WEEKES 250 flint and chalk inclusions and pottery, which was dated not later than c.1475-1550. A range of copper alloy and iron finds was also recovered from these deposits, including a residual late Roman coin, a brooch, a bracelet fragment, a buckle, two pins, a blade and various fittings, buttons, mounts and unidentified objects. To the north and north-west of this area pits with pottery dating to this period [G94: c.1375-1525; G211: c.1475- 1550] were only partially seen in sections. It is possible that these groups represent a continuation of the general soil horizon, and it is worth noting that these contexts included similar small finds: such as a copper alloy pin [G94] and a buckle [G211]. Isolated structural features [G104, including fragments of postmedieval brick], a possible pit [G105] and dumped deposits [G106] containing pottery dated at the latest to c.1525 respectively cut and overlay the soils in the southernmost corner of the site, these in turn truncated by successive structures and demolition [G107-9]. Pottery from the earlier of these groups [G107-8] was dated c.1375-1525. Evidence for the latest structure here [G109] consisted of five irregular and truncated floor layers [two of chalk, two of clay and one of mortar] associated with a flint and chalk east-west aligned structure, either the foundation of a small wall or cobbled border for a path or work area. Pottery from one of the chalk surfaces dated to c.1525 at the latest, although possibly nearer c.1475. A general date of activity during the Tudor period is therefore suggested, perhaps contemporaneous with the soil build-up in this area. The demolition deposit of the latest structure [G113] comprised a layer of ash, burnt clay and carbon, with pottery dated at the latest to c.1475- 1550. A broadly contemporary development seems to have occurred a little to the north-west. In this case, again in an area of the soil used at least initially for the cutting and backfilling of pits containing demolition material [G116; G124], a substantial structure was built, represented by a (truncated) north-west/south-east flint wall some 10m long, with associated clay floors and two post-holes to the north-east. The northern extremity of the clay floor was cut by a probably square pit with interesting contents [G117]. The highly organic basal fill of the feature was sealed by a layer comprised mostly of fish bones (approximately 85% of the deposit), this in turn being sealed by capping deposits of demolition and hearth waste. The upper fill was a dump of domestic refuse and probable cess. The use of this feature, which while it cut a clay floor seems to have respected the adjacent wall building in its alignment, is suggestive of commercial activity. Environmental sampling of the fish bones showed a preponderance of herring heads, consistent with processing. Pottery from the capping and later cessy deposits dated to c.1475-1550 and c.1450-1550 respectively. Further possible elements of the structure or its associated activity lay some 5m to east [G157], with patches of a metalled surface, EXC AVATIONS AT ROSE LANE , CANTE RBU RY, 2002-4: NE AR MARLOWE ARCADE 251 three post-holes backfilled with similar material, and a possible levelling layer of burnt and unburnt clay thought to represent either a clay floor or hearth remains. Slightly later archaeology focused in the northern corner of the site, where another sequence of building, occupation and demolition was recorded. Early layers representing either internal or external surfaces [clay floors and a cobbled area: G78] were only partially seen at the limit of excavation within a deeper service trench, but produced pottery dated to c.1525-1625. The surfaces were possibly associated with a mortared flint and chalk wall [G80], forming a sub-square structure (2.6 x 2.7m) with the northern wall missing. An extensive area of unexcavated mixed deposits comprising soils and intercutting pits [G81] surrounded the structure and was probably contemporary with it. The structure itself and some of the surrounding soils were sealed by large amounts of demolition debris [G82] mainly of chalk rubble but including some peg tile. Further soil build-up [G83] was noted as having been cut through by the adjacent pipe trench, also sealing the demolition layers, indicating that the ‘garden’ context of the building continued following its destruction. A cluster of nine post-holes [G84] had then been cut into these garden soils and demolition deposits. Removal of this timber structure took place c.1550- 1650 at the earliest, but probably later. To the south-west the soils were cut by a sequence of pits [G85] for disposal of domestic refuse. Evidence of a further structure [G87] was seen to cut earlier demolition deposits [G82] to the north-east, and a tank [G88, perhaps for cess] was recorded at the limit of excavation, just to the south-east of this group. This consisted of three sides of a probably rectangular structure, approximately 1.5m by 0.8m remaining, with semicoursed walls of roughly shaped chalk blocks in a yellow sandy mortar. Pottery dated to c.1575-1650 was associated with this structure. More evidence of buildings of this approximate date was partially seen in the deeper sections of the large service trench in this area [G95], including make up and levelling deposits overlain by a metalled surface, perhaps contemporary with a compact crushed chalk surface and its bedding, and also to the south in the form of a remnant clay floor [G90]. Large truncated areas of intercutting pits were observed at formation level to the south and south-west of the buildings [G89; G91], and further pits in the