The Roman Villa at Minster-in-Thanet

THE ROMAN VILLA AT MINSTER-IN-THANET.

PART 5: THE MAIN HOUSE, BUILDING 1

keith parfitt, dave perkins, emma boast and gerald moody

This is the fifth paper detailing results of the Society’s excavations, undertaken jointly with the Trust for Thanet Archaeology and the Thanet Archaeological Society, at the Roman villa site on Abbey Farm, Minster-in-Thanet (NGR TR 3135 6463, centred). The report describes the main villa house, designated Building 1. This was the first structure to be investigated, under the direction of Dave Perkins, between 1996 and 1999 (Perkins 1996; 1999; Plates I-III). Although the work extended over four seasons in all, most of the building was excavated during 1996 and 1997. It was the largest structure examined and detailed analysis and definition of the building sequence recorded, on what was in the early years, a new and unfamiliar site has required considerable study and thought, hence some delay in reporting.

Building 1 was the principal dwelling of the villa complex and consisted of a ‘winged-corridor’ house of well-known Romano-British type. It stood on a gentle slope at an elevation of 16-17m od and was constructed on an east-west axis, facing downhill to the south. This arrangement would have provided its residents with panoramic views across the nearby Wantsum Channel, with the Kentish mainland beyond (Perkins 2004, figs 1 and 2). Immediately to the west of Building 1, set neatly in line with the end of its west wing, was a detached bath-house (Building 3). This has been described previously (Parfitt in Perkins 2004). The two buildings occupied prime position within the upper part of a large walled compound (Figs 1 and 2). Other substantial buildings lay outside the south wall of this enclosure (Fig. 1, Buildings 4 and 6; Parfitt 2006; 2007).

During the course of the excavations at Minster a number of early features were located below subsequent Roman buildings. These clearly indicate previous settlement on the site, although they are unlikely to all belong to the same period. Traces of several such features were located under Building 1, sealed by a subsoil layer that was cut through by the foundation trenches for the building. Of particular interest was a section of ditch revealed below Room 1 (not shown on Fig. 2). Its filling yielded 23 sherds from a jar of Thompson (1982) Type C3. This vessel may be dated to c. 50 bc-ad 50 (Lyne 2002) and seems to be associated with a late Iron Age/early Roman settlement that predated the villa complex. Another ditch broadly datable to this period was located below Building 6 (Parfitt 2007, 263) and two others found to the south of Building 4 may also belong here (Parfitt 2006, 117). A more detailed consideration of the pre-villa activity on the site will be given in a future report.

description of the main house

The outline of Building 1 was first revealed on an air photograph taken in 1979 (Perkins 1996, 325; 2004, 25). Subsequent excavation demonstrated that the structure had been thoroughly robbed after its abandonment and in more recent times had suffered from extensive (and continuing) plough-damage, so that virtually no floors, occupation levels or associated stratification survived anywhere outside the basemented rooms. All the main walls of the building had been removed to foundation level with some sections totally destroyed. Little structural detail beyond the ground-plan thus remained in situ (Plates I-III). The plan recovered, however, is essentially complete and from this it can be seen that the building’s layout was intended to be closely symmetrical. Nevertheless, the original surveyors had created several imprecise right-angles during the initial laying-out stages which slightly skewed the structure (Figs 1 and 2).

A significant number of additions and modifications were identified within the surviving fabric, implying that the final building was the product of a quite complex structural history. The construction sequence set out below (Phases 1-7; Fig. 2) represents the most likely arrangement, in as much as it can be determined from the surviving remains, but it must be treated as tentative.

At its most extensive, Building 1 measured some 43m (e-w) by 33m (n-s) and contained over 20 separate rooms and corridors (Room Nos. 1-19, 28-31; Tables 1 and 2). These were laid out in the form of a central range with two southward projecting wings, surrounded on all sides by broad corridors. Unlike many comparable villas where they were later additions, it is clear that Minster was designed with projecting wings from the first, although the enclosing corridors appear to have been subsequent additions (Fig. 2).

A centrally placed entrance porch (Room 30) extended from the south side of the Central Range and eventually a boundary wall constructed be-tween the ends of the wings created a small private courtyard immediately in front of the house. There was a projecting apsidal room centrally placed along the north side (Fig. 2, Room 10; Plate IV), with another at the end of the West Wing (Room 28). Both were hypocausted and both seem to have been later additions to the original structure (see below, Phase 6).

Phase 1, Central Range and Wings

As originally constructed, Building 1 consisted of a substantial, winged villa containing eleven rooms (Fig. 2). Only the foundations remained. There was no evidence for any enclosing corridors at this early stage, unless these were of timber construction and had been destroyed by later developments. The foundations were carefully constructed from alternate layers of rammed chalk rubble and small black flint pebbles. They were between 0.55-0.95m wide and up to 0.40m deep. In a few places traces of cream-white mortar occurred on the surface of the foundations and this must represent the bedding of the lost wall masonry.

A row of seven unheated rooms (Rooms 1, 5, 8, 9, 12, 13 and 14) con-stituted the Central Range. The middle three (Rooms 8, 9 and 12) were large, rectangular and of broadly similar sizes (Table 1). They were flanked on either side by a narrow cross-passage or stairwell (Rooms 5 and 13). Beyond, the end-room at the east (Room 1) was of roughly similar size to the three middle rooms, whilst that at the west (Room 14) was slightly smaller and square in plan. Its smaller size is accounted for by the presence of a correspondingly larger room in the adjacent West Wing (Room 15a).

