The Roman Villa at Minster-in-Thanet

315 THE ROMAN VILLA COMPLEX AT ABBEY FARM, MINSTER IN THANET. PART 7: BUILDING 7, A LATE ROMAN KILN AND POST-BUILT STRUCTURES gerald moody In the 2003 season of excavation of the Roman Villa at Abbey Farm, Minster, five east-west orientated trenches were excavated across the open area south of the villa house (Building 1), within the area enclosed by the boundary wall. In one of these trenches (Fig. 1; 2003 Trench 2) the upper fills of a large cut feature were exposed, on the gentle, west facing slope, 30m south of Building 1. The dark brown fills of the feature were clearly distinguished from the surrounding natural deposit in the base of the trench. The feature was investigated within the limits of the trench and some of the sequence of upper deposits of remaining topsoil and subsoil [AFM 03, deposits 7001, 7002 respectively] that had settled in the feature were excavated along with part of the upper fill, to define the feature further [7009 and 7017]. Late Roman pottery and a late fourth- century coin (Valens 367- 378; Holman and Parfitt 2005, 225) were recovered from the subsoil close to the feature. Within the upper fill [7009], pottery dating to the late fourth century and five further late fourth-century coins (three House of Constantine between 335 and 365, one Valentinian I 364-375, one Gratian 367-375; ibid.) were recovered. In the earliest deposit sampled [7017] another late coin (Gratian 367-375) was found along with fourth-century pottery (ibid.; Lyne 2001). When these upper deposits were removed, chalk blocks and patches of burnt clay and ash were visible in the upper surface protruding from the feature below. Rare evidence of late Roman phase of occupation on the Abbey Farm site had been encountered. The resources of the 2003 excavation were not sufficient to investigate the feature properly and therefore further investigations were planned for the next season of excavation in 2004. Excavation of Building 7 in 2004 In the summer of 2004, the final excavation season at Minster (Pout GERALD MOODY 316 Fig. 1 General plan of the villa complex showing the location of trenches excavated in 2003 and 2004 to expose Building 7 and associated features. ROMAN VILLA, MINSTER-IN-THANET. PT 7: BUILDING 7, KILN, POST-BUILT STS. 317 2004), the feature in Trench 2 was targeted in a more extensive excavation area (Fig. 1, 2004 Trench 9). The full extent of the cut feature was exposed, along with gullies and post holes not encountered in 2003. The cut feature was sampled in a grid of alternate segments; the sections of these segments have been reconstructed as a continuous cross section (Fig. 2, Sections 1 and 2) by reversing some of the drawn sections. Excavation of the segments demonstrated that the feature was the subterranean element of a kiln or oven. Later demolition deposits were removed to expose the material created by the last use of the kiln; in some segments, the original structural cut of the kiln was exposed. Eventually the deposits in the unexcavated areas of the grid were removed so that the kiln could be planned and photographed without any later demolition or backfill material present (Plate I). The alignments of post holes and gullies associated with the kiln suggest it was probably protected by a post-built enclosure without a roof, or a light leanto or barn like building (designated Building 7; Plate II). The Structure of the Kiln The primary cut for the main chamber of the kiln structure was a fairly West facing view of the kiln from the stoke chamber Plate I GERALD MOODY 318 Fig. 2 Sections across the main body of the kiln and stoke chamber. ROMAN VILLA, MINSTER-IN-THANET. PT 7: BUILDING 7, KILN, POST-BUILT STS. 319 regular rectangle in plan, measuring 4.66m long on its north-south axis and 2.92m wide along the east-west axis (Fig. 3, F.7692). The undisturbed edges of the chamber were near vertical, penetrating the natural geology to an average depth of 0.96m (Fig. 2, Section 1). The main chamber was only excavated to natural deposits in two places (Fig. 2, Section 1, F.7692); the evidence in these locations indicates that base of the chamber was relatively flat when it was cut. A stoke chamber for the kiln was formed by a second cut that intersected the eastern side of the main chamber, forming an opening 1.9m wide. The width of the chamber remained broadly consistent along its length (Fig. 2, Sections 2 and 3; Fig. 3). The floor of the stoke chamber sloped upwards gradually for 1.76m, away from the floor level of the main chamber, before rising steeply at the eastern end. A more gently sloping hollow at the eastern end of the stoke chamber provided access from the Roman ground level, perhaps half a metre or more above the level of the upper edge of the surviving cut. The sides of the main chamber of the kiln were lined with courses of roughly squared chalk blocks, set in a bonding of sandy clay (Fig. 3). Generally the walls only survived to a height of