Iron Age and Romano-British Settlement at Bishop's Avenue, North Foreland, Broadstairs

197 Iron Age and Romano-British settlement at Bishop’s Avenue, North Foreland, BroadstaiAIrs gerald moody T he Trust for Thanet Archaeology carried out excavations on two adjacent development sites located to the north and south of Bishop’s Avenue, North Foreland, Broadstairs (Fig. 1). Iron Age and Romano-British features and artefacts were found on both sites; a small Roman building was located on the southern site. In July and August 2004 an excavation was carried out on land formerly within the grounds of Hamilton Lodge (Site Code BAB 04; NGR 39684 68890; Moody 2005a) in advance of the construction of four detached dwellings. The second development fronting onto Bishops Avenue involved the construction of a single detached dwelling on land separated from the garden to the rear of 103 Stone Road (Site Code SNsnB 04; 39704 69858; Moody 2005b). Archaeological excavations and a Watching Brief on the excavation of foundations for the development were carried out between March and December 2004. In this report the two sites will be referred to as BAB and SNsnB respectively. T he two sites are located at elevations between 30m and 25m above od on the south-east facing slope of a valley that falls from the Isle of Thanet’s central chalk plateau to the sea at Stone Bay. The two sites were divided by a road established in the late nineteenth century and represent elements of an extensive prehistoric to Roman settlement explored in some detail on a number of developments and voluntary archaeological projects (Hogwood 1995). Iron Age features BAB Three segments were excavated through an east-west aligned gully which terminated at the western end (Fig. 2, Ditch 1). Sherds of an Early to Middle Iron Age rusticated storage jar (550-350 bc) were found in the fill of one segment of the otherwise sterile ditch. The majority of the features on this site were dated to the Late Iron Age (150/100-50 bc). An GER ALD MOOD Y 198 Fig. 1. Site location plan. IRON AGE & ROM ANO -BRITISH SETTLEMENT AT BISHOP ’S AV., BRO ADST AIRS 199 east-west aligned ditch with a square ended terminal on the western end (Fig. 2, Ditch 2) parallel with Ditch 1 was sectioned in four places. All of the sections had a flat bottomed profile with a sharp break at the base and steep sides. Ceramics from the fills suggest the ditch was cut in the later Iron Age period although some intrusive Roman elements were present and it may have remained open into the Roman period. T wo typical Iron Age four-posthole groups were excavated (Fig. 2; 4 posts 1 and 2). The first consisted of four equally spaced circular postholes arranged into a regular square, each filled with a single deposit. A small fragment of copper-alloy sheet and a Potin coin were recovered from the fill of one posthole (Fig. 2, A). The second structure is represented by a regular group of three postholes with a fourth presumed to lie outside the excavation area. Two of the postholes were circular and filled with single deposits, the third consisted of a group of three cuts backfilled with the same deposit suggesting the remodelling or strengthening of the structure on this corner (Fig. 2, B). The coin and ceramics from the posthole fills date both structures to around 150-50 bc. SNB A shallow curving profiled cut into natural brickearth sectioned at the western end of this site (Fig. 2, Terrace) may represent an artificial terrace, possibly formed by ploughing along the contours of the slope. Further sections excavated toward the east showed a shallow slope progressing east toward the main area of excavation. Soil deposits overlying the terrace became progressively deeper and contained ceramics with an extensive later Iron Age component indicating that the material was derived from cultivated soils of that period washed down the slope of the valley. The ditches described above may have been field boundaries or enclosures associated with this cultivation. R oman features BAB Five segments were excavated through a narrow north-south aligned gully or small ditch demonstrating a steep sloping profile with a distinct slot in the base (Fig. 2, Ditch 3). Further north the slot was all that remained of the ditch as the site was truncated by later activity. Another well defined east-west aligned linear feature (Fig. 2, Ditch 4) intersected a narrow flat-based gully of the same date (Fig. 2, Ditch 6) and terminated in a cluster of pits at the eastern edge of excavation. A deep, straight-sided pit, circular in plan with a flat base (Cut 117) contained four fills, the earliest containing early Roman ceramics and a single sherd of Colchester beaker. Residual Late Iron Age material including a pottery bead and an Iron Age pottery spindle whorl were distributed throughout the fills. Large flints and roof tile recovered from the pit may be debris from the construction or demolition of buildings found at the SNsnB site. GER ALD MOOD Y 200 Fig. 2. Site and feature location plan. IRON AGE & ROM ANO -BRITISH SETTLEMENT AT BISHOP ’S AV., BRO ADST AIRS 201 O ther features, dated to the later second century, included a large pit separated into three or four features (possibly including cut 159) as a result of truncation by later activity (Fig. 2, C) and two irregular inter-cutting pits at the western ends of the linear ditches (Fig. 2, D). Although a large assemblage of Late Iron Age ceramics was recovered from the fills of the gullies and pits the features date from the first to the mid second century ad. The profuse Iron Age material suggests that much residual occupation debris was disturbed by the Roman settlement. E nvironmental samples taken from the Roman pits on the BAB site contained the charred remains of typical Roman cereals, spelt wheat, barley and emmer. Some of the remains may represent the waste product from the final stages of crop processing when most of the chaff, smaller weed seeds and tail grain would have been removed. Seeds from arable crop weeds were also present. T wo shallow isolated pits (Cut 123, Cut 148) were excavated but contained no artefacts or dating evidence. SNB The earliest features possibly dating to the Roman period were a number of infant burials discovered on the interface with the hillwash deposit. Stratigraphic relationships were limited and two scenarios can be proposed for the deposition of the infants. An existing burial group, possibly of Iron Age date may have been disturbed