Based on its positioning, in the middle of the Central Range, Room 9 may perhaps be identified as a summer triclinium (dinning room), provided with views through the main entrance and out across the walled courtyard, in true Roman style. Within this room a pit showing evidence of burning occurred (Fig. 2). This clearly represented a hearth pit but whether it was contemporary with the use of the villa or an earlier feature could not be determined from the surviving remains.

The East and West Wings were of equal length and projected some 11m forward (i.e. south) of the main range (Fig. 2; Plates I-III). Each wing contained two unheated rectangular rooms (East Wing, Rooms 2 and 3; West Wing, Rooms 15a and 16). At the ends of the wings, Rooms 3 and 16 were of broadly similar size (Table 1) but Room 15a in the West Wing was substantially larger than the corresponding one (Room 2) in the East Wing. That Room 15a had some special importance seems to be indicated by its size and by the fact that at a later date it had been remodelled to include a hypocaust (see below, Phase 6, Room 15b).

Phase 2, Room 31

Sometime after the Phase 1 house had been completed, a new, unheated, rectangular room was added to the structure (Room 31). This was somewhat oddly positioned in the internal angle between the Central Range and the East Wing (Fig. 2). It was represented by shallow foundations of rammed chalk rubble, 0.30-0.60m wide and 0.05m deep. These stopped just short of the existing foundations of the main (Phase 1) building and had subse-quently been cut across by the walls of a new south corridor (see below).

Room 31 measured internally some 4.50m (e-w) by 6.00 m. (n-s) and perhaps provided additional domestic accommodation. Internal access was most probably via the cross-passage situated towards the eastern end of the Central Range (Room 5). Whatever its exact purpose, use of the room seems to have been quite short-lived because it was soon demolished to make way for a new corridor built along the south front of the house during Phase 3 (see below).

Phase 3, The Inner Corridor

Butting onto the outer corners of the main wings and cutting across the foundations of the by then demolished Room 31 (Phase 2, see above), a broad corridor had subsequently been added to the south side of the building (Fig. 2). Only the mortared chalk rubble foundations survived. These were 0.55-0.65m wide and were at least 0.10m deep. They seem to have supported a covered walkway about 2.50m wide, which extended around the entire south front of the house, providing access to all the principal rooms. From the surviving remains it cannot be certain if this corridor took the form of an open, colonnaded walkway or a fully enclosed gallery.

Phase 4, The Outer Corridor

Sometime after the construction of the corridor along the south front (Phase 3, see above), an extensive new outer corridor was added along the west, north and east sides of the house. This has been designated here Phase 4 (Fig. 2), although its exact relationship to the works associated with what have been termed Phase 5 (see below) cannot be certain. It is possible that the order of these two phases could be reversed, or they might even be contemporary.

Again, only the chalk rubble foundations of the new corridor wall survived. These were 0.50-0.75m wide and 0.15m deep and seem to have formed a walkway that was of a similar width to the existing one along the south side. Presumably, new doorways had to be cut into the two end-walls of the extant south corridor to allow through-access. When completed, however, a broad covered passage had been created all the way around the outside of the building.

Phase 5, Central Porch (Room 30) and the Enclosed Courtyard

Perhaps as a continuation of the Phase 4 works (see above), an almost square projecting central porch (Room 30) was added onto the outside of the south corridor (Phase 3). This must have been built around the same time as a walled inner courtyard was created in front of the house (Fig. 2).

The Central Porch (Room 30)

Butted onto the south corridor wall were two parallel footings of chalk rubble. These clearly relate to a centrally positioned entrance porch (Room 30) added to the front of the house. The western foundation was 0.45-0.85m wide and 0.25m deep; the eastern one was about 0.54m wide and 0.33m deep. Open on the south side, the new porch originally seems to have measured internally around 3.30m (e-w) by 3.90m (n-s). A spread of chalk rubble noted between the foundations (Fig. 2, 3089) may represent the base of the original flooring. Immediately in front of the porch lay an area of rough pebble metalling (Fig. 2, 3086) which must be the remains of a path leading up to the entrance.

The Inner Courtyard

The construction of a new boundary wall between the outer corners of the south corridor created a small private inner courtyard or garden in front of the new entrance porch, set between the wings of the main building. The area enclosed measured about 16m (e-w) by 10m (n-s). The foundation for the new boundary wall had subsequently been partially robbed away but where surviving it was about 0.60m wide and up to 0.35m deep, constructed of mortared flint nodules partially laid in herringbone fashion.

Phase 6, New Heated Rooms and Other Additions

During the later life of Building 1 more changes and additions were made to the structure. These need not all be exactly contemporary but they have been grouped together here to form Phase 6 (Fig. 2). Works assigned to this phase include the rebuilding of Room 5 in the Central Range, the creation of two, probably three, new hypocausted rooms (Rooms 10, 15b and 28) and the addition of a small room in the south-west corner of the inner courtyard of Phase 5 (the Gatekeeper’s Lodge, Room 29). These developments represent continuing architectural aggrandisement of the villa building and presumably reflect the rising social status of its occupants. Collectively, they seem to represent the peak of architectural achievement in Building 1.