three or four courses; but in the best preserved section on the southern edge of the chamber, six courses of blocks remained standing to a height of 0.5m. It is likely that any superstructure for the Kiln was constructed with similar materials, continuing the existing structure upwards. Overview of the kiln and associated post holes and gullies viewed from the north-east Plate II GERALD MOODY 320 Fig. 3 Detailed plan of the structural elements of the kiln. ROMAN VILLA, MINSTER-IN-THANET. PT 7: BUILDING 7, KILN, POST-BUILT STS. 321 A central dividing wall, also built of chalk blocks with clay bonding, separated the chamber into two parts. At the eastern end of the central wall, a short length of the structure was formed of large rounded cobbles set in clay bonding and incorporating two fragments of quern or millstone (Fig. 2, Section 3; Fig. 3, 7715). This may have been a strengthened column base to support a post or arches spanning the entrances to the chambers at the western end of the stoke chamber. Two sub-divisions, also built of chalk blocks bonded with clay, extended at a slight angle from the point of intersection between the stoke hole and the main chambers. The divisions split each half of the chamber to form two long passages, each with a right-angled return at the western end, doubling back and widening as they approached the eastern end (Fig. 4). Although no structural evidence existed, a vent or chimney shaft in the superstructure on either side of the stoke chamber must have existed, to allowed the hot gasses to escape from each passage when it was in use. The inner surfaces and elevations within the two passages of the main chamber had been rendered throughout with a thick lining of smoothed sandy clay (Fig. 3; Fig. 2, Section 2, 7678 and 7710). In the wider section of the northern chamber, in the north-east corner of the kiln, a deeper hollow had been scooped out of the flat base (Fig. 2, Section 1, F.7723; Fig. 3). A line of stake holes was exposed in the base of the scoop (Fig. 4, composite group 7717). On the southern side of the kiln, a post hole [F.7712] filled with grey-brown silty sand [7711] and capped with two fragments of flat tile was cut into the base (Fig. 2, Section 1). These structural features suggest that timber supports may have been used as a frame when the building was constructed or perhaps for repairs when the kiln was in use. If the superstructure of the kiln was formed of fired clay, as indicated by the large quantities of structural daub present in the fill, the stakes may have been used as a former to support the clay capping the passages. The timber formers may have been burnt away when the kiln was in use. A rounded cut formed an indent in the upper edge of the foundation cut for the stoke hole [F.7724]. This feature was mirrored on the northern side by a similar rounded indent [F.7725]. At the entrance to the southern passage within the stoke chamber, the southern edge was formed by the scorched vertical surface of natural clay. At the eastern end of the open section of the stoke chamber, the edge was retained by a few irregular chalk blocks, only two course thick at most (Fig. 3, 7673), set in the lower reaches of the indent cut F.7724. These blocks were supported on a clay mass that lay behind the burnt inner face, continuous with the scorched natural clay further along the neck. On the northern side of the stoke chamber, the north face within the northern passage was formed by the vertical face of the scorched natural clay. This scorched face continued toward the east where it formed GERALD MOODY 322 Fig. 4 Plan of gullies and post holes of Building 7 enclosing the kiln shown with lining reconstructed. ROMAN VILLA, MINSTER-IN-THANET. PT 7: BUILDING 7, KILN, POST-BUILT STS. 323 the outer elevation of a mass of irregularly coursed chalk rubble and occasional cobbles (Fig. 3, 7719), laid within the indented cut F.7725. At the western end, three courses of chalk blocks (Fig. 3, 7716), continuing the line of the internal dividing wall in the northern section of the main chamber, were pressed against the vertical face of the cut, supported by the body of clay behind the scorched crust below. Two large flint cobbles spanned the gap between the coursed chalk blocks [7716] and irregular mass of chalk rubble [7719], possibly forming the base of a timber post setting incorporated within the masonry retaining structure. As they overlay the scorched clay crust, it is also possible the cobbles were inserted as a temporary repair to the retaining wall during the life of the kiln In use it appears that the masonry lining of the stoke chamber would have formed vertical walls on the interior face, creating a regular, rectangular entrance to the chamber from the ground level above. It is possible the indented cuts on either side of the chamber, once packed with rubble, formed pads to support the columns for a lintel or arch spanning the entrance to the stoke chamber. The Kiln in Use