The Roman Villa at Minster-in-Thanet

261 THE ROMAN VILLA AT MINSTER-IN-THANET. PART 4: THE SOUTH-WEST BUILDINGS, 6A AND 6B keith parfitt Continuing our series detailing the results of the Society’s investigations, undertaken jointly with the Thanet Archaeological Society and the Trust for Thanet Archaeology, on the Roman villa site at Abbey Farm, Minster-in-Thanet, we here present our fourth report. This describes the substantial complex, designated Building 6, which was excavated at the south-western corner of the site. The villa lies on a gentle south-west facing slope at an elevation of 13-17m above od, overlooking the modern village of Minster and the ancient Wantsum Channel (Perkins 2004, figs 1 and 2). NGR TR 3135 6463, centred. A series of Roman structures, most occupied between the late first and early third centuries ad were set either inside or around a large rectangular walled enclosure (Perkins 2004; Fig. 1, B1-B6). The principal dwelling (Building 1) consisted of a classic ‘winged-corridor’ villa, with a small detached bath-house immediately adjacent (Perkins 2004, Building 3). Both these structures were placed within the higher, northern end of the walled enclosure. Downhill to the south, a central gateway led through the enclosure wall. Immediately outside the enclosure, at its southern corners lay two more substantial buildings, Building 4 at the south-east and Building 6 at the south-west. Building 4, a stone corridor house of several phases, has been previously reported (Parfitt 2006). Building 6 was first located during extensive trenching undertaken across the site in 2002. It was fully excavated over the next two seasons, with the northern half being examined in 2003 (Plate I) and the southern half in 2004 (Plate II). The building remains in this area were the most complicated encountered anywhere at Minster and it eventually emerged that they actually related to two separate, successive, structures occupying the same plot (Fig. 2). These are described below as Building 6A and Building 6B. The latest, Building 6B, was a simple one phase structure but the earlier Building 6A was larger and rather more complicated, probably with at least three phases of development (Phases 1-3), and included a bath-suite (Fig. 3). KEITH PARFITT 262 Fig. 1 General plan of the villa complex showing the position of the Building 6 in relation to the villa enclosure wall, Building 4 and main house (Building 1). THE ROMAN VILLA AT MINSTER-IN-THANET. PART 4: BUILDINGS 6A AND 6B 263 building BUILDING 6A Building 6A was a stone structure aligned n-s, following the natural fall of the ground between the 12 and 13m contours (Perkins 2004, fig. 3). It was the earlier of the two Roman buildings here but, like Building 4 at the south-east corner, it does not appear to belong to the primary phase of villa construction because it was butted onto the outside of the pre-existing villa enclosure wall (Fig. 1). It was also partially built across an infilled ditch (F. 7197/7284, see below). Early Ditch, F. 7197/7284 Traces of this substantial ditch, perhaps marking an early boundary, were located below the eastern side of Building 6A, its filling cut through by the earliest foundations (Fig. 2). The ditch ran roughly n-s and was traced for a minimum distance of 14.75m, under Rooms 33 and 40. It was at least 1.50m wide and 1.20m deep, with steeply sloping sides and a dished base. Examination of the profile suggested that it had probably been re-cut at least once. The soils filling the feature produced significant quantities of pottery, animal bone and shell, together with fragments of Roman tile which could be intrusive. More than 500 pot sherds, all PlateLATE I Building 6 – general view of excavations in 2003, looking south 264 Fig. 2 Building 6, general site plan. THE ROMAN VILLA AT MINSTER-IN-THANET. PART 4: BUILDINGS 6A AND 6B 265 dating to the first century ad, were recovered; the earliest pieces are pre-Conquest and there is nothing which need be dated later than c.ad 60 (see below, Assemblage A). The quantity of domestic rubbish recovered from the ditch is clearly suggestive of habitation close-by. This occupation would seem to be earlier than the first villa buildings and thus provides evidence that the villa complex was established on an existing settlement site, seemingly of pre-Conquest origin. The occurrence of over-fired wasters within the pottery assemblage indicates that pottery was being manufactured at this early settlement. Detailed Description of Building 6A In its final form, Building 6A contained a total of thirteen separate rooms (Figs 2 and 3). For ease of reference these have been numbered (Room Nos 32-44); the sequence continues on from room numbers previously allocated to Buildings 1 and 3. Several phases of development are clearly represented (Fig. 3) but the extensive and systematic robbing of the structure that occurred prior to the erection of the succeeding Building 6B has limited our detailed understanding of how Building 6A evolved. All the walls had been taken down to foundation level and virtually no associated floor-levels or occupation layers survived, although some PlateLATE II Building 6 – general view of excavations in 2004, looking west. Scale, 1 metre KEITH PARFITT 266 Fig. 3 Phase plan showing the development of Building 6A. THE ROMAN VILLA AT MINSTER-IN-THANET. PART 4: BUILDINGS 6A AND 6B 267 useful assemblages of finds were recovered from the fillings of the robbed hypocaust basements. The construction sequence set out below (Fig. 3, Phases 1-3) represents the most likely arrangement insofar as it can be determined from the surviving remains and must be treated as tentative. Phase 1: Accommodation Block (Rooms 32 and 33) As originally constructed, Building 6A consisted of a simple, rectangular, two-roomed structure butted directly onto the outside of the pre-existing south villa enclosure wall (Figs 2 and 3). Its south gable was positioned the same distance forward of the enclosure wall as the south end-wall of the central range of Building 4, opposite (Fig. 1). Unlike Building 4, however, no gap was left between the boundary wall and Building 6A. This effectively made Building 6A longer in comparison with its south-eastern counterpart, although the two structures were of comparable width (Table 1). TableABLE 1. COMPARATIVE ROOM SIzES IN BUILDING 4 AND BuiUIlDING 6A (Phase 1) Building Room Internal Sizes (m) 4 Central Range, North room 5.85 (n-s) x 6.00 (e-w) 4 Central Range, South room 5.50 (n-s) x 6.00 (e-w) 6A North Room (Room 32) 5.65 (n-s) x 6.00 (e-w) 6A S outh Room (Room 33) 8.25 (n-s) x 6.00 (e-w) Internally, Building 6A measured some 14.50m (n-s) by 6.00m (e-w). A cross-wall divided this block into two rooms of unequal size (Rooms 32 and 33). The north room (Room 32) was 5.65m (n-s) by 6.00m (e-w), whilst the south room (Room 33) was 8.25m (n-s) by 6.00m (e-w) (Table 1). These rooms were delimited by foundations between 0.60-0.70m wide and up to 0.50m deep. The foundations were constructed from alternate layers of rammed chalk rubble and small black pebbles with crushed marine shell (Plate I). Only very slight traces of the south foundation remained as it had been almost completely removed by the subsequent foundation trench for Building 6B (see below). Traces of buff mortar with small pebbles survived on the surface of the foundations, particularly along the east wall. This clearly represents the bedding for the robbed wall above. The scars of robbed wall stones could be discerned and traces of the original mortar edges of the missing walling survived in a number of places. The mortar deposit along the west wall KEITH PARFITT 268 was not well preserved but three large flint cobbles near the south-west corner were still attached to the foundation and provided some evidence for the nature of the main wall above. The east foundation had been inaccurately laid out by the Roman builders, forming a poor right-angle with the villa enclosure wall. On the evidence of the line of the mortar bedding, the mistake seems to have been rectified when the main wall was constructed above, but this caused a progressive mismatch between the wall and its underlying footing so that at the south-east corner of Room 33 only about half the width of the wall had actually been supported on its foundation. A sub-rectangular area of burnt soil (7167) occurred within Room 32. This could represent a hearth relating to the present building but it might equally well be connected with the succeeding Building 6B (Fig. 2; see below). The exact purpose of Building 6A during Phase 1 remains uncertain. Like Building 4, it seems most probable that it was a simple two-roomed domestic structure, perhaps for the accommodation of slaves or other