Room 5

At some stage, Room 5, the cross-passage at the eastern end of the Central Range, underwent extensive re-building, with the original Phase 1 chalk and pebble footings being replaced by ones of more substantial mortared flint construction. These were 0.60-0.65m wide and 0.50m deep. The new foundations were continued on into the adjacent parts of Rooms 1 and 8, implying a major re-construction of this part of the building (Fig. 2). Since the focus of this rebuilding was an existing pass-age, it seems likely that the work was associated with the insertion of a new staircase, either as a replacement to an earlier, less solid affair, or the provision of such a structure for the first time. Either way, the work seems to imply the existence of an upper storey to the building. The corresponding cross-passage at the western end of the range (Room 13) was left untouched.

Rooms 10 and 11

A detached, hypocausted apsidal room (Room 10) was located on the far side of the outer corridor, centrally positioned to the north of Room 9 (Plate IV). The associated stoking-chamber (Room 11) was somewhat curiously placed on the south, within the corridor. The exact purpose of Room 10 occasioned much discussion and speculation during the course of the excavation and continues to be enigmatic.

From the surviving remains, it was not entirely clear whether the corridor (assigned to Phase 4, see above) had been butted onto a pre-existing, detached structure or whether Room 10 was cut through the foundations of the outer corridor wall. On general grounds it seems most probable that the existing corridor had subsequently been cut into by Rooms 10 and 11, indicating that these additions came sometime after Phase 4. If this is correct, it would appear that the intention had been to create a small, private, heated room set away from the main parts of the building.

Room 10 was D-shaped in plan with its straight wall on the south side abutted by basemented Room 11. Internally, the apse was 3.75m wide (e-w) and 1.80m deep (n-s). Its walls were around 0.65m thick and survived to a height of between 0.30 and 0.66m above sub-floor level (see below). They were constructed of flint cobbles with some red tile set in coarse cream mortar, and were rendered internally with opus signinum (Plate IV). The remains of a 0.50m wide, tile-edged flue arch survived in the middle of the south wall, confirming that the room had been hypocausted (Fig. 3, a). The top of the arch had subsequently been destroyed and no traces of the original suspended floor or supporting pilae stacks remained within Room 10. The hypocaust sub-floor was composed of opus signinum but a careful examination of this failed to reveal any clear mortar pads or scars relating to the missing pilae. There seems no doubt that the hypocaust had been thoroughly robbed for its materials not long after it had gone out of use. Amongst the soil and rubble back-filling the room were large quantities of broken tile, painted plaster and about 90 pot-sherds datable to the period c.180-270.

Room 11, on the south side of Room 10, was a basement that served as the stoking-chamber for the hypocaust under Room 10. It consisted of a narrow, deep room set within the line of the Phase 4 corridor. Internally, this room was about 1.00m wide (n-s) and 5.70m long (e-w). It was entered via a ramp leading down from outside the building to the north, along the eastern side of Room 10. The floor of Room 11 consisted simply of the trodden surface to the natural and lay about 1.30m below surviving wall-top level and about 0.50m below the base of the furnace arch (Fig. 3, b). It is conceivable that this basement chamber might have been originally floored-over at head-height in order to allow un-interrupted access along the corridor but this would have created a very dark, cramped and poorly ventilated furnace room, that would have been difficult to operate. More probably, the corridor here was now reduced to about half its original width and skirted around the new Room 11.

An interesting constructional detail within Room 11 was the use of unfired mud bricks for the retaining wall on the south side (Fig. 3, b). The bricks had been set in a coarse cream mortar and were rendered over with a thick layer of similar material. Mud brick is seldom encountered in Romano-British buildings, presumably due to the country’s generally damp climate (Davey 1961, 24) and the choice of this material for lining a subterranean chamber at Minster seems rather odd. Presumably, the heat from the hypocaust furnace meant that the room was always kept reasonably damp-free. The individual bricks had been laid end-on and generally measured about 290 x 400 x 80mm high (11½ x 15¾ x 3¼in.). These dimensions may be compared with the sizes given by the Roman architect Vitruvius, equivalent to 11½ x 17½ x 3in. (quoted in Davey 1961, 25).

Once the hypocaust system had gone out of use, Room 11 became filled with layers of rubble and soil. These deposits contained quantities of broken tile, painted plaster, mosaic fragments and loose tesserae, together with some domestic rubbish in the form of pottery, animal bone and marine shell. The mosaic fragments and tesserae recovered could be derived from either the corridor above, Room 9 or Room 10, and perhaps all three – a detailed study of the fragments has yet to be undertaken. The pottery recovered from the chamber totals 236 sherds. These include fresh, conjoining pieces from a number of different vessels, implying deliberate dumping here. A date of c.200-230 seems likely for the deposition of this material (Lyne 2002).

The provision of the stoking-chamber for the hypocaust of Room 10 inside the north corridor, rather than as an outside appendage, is unusual but it can be matched by similar arrangements in the West Wing (see Rooms 15b, 19 and 28), and also in Building 4 at the south-east corner of the site (Fig. 1; Parfitt 2006). Presumably, it was the protection from the weather afforded by placing these stoking-chambers inside the corridors that determined their positioning. It tends to suggest that at least these sections of corridor were open colonnades, rather than being enclosed, thereby providing good air circulation for the working of the systems and allowing smoke and fumes from the furnaces to easily disperse.