The narrowest openings of the two passages were located on either side of the central division in the stoke chamber (Fig. 2, Section 2). The fire was probably set close to the entrances with hot air passing through each passage warming the roof, heating the underside of a solid floor surface exposed in the space above. As the kiln operated, the clay rendering within the chambers and at the seat of the fire in the stoke hole had been scorched and baked hard. The scorched faces and soot blackened remnants of the lining were present in both chambers, usually only surviving in place in the lower reaches of the feature, at the base of the remaining chalk block wall core. The lining of the southern chamber survived to a greater extent, nearly reaching the top of what remained of the supporting chalk block wall. The scorched clay face was covered throughout the kiln with a thin deposit of fine soot [7628], the remnant of the material deposited up to the last firing of the kiln. A sample taken from the soot contained a few charred grains of bread wheat. Excavation of the feature usually stopped at the point where the scorched crust and final deposit of soot were revealed; however, in two sections of the north and south chambers the baked crust was removed revealing a thick deposit of laminated soot (Fig. 2, Section 1, 7722 and 7691) accumulated from earlier firings overlying the natural clay. This deposit is assumed to underlie all the areas where excavation did not exceed the depth of the ashy crust. GERALD MOODY 324 Post abandonment deposits A mound of debris (Fig. 2, Section 1, 7627 and Fig. 3), composed predominantly of chalk blocks, with flint cobbles and occasional fragments of roof tile, had built up against the western wall within the northern chamber, extending into the passage on either side. The material in the debris is similar in composition to the structure of the kiln and may have derived from the collapse of a significant section of the superstructure into the passage some time after the kiln had been abandoned, before the chambers had collapsed more extensively. Within the channels of the main body of the kiln, and in the stoke chamber, large quantities of scorched clay that had fallen from the lining of the walls as they decayed was found tumbled into the chambers, falling in approximately horizontal bands (Fig. 2, Section 1-3, North chamber, 7690, 7626, 7625, 7622; South chamber 7684, 7623, 7616; Stoke Chamber, 7672, 7629, 7670, 7671, 7621). Tips of burnt debris alternated with deposits of clean sandy clay derived from the bonding of the chalk block walls. Large quantities of daub from the superstructure were also present within the deposits, occasionally very large fragments were recovered showing impressions of the rods and sails of a timber framework. Above the deposits largely composed of the debris from the structure of the kiln there was a shallow, horizontal interface, sloping slightly toward the middle of the feature. Above this interface a deposit of fine loam [7604] filled the upper limits of the eroded construction cut for the body of the kiln. This was a continuation of the deposits sampled in 2003 and described earlier [7609 and 7617]. The remnants of the deposit were excavated in 2004 recovering more sherds of late Roman pottery contemporary with the finds made in 2003. Gully Features Several gullies were located in the central area of the north side of the excavation. One 3m-long gully, slightly curved in plan with rounded terminals, was located close the northern limit of excavation [F.7603]. Segments excavated through the silty fill [7602] showed the cut to have a smoothly curved profile (Fig. 5, Sections 1 and 2). Although two sherds of late Iron Age pottery were recovered from the fill, daub and nail fragments suggest that the feature is contemporary with the group of linear gullies further to the south (Fig. 4; F.7681, F.7682, F.7683) which contained similar deposits of daub. The Iron Age pottery appears to be residual in a much later feature. A second gully [F.7681] was located 2.5m further south, parallel to F.7603. The eastern terminal of this ditch was the deepest part, measuring 0.4m (Fig. 5, Section 3). In this location, the profile of the feature was ROMAN VILLA, MINSTER-IN-THANET. PT 7: BUILDING 7, KILN, POST-BUILT STS. 325 Fig. 5 Sections of gullies and post hole groups. GERALD MOODY 326 steep sided, with a curved base, breaking slightly to a shallow shoulder on the southern edge, probably as a result of disturbance at some later period. In a segment further west, the feature was shallower at 0.25m (Fig. 5, Section 4), with a similar steep sided curving profile. The fill of this gully [7619] was predominantly composed of fragments of structural daub in a matrix of brown, clayey sand. The compaction of this material suggests it may have been deliberately placed in the ditch to create a firm foundation or surface. To the east of F.7681 another spread of daub was present [F.7677]; this deposit was less than 0.3m deep and it was difficult to define the cut that contained it except by the interface between the daub and natural sandy clay surrounding it (plan only shown on Fig. 4). A single sherd of early Anglo-Saxon pottery was present in this feature; however, the amorphous edges of the deposit suggest it could be derived from daub disturbed from one of the other features at a late period and does not provide convincing evidence of a date for the whole group of gullies. On the western side of F.7681 there was a slightly better defined gully [F.7682], 2m in length. The eastern terminal of the feature had a steep sided, rounded profile 0.2m in depth (Fig. 5, Section 5). Further west the terminal was truncated and the feature became shallower, the profile at the western end being 0.15m deep (Fig. 5, Section 6). The fill [7675] contained a similar compacted deposit of daub fragments, as well as fragments of Roman brick or tile in a sandy clay matrix. Slightly further to the west, another distinct cut with a rounded profile (Fig. 5, Section 7, F.7683), 0.22m deep, was filled with a similar deposit of burnt daub, with occasional flints and chalk fragments [7686]. The feature became irregular and less defined further west until only a thin scatter of daub fragments, possibly dragged from the feature by later ploughing, was spread at an angle toward the south east. The most distinct elements of the group of daub-packed gullies formed a broad alignment, approximately 8m in total length. The deepest elements may have formed post settings for a structure. The daub in the fills may have derived from the clay render for such a structure falling into an adjacent gully as it decayed; however, it is not clear how this material could have been fired while it was in place. The daub may have been gathered from another area to pack into the gully to provide a firm base for some other purpose. The fill of the most substantial of the gullies [F.7681] was cut by a post hole [F.7704] belonging to a later phase of post-built structures. It is possible that the daub-filled gullies represent some form of preparatory foundation work for the post-built structures, or may perhaps be entirely unrelated to the later building. The Post-built Structures The post holes and linear gullies associated with the kiln structure ROMAN VILLA, MINSTER-IN-THANET. PT 7: BUILDING 7, KILN, POST-BUILT STS. 327 suggest that it occupied the eastern end of a barn or shed, constructed on a foundation of opposed pairs of timber posts, with a span of approximately 6-7m between the rows and a length of at least 15m (Fig. 4). The plan of the post hole settings demonstrates a degree of irregularity in the layout of the structure, which might be expected in a timber barn but could also support an alternative explanation involving separate structures. Later truncation of the posts frequently left only a cluster of rounded flint cobble packing in place within a shallow remnant of the original cut and it is difficult to estimate the dimension of any structural timbers that occupied the post settings. Four principal groups of linear alignments of roughly circular post settings can be identified (Fig. 4, Groups 1-4; Fig. 5): Group 1 consisted of four posts, three [F.7688, F.7689, F.7611] were similarly proportioned at approximately 0.35m in diameter, each had been heavily truncated (Fig. 5). At around 0.4-0.5m diameter, the easternmost setting of Group 1 [F.7695] was larger than the rest of the group, the round based cut survived to a depth of 0.2m. The packing of the post hole included flint cobbles, sandstone and tile in a sandy clay matrix, some of which had been dragged by later ploughing out of the post hole toward the south. Group 2 mirrored the arrangement of Group 1, the easternmost post being represented only by a cluster of rounded flint cobbles pressed into the natural, with nothing of the cut discernible [7721]. A better preserved post hole [F. 7699] with a steep sided and flat based cut, 0.25m in diameter and 0.2m deep, was located at the centre of the group. Tile and cobble from the packing were present in the sandy clay fill [7698]. The third post hole of the group [F.7697] at the western end was originally approximately 0.4m in diameter. The profile was steep sided but irregular, the base sloping from the north to the south. The sandy clay fill [7696] contained sandstone, tile and cobble from the original packing of the post. A shallow sloping cut extended from the southern side of the post hole cut, where later ploughing seems to have broken down the southern edge and pulled part of the packing from the post hole. Group 3. Another large post hole [F.7702] was located at the eastern end of Group 3. The profile sloped steeply toward the central point at the base, at a depth from ground level of 0.2m. The packing was again composed of sandy clay [7701] and occasional larger flint cobbles. The central post of this group [F. 7685] was