workers who needed to be housed outside the main villa enclosure. Whatever the original function of the building, it is clear that an extensive refurbishment associated with a major change of use subsequently took place. This constitutes Phase 2 (see below). Phase 2: Bath-Suite (Rooms 34-38, 41a and 44) At some later stage, the Phase 1 structure underwent a major programme of alteration, which saw the addition of several extra rooms (Fig. 3, Rooms 36, 37, 38, 41a and 44) and the construction of a new hypocaust system within the original building (Rooms 34 and 35). These changes were apparently associated with the conversion of Building 6A into a bath building, provided with a small heated bath-suite and other rooms which need not have been directly connected with the new baths. Two separate rooms were added on the eastern side (Rooms 36 and 37), with two more along the south (Rooms 38 and 44). There was some slight evidence, in the form of a shallow robber-trench (Figs 2 and 3, F. 7308) for another room (Room 41a) lying on the south side of Room 44 but this had been largely removed by a later basemented room built in the same place (Phase 3, Room 41, see below). The additions of Phase 2 significantly enlarged the original building (Fig. 3). The main entrance into the remodelled structure may have been placed in the middle of the east wall, facing Building 4 and been reached via a small courtyard now created between Rooms 36 and 37 (Fig. 3). A small patch of rammed chalk rubble here (Fig. 2, 7175) could relate to the surfacing of this yard. Full details of the workings and layout of the bath-suite cannot be deduced from the surviving remains but it does seem clear that the new THE ROMAN VILLA AT MINSTER-IN-THANET. PART 4: BUILDINGS 6A AND 6B 269 hypocausted rooms (Rooms 34 and 35), together with their associated stoke-pit, lay entirely within the original building (Fig. 3). The four rooms added along the south and east sides were all surface built, without any evidence for hypocausts. Robbing of their shallow footings had been almost total, with only three lengths of foundation remaining in the south-eastern quarter. No definite floors survived within these unheated rooms but spreads of mortared pebbles in Room 38 (7288) and Room 44 (7306), together with small areas of chalk rubble in Room 37 and Room 44 (7204) might represent the last remnants of these (Fig. 2). Alterations to Room 32: the south end of this Phase 1 room had been cut away by the excavation of the stoke-pit for the new hypocaust system (Plate I). The original dividing wall between Rooms 32 and 33 was partially removed by this work (Fig. 3). It remains unclear whether the original wall had been replaced by another in a different position, of which no traces survived, or whether the entire room now became part of the new praefurnium. If so, the untouched area beyond the stoke-pit could have served as a fuel store. Alterations and additions to Room 33: the western side of this Phase 1 room was now taken up by the newly inserted stoke-pit, hypocaust furnace and two new heated rooms, Rooms 34 and 35 (Fig. 3; Plates III and IV). The remaining portion of the original room to the east was left PlateLATE III S toke pit and flue of Phase 2 hypocaust, Building 6A, cut through by cobble foundations of Building 6B, looking south-west. Scales, 2 metres and 50cm KEITH PARFITT 270 untouched and perhaps now became a narrow corridor providing access between the main entrance and Rooms 37 and 38 to the south. The stoke-pit for the new hypocaust system (Fig. 2, F. 7107) was roughly oval in shape and as excavated measured at least 3.40m (e-w) by 3.08m (n-s). It was 0.77m deep with steeply sloping, unrevetted sides (Figs 2 and 3; Plate III). In its base was a sub-rectangular ‘ash pit’, situated immediately in front of the furnace. The site of the furnace, itself, was represented by two 1.25m long spur walls built out from the north side of Room 34, bounding a central flue 0.40-0.50m wide. The spur walls survived to a height of 0.66m. They were constructed from tegulae set in a coarse opus signinum and showed evidence of intense heating (Plates III and IV). Room 34: this was a hypocausted room inserted into the western side of the Room 33. It lay immediately south of the furnace and represented the smallest room added to the new complex (Fig. 3; Plate IV). The associated walls had subsequently been extensively robbed but the original internal dimensions of the basemented area seem to have been about 1.20m (n-s) by 2.40m (e-w). The north and east walls were about 0.70m thick and survived to a maximum height of 0.60m (Fig. 4, section A, 7223). An opening in the PlateLATE IV Building 6A, Phase 2, Room 34 and hypocaust flue, looking north. Room 35 in foreground THE ROMAN VILLA AT MINSTER-IN-THANET. PART 4: BUILDINGS 6A AND 6B 271 Fig. 4 Building 6; sections A, B and C. KEITH PARFITT 272 middle of the north wall, 0.50m wide, connected with the furnace (see above). The walls were constructed with a central core of large flint cobbles faced with courses of broken tile, all bonded in a cream mortar containing frequent small black pebbles. The south wall, dividing Room 34 from the adjacent Room 35, was of similar width and construction but had been robbed down to floor-level. The base of a central aperture could just be identified – this would have allowed heated air to pass into Room 35 (see below). Apart from one small portion (Fig. 4, section A, 7222), the west wall of the room remained unexposed, sealed below the later foundations of Building 6B. The basement floor of Room 34 survived largely intact and was made of coarse opus signinum. Two mortar pads that had once supported subsequently robbed pilae were recorded on this. From its small size and location adjacent to the furnace, Room 34 may be readily identified as a hot plunge-bath. Its dimensions (see above) may be directly compared with those of the hot bath in Building 3, northern bath-house, which measured 1.15 by 2.12m (Parfitt in Perkins 2004, Room 25). Room 35: this was a larger hypocausted room situated immediately south of Room 34. Originally it seems to have been about 3.00m square, although the walls had been thoroughly robbed. Virtually nothing survived above basement floor level. The north and east walls appear to have been of a similar width and construction to the walls of Room 34 (Plate IV) but no trace of the west wall survived in the exposed area, its line again being largely hidden by later foundations. Inserted against the original Phase 1 foundation of Room 33, very little of the south wall remained, its line subsequently being followed by the end wall of Building 6B. However, a small fragment survived near the east corner; this was built of tile fragments set in a coarse opus signinum. It seems to have been quite narrow and butted onto the east wall. There can be little doubt, however, that these two walls belong to the same period and the butt joint between them must reflect the sequence of constructing a new basement inside an existing surface-built room. The basement floor of Room 35 survived intact and was again of coarse opus signinum. It had been laid at the same level as that in Room 34 and a few traces of mortar pads that had supported subsequently robbed pilae remained. Room 36: this unheated room had been added onto the eastern side of the Phase 1 building adjacent to Room 32, positioned on the north side of the possible entrance and butting the outside of the villa enclosure wall. No original floor surfaces survived and the east and south walls had been completely robbed. From the surviving foundations and robber trenches, the room seems to have had internal dimensions of about 3.10m (e-w) by 6.00m (n-s). The positioning of this room, at the north end of the building, might suggest that it did not form an integral part of the southern bath-suite, THE ROMAN VILLA AT MINSTER-IN-THANET. PART 4: BUILDINGS 6A AND 6B 273 although it could have provided accommodation for attendants working at the baths. In this latter context a large hearth (Fig. 2, 7174), located adjacent to the east wall may be significant. This was formed from tiles and clay and showed evidence of intense burning. Cut by the robber trench for the east wall, however, it cannot be certain whether the structure was contemporary with the use of Room 36 or an earlier feature. Traces of a preceding hearth occurred directly below, indicating that this fireplace had been used over a significant period of time. A few undiagnostic Roman pot sherds, broadly datable from the mid-first to the mid-third centuries, were recovered from the upper hearth. Room 37: this was another unhypocausted room on the eastern side of the building, placed on the south side of the