The full significance of Room 10 is still under consideration and more research is required. Its positioning astride the main axis of symmetry through the building, its apsidal form, the provision of under-floor heating and perhaps a mosaic, all clearly imply that this was an important, albeit small, room. At present, it can perhaps be provisionally suggested as being a small private chamber used for study or relaxation – possibly it was the villa owner’s personal office, positioned a little away from the bustle of the main house but still close to its principal rooms.

Room 15b

At some stage, the large room in the West Wing, Room 15a, was extensively remodelled, apparently to allow the insertion of a hypocaust (designated here Room 15b). Later, this room had been thoroughly robbed, leaving details of the new arrangements uncertain – the hypocaust fittings, any supporting basement sub-floor and much of the inserted walling had all been removed.

It was clear that initially most of the floor area within the original Phase 1 room had been dug out to a depth of at least 0.35m in order to create a basement. Retaining walls had then been inserted as a lining to this sunken chamber. On the south and east sides the new retaining walls were built directly inside the Phase 1 foundations (Plate III). On the north side, however, a gap of about one metre separated the old from the new work (Fig. 2). As a consequence, the size of the room was slightly reduced. The remaining gap to the north was probably just wide enough to accommodate a narrow cross-passage, linking the new stoke-chamber in the west corridor (see below) with Room 13, another passage.

Presumably to compensate for the reduction in the size of the room on the north, the west wall foundation of the original Phase 1 room was removed in order to allow the new hypocausted room to be extended out into the west corridor. The extension must now have blocked the central sector of this corridor and prevented through-access (Fig. 2). Within the corridor extension, the surviving remains showed that the inserted north wall had been set at an angle to the main wall line (Fig. 4, 1039). Unfortunately, no corresponding wall on the south side survived but the remains may just be sufficient to suggest that the western end of the new room incorporated a semi-hexagonal apse around 2m deep (e-w). As remodelled, the room probably had overall internal dimensions of about 4.50m (n-s) x 8m (e-w).

Where surviving, the newly inserted walls were all found to be of similar construction and mainly consisted of large flint cobbles and some tile fragments set in a cream-white gritty mortar containing crushed marine shell. They were generally between 0.50 and 0.65m wide; a few sections survived to a maximum height of about 0.30m but in many places they had been largely or totally removed.

The stoking-chamber for the new hypocaust system of Room 15b had been placed in the west corridor and was represented by a 0.30m deep pit with a heavily burnt base (Fig. 4, F. 1093), set on the outside of the angled north wall. On the east side of this pit a section of walling (1092) must have formed one side of the furnace flue. It also showed evidence of intense heating. The wall was constructed from flint cobbles and tiles bonded in a similar mortar to the other inserted walls; no traces of its western counterpart had survived.

On the north side, the stoke-pit and surviving flue wall had subsequently been truncated by the new south wall of Room 17 (Fig. 4, 1088) assigned to Phase 7 (see below). This must have prevented access to the stoking-chamber, and probably implies that the hypocaust in Room 15b had gone out of use by this time.

After it had been robbed of its re-usable materials, the basemented area of Room 15b was back-filled with soil, building rubble and some domestic rubbish (Fig. 5, 1021). The significant quantities of broken tile and painted plaster recovered from this deposit are presumably derived from the demolished room above. The pottery, animal bone and marine shell must represent dumped domestic rubbish. A total of 225 sherds of pottery, including five samian fragments, were recovered. The latest pieces within this assemblage date to the early third century.

Rooms 19 and 28

Room 28 had been added to the south end of the West Wing, separated from the main range of rooms by being positioned on the outside of the enclosing corridor (Fig. 2). As with Room 15b, it had subsequently been extensively robbed and no wall foundations remained except for a 0.60m wide fragment of mortared flintwork on the west side. The rest of the outline of the room was represented entirely by robber-trenches about 0.30m deep.

From the surviving foundation and robber-trenches it seemed reasonably clear that the room had originally consisted of two elements – a main rectangular chamber, measuring internally about 5.50m (e-w) by 2.60 m. (n-s), with a smaller, projecting apse on the south side, overlooking the outer courtyard. The apse would have been of a similar size to Room 10, measuring about 3.20m wide (e-w) and 1.80m deep (n-s). Within the main room, traces of a layer of flint-cobbles set in yellow-buff mortar appeared to represent the eroded base of the original sub-floor.

In Roman buildings, rooms with apsidal ends have frequently been equated with dining-rooms (triclinia) but this is not invariably the case. The present room would appear to have been too small to have functioned as such (Witts 2000, 295-301) and its use as some sort of private reception room, situated at the end of the building seems more likely. Either way, it indicates a certain level of sophistication in the later architectural arrangement of Building 1, although its addition to the West Wing now disrupted the symmetry of the building’s original front elevation, because no corresponding addition was made to the East Wing.

Immediately to the north of Room 28, lay Room 19 (Plate III). This consisted of a narrow, sunken room positioned inside the corridor enclosing the main West Wing (see above, Phases 3 and 4). Mortared flint walls between 0.20-0.30m wide delimited a chamber with internal dimensions of 4.95m (e-w) by between 0.80 and 0.90 m. (n-s). As surviving, the room was about 0.70m deep, the floor consisting simply of the trodden surface to the natural. It was entered from outside the building by a set of steps leading down from the south, along the western side of Room 28.