north-west edge of excavation [G92-3]. Backfills of pits in group G91 contained late sixteenth- to early seventeenth-century pottery and clay pipe stems of c.1625-1675, associating them more closely with the next phase in the development of the northern half of the site, a general and extensive build up of garden soils [G96] that sealed all features and deposits so far described in this area. That this at least began to occur not later than the seventeenth century (and perhaps before 1650) is evidenced by the latest pottery and clay JAKE WEEKES 252 pipes from these wide ranging contexts, yielding dates such as c.1580- 1630, 1590-1620 and 1610-40. A number of pits cut within these soils [G138-9; 142; 160-1; 125; 221] across the site had been used for disposal of domestic waste including a notable assemblage of pot dating to between 1580 and 1675. Isolated structural and garden features were also represented [G133-5; 140-1]. At some time in the late seventeenth and early eighteenth century (Fig. 6), a brick building [G128] was constructed in the northern corner of the site. A north-east/south-west aligned wall, 6 x 0.5m and standing to six courses was recorded here, lined by post-holes, these probably for scaffolding either for construction or demolition. The latest pot from deposits associated with this structure dated to c.1680-1710. Adjacent garden soils continued to be used for garden features [G131-2; 137; 143] and for pits for refuse and other disposal [G130; 136]. Pottery from these contexts was dated mainly to the second half of the seventeenth century, although some sherds may have been as late as 1750 and pipe stems as late as 1775. To the south, demolition material in layers and pits [G119] produced pottery dated c.1725-1800, while sherds from a vestigial chalk surface further to the south [G159] and within intercutting refuse pits adjacent to the north-west section [G123] was uniformly of eighteenthcentury date, and as late as 1780. In the southern corner of the site a north-east/south-west aligned walled cellar cutting through a general soil horizon may have first been built as early as the seventeenth century. The first phase of the structure [G120] was of flint, chalk and brick construction, the north-western inside edge faced and partly rendered. A robber cut backfilled with demolition debris lay at right angles. The cellar had been rebuilt or renovated twice in its history [G225; 226], and its overall use is tentatively dated c.1680- 1800. Nearby pits containing demolition material [G112], burnt hearth residues [G121; 144] and cess [G146] and isolated structural features [G153], patches of floor and surfaces [G114; 170], disturbed demolition [G145] and a disused well [G154] could be broadly contemporary. Dating is also speculative when it comes to further partially seen and isolated structures, floors and demolition [G155-6; 180], observed across the central, north-east and eastern areas of the site. An earliest date of c.1800 is generally proposed for most of these features, mainly on the basis of brick and mortar type, although a small cellared structure [G180] was of stone construction and could well be earlier. A flint-lined well with associated surfaces and construction features [G60], was revealed in section in the eastern corner of the area removed by a modern basement (and truncated by it, see Fig. 3). The well may have been used from the medieval period but, if so, was probably given added structure during the post-medieval. This was composed of 90-95% flint with some brick and peg tile, bonded with mid yellow gritty mortar EXC AVATIONS AT ROSE LANE , CANTE RBU RY, 2002-4: NE AR MARLOWE ARCADE 253 with very small chalk, carbon and flint inclusions, the inside face roughly coursed. This utility seems to have served a number of adjacent postmedieval buildings, clearly representing several phases of construction, use and demolition, but all sadly undatable with any certainty. Following a sequence of abandonment, burnt destruction, and demolition deposits [G63], a number of floor surfaces [G61-2] appear to have abutted the Fig. 6 Post Medieval (1700-) phase plan (scale 1:250). JAKE WEEKES 254 well superstructure. These surfaces could have been variously internal or external, with areas of clay floor, tiling and cobbles all represented along with associated levelling and occupation deposits. A new building was then constructed next to the well [G64], with at least two rooms separated by a north-east/south-west aligned wall. It is unclear whether this building incorporated the well, but it seems to have re-used an extant section of wall from the later medieval building here [G58], suggesting some continuity. Pottery evidence from deposits resulting from the demolition of this building [G65] suggests that this probably did not occur later than the middle of the eighteenth century. To the south was a clay floor with a north-east/south-west aligned beam slot or robbed wall in line with two post-holes [G166], all filled with demolition material. This would appear to be early post-medieval in date on the basis of construction method, as would another building, to the south-west, evidenced by two clay floors associated with flint wall remnants, a post-hole and a group of stake-holes [G51]. Overlying demolition deposits [G107] for this structure produced pottery of the date range c.1750-1825. Various pits [G21; 42; 45; 46; 49], isolated structural features [G30-1; 38], soil build-up [G39] and other deposits [G44] seen in the basement section probably relate to the later phases of these buildings. The latest contexts recorded in a small slot excavated in the southwest area of the site