also better preserved, with a curved profile surviving to a depth of 0.2m and a diameter of 0.45m. The sandy clay fill [7605] contained only a few flint cobbles from the original packing. The western end of the group was marked by a pair of post settings, the most substantial [F.7686] being aligned with the post hole adjacent to the east and, at 0.4m diameter and 0.2m depth, was comparable in dimension to the other post holes in the group. Remains GERALD MOODY 328 of the cobble and tile packing survived in the fill [7606]. The smaller adjacent post hole [F.7687] was 0.25m in diameter and only 0.1m deep. The sandy clay fill [7607] contained only fragments of tile and a few flints. It is likely that this feature formed an additional structural support or repair to the adjacent post. Group 4. The easternmost post setting of the three forming Group 4 [F.7680] was reasonably well preserved, the cut measuring 0.3m in diameter and 0.15m deep with a similar steep sided, curved base to the other better preserved post hole cuts in the structural groups. Several flint cobbles from the packing survived on the upper surface of the fill [7679]. The central post in the group was less well preserved, surviving as a cluster of packing cobbles [7713] with little indication of the edges of the original cut [F.7714] which was largely defined by the limits of the cobble spread. The third, westernmost, post [F.7704] survived to a much greater depth and was probably originally a more substantial setting than the others in this group. The original diameter was approximately 0.25m, with depth between 0.3 and 0.15m. The profile was irregular with a very steep edge on the north side, where the post was cut through the daub packing of a gully below. The base on the south side sloped down to intersect with the vertical northern edge. The southern edge had been disturbed by later ploughing which seems to have torn out the edge of the post hole and pulled out some of the packing from the silty clay fill [7703]. Fragments of sandstone and flint cobbles survived in the main body of the fill, and a block of chalk was located at the base. The plough damage had pulled a fragment of broken millstone, probably used to pack a post, flints and chalk [7707] into a narrow gouge [F.7708] extending from the original edge. Group 1 and Group 2 on the southern side were approximately aligned with each other and may represent a continuous series of posts. The gap separating the two largest diameter posts [F.7695 and F.7697] may have been an entrance to the interior of the structure. On the northern side post hole Groups 3 and 4 did not form a continuous alignment, with Group 4 angled slightly from the alignments of the other groups. Again it is possible that the gap between the two groups, formed by the largest post of Group 3 (F.7702) and westernmost post of Group 4 (F.7704), formed an entrance of similar width, opposite to the one on the southern side. The similar number of posts in each alignment of groups suggests that they may form opposed pairs between the north and the south sides, which could have supported some sort of roof structure if they were joined by rails along the alignments and a tie beam with a span of approximately 7m between the pairs. As the flow of air to the stoke hole of the kiln would have been affected by the prevailing wind, it is also possible that the posts formed some sort of frame for a protective wall or mobile baffles ROMAN VILLA, MINSTER-IN-THANET. PT 7: BUILDING 7, KILN, POST-BUILT STS. 329 to regulate the flow of air to the fires in the stoke hole. A small, rounded rectangular post hole (Figs 3 and 4; F.7720) located on the southern edge of the stoke chamber of the kiln, could have supported a cover or screen that protected the entrance to the chamber. The misalignment of post hole Group 4 suggests that this building would have been fairly irregular and perhaps only intended to be of temporary duration. A fourth truncated post hole at the western end of Group 1 [F.7688] suggests that the structure may have extended by several metres further to the west. Either an opposing post, which would continue Group 3 to the west, has been entirely truncated, or the asymmetry lends weight to the idea of several free standing screens. The Date of Building 7 The accumulation of a large assemblage of late Roman pottery, in association with some of the latest Roman coins to be discovered on the Abbey Farm site, is obviously of considerable value to understanding the development of the villa complex over time. Caution is needed before assigning the whole structural complex of Building 7, the kiln, post-built structures and the gullies to this late date. The truncation of the structural remains such as the post holes and gullies associated with the kiln creates an obvious problem in dating and interpretation. The presence of a range of pottery dating from the Late Iron Age, within the later deposits that settled over the hollow of the