The size, shape and form of Room 19 is very similar to Room 11 on the north side of the building (see above). That room had formed the stoking-chamber for the adjacent hypocausted Room 10. A similar function for Room 19 thus seems to be suggested. Certainly, the structure appears too narrow to have served as a useful cellar. As with Room 11, it would appear that the enclosing corridor here must have now been reduced in width to skirt around the outside of the newly inserted chamber. No remains of any furnace flue survived in Room 19 but the upper courses of the remaining basement walls had been robbed in places and the evidence may have been destroyed. Nor was there any good evidence that the adjacent Room 28 had ever been basemented to take a hypocaust. However, this might be explained if the hypocaust had actually stood above Roman ground level, reflecting the architect’s desire to maintain a constant level throughout the length of the West Wing and leading to the raising of the floors at the southern (downhill) end of the range.

At some stage during the use of Room 19 a layer of cream-white mortar was laid at the foot of the entry steps, perhaps to form a new threshold. This surface lay about 0.35m above the original floor level and seems to imply that the chamber had been re-floored. A series of ashy soil deposits relating to the early use of the room were preserved under this mortar layer. A primary deposit of trodden soil and ash lying over the original floor produced two undated pot-sherds and a fragment of ceramic water-pipe (see Parfitt 2007, pp. 283-4, for comments on water-pipes). Subsequent deposits produced about 180 pot-sherds datable to between c.120 and 250. The latest of these imply that the chamber was in use during the first half of the third century.

Room 29, the Gatekeeper’s Lodge

A section of the Phase 5 courtyard wall (see above) had subsequently been demolished to allow the insertion of a small new room (Fig. 2, Room 29). This was placed at the south-western corner of the courtyard, adjoining the West Wing. The new room had subsequently been robbed but from the surviving remains it would seem to have originally been almost square in shape and measured internally about 1.60m across. The surviving foundations were between 0.65 and 0.70m wide and consisted of mortared flint nodules up to 0.12m deep.

A small room similarly positioned at the Gadebridge Park villa in Hertfordshire was interpreted as a porter’s lodge (Neal 1974, 55; room 36) and a similar function at Minster seems quite possible; it has been tentatively identified here as a gatekeeper’s lodge.

Phase 7, Room 8 and sub-division of the Outer Corridor

Perhaps the latest changes to be made to Building 1 are represented by a new wall reducing the size of Room 8 in the Central Range, together with a number of cross-walls constructed within the corridors, apparently to create additional rooms (Table 2). The east and west outer corridors were both sub-divided in this way, creating two long rooms (east side, Rooms 6S and 6N; west side, Rooms 17 and 18). A new wall built across the south corridor at the end of the East Wing formed another small room there (Room 4).

Room 8

Sometime after the Phase 6 re-build of Room 5 (cross-passage, see above) a new wall had been inserted on the southern side of Room 8, reducing the length of the original room to 5.40 m (n-s). The surviving flint foundation for this new wall was about 0.70m in width. The strip to the south of the new wall could now have formed a 1.40m wide passage connecting with Room 5. There was some evidence to show that the foundation to the north wall of Room 8 had also subsequently been widened and strengthened, suggesting that the late changes made to this room might have been quite extensive (Fig. 2). A final demolition deposit had been preserved within Room 8. This yielded about 200 pot-sherds, including eleven pieces of samian ware, datable to the period c.180-270.

Sub-division of the Outer Corridor (Rooms 4, 6S, 6N, 7, 17 and 18)

There was clear evidence that the outer corridor of the villa house had subsequently been sub-divided to provide some extra rooms (Table 2). This feature has been noted at a number of other Romano-British villas. It also occurred in Building 4 at Minster and has been recorded elsewhere, at such sites as Keston (South Masonry Building; Philp et al. 1991) and Gadebridge Park, Herts. (building A; Neal 1974).

Little can be deduced about the function of most of the new corridor rooms created within Building 1, beyond the inference that further accommodation, work rooms or storage space was required, in preference to through-walkways. If the corridors had originally been open colonnaded walkways they must now have been fully enclosed, at least in certain sections. In the north-west corner of the building, the newly created Room 17 contained evidence to suggest that it had been used as a blacksmith’s workshop (see below). Because it was basemented, this room had preserved the remains of internal structures and subsequent infill deposits. It is described in more detail below.

Room 17

Room 17 had been formed by walling off the north end of the west corridor to create a long, narrow room with internal dimensions of 2.50 (e-w) x 9.30m (n-s) (Fig. 4; Table 2). The new walls were mainly composed of flint nodules set in cream gritty mortar. Access into the room seems to have been from Room 7 (North Corridor), via a central doorway in the new east wall. The south wall for the room had cut through the earlier stoking-chamber (Fig. 4, F. 1093) associated with the hypocaust inserted into Room 15b (Phase 6, see above).

The floor of Room 17 was set somewhat below contemporary ground-level (at least 0.35m deep) to give a shallow basement (Fig. 5, a, b). Indeed, this basement may have been created by extending the original hypocaust stoking-chamber associated with the adjacent Room 15b, which presumably had gone out of use by this time. On the western side, the new basement cut through an earlier pit (Fig. 4, F. 1116) that had probably originally been dug into the floor of the Phase 4 corridor. Its brown clay loam filling produced small quantities of animal bone and marine shell, together with two pieces of window glass.