consisted of dumped demolition material [G179], probably within pits, yielding pottery suggesting deposition after 1815. This disposal would have been carried out to the rear of nineteenth- and twentieth-century occupation clearly focussed in the eastern corner and along the south-east edge of the site. Complex stratigraphy continued at these levels, with various activities and changes of use represented by garden soils [G47; 164], demolition layers and deposits [G150; 163; 168; 172], pits [G152; 162; 164; 171], isolated structural features [G48; 149; 170] and brick buildings, paths, drains and a well [G148; 151; 167; 169; 173]. Cellars and related material in the eastern corner of the site [G201; 222-3] bore signs of demolition and the well seen in section in this area had already been filled in with demolition rubble [G224] prior to its truncation by a modern basement. Post-Roman Finds A wide range of small finds, most probably residual, were recovered from the excavation (Bevan undated), many perhaps dating to the late medieval period but most of which probably lost or discarded during the early post-medieval period and later. These items, variously testifying to personal ornament and costume, crafts and some gardening activities, included: (in copper alloy) a plate brooch, a possible child’s bracelet, a EXC AVATIONS AT ROSE LANE , CANTE RBU RY, 2002-4: NE AR MARLOWE ARCADE 255 notable assemblage of pins (40 complete as well as further fragments), buckles, hooked fasteners, a button, a chain fastener, a possible purse fitting, a lace tag tweezers, a thimble, needles, miscellaneous fittings and mounts, as well as several iron knives (one with bone handle), a bucket handle, a pitchfork and four fragmentary horseshoes. The post-medieval pottery assemblage, mainly derived from refuse pits, was of good quality and would certainly contribute to further comparative study (Cotter 2004, passim). The animal bone assemblage from this period included a range of butchered (cattle, sheep, pig, fallow deer and rabbit) as well as probably domestic or feral animals (dog and cat) and horses (Bendrey 2006, 3). Some of the butchered bones (e.g. horn cores) bore the marks of further processing, probably associated with craftwork. discussion Many of the early deposits recorded in the borehole logs at Rose Lane most likely precede the Rose Yard Romano-British building and should be considered as approximately equivalent to later pre-Roman Iron Age and early/mid- Roman features and layers discovered in the adjacent Marlowe Arcade excavations (late first century bc - c.ad 300; see Blockley et al. 1995, 27-169). These findings and other potential early features seen during the recent excavations are too indeterminate, however, to add much to our knowledge of the area at this time. The neatness of Roman period Walls 1 and 2 observed in the evaluation of the CEC trench suggests that both were originally exposed elevations, and it may well be significant that they lie in the projected path of the outer wall of the adjacent public baths palaestra. A section of this southwest/ north-east aligned wall had already been seen in the 1940s to the south-west, and the south-west corner and alignment of the wall were again recorded during the early Marlow Arcade excavations (Blockley et al. 1995, 97-100; fig. 42). The distances between centres of buttresses on this wall were recorded as approximately 20 Roman feet (5.91m) apart (Simon Pratt, pers. comm.); it is suggestive therefore that Walls 1 and 2 from the recent evaluation fall approximately at a multiple of this spacing (x 4 = 23.64m) being about 23.75m from the nearest known buttress to the south-west. Walls 1 and 2 could therefore have originally formed an external buttress on the palaestra wall, later re-used in the construction of the Rose Yard building. The Rose Yard building itself and building R26 from the Marlowe excavations are likely have been built at the same time, around ad 300, along with their associated lane (Blockley et al. 1995; 174-5; 210; 243- 4). It seems increasingly likely, however, that the Rose Yard section at least was destroyed by fire soon afterwards. It was already suspected that the Rose Yard building was out of use by the middle of the fourth century, JAKE WEEKES 256 possibly following a fire (Blockley et al. 1995, 233), and evidence from the recent excavations lends further weight to the suggestion. The destruction by fire of structures seen in the basement section just a few metres to the east and south-east [G5-7] might be significant, and ashy and carbon rich demolition deposits found strewn across the lane [G207] present more localised evidence. These and floors and demolition deposits within Room 4, could not be dated beyond c.ad 300. The recent excavations demonstrated that a timber-framed structure was subsequently built into levelled demolition deposits above the lane, as late as 400. Later deposits within the area of Room 4 appear to suggest that the new metalled yard associated with this structure extended further to the south, towards adjacent building R26, which was itself being robbed of materials and subjected to squatter occupation at this time (Blockley et al. 1995, 226-7). Beyond structural evidence there are hints within the material culture of what we tend to call ‘everyday life’ in the Roman period. Findings of tessarae, painted plaster, bone pins, beads and possible jewellery settings, slag and hammerscale and bone refuse all link to important themes, such as construction of identity and status through building interiors and dress accessorising, the perhaps surprising proximity of light industry to the urban centre, or