possible entrance, adjacent to Room 33 (Fig. 3). Again, no original floors survived, unless a small patch of chalk rubble near the south-east corner was related. The north and east walls had been completely removed but a short section of the south wall foundation, dividing it from Room 38, remained. This was 0.75m wide and consisted of large flint cobbles set in a cream gritty mortar containing frequent small black pebbles. An extensive spread of cream mortar immediately outside the line of the east wall (Fig. 2, 7202) could represent an associated construction layer. Originally, Room 37 seems to have had internal measurements of 3.10m (e-w) by 4.55m (n-s). It was thus slightly shorter than Room 36 opposite but nevertheless appears to have been intended to be one of a pair of rooms, situated on the eastern side of the building and possibly flanking its main entrance. Room 38: this unheated, rectangular room lay at the south-eastern corner of the building. It had internal dimensions of about 3.00m (n-s) by 5.15m (e-w). Much of its walling had been robbed, although foundation fragments remained on the north, west and south sides. These foundations were all of similar construction and consisted of large flint cobbles set in a coarse cream mortar with frequent small flint pebbles and crushed marine shell. A mortared pebble layer (Fig. 2, 7288) across the western end of the room may represent the base of its floor. Room 41a: a shallow robber-trench (Figs 2-4, F. 7308) located to the south of Room 41 (assigned to Phase 3) appeared to be unrelated with that basemented room and hints at the existence of an earlier room on the same site, subsequently removed by the sunken structure (see below). If this is correct, it would seem that the earlier room was surface-built (unheated). It lay at the south end of the bath-suite, immediately south of Room 44 and was perhaps almost the same size. During Phase 3, Room 41 seems to have been a plunge bath (see below) and this might be a reflection of the function of Room 41a during the preceding phase. Room 44: this unheated room adjoined Rooms 35 and 38, perhaps with another (Room 41a) on the south side (see above). Its south and KEITH PARFITT 274 west walls had been completely robbed away (Fig. 5, F. 7327) but the foundations of the east wall, dividing it from Room 38, survived (see above). The room seems to have had internal dimensions of about 2.65m (n-s) by 3.30m (e-w). Spreads of chalk rubble (Fig. 2, 7204) and mortared pebbles (Fig. 2, 7306) within the eastern half of the room may represent the base of its floor. Most of the additions made during Phase 2 (Fig. 3), including two new hypocausted rooms with stoking-chamber at the northern end, are connected with the conversion of a pre-existing structure, originally built for a different purpose, into a bath building. With additional covered space to the north, the new bath-suite had overall dimensions of about 10m (n-s) by 11m (e-w). This is slightly larger than the villa’s other excavated bath-house, Building 3, which measured 7.15m (e-w) by 9.55m (n-s), with a projecting apse on the west side (Parfitt in Perkins 2004, 33). Within the Building 6A baths, functions for the individual rooms may be suggested with varying degrees of certainty. Thus, Room 34 immediately adjacent to the furnace, was clearly the Hot Bath, whilst the slightly larger adjacent Room 35 must have served as the Hot Room (caldarium). To the south of this, Room 44, although unheated, could perhaps have been the Warm Room (tepidarium), with Room 38 to its east as the Cold Room (frigidarium). To the north, Room 37 might have functioned as the Undressing Room (apodyterium), perhaps situated adjacent to the main entrance into the building. Accepting the existence of Room 41a, this could have contained a projecting cold plunge-bath (surface-built). Phase 3: Extension of the Bath-suite (Rooms 39­-43) A second enlargement of Building 6A subsequently occurred with the addition of a range of five more rooms along the south and west sides of the existing structure (Fig. 3, Rooms 39-43). Again, there had subsequently been extensive robbing of these rooms. Three of the new rooms were sunken (Rooms 41-43) and must relate to another hypocaust system, unconnected with that previously constructed during Phase 2. Of the sunken rooms, Room 41 might have been placed on the site of a pre-existing, surface-built room (Phase 2, Room 41a, see above). Room 42 at the south-west corner, at least in its final form, represented the stoking-chamber for the new heating system, with furnace flues here serving two adjacent rooms (Rooms 41 and 43). Room 39: this was an almost square, unheated room situated at the south-east corner of the new range, abutting Room 38 of Phase 2. It was less than 2.00m across and seems rather too small to have served as anything other than a store-room. A short length of its south wall foundation survived. This was 0.65m wide and consisted of a single course of large THE ROMAN VILLA AT MINSTER-IN-THANET. PART 4: BUILDINGS 6A AND 6B 275 Fig. 5 Building 6; section D across Room 43. KEITH PARFITT 276 flint cobbles set in coarse cream mortar with frequent small flint pebbles and crushed marine shell. The remainder of the foundations had been completely robbed away. Sometime after the robbing occurred, the south-west corner of the room had been cut by an oval pit (Fig. 2, F. 7253), the filling of which produced a coin of Faustina I (deified), dated to ad 141-145 (Holman and Parfitt 2005, coin 2003/29; see below), together with over 200 sherds of pottery datable to the period, c.ad 120-250. Room 40: this was another small unheated room abutting Room 38. It was rectangular in shape with internal dimensions of about 1.55m (n-s) by 3.25m (e-w); again the structure had been extensively robbed. Traces of a surface laid opus signinum floor survived along the north side, abutting the (Phase 2) wall of Room 38 but most of this floor had been removed by a rectangular robber pit (Fig. 2, F. 7307) that occupied much of the room’s interior. The soil and rubble filling this pit produced broken tile and a few pot-sherds of first/second-century ad date, which are all likely to be residual. A second substantial robber pit outside the south wall (Figs 2 and 3) probably marks the position of an external buttress. Room 41: the evolution of this basemented room was complex, although subsequent robbing had removed many of the structural details. Most of its walls had been demolished to base level. A shallow, high-level robber- PlateLATE V Room 41, showing wall foundation 7291 sealed by floor 7289. Scale, 50cm THE ROMAN VILLA AT MINSTER-IN-THANET. PART 4: BUILDINGS 6A AND 6B 277 trench (Figs 2-4, F. 7308), recorded on the south side, however, seemed to be unconnected with the main structure and possibly related to an earlier (Phase 2) room that had occupied the same site (Room 41a, see above). During Phase 3 the site of this inferred early room was cut away to create a new deep room (Room 41). Room 41, itself, lay on the south side of Room 44 (assigned to Phase 2). It was rectangular in shape with internal dimensions of about 1.50m (n-s) by 2.60m (e-w), which was perhaps slightly smaller than any Phase 2 predecessor. The primary basement floor to Room 41 consisted of a cream-pink mortar layer with a well-defined surface surviving in places (Fig. 4, 7290; Plate V). Apparently associated with this original floor at the south-west corner of the room was the base of a low-level drain passing through the south wall and discharging into an open ditch. This ditch (Figs 1-3, F. 7246) had possibly first been cut during Phase 2 to drain Room 41a. Inside Room 41, the drain outlet was formed from two imbrex tiles set in opus signinum (Plate VI). Three pot-sherds, caked in mortar but datable from the first to early second century came from its filling. The small size of the room overall, together with the presence of the drain, leaves little doubt that this was originally used as an unheated (cold) plunge bath. Subsequently, Room 41 underwent major changes, although not all the PlateLATE VI Detail of original drain outlet through south wall of Room 41. Scale, 50cm KEITH PARFITT 278 details are clear. On the north side, the primary floor was now used to support a new wall of mortared flint cobbles, about 0.53m wide (Figs 2-4, 7291; Plate V). This would have reduced the width of the room considerably but the construction of a new wall on this side may have been necessary in order to strengthen the earlier foundations here. The original excavation for the basement had extended about 1.00m below the bottom of the foundations of the adjacent surface-built Room 44 but does not seem to have included any new supporting wall. Perhaps the