The primary floor/levelling deposit in the base of the new room consisted of a 0.05m thick layer of orange-brown sandy loam resting directly on the truncated surface of the natural. This deposit contained a few oyster shells and 27 residual pot-sherds datable to the period c.70-140. It was overlain by a patchy occupation layer of black ash and carbon (Fig. 5, b, 1109). This contained no datable material and was soon sealed by a thin deposit of flint and mortar rubble (Fig. 5, b, 1101), which contained a significant quantity of painted plaster, quite probably derived from the walls of the original corridor in this area. Thirty-six sherds of pottery were recovered from this layer and the latest of these belong to the first half of the third century.

The top of the rubble layer was well-trodden clearly indicating that it had served as a rough walking surface. It supported two tile- and stone-built structures (Fig. 4, 1041 and 1099) which seem to have had an industrial function. In the south-east corner of the room were the remains of what appeared to be the flue of a large oven or furnace (1041). As surviving, this consisted of a section of walling measuring 1.82 (e-w) x 0.78m (n-s) abutting the outer walls of the room. It was constructed mainly of large flint cobbles, with some tile fragments, set in light brown clay and survived to a maximum height of 0.42m (Fig. 5, b). The north face of this wall showed evidence of intense heating. There can be little doubt that it represents the southern side of a larger structure which had not survived, perhaps being deliberately dismantled.

Built against the east wall of the room further to the north was a sub-rectangular base about 0.90m across (1099). This was constructed of flint cobbles and broken tile fragments set in light brown clay. It survived to a height of about 0.40m. A few sherds of pottery had been incorporated into this structure and the date of these suggests that it could not have been built before c.180. The structure again showed evidence of burning and it appears to represent some sort of raised hearth. An industrial use again seems to be implied.

A thick deposit of black ashy loam and carbon accumulated over the floor around these two structures (Fig. 5, a, b, 1007). This material must largely represent rake-out from regular firings of the furnace structure and working on the raised hearth. The layer was up to 0.35m thick and produced a considerable number of nails and other iron fragments, which is perhaps consistent with black-smithing activity in this room. Additional finds from this extensive deposit included a coin, a bone needle, a copper-alloy pin, pottery, tile fragments, over 30 pieces of window glass, animal bone and marine shell. The coin is an unworn issue of Antoninus Pius dated to ad 140-144 (Holman and Parfitt 2005, table 2, coin ref. 1997/1). Some 450 sherds of pottery were discovered and these mostly date to the period c.120-250. The glass recovered is probably derived from windows set in the outer walls of the room.

In order to enlarge the original structure, a smaller tile platform (Figs 4 and 5, b, 1100) was subsequently added to the western side of the raised hearth (see above). This extended across the top of the ashy layer (1007) implying that the surface of this deposit had served as the floor for later activity in the room. It seems probable that the southern furnace structure had gone out of use and been partially dismantled by this stage.

On the evidence of the preserved internal structures, the thick ash deposit and the iron fragments that it contained, it seems quite possible that Room 17 had served as a blacksmith’s workshop. The raised hearth 1099/1100 may well represent the base of the actual forge. Thus, during the later life of the villa, it may have been in this small room that the estate blacksmith made and repaired the variety of iron tools and other items that would be regularly needed on a busy working farm.

The room eventually went out of use and the basemented area was in-filled with soil, tile, flint and white mortar rubble (Fig. 5, a, b, 1022). This final levelling deposit also included some painted plaster and 25 pieces of window glass, together with quantities of domestic rubbish in the form of pottery, vessel glass, animal bone and marine shell. Nearly 350 sherds of pottery were recovered. This material ranges in date from the second to fourth centuries but much of it is clearly residual. It would seem that the final in-filling of this room had occurred during the later fourth century.

the finds

The excavations in the area of Building 1 produced a considerable quantity of finds but as with previous site reports, details of most of these must be held over. Some notes on the coins and pottery, however, have been given below since these provide the main dating evidence for the structure.

Coinage from the area of Building 1 (not illustrated)

Only three coins were discovered in the area of Building 1. These have been described previously (Holman and Parfitt 2005, table 2). Of most significance was an as of Antoninus Pius, dated to ad 140-144, found within the filling of Room 17 at the north-west corner of the building (Context 1007; coin ref. 1997/1). This is in very good con-dition, suggesting little circulation. Based on its own merits, a date of deposition before ad 150 could be suggested for this coin, but the same dark, ash-rich layer also produced over 450 pot-sherds which fall within the date-range c.120-250. The coin and much of the pottery is therefore likely to be residual in its excavated context and the layer would seem to incorporate a proportion of re-deposited, earlier midden material, perhaps included as levelling.

The other two coins were recovered from a general soil layer outside the building to the north-east (Context 822). These consisted of an issue of Trajan, dated ad 114-117 (coin ref. 2002/28) and one of Theodora dated 337-340 (coin ref. 2002/5).

Pottery from the area of Building 1 (not illustrated)

Around 5,000 sherds of pottery came from the immediate area of Building 1. The largest assemblages were contained in the fillings of the basemented rooms (Rooms 10, 11, 15b, 17 and 19). Together, these produced almost one-third of the total pottery recovered. A demolition deposit surviving within Room 8 yielded another 203 sherds and over 200 more came from the ploughsoil and general deposits sealing the building. From Malcolm Lyne’s analysis (2002), it would seem that virtually all the pottery recovered from this area is of later first, second or early third-century date, with only a small number of sherds dating to the late Roman period. There is also a little pre-Conquest material. Full details of the pottery will be presented in a future report.