abandoned kiln indicates the potential for the residuality of ceramic material to complicate the dating of the features. There was a clear horizontal interface dividing the soft loam fills that overlay the area of the kiln from the distinctive deposits of burnt clay, daub and masonry filling the chambers of the structure. Other than some roof tiles no datable ceramics were present in the lower sequence of deposits. The finds within the upper loam deposits contained residual ceramics spanning the life of the settlement on the site, but encapsulated coins that must have been deposited in the last quarter of the fourth century at least and ceramics also consistent with a date in the last quarter of the fourth century or early fifth century. This clear separation between the lower and upper deposits suggests that the pattern of deposition was more complex than simply reflecting the abandonment and decay of the kiln. The structure was filled almost exclusively with material derived from its own superstructure before the finer loam deposits were formed. It is possible that the hollow of the decayed kiln body formed the focus of a distinct, later phase of deposition, with the hollow perhaps being used as a rubbish tip, or forming a shallow pond into which small change was thrown. Perhaps the post-built structure enjoyed a longer life than the body of the kiln and the loam deposits represent the debris of intensive occupation within the structure after the kiln was abandoned. GERALD MOODY 330 discussion In concept and structural form, each of the pair of channels within the kiln was not dissimilar to the right-angled stoke chambers and channels in the main villa complex that warmed the apses on the north range, and the southern end of the western range. These chambers, cut below ground level, must also have been reached by external entrances formed in the foundations of the villa ranges that must have been spanned by lintels or arches. There are also similarities in the layout and operation of the kiln with those originally constructed in the bath buildings, or later inserted into the buildings of the main villa complex, particularly those inserted into Building 6. The technology of Building 7 is a variation on a familiar structural component within the architecture of the villa complex, although it was reduced to its essential components in a free standing structure, possibly segregated from the main centre of occupation as it presented a fire hazard. Comparable timber buildings exist at the villa complex at Keston with one striking parallel in the North Timber Building, the postbuilt internal structure of which has a similar span and post spacing as Building 7 although the overall length is greater probably due to better preservation (Philp et al. 1991, 81). At first glance the structure of the kiln at Abbey Farm appears crude in comparison to the masonry structures used in the furnaces of the main villa complex. Rather than being built of chalk blocks, the sub-structure of the kiln was formed of a composite material of clay tempered by a core of chalk blocks. In use, these materials were essentially baked into a solid structural mass. Clay bonded chalk blocks compensated for the lack of building stone in the area from which similar structures were built in other Villa estates (Morris 1979) such as the three masonry corn drying ovens that were inserted in the western end of the North Timber Building at Keston (Philp et al. 1991, 81). With the conjectured superstructure in place, the kiln would have formed a sophisticated apparatus for the management and regulated distribution of heated air, derived from the fire set in the central stoke chamber. Perhaps it would be expected that if the main building had been abandoned by this time it would have provided a ready supply of tiles and brick to create a more robust structure. Only a few scraps of stone and occasional roof tile were utilised in the packing of the post holes. Debate continues about the functions of similar structures, generally referred to as corn driers, with the two leading interpretations favouring parching of grains prior to milling (as opposed to the drying of large quantities of grain prior to storage) and the use of the warm floors to germinate grains as a precursor to extracting the sugars produced in the grain for brewing. However, seasonal demands and variations in technique may have resulted in the same technology being adaptable to ROMAN VILLA, MINSTER-IN-THANET. PT 7: BUILDING 7, KILN, POST-BUILT STS. 331 both principal processes and perhaps many others associated with the requirements of a farm. Throughout the Roman occupation of the Abbey Farm site at Minster, structures were created that harnessed the supply of fresh water from the spring that rose locally in the valley, utilising a bypass system and gravity to feed water through various processes and systems of storage. To the management of the water supply was added a series of heating processes