threat of collapse or subsidence in the existing structure now prompted the addition of this retaining wall, which need not have extended to the full height of the room. Later, however, the added north wall was demolished and sealed by a new floor (Plate V). This was composed of cream-pink mortar supported on a layer of flint cobbles (Figs 2-4, 7289). Much of the floor had subsequently been destroyed and it only survived reasonably intact at the eastern end. It seems most likely that it served as the base for a hypocaust, although no evidence for the suspended floor or the supporting pilae had survived later robbing. Nevertheless, clear evidence that Room 41 had once been heated is provided by the presence of a furnace flue, located on the eastern side of the adjacent Room 42, which can only have served PlateLATE VII Room 42 containing the remains of two hypocaust furnace flues THE ROMAN VILLA AT MINSTER-IN-THANET. PART 4: BUILDINGS 6A AND 6B 279 PlateLATE VIII High-level external drain leading from Room 41, looking north. Scale, 1 metre the present room (Plate VII). Also, the drain outlet at the south-western corner of the room had been re-modelled and was now set at a higher level. Immediately outside the building a new stone-lined drain with a tile base was constructed for a distance of about 2.65m (Figs 2-4, 7247; Plate VIII), before it discharged into the open ditch. Later, this exterior channel was lengthened by the addition of tiles and a second-hand ceramic water-pipe (Plate IX). Excavation and wet-sieving of the dark grey-brown gritty loam filling the stone-lined channel (Fig. 4, section C, 7282) produced little of significance. There were 58 potsherds, datable to the period c.ad 70-170. Inside Room 41, the base of the new drain outlet was about 0.70m higher than the original outlet and there seems little doubt that the intervening space had been occupied by a new hypocaust chamber. The evidence for the addition of a hypocaust, the raising of the drain outlet and the continued small size of the room can leave no doubt that the original cold bath had now been converted into a hot bath, with its floor set at a higher level. Such a fundamental change in the use of Room 41 may imply that there were other significant changes made in the arrangement of the bath-suite during Phase 3, although these are largely impossible to determine from the surviving remains (but note also changes to Room 42). KEITH PARFITT 280 PlateLATE IX Detail of ceramic water-pipe and tiles forming extension to high-level drain from Room 41. Scale, 50cm Room 42: this was a small sunken room at the south-western corner of the building, measuring internally 1.60m (n-s) by 2.75m (e-w). In its final form it had functioned as the stoking chamber for the hypocausts in Rooms 41 and 43 and contained a tile-built flue for each (Figs 2 and 3; Plate VII). However, there is a possibility that this room did not begin life as a praefurnium. The room seems very small to have accommodated two separate furnaces, if these both worked together. Moreover, the outer walls of the room, although heavily robbed had some appearance of being later inserts, rather than integral with the adjacent walls. Also, a deep n-s drain trench (Fig. 2, F. 7302) found outside the building led away from the south-west corner roughly at basement floor level. This cut did not extend through the wall and appeared to be earlier, and unrelated to the final structure. These details combine to suggest that Room 42 as a stoke chamber was created out of an earlier room of uncertain form and function. The external drain trench could be suggestive of its former use as a sunken bath but this is far from certain. A perforated lead drain cover came from the base of the drain trench, although clearly it was not in situ. Room 43: this basemented room was the largest of those added during Phase 3. It had been butted onto the western side of Room 44 and measured internally about 2.10m (e-w) by 3.05m (n-s), although its walls and floors THE ROMAN VILLA AT MINSTER-IN-THANET. PART 4: BUILDINGS 6A AND 6B 281 had been almost completely robbed away (Fig. 5, section D). This appears to have been a hot room, directly heated from Room 42 (Plate VII). There could have been no direct access into this chamber from the other hot room located to the north-east (Room 35), and if there was an entrance into Room 43 from within the main building this must have opened from Room 44 (Fig. 3). Alternatively, access could have been via a door set in one of the outside walls, on the north or west sides. If so, this might imply that the room formed a self-contained unit, perhaps a laconicum (a very hot, dry room). TableABLE 2. BUILDING 6A, INTERNAL ROOM SIzES IN METRES Room No. Size, N-S Size, E-W Phase Location and Notes 32 5.65 6.00 1 North end 33 8.25 6.00 1 Contains Rooms 34 and 35 34 1.20 c.2.40 2 Hypocausted, Hot Bath 35 3.00 c.3.00 2 Hypocausted, Hot Room 36 6.00 3.10 2 East side 37 4.55 3.10 2 East side 38 3.00 5.15 2 South end 39 1.50 1.60 3 Small store ? 40 1.55 3.25 3 ?External buttress on S wall 41 1.50 2.60 3 Sunken bath 42 1.60 2.75 3 Stoke chamber 43 3.05 2.10 3 Hypocausted, Hot Room 44 2.65 3.30 2 South end Water Supply and Drainage for Building 6A Some evidence for the water supply and drainage works associated with the Building 6A baths was recorded (Figs 1, 2 and 3). Of two large pits (Fs 7105 and 7128) located a short distance to the west of the building (Fig. 2), one was certainly a timber-lined well and the other probably so. Waste-water was taken from the baths by means of a ditch, remodelled several times and at some stage containing lengths of stone-lined channel (see above) and timber pipes. The Wells, Fs 7105 and 7128 A well shaft (F. 7105) located some 5m to the west of Building 6A (Fig. 2) was excavated to the top of the current water-table – a depth of about 4.60m from present ground-level. On the surface, the feature appeared KEITH PARFITT 282 as a large oval pit measuring 3.85m (e-w) by 3.50m (n-s). At a depth of 2.20m, however, the main pit narrowed into a vertical shaft, sub-rectangular in plan and measuring 0.80m by 1.05m. A specialist team was able to excavate this shaft for 2.40m when the top of the water-table was reached and safety considerations precluded any further work. No traces of any original lining survived within the shaft but the occurrence of vertical grooves at each corner would be consistent with the former presence of a timber plank lining of typical Roman construction. The well had been deliberately backfilled with large amounts of pottery and other domestic rubbish, together with quantities of tile and building rubble. Over 1,200 pot-sherds were recovered. The bulk of this material dates to the second century, with some residual pieces from the first century and a few sherds which could be as late as the early third century (see below, Assemblage D). A Celtic coin of Eppillus (c.ad 1-15) was also recovered (Holman and Parfitt 2005, coin ref. 2003/25; see below), which must be contemporary with the earliest dated (residual) pottery. About 1.50m to the north-east of F. 7105, the top of a slightly larger shaft was partially investigated (Fig. 2, F. 7128). It seems fairly certain that this represents a second well, although it could not be fully excavated. The shaft was half-sectioned to a depth of about 2.00m and like F. 7105 was also found to have been backfilled with large amounts of domestic rubbish, tile and rubble. More than 2,800 pot-sherds were recovered from its filling. As with the other shaft, most of the pottery dates to the second century with some residual pieces of the first century (see below, Assemblage E). A number of metal items also came from the filling of this feature. These included objects of copper-alloy and lead, together with iron nail fragments. Of particular importance were three similar, leaf-shaped, copper-alloy pendants. These have been identified as pieces of military equipment (Chris Tucker pers. com.) and are presently being studied in detail. Eventual subsidence of the shaft filling caused a surface hollow to form. This had been deliberately infilled with laid cobbles and tile fragments. Given the close proximity of a natural stream in the little valley below, the discovery of at least one definite well on the hill-top occupied by the villa complex is of particular interest. There seems little doubt that the close proximity of the bath-suite in Building 6A, with its requirement for large amounts of water, had dictated the positioning of the well, although nothing similar was identified in the area of the Building 3 baths (Parfitt in Perkins 2004). This may perhaps be explained by the fact that the area of the northern baths stood at an elevation about 3.50m