Other Finds

Other finds from the area of Building 1 included large quantities of broken brick and tile, significant groups of painted plaster, tesserae and mosaic fragments, vessel and window glass, a range of small finds and kitchen waste, comprising animal bone and marine shell, mostly oysters. Details of these various groups will be set out in separate future studies.

dating of the main house

As with the other major buildings excavated at Abbey Farm, a general lack of material from well-stratified deposits directly associated with the use of Building 1 hampers any attempt to closely determine its period of occupation. The largest assemblages of datable finds from within the building came from late infill layers preserved within the various sunken and basemented rooms.

Some limited evidence as to the initial construction date of the structure is provided by material contained within the subsoil layer cut through by the foundations of Building 1. This produced small quantities of pottery, mostly datable to the prehistoric period or the first century ad, with a few later sherds which appear to represent intrusive material derived from subsequent activity on the site. Below this subsoil, the filling of a ditch under Room 1 produced sherds dated to c. 50 bc - ad 50. Taken together, the available evidence would seem to suggest that Building 1 cannot have been constructed before the last quarter of the first century ad at the earliest, even if there had been occupation on the site prior to this.

Two of the three coins and most of the pottery found in the area of Building 1 date to the second or early third century and there seems little doubt that this was the main period of use of the structure. It would appear that Building 1 had been abandoned around the middle of the third century. After it ceased to be used, the structure was thoroughly robbed for its stone, tile and other useful building materials. This was also found to be the case with the adjacent bath-house (Building 3) and with the two southern structures, Buildings 4 and 6B. The evidence for robbing in Building 1 was most apparent in the hypocausted rooms (Rooms 10 and 15b), where virtually everything had been systematically stripped-out before the unwanted hollows were backfilled with soil, rubble and old domestic rubbish.

Clear wall robber-trenches, such as were recorded in Buildings 4 and 6B, however, were not observed across much of Building 1. Many of the surviving wall footings, especially those belonging to Phase 1, were of rammed chalk rubble and/or small pebbles, and were presumably left in place because they were of little value for recycling.

As has been previously noted with the other excavated buildings, the dating of the pottery from the robber-trenches and backfilled basements of Building 1 suggests that the robbing of the structure took place during Roman rather than medieval times, most probably during the fourth century.

discussion

A full discussion of Building 1 and its significance is best left until details of the Minster villa complex have been fully published but a few preliminary observations and comments may be made here. From the available evidence, it would seem that Building 1 had been first erected towards the end of the first century ad (Fig. 2, Phase 1). Even though little more than foundations survived, it is readily apparent that this building had once been a structure of some architectural quality. It had evolved through several structural phases (Fig. 2, Phases 2-6, see above), apparently reaching its peak of architectural sophistication at the end of the second century. The presence of apsidal and hypocausted rooms, the evidence for mosaic floors, painted walls and glazed windows, all indicate that this was a structure of some considerable status, even if not to be included amongst the very grandest of Roman Britain’s country villas. The final changes, associated with Phase 7, however, seem to suggest that the once fine house was declining in status during the third century.

As originally laid out, Building 1 took the form of a substantial winged villa, with nine principal rooms, all unheated (Fig. 2). There was no evidence that this winged form had evolved from a smaller dwelling, as is often found to be the case, although the enclosing corridors were a later refinement. The addition of corridors created a villa of very well known Romano-British winged-corridor type. There are perhaps a dozen such examples known in Kent but the period 3 villa house at Gadebridge Park in Hertfordshire (building A) provides a remarkably close parallel to Minster in terms of its date, size and layout (Neal 1974). Neal has stressed the importance of the symmetry apparent in the design of that building and points out that ‘Symmetry was one of the badges of Romanization’ and that it is ‘a likely indicator of the Romanized attitude of mind of the man who lived in it’ (Neal 1974, 92). These comments must apply equally well to Building 1 at Minster. Indeed, efforts to create a broadly symmetrical layout can be seen throughout the Minster complex (Parfitt 2007, 294).

The provision of a bath-suite was essential for any villa of status. There was no integral bath-suite within Building 1 at Minster and a separate adjacent structure served this purpose. Building 3 clearly constituted the bath-house during the later phases but its positioning suggests that it may not be contemporary with the original construction of Building 1 (Fig. 2). It seems to have been fitted in subsequently, utilising available space on the western side of the walled enclosure. The south corridor of Building 1 (Phase 3) would appear to have been in existence by this time because the bath-house was carefully aligned with it. All this might suggest that an earlier detached bath building associated with the primary phases of the villa house still awaits discovery (Parfitt in Perkins 2004, 48).

There can be no doubt that Building 1 formed the principal dwelling on the Abbey Farm site but it was not the only house that existed within the complex. Further south, outside the main walled enclosure, Building 4 seems to represent a second dwelling of rather more modest proportions (Parfitt 2006), whilst Building 6 also apparently provided accommodation at certain times (Parfitt 2007). The position of these two southern structures, outside the walled enclosure (Fig. 1), seems to suggest a lower social status for the individuals who occupied them (Parfitt 2007, 295).