notable idiosyncrasies of animal husbandry and diet. Even in periods where detailed documentary records are increasingly available, archaeology can begin to fill in the material details of lives often only hinted at by history. The Rose Lane excavations present a mass of such minor details that add up to a generalised picture of the area’s gradual development throughout the medieval and post-medieval periods. This sequence begins with typical evidence of robbing of Roman period structures, but occupation and light industry was well underway by the middle of the thirteenth century and probably much earlier, at least along the north-eastern boundary of the site. This continued into the early fifteenth century with the operation of furnaces, probably associated with iron working, near the south-eastern site boundary. The activity was no doubt carried out by the occupiers of tenements along the western side of the early alignment of Rose Lane, then called ‘Dungeon Lane’. Soil build-ups on the site during the late medieval period could be interpreted as gardens, but finds evidence suggests that various activities were taking place to the rear of the tenements, as we might expect, a pattern that continues in the post-medieval period. Certainly processing of herring on a commercial scale is evidenced during the late Tudor period, although what purposes the increasingly complex mass of later structures served here becomes increasingly unclear. The assemblage of early postmedieval personal ornament-, clothing- and craft-related items from the area as a whole could not be adequately contextualised, but is certainly suggestive of activities other than commercial processing, such as the use EXC AVATIONS AT ROSE LANE , CANTE RBU RY, 2002-4: NE AR MARLOWE ARCADE 257 of the area as yards to the rear of hostelries, for example. Much of the discarded pottery from this period could also derive from such a context, and it may be that at least some of the material was used in historically attested taverns nearby, or their antecedents. A brick building in the north corner of the site surmounts continual occupation and successive buildings there from the thirteenth to the eighteenth centuries, and continual occupation is also particularly apparent along the south-east limit of excavation where, from the late medieval period on, external working areas, specialised flint buildings, clay floors, wells and cellars are all represented. The archaeological evidence is entirely commensurate with what we know of the development of Pillory/Rose Lane during the period, as an area serving the local market (in particular with slaughterhouses) and providing services for those arriving and staying in the city: the Rose Tavern from 1674, and, from c.1800, coach and store houses, stables, granaries and yards. Many of the later features in the eastern area of site clearly relate respectively to the establishment of the Rose Tap, war damage, and subsequent demolition and redevelopment. The Rose Lane excavations were carried out in often difficult conditions in areas that afforded only tantalising glimpses of earlier stratigraphy. Even where previous findings offered hope of a developing understanding of the site, little could be done to relate new findings to them. Establishment of a clear understanding of the relationship of new evidence to existing knowledge should be a basic tenet of any future work in the area, since opportunities for this will decrease with every new development. However, despite typical limitations of the archaeological project in terms of scope and funding, there is little doubt that the most recent excavations at Rose Lane produced significant new data for an important area of Roman, medieval and post-medieval Canterbury. For the Roman period there was corroborative evidence that adds a few but nonetheless notable details to a constantly developing narrative. For the medieval and post-medieval periods the new material evidence produced by these excavations should feed into future interdisciplinary research. acknowledgements Analysis and interpretation of this site was greatly aided by the relational database and stratigraphic report prepared by Dan Barrett and Jude Westmacott, along with further work by Crispin Jarman, Damien Boden and Christopher Sparey-Green. Thanks are equally extended to all those who worked on the site in often difficult conditions, to Simon Pratt for his assistance in discussions relating to the location of Roman period remains, and to Peter Atkinson for skilfully producing drawings for publication. JAKE WEEKES 258 bibliography Barrett, D. and Westmacott, J., 2005, ‘6-8 Rose Lane, Canterbury: Stratigraphic Report’, unpubl., CAT. Bendrey, R., 2006, ‘The mammal bones from archaeological excavations at 6-8 Rose Lane, Canterbury, Kent’, unpubl., CAT. Bevan, L., undated, ‘Rose Lane Small Finds’, unpubl., CAT. Blockley, K., Blockley, M., Blockley, P., Frere, S.S. and Stow, S., 1995, Excavations in the Marlowe Car Park and Surrounding Areas. The Archaeology of Canterbury, Volume 5, Canterbury: CAT. Branch, N.P., Green, C.P. and Swindle, G.E., undated, ‘Geoarchaeological investigations at 6-8 Rose Lane’, unpubl. archive report, Dept of Geography, Royal Holloway, University of London. Cotter, J., 2004, unpubl. archive notes on post-medieval pottery: CAT. Jarman, C., 2002, ‘Archaeological evaluation on land to the rear of Nos 6-8 Rose Lane, Canterbury’, unpubl., CAT. Jarman, C., 2003, ‘Test pits at Nos 6-8 Rose Lane, Canterbury’, unpubl., CAT. Sweetinburgh, S., 2002, contribution in C. Sparey-Green, ‘Proposed Redevelopment at 6-8 Rose Lane, Canterbury: Archaeological Desk Study’, unpubl., CAT.