that provided warm baths for the occupants. If the driers were used to produce malt, then it is likely that evidence for brewing might be found nearby. Traditional breweries still use a sequence of gravity fed processes supplied by fresh water springs. The furnaces within the villa building supplied water for baths but also potentially the warm water required for brewing and several large pools and tanks would have been available on the site. Throughout the period of Roman occupation at the site, there would have been a continuous demand for fuel, predominantly charcoal, from the surrounding countryside which would have changed little in the transition from a domestic to an industrial process. The cobble-packed post holes of the post-built structure associated with the kiln were not unique on the Abbey Farm site. An alignment of similar, cobble-packed posts holes were discovered at the northern limit of the site, cut into the foundations of the boundary wall. These may represent a phase of timber fences or buildings that could be contemporary with the kiln. The publication of a series of buildings close to the ancient route following Thanet’s central chalk ridge near Minster (Hicks A. in Bennett et al. 2008) has added considerable information on the range and function of Roman timber and basemented structures that could be contemporary with buildings at Minster built in masonry, tile, brick and concrete. Occupation within this major complex of structures reached its peak in the late third century and in a striking parallel with the upper fills covering the kiln at Minster, late Roman pottery of 4th to early 5th century date was found in the upper deposits formed after the buildings had been abandoned. The excavation of the kiln and associated structures at Minster in 2004 has added depth to our understanding of the structures and technology that were in simultaneous operation in the Roman period and have raised questions about the social mix implied by the co-existence on one site of structural remains of varying sophistication which have not been fully explored to date (Moody 2008). acknowledgements With the publication of this, the seventh, report on the excavations at the Abbey Farm Villa the excavation team must once again acknowledge GERALD MOODY 332 with gratitude the officers of the Kent Archaeological Society who organised and funded the excavations in the 2004 season. The support of the then president Paul Oldham and the organisation of the students and facilities by David Bacchus and Chris Pout once again allowed members of the Kent Archaeological Society and the Isle of Thanet Archaeological Society to make further important discoveries in the final excavations at the Abbey Farm site. Of course our gratitude will always extend to Mr Jack Clifton on whose land the excavations took place and whose considerable interest in the progress of the work allowed us to return to excavate each year. The initial excavation of the deep cut feature in 2003 was carried out by Tim Allen and the full excavation of Building 7 in 2004 was supervised by Ges Moody while Keith Parfitt was responsible for the overall direction of the site in both years. The work took place in very difficult conditions and thanks are extended to the volunteer excavators who braved the rain and mud to reveal such an interesting aspect of the story of the development of the Villa in the Roman period. The continued support of the volunteers from the Isle of Thanet Archaeological Society in processing the artefacts from the site, and the continued support and encouragement of the Fieldwork committee of the Kent Archaeological Society have contributed greatly to the process of publication of the site, which is inevitably a complex and sometimes a slow process. The text for this report was produced by Ges Moody and the illustrations by Emma Boast. The author is grateful for comments on the text by Keith Parfitt. bibliography Bennett, P. et al., 2008, At the Great Crossroads. Prehistoric, Roman and Medieval discoveries on the Isle of Thanet 1994-95, CAT Occas. Paper No. 4, Canterbury. Holman D.J. and Parfitt K., 2005, ‘The Roman Villa at Minster in Thanet. Part 2: The Iron-Age, Roman and Later Coinage’, Archaeologia Cantiana, cxxv, 203-228. Lyne, M., 2001, ‘The Pottery from the Abbey Farm, Minster, Roman Villa’, unpubl. archive report. Moody G., 2008, The Isle of Thanet from Prehistory to the Norman Conquest, The History Press, Stroud. Morris. P., 1979, Agricultural Buildings in Roman Britain, B.A.R., Oxford. Philp, B., Parfitt, K., Wilson, J., Dutto, M. and Williams, W., 1991, The Roman Villa Site at Keston, Kent. First Report (Excavations 1968-1978), Kent Archaeological Rescue Unit, Dover. Philp, B., Parfitt, K., Willson, J. and Williams, W., 1999, The Roman Villa Site at Keston, Kent. Second Report (Excavations 1967 and 1979-90). Kent Archaeological Rescue Unit, Dover. Pout, C., 2004, ‘Final Excavation at Abbey Farm’, KAS Newsletter, 62, 1-2.