above that of Building 6A, so that consequently any well there would need to be deeper and be more difficult to excavate. If, as appears likely, F. 7128 was a second well, it seems most probable that these were successive shafts supplying water for Building 6A. THE ROMAN VILLA AT MINSTER-IN-THANET. PART 4: BUILDINGS 6A AND 6B 283 H owever, the pottery recovered from their filling is of a broadly similar date range, which could imply that they were both back-filled at the same time. If so, they may have been open together. Perhaps a second ?deeper well became necessary because the amount of water required in the baths increased as the suite was enlarged during Phase 3 (see above). It may be reasonably concluded that both wells had been filled-in at the end of the second century ad, after the adjacent bath-suite had gone out of use, as part of the redevelopment of the area associated with the construction of Building 6B (see below). Drainage Waste-water from Room 41 (plunge bath) discharged into an open ditch (F. 7246, see above). This ran downhill to the south-west and was traced for a distance of some 12.00m in the excavation (Figs 1-3). It was about 1.35m wide and ranged in depth from 0.84 to 1.02m. Steps and ledges recorded in its profile suggested that the ditch had been re-cut on several occasions and this may be related to changes in the level of the main outlet from Room 41 (see above). Of particular interest was the discovery of three in situ wooden water-pipe fittings towards the north-eastern end of the ditch. There were two iron collars for connecting individual lengths of pipe; they were found about 2.00m apart and lay just above the base of the ditch. About 0.40m beyond the north-eastern collar a vertical end-cap of lead, with a central hole around 45mm in diameter, was discovered. This had been fixed to the end of the wooden pipe with iron nails and presumably marked the starting point of the pipe-line, beginning some 4.00m out from the south wall of Room 41. Since the fittings did not lie directly on the base of the ditch it would seem that the pipe-line had been a later improvement, although there was no clear indication that the open ditch had then been backfilled to enclose the new pipe-line. Wooden water-pipes are well known on many villa sites but they are invariably connected with incoming water supplies, as at Keston (Philp et al. 1991, 104-9), rather than waste-water removal. Their employment as an outlet pipe at Minster may reflect the re-use of a spare length of second-hand wooden pipe that had conveniently become available. The absence of any more iron collars in the ditch could suggest that this amounted to only a short length, although further evidence may have been destroyed by subsequent re-cuttings of the ditch (see below). Evidence for a Piped Water-supply to the Villa The re-used wooden water-pipes contained within ditch F. 7246 (see above) and the second-hand ceramic pipe employed at the end of the KEITH PARFITT 284 adjacent stone-lined channel (Plate IX), must ultimately derive from a piped water system. It could be argued that odd lengths of old pipe had been carried to the Minster site from some other settlement where such a supply existed (e.g. Canterbury). However, further fragments of ceramic pipe came from the filling of the later re-cutting of this same ditch (F. 7233, see below), and also from the general soil deposit covering Building 6, the filling of boundary ditch F. 5014 adjacent to Building 4, demolition deposits associated with the Building 3 baths (Parfitt in Perkins 2004, 45) and from Building 1. There now seems to be sufficient material (one complete pipe and at least 22 fragments) to suggest that all these pieces derive from a formal system of piped water that once served the Minster villa complex directly. The most obvious source of fresh running water would have been the little stream in the valley directly below the site but the problems of raising water uphill to the villa would have been very considerable (see Perkins 2004, fig. 3). More probably, a local spring issuing from the foot of the rising ground to the north of the villa was tapped, the water flowing down to the villa site by gravity. If so, a conduit of at least 400m is likely to have been necessary. No in situ traces of such a conduit have ever been discovered but ploughing over the centuries may well have destroyed much of the evidence. Demolition and Levelling of the Building 6A complex Once the use of Building 6A had come to an end, the structure was systematically robbed of its usable stone, tile and brick, presumably to provide the raw materials needed for new building work. The hypocausts were stripped to basement level, although the useless, heavily fire-damaged flue walls adjoining Room 34 were left. When everything had been removed, the resulting wall robber-trenches and open basements were backfilled with rubble, soil and domestic rubbish. The adjacent well shafts were also infilled and it would seem that the entire area had been carefully levelled in order to create a flat site; soon after, Building 6B was laid out here (see below). S ignificant quantities of finds were recovered from the backfilling of the Building 6A hypocaust basements and the adjacent well shafts, with smaller amounts coming from the wall robber-trenches. This material included large amounts of broken tile and brick, painted plaster, animal bone, shell and some metalwork. Most of this is likely to be derived from the occupation of Building 6A itself or the pre-villa settlement represented by the early boundary ditch (F. 7197/7284, see above) but further material could have been brought in from other parts of the site. Two dead dogs had been deliberately placed just above the base of Room 43 (Fig. 5, section D, 7281; Plate X, 7298). THE ROMAN VILLA AT MINSTER-IN-THANET. PART 4: BUILDINGS 6A AND 6B 285 The soil and rubble filling the Phase 2 stoke-pit and the adjacent hypocaust flue yielded more than one thousand pot-sherds ranging in date between the late first and second centuries (see below, Assemblage C). The fill of the basements to Room 34 (Fig. 4, section A) and Room 35 produced over 500 more sherds, again with a late first- to second-century date-range. The infilled hypocaust basements associated with the Phase 3 rooms along the south and west sides of Building 6A (Fig. 4, section B; Fig. 5, section D) also produced large quantities of dumped material. Altogether, almost 1,400 pot-sherds were also recovered from these deep rooms. The upper filling of Room 43 (Fig. 5, section D, 7206) also yielded a coin of Hadrian dated ad 128-138 (Holman and Parfitt 2005, coin ref. 2003/34). Again, virtually all the pottery is datable to the later first or second century but there are two or three intrusive pieces from the upper filling which date from the late third-fourth century. Dating of Building 6A As with Building 4, a general lack of occupation layers and other well-stratified deposits associated with Building 6A limits any attempt to closely determine its period of use. However, the date of the pottery from the filling of the large ditch (F. 7197/7284, see Assemblage A) cut through by the Phase 1 walls indicates that the original structure could not have been PlateLATE X Room 43, dog burial, 7298. Scale, 50cm KEITH PARFITT 286 erected until after c.ad 60. Moreover, the fact that the building had been butted onto the pre-existing villa enclosure wall suggests a significantly later date than this. The large collection of dumped pottery recovered from the robbed basemented rooms and wall trenches (over 3,000 sherds) appears to provide the best guide as to the period of use of the building, although all of it is clearly re-deposited. Virtually all this material dates to the first or second century ad, with only two or three sherds that are later and apparently intrusive. The earliest pieces must relate to the period of pre-villa occupation identified. Overall, it would seem that the construction, occupation and abandonment of Building 6A falls entirely within the second century. The structure was probably finally demolished around ad 200. building BUILDING 6b Shortly after the area of Building 6A had been cleared and levelled, an entirely new structure was laid out across the site, Building 6B. This was of broadly similar size to the original Building 6A and occupied virtually the same position, again butting onto the outside of the main villa boundary wall (Fig. 6). There was, however, a slight shift in the positioning of the new building, so that its side walls lay about 1.30m to the east of the corresponding walls of the original Building 6A (Fig. 2; Fig. 4, section A; Plate I), whilst the successive south walls occupied much the same position (Plate II). The new structure appears to have included no dividing cross-wall and thus consisted of a single large room (Fig. 6; Plate I), measuring internally 14.55m (n-s) by 6.10m (e-w) – almost exactly the same size as Phase 1 of the preceding Building 6A, which was 14.50m (N-S) by 6.00m (e-w). Buried at only a shallow depth, preservation of Building 6B was generally poor and again no associated floors or occupation layers had survived. Moreover, the structure had eventually also been thoroughly robbed of its stonework (see below). Cutting through the footings of Building 6A, the outer wall foundations of Building 6B were 0.60-0.90m. wide and 0.30-0.45m deep (2-3 courses) but had been completely robbed away at the south-west and north-west corners (Fig. 6). No traces of the main walling above remained at any point. The foundations were mostly composed of fist-sized flint cobbles carefully laid in brown sandy clay (Plates I-III). Many of the flints showed traces of adhering mortar indicating that they were re-used. At one point in the eastern foundation, a considerable thickness of clay separated the individual flint courses. Along much of the western side the foundations had been taken to a greater depth to compensate for the unconsolidated back-filling of the Building 6A hypocaust basements (Fig. 4, section A). THE ROMAN VILLA AT MINSTER-IN-THANET. PART 4: BUILDINGS 6A AND 6B 287 Fig. 6 Plan of Building 6B and associated features. KEITH PARFITT 288 Very few internal features may be directly associated with Building 6B, although a sub-rectangular ash-filled pit (F. 7170), centrally placed adjacent to the south wall, probably relates to a hearth or oven within this building (Figs 2 and 6). It was 0.39m deep and the filling produced 73 pot-sherds dating to the period c.ad 170-250, together with fragments of tile and shell. Another area of burning at the north end (7167) represents a hearth (undated), which could relate to either Building 6A or 6B (Figs 2 and 6, see above). There can be no doubt that Building 6B was erected soon after the levelling of Building 6A (see above). Indeed, much of the stone-work from the earlier structure is likely to have been re-used in the later building (see above). The date of the pottery recovered from the infilling of the hypocaust basements and robbed wall trenches associated with Building 6A implies that 6B must have been constructed around the start of the third century. This is broadly consistent with the dating of the pottery recovered from the ash-filled pit, F.7170, discovered within the later building, which seems to have been infilled sometime after ad 200. As with Building 6A, following its abandonment the walls of Building 6B had been extensively robbed. The filling of the associated robber-trenches produced about 250 sherds of Roman pottery. The bulk of these are again of late first or second century date and must be residual but a few later pieces indicate that the trenches could not have been backfilled until after c.240. On balance, it would seem that the robbing of Building 6B occurred sometime during the second half of the third century. This was also the period when the abandonment and robbing of Building 4 seems to have occurred (Parfitt 2006, 131; see below). Re-cut Ditch, F. 7233 At some late stage, the original drainage ditch leading away from Room 41 of Building 6A was again re-cut (as F. 7233). The north-eastern terminal of the new ditch lay some distance from the building and clearly it cannot now have taken any waste-water away from the structure (Fig. 2). The re-cut ditch was quite shallow (0.49m) and was presumably intended to perpetuate the line of the original, which by this time must have become a well-established boundary running across the site. It seems most likely that the final re-cutting of this ditch was broadly contemporary with the use of Building 6B, rather than Building 6A (Fig. 6). Eventually, the re-cut ditch had been backfilled with a dark grey clay loam containing quantities of tile, pebbles and chalk lumps, together with much domestic rubbish including pottery, animal bone and oyster shell. Over 700 sherds of pottery were recovered. Most of this was residual material of late first and second century date but it also included a few later pieces dating to the third century. There were also three fragments THE ROMAN VILLA AT MINSTER-IN-THANET. PART 4: BUILDINGS 6A AND 6B 289 of ceramic water-pipe which might be derived from an extension of the wooden pipe-line laid in the preceding ditch (F. 7246, see above). the THE FINDS Excavations in the area of the Building 6 complex produced a considerable quantity of finds but details of most of these must be held over for future reports. Notes on the coins and pottery recovered are given below since these provide the main dating evidence. Coinage from the Area of Building 6 Four coins were discovered in the area of Building 6. These have been detailed previously (Holman and Parfitt 2005, tables 1 and 2). Three are Roman; a coin of Nerva (ad 96-97) came from general soil over the building (coin ref. 2003/1), one of Hadrian (ad 128-138) was recovered from the upper filling of Room 43 (coin ref. 2003/34) and one of Faustina I (deified, ad 141-145) was found in the late pit F. 7253 cutting the robbed walls of Room 39 (coin ref. 2003/29). There was also a residual Celtic coin of Eppillus, c.ad 1-15, from the filling of well, F. 7105 (coin ref. 2003/25). Another corroded disc from the same feature (coin ref. 2003/11) is not certainly a coin. The date-range of this small coin assemblage is broadly comparable with that of the pottery recovered (see below) and seems to reflect the main period of use of this part of the site. Pottery from the Area of Building 6 (not illustr.) based on a report by Malcolm Lyne Around 10,000 sherds of pottery were recovered from the area of Buildings 6A and 6B, although little material can be directly associated with Building 6B. Significant assemblages came from the filling of the early ditch, F. 7197/7284, pre-dating Building 6A (Assemblage A); the filling of bath-suite drainage ditch F. 7246 (Assemblage B); the filling of hypocaust stoke-pit F. 7107 (Assemblage C); and the filling of the two well shafts west of the buildings (Assemblages D and E). Further substantial groups were recovered from the robbed basements of the hypocausted rooms associated with Building 6A. Virtually all the pottery recovered is of first- or second-century date, with some pre-Conquest material. Only a small number of sherds date to the later Roman period, which strongly implies that there was little activity on this part of the site during the late third and fourth centuries. The principal pottery assemblages are described below, ahead of their more detailed publication in a future report (Lyne 2006). The fabric codings are KEITH PARFITT 290 those used by the Canterbury Archaeological Trust (Macpherson-Grant et al. 1995). Assemblage A: from the fills of early ditch F. 7197 beneath Building 6A (contexts 7194, 7195, 7196 and 7201) A total of 519 sherds (9,448g) of pottery came from this feature. The assemblage has vessels in the local silt-tempered ‘Thanet Dry’ fabric (BER16) making up three-quarters of it by Estimated Vessel Equivalent (EVE). Of these vessels, the overwhelming majority are jars, with only small numbers of Gallo-Belgic platter copies and butt-beakers being present. Jars are almost equally divided between bead-rimmed and necked vessels. Necked-jars and storage-vessels in ‘Belgic’ grog-tempered fabrics B2, B2.1 and B2/R1 make up a further 22 per cent of the assemblage and there are bodysherds from a jar in chalk-tempered fabric B4, a Gallo-Belgic Whiteware butt-beaker in fabric BER7 (c.10 bc-ad 60) and a biconical vessel in wheel-turned fine Upchurch greyware fabric R16 (c.43-100). Many of the sherds in fabrics B2/R1 and BER16 are discoloured and overfired wasters and indicate nearby production of vessels in both of these fabrics, with the grog-tempered fabric B2/R1 being mainly reserved for storage jars and ‘Thanet Dry’ fabric being used for the rest of the products. The presence of the vessel in calcined-flint-tempered fabric P2 and body sherds in the ‘Thanet Dry’ fabric variant with silt and sparse calcined-flint filler suggest that this assemblage began to accumulate during the 50 years or so before the Claudian invasion. The biconical fragments in Upchurch fine greyware in turn indicate continued accumulation after ad 43, although a near absence of other sherds in Romanised fabrics suggests that the assemblage had ceased to be added to by c.ad 50/60. Assemblage B: from the fills of ditch F. 7246 draining the bath-suite attached to Building 6A (contexts 7237, 7239, 7243, 7245 and 7322) The constructional