Along with the other principal structures at Minster, Building 1 seems to have been finally abandoned around the middle of the third century. Not long after, all these buildings had been thoroughly and systematically stripped of their re-usable materials. This seems to imply that some extensive new construction works were being undertaken on another site not too far away. Perhaps one explanation to consider is that the military authorities had acquired the site with the intension of using it as a quarry for the building materials needed in the construction of the late third century fort at nearby Richborough (see Perkins 2004, fig. 1 for location map). Both sites being situated close to the shores of the Wantsum would have readily allowed heavy loads of salvaged timber, tile and stone to be moved between the two mainly by water – a distance of just under 5km. The available dating evidence from Minster, however, might suggest that the robbing occurred slightly later, during the fourth century, after the main walls at Richborough had been completed. As an alternative idea, it may be suggested that these re-usable materials were needed for a more local building project, conceivably a new villa that was to supersede the Abbey Farm complex. If this is so, the new site has yet to be located.

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 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. Members of Thanet Archaeological Society have spent much time processing and cataloguing the finds.

The Project Director (Dave Perkins) would like to thank the Supervising Archaeologists, Emma Boast, Tim Allen, Keith Parfitt and Ges Moody, for their support and enthusiasm during all phases of excavation and post-excavation work. Susan Deacon, Emma Boast, Oliver Gardner and Ges Moody prepared the site drawings for publication. Finally, the hard work of all the excavators must be recorded – little would have been achieved without their efforts.

bibliography

Davey, N., 1961, A History of Building Materials (London).

Holman, 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.

Lyne, M., 2002, ‘The Pottery from the Minster Roman Villa’ (AFM 1996-2002) (archive report commissioned by KAS).

Neal, D.S., 1974, The Excavation of the Roman Villa in Gadebridge Park, Hemel Hempstead, 1963-⁸, Soc. of Ants Res. Rep. No. XXXI.

Parfitt, K., 2006, ‘The Roman Villa at Minster-in-Thanet. Part 3: The Corridor House, Building 4’, Archaeologia Cantiana, cxxvi, 115-133.

Parfitt, K., 2007, ‘The Roman Villa at Minster-in-Thanet. Part 4: The South-West Buildings, 6A and 6B’, Archaeologia Cantiana, cxxvii, 261-296.

Perkins, D.R.J, 1996, ‘The Abbey Farm Villa Training Excavation’, Archaeologia Cantiana, cxvi, 325-9.

Perkins, D.R.J., 1999, ‘Roman Villa Site (Abbey Farm): Minster-in-Thanet’, Archaeologia Cantiana, cxix, 393.

Perkins, D.R.J., 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.

Thompson, I., 1982, Grog-tempered ‘Belgic’ Pottery of South-eastern England, BAR Brit. Series 108 (Oxford).

Witts, P., 2000, ‘Mosaics and room function: the evidence from some fourth century Romano-British villas’, Britannia, XXXI, 291-324.

PLATE I

View across the excavated East Wing, 1996, looking south

Fig. 1 General plan of the villa complex showing the position of the main house (Building 1) in relation to other buildings and structures.

PLATE II

View across the excavated Central Range and West Wing, 1997, looking west

Fig. 2 General plan of Building 1 showing room numbers and structural phases (Phases 1-7).

PLATE III

View across the West Wing, 1997, looking north

PLATE IV

Rooms 10 and 11, looking north-west

Table 1. Details of original (Phase 1) rooms within Building 1

Room	Internal size (m) e-w x n-s	Notes
1	5.60 x 7.50	Central Range, east end
2	5.60 x 4.20	East Wing
3	5.50 x 5.50	East Wing
5	1.50 x 7.50	Central Range, east cross passage
8	5.10 x 7.50	Central Range
9	5.75 x 7.50	Central Range, central room
12	6.00 x 7.50	Central Range
13	1.10 x 7.50	Central Range, west cross passage
14	5.00 x 5.10	Central Range, west end
15a	5.30 x 6.75	West Wing, hypocaust added (Rm 15b)
16	5.10 x 5.10	West Wing

Fig. 3 a) External elevation of south wall of Room 10; b) Section through Rooms 10 and 11.

Fig. 4 Detailed plan of Room 17 and part of Room 15.

Fig. 5 a and b) Sections across Room 17; c) Section across Room 15b.

Table 2. Details of added rooms within Building 1

Room	Internal size (m) e-w x n-s	Phase	Notes
4	5.50 x 2.50	7	Added to East Wing
6S	2.50 x 14.00	7	Outer Corridor, east side (south)
6N	5.00 x 10.85	7	Outer Corridor, east side (north)
7	33.10 x 2.50	7	Outer Corridor, north side
8 (late)	5.10 x 5.40	7	Reduced to create south passage
10	3.75 x 1.80	6	Apsidal room, north side
11	5.70 x 1.00	6	Basement stoke chamber for Room 10
15b	8.00 x 4.50	6	Hypocausted room within Room 15a
17	2.50 x 9.30	7	Outer Corridor, west side (north)
18	2.50 x 16.50	7	Outer Corridor, west side (south)
19	4.95 x 0.90	6	?Basement stoke chamber for Room 28
28	5.50 x 4.40	6	West Wing, room with apse
29	1.60 x 1.60	6	‘Gatekeeper’s Lodge’
30	3.30 x 3.90	5	Entrance porch
31	4.50 x 6.00	2	Added room, inner courtyard