contexts associated with Building 6A and the bath-suite added to it lacked ceramic dating evidence and internal occupation deposits were absent. The fills of ditch F. 7246 draining the putative plunge bath in Room 41 did, however, yield 135 sherds (1,779g) of pottery contemporary with the earlier part of the life of the building. This assemblage was totally lacking in sherds from BB2 vessels but had ‘Thanet Dry’ fragments accounting for a modest 12 per cent by sherd count and ‘Belgic’ grog-tempered sherds for a somewhat greater 28 per cent. The 40 sherds in fine North Kent greyware fabric R16 include fragments from a beaker of Monaghan (1987) Class 2A1 (c.ad 70-100) THE ROMAN VILLA AT MINSTER-IN-THANET. PART 4: BUILDINGS 6A AND 6B 291 and a necked-bowl of Class 4J1 (c.ad 43-110/120). The samian includes fragments from a South Gaulish Dr 36 (c.ad 70-110) and a similar vessel in Martres-de-Veyre fabric (c.ad 90-120). Sandy Canterbury greywares in fabric R5 make up a mere 12 per cent of an assemblage which can probably be dated c.ad 70-120. Assemblage C: from the fills of stoke-pit F. 7107 and ash-pit F. 7219 for the hypocaust in Building 6A (contexts 7106, 7116, 7125, 7138 and 7145) The 1,069 sherds (12,921 g) of pottery from these features date to between c.ad 70 and some time during the last quarter of the second century. The make-up of the assemblage is similar to those from wells, Fs 7105 and 7128, and includes fragments from Central Gaulish samian forms Dr 18/31 (c.120-150), Dr 27 (c.120-150), Dr 31 (c.150-200), Dr 33 (c.120-200), Dr 36 (c.120-200), Dr 37 (c.120-200) and Curle 11 (c.120-140). Thameside products include BB2 bowls of Monaghan (1987) Classes 5D4 (c.110/20-200) and 5D5 (c.110/20-150/70), jars of Class 3J1 (c.110/20-150/90) and a necked-bowl of Class 4A2 (c.170-230). Fine Upchurch greywares include fragments from a Class 1B7 flask (c.120/50-180/90), biconicals, and there are fragments from two Class 3H8 jars in Thameside greyware fabric LR2.1 (c.170/90-210/30). Assemblage D: from the fills of well F. 7105 (contexts 7104, 7120, 7121, 7129, 7143, 7157, 7177, 7178 and 7198) The 1,211 sherds (25,050g) of pottery from this feature were quantified by EVEs. This assemblage has much in common with that from the adjacent well, F. 7128, but also includes elements of post- ad 150 date. Thameside products make up a larger percentage (40 per cent) than of the assemblage from the other well. The BB2 wares include many examples of the c.110-180 dated bowl forms with latticed decoration as well as one undecorated example of ad 150/170-250 date. The pottery tends to be fresher than that from the other well and includes large fragments from jars and bowls in Transitional ‘Belgic’-Native Coarse Ware and jars in developed Native Coarse Ware of post- ad 170 date. There are fewer Canterbury products than from the other well but these include fresh fragments from a number of flagons of late second- to early third-century date. Other wares include fragments from Central Gaulish samian forms Curle 11 (c.120-140) and Dr 38 (c.140-200), a Colchester Whiteware wall-sided mortarium (c.170-250) and a neck-cordoned bowl of Monaghan (1987) Class 4A2 in blue-grey Thameside greyware with burnished latticing on its shoulder, c.120-200. KEITH PARFITT 292 Assemblage E: from the fills of well F. 7128 west of Building 6A (contexts 7112, 7127, 7130, 7131 and 7165) Although only its upper levels were excavated, this feature yielded 2,890 sherds (47,800g) of pottery. The assemblage is one of the largest from the entire site. Amounts of handmade native and derived wares in fabrics B2/R1, R1 and BER16 are comparatively insignificant (15 per cent) in an assemblage with a wide range of fabrics. The most important source of pottery remained the Canterbury kilns, with sandy greyware fabric R5 accounting for a third of the pottery and oxidised fabrics R6.1 and R6.3 for a further 6 per cent. Most of the vessels in fabric R5 are jars and carinated bowls, which appear to have been supplied to Minster in roughly equal numbers: lids are the only other significant form in this fabric. Thameside products in Fabrics R14, R16, R17, LR2.1 and LR2.2 make up another third of the assemblage and include BB2 everted-rim cooking-pots (c.110/20-170), bowl types 5D2 (c.110/20-150/80) and 5D0.2 (c.120-150/80), as well as dish types 5E2.4 (c.110/20-210/300) and 5F3 (c.130-270) in fabric R14. Vessels in fine North Kent greyware fabric R16 include types 2A3 (c.100-150), 2A4 (c.130-170), 3B1.2 (c.50/70-100), 4H1 (c.90-130), 4J1.7 (c.43-120), 5B3.1 (c.70-130), 7A2.3 (c.43-120/40) and 7A3.1 (c.43-100). Fragments in oxidised Hoo fabric are mostly from flagons: the Thameside greywares include sherds from a Monaghan (1987) Class 4F4 jar (c.70-120) and a flask of probable second-century date. Other finewares include fragments from Central Gaulish samian forms Dr 18/31 (c.120-150), Dr 27 (c.120-150), Dr 42 (c.120-150) and Dr 44 (c.130-200) and East Gaulish Samian form Dr 37 (c.140-200). Most of the other fineware fragments are from roughcast cornice-rim beakers in colour-coated Central Gaulish, Sinzig and Cologne fabrics (c.60-140, 130-250 and 130-250, respectively). Fragments from a white-slipped poppyhead beaker in Highgate Wood C fabric are also present. Other Finds Other finds from the area of Building 6 included large quantities of tile and brick, painted plaster, a range of small finds and food debris in the form of animal bone and marine shell (mostly oysters). The two dog skeletons recovered from the filling of Room 43 are of some interest and may represent ritual depositions. As with previous site reports, details of these various groups of finds will be set out in separate future studies. THE ROMAN VILLA AT MINSTER-IN-THANET. PART 4: BUILDINGS 6A AND 6B 293 general GENERAL DISCUSSION Building 6A: although heavily robbed and poorly preserved, a number of general points about the nature of Building 6A can be made. Firstly, it appears to have been a subsequent addition to the original villa complex, although it was constructed in an area where there had been some earlier, pre-villa occupation. It may have originally been built as accommodation for estate workers but there can be no doubt that during Phase 2 Building 6A had been converted into a bath-suite, which was enlarged and modified during Phase 3 (Fig. 3). It seems quite clear that the baths had been inserted into and added onto a pre-existing structure. The extent of the work indicates that the changes amounted to the complete conversion of an earlier building to a new use, rather than the addition of a bath-suite to a structure that otherwise continued to function much as before. There now seems to have been little, if any, space within the building for domestic accommodation. Thus, the new bath-house would appear to have stood as a detached structure, outside the main villa enclosure and not directly connected to any domestic complex. Another bath building (Building 3) has been previously excavated on the site. This stood inside the walled compound, immediately adjacent to the main villa house, some 80m uphill to the north of Building 6A (Parfitt in Perkins 2004, 33-45). The date of construction of neither of these two baths is known precisely but the available pottery and coin dating evidence indicates that both were in use during the second century and they must have been broadly contemporary. Both were additions to the original villa complex. From its close proximity to Building 1, it is immediately obvious that Building 3 was intended for use by those who lived at the main house. In contrast, the Building 6A baths lay outside the villa enclosure. The implication seems to be that this second bath-suite was intended for the use of estate workers, whose presence within the villa owner’s private walled compound was not acceptable. Also situated outside the villa enclosure, some 50m to the east of Building 6A, was Building 4. This structure went through several phases of expansion but seems to have functioned as a modest dwelling throughout. It was perhaps the villa bailiff’s house (Parfitt 2006, 132). Significantly, this building was never provided with its own bath-suite and it now seems highly likely that Building 6A, opposite, was subsequently modified to provide this facility for the residents of Building 4 and possibly others. Once Building 6A had ceased to function as a bath-house, at the end of Phase 3, it was demolished and quickly replaced by a completely new structure, Building 6B, which very clearly was not a baths (see below). Obviously, the requirement for a bath-suite on this part of the site had been confined to a specific time period. With the demolition of the Building 6A baths, Building 4 would seem to have been left without any KEITH PARFITT 294 such facilities, even though the structure itself had been enlarged and improved. This may imply that the northern baths were now available for use but it must also raise the question as to whether a new baths complex, whose location is presently unknown, had been constructed. Neither the Building 3, nor Building 6A baths belong to the earliest phase of villa occupation and it has been previously speculated that an earlier bath-suite might have existed for the occupants of the main house (Parfitt in Perkins 2004, 48). Yet another bath-suite, replacing the worker’s baths in Building 6A may now also be implied. One possibility is that an undiscovered bath-house lies by the stream in the valley below the site, hidden under a significant accumulation of down-washed soil. Building 6B: this was probably erected at the start of the third century. The substantial foundations indicate that this was again a masonry structure and much of the raw material used in its construction is likely to have come from the previous building. What is of particular interest is that the replacement building was virtually the same, in terms of size, shape and positioning as Building 6A in its earliest phase (see Fig. 2). This might suggest that these two structures had a broadly similar function. The new building apparently consisted of a single large room, provided with at least one hearth/oven pit. This would be consistent with a domestic function, as previously suggested for the earliest phase of Building 6A (see above). The new building contained no hypocausted chambers. In its turn, Building 6B had been demolished and thoroughly robbed for its stone. The date at which this later phase of robbing occurred cannot be closely determined but the exact, methodical work is very similar to the way in which the earlier building had been dismantled, and in the absence of any associated post-Roman finds, the implication is that this, too, occurred during the Roman period. General Consideration of the Southern Buildings From the positioning, outside the south-western corner of the walled villa enclosure, it seems clear that the Building 6 complex was intended to mirror the arrangements at the south-east corner, as represented by Building 4 (see Fig. 1). Indeed, the size and plan of Building 6A during Phase 1 is closely comparable to the earliest phase of Building 4 (Central Range) and there seems little doubt that balance and symmetry to the southern elevation of the villa complex, in true Classical style, was desired in order to impress visitors as they approached the site up-slope from the south, or sailed past the estate through the Wantsum Channel (Fig. 1). However, Buildings 4 and 6A were not exactly the same and it seems unlikely that they represent a matched pair of buildings laid out together, even if their period of use covered more or less the same time-span. THE ROMAN VILLA AT MINSTER-IN-THANET. PART 4: BUILDINGS 6A AND 6B 295 A number of minor differences in constructional detail between the two buildings may be noted; thus Building 6A was butted directly onto the outside face of the south villa enclosure wall, whereas Building 4 initially stood detached from it. There was seemingly a greater distance left between the corner of the villa enclosure wall and the site of Building 6A, than from the opposite corner to Building 4. The foundations of the two buildings were of differing construction. Although both Buildings 4 and 6A contained two large rooms, the overall length of Building 6A was more than the Central Range of Building 4 because it butted directly onto the villa enclosure wall (Fig. 1; Table 1). All of this seems to suggest that Buildings 4 and 6A were built at slightly different times, although it is not possible to determine which was the earlier from the available evidence. Nevertheless, there cannot be much difference between their construction dates, and both are certainly later than the erection of the villa’s walled enclosure. The implication seems to be that some overall architectural scheme for the complete villa complex existed ‘on paper’ but that this was only implemented on the ground piecemeal, with individual buildings being erected following the planned arrangement as and when circumstances allowed. In detail, building development at the south-west corner was rather different to that at the south-east. In terms of location, however, Buildings 4 and 6 are linked because they were both placed outside the villa enclosure, closer and more easily accessible to each other than to the main house further up-slope within its private walled compound. Their positioning suggests that both structures were for the use of estate workers whose presence within the main enclosure was not wanted. In their earliest form, the two buildings each apparently consisted of a two-roomed dwelling, although Building 6A was never enclosed by corridors like Building 4. Both buildings had subsequently been modified to include hypocausted rooms. With Building 6A the changes were extensive and amounted to the insertion of an entire bath-suite, whilst in Building 4 only one small heated room was constructed in the outer corridor, perhaps as an office (Parfitt 2006, 132). It seems most likely that the Building 6A bath-suite was primarily intended for the use of the occupants of Building 4. During their final phases, Buildings 4 and 6 were rather different. In Building 4 the main domestic range remained unchanged but a new outer corridor replaced the earlier one. This now enclosed all four sides of the building and extended across the line of the villa’s south enclosure wall, perhaps allowing easier access into the main villa compound, with its own bath-house. In contrast, Building 6A, with its extended bath-suite, was eventually demolished, to be replaced with an entirely new building (6B) of very similar size and shape to the original pre-baths structure. It would thus appear that the building had reverted back to being an accommodation block, now probably of inferior status to the enlarged Building 4. KEITH PARFITT 296 Sometime during the later Roman period both the enlarged Building 4 and Building 6B were demolished and extensively robbed for their tile and stone. The open robber-trenches and basements were filled in and the entire area seems to have been tidied up, levelled and left unoccupied. Clearly, some major changes in the overall layout of the site had occurred, leaving this lower part of the villa complex abandoned. acknowledgements The Abbey Farm Villa excavation team again wish to express their gratitude to the Kent Archaeological Society for funding and administrating the excavations. In particular, the then President, Paul Oldham must be warmly thanked, together with David Bacchus and Chris Pout of the Fieldwork Committee. Farmer Mr Jack Clifton again arranged access to the site and keenly followed developments throughout the course of the work. Thanks are also due to Minster Parish Council, and the staff of Minster Agricultural Museum for their kind co-operation. Dave Perkins, although unable to participate in the fieldwork, has continued to take a keen interest in the project. Members of Thanet Archaeological Society have spent much time processing and cataloguing the finds. The Project Director (Keith Parfitt) would like to thank the Supervising Archaeologists, Tim Allen, Emma Boast, Barry Corke and Ges Moody, for their support and enthusiasm during all phases of excavation and post-excavation work. Susan Deacon, Oliver Gardner and Ges Moody prepared the site drawings for publication. Finally, the hard work of all the many excavators must be recorded here – as always, without their efforts little would have been achieved. bibliography H olman, D.J. and Parfitt, K., 2005, ‘The Roman villa at Minster-in-Thanet. Part 2: The late Iron-Age, Roman and Later Coinage’, Archaeologia Cantiana, cxxv, 203-228. L yne, M., 2006, ‘The Pottery from the Abbey Farm Minster Roman Villa’ (archive report commissioned by KAS). Macpherson-Grant, N., Savage, A., Cotter J., Davey, M. and Riddler, I., 1995, Canterbury Ceramics 2. The Processing and Study of Excavated Pottery. Monaghan, J., 1987, Upchurch and Thameside Roman Pottery. A ceramic typology for northern Kent, first to third centuries A.D. (BAR 173). Parfitt, K., 2006, ‘The Roman Villa at Minster-in-Thanet. Part 3: The Corridor House, Building 4’, Archaeologia Cantiana, cxxvi, 115-133. Perkins, D., 2004, ‘The Roman Villa at Minster-in-Thanet. Part 1: Introduction and Report on the Bath-house’, Archaeologia Cantiana, cxxiv, 25-49. Philp, B., Parfitt, K., Willson, J., Dutto, M., and Williams, W., 1991, The Roman Villa Site at Keston, Kent. First Report (Excavations 1968-78), KARU, Dover.