by the excavation of the cellared building (see below). Alternatively the infants may have been sealed under the rammed chalk deposit laid to form the floors of the building. This could have happened later in the use period of the structure as several patches and additional chalk deposits were observed. Human infant burials have been discovered incorporated into the structure of buildings such as at the East Barracks building at Reculver (Philp 2005, 75). A full term infant (Fig. 3, SK 1) along with the disturbed remains of a second full term infant had been placed tightly against the eastern edge of an elliptical pit (Cut 2031) with steep sloping sides backfilled with soft sandy silt. The pit was sealed by the rammed chalk and mortar foundation platform of the north wall of a Roman building providing firm stratigraphic evidence that it predated the structure. At the interface between the floors of a later Roman building two separate articulated infant inhumations were excavated (SK 2 and 3) along with bones from two further individuals. The inhumations were close together, flattened on to the horizon of Iron Age plough soil. A large cobble in a shallow depression to the east of the burials may represent a grave marker. Five sherds of Later Iron Age pottery were recovered from the silty matrix between the bones but these may have been derived from the earlier hillwash on which the group was placed. GER ALD MOOD Y 202 Fig. 3. Building 1, showing detail of Kiln 4 and Kiln 5 (insets). IRON AGE & ROM ANO -BRITISH SETTLEMENT AT BISHOP ’S AV., BRO ADST AIRS 203 Building 1 – Cellared structure T he northern and western edges of Building 1 were exposed in the area excavated in advance of the development. The rectangular structure (Fig. 2, Building 1) measuring 4.6 x 5m, appears to be the remnant of a cellar cut into the hillwash deposit. The construction cut with a shallow curving profile was exposed in a section across the building. The cellar was lined with cobble walls composed of large flint nodules in a bonding of rammed clay and chalk. A maximum of four courses survived in the north-western corner of the structure where the building was preserved to the greatest extent. Later terracing appeared to have truncated the building toward the southern end where no more than one or two courses survived. The rammed chalk flooring associated with the structure continued through a break in the southern wall of the building suggesting this was an entrance into the structure (Fig. 3). Stairs or a gap in the wall may have existed at this point although the evidence is not conclusive. A linear gully (Cut 2007) in the beaten chalk floor may have been a beam slot containing a structure associated with the entrance. The beam slot was revealed after the removal of a demolition dump late in the stratigraphic sequence possibly lodged in the void left by removing a door or staircase. At the south-eastern corner the lower section of a conglomerate quern had been laid flat with its central hole facing upward. It is possible that this quern was in situ re-used to act as a pivot for a door. T he north-eastern part of the structure was only revealed in plan when the trench was extended by mechanical excavator prior to backfilling. Although a short north-south aligned return was revealed this terminated at an offset platform of cobbles, which in turn appeared to have been cut away by recent terracing. This arrangement may be evidence that the cellar was excavated as an abutment to the western side of another building that has now been entirely lost. The area to the south and west of the site has been extensively terraced for the platforms of the adjacent bungalow although Roman finds described as pottery and earthworks are recorded on the Ordnance Survey within a metre or two of the eastern limit of the building. Building 2 Building 2 (Fig. 2) was represented by only a single course of cobble wall foundation forming a right angle located 10m to the west of Building 1. A modern cut had destroyed the corner at the southern end of the building. The cobble footing was contained within a shallow cut, filled with a compact demolition deposit of rammed chalk and silt containing two sherds of pottery dated to the early Roman period. It is possible that Building 2 was the truncated base of another cellar with only the last courses of the GER ALD MOOD Y 204 lining remaining in situ and the cut filled with redeposited demolition material. Building 2 was constructed directly over the possible Iron Age lynchet. Natural brickearth was present directly below the foundation at the western extreme but under the eastern edge of the structure a shallow remnant of the Iron Age hillwash deposit was present. Cobble walls east of Building 2 T wo parallel walls with evidence of at least two courses of cobble and associated spreads of tumbled flints were set within the fine soil of the hillwash between Buildings 1 and 2. There was no clear evidence of a construction cut but it seems likely that both walls represented foundation courses laid in a trench that has been obliterated by the gradual movement of the fine soil. The structures may represent the foundations for outer porches or corridors associated with the structure of Building 2 but seem poorly founded to be structural footings. They could represent low revetments designed to retain an artificial terrace possibly associated with Building 1. Pottery recovered from the soil deposits on which the walls were founded and from the interstices of the cobbles themselves span a broad range from the early Iron Age to the late Roman period. Some intrusive later wares were also recovered. T he earliest feature associated with Building 2 was a small circular pit with a gully forming a keyhole-shaped feature (Fig. 2). The sides of the pit were very steep, breaking sharply to a flat base. The fill of the pit (3012) contained seven sherds of Middle Iron Age flint-tempered pottery and a single sherd of second-century Roman pot contemporary with Building 1. It is possible that this feature was a relic of the occupation of Building 2, surviving below the level of the truncation by the robber trench that had removed the upper elements of the structure. Internal features and Kilns within Building 1 T he earliest feature within the structure was a small pit (Cut 2024) with a curving bowl-shaped profile filled with a single ashy silt deposit and containing sherds of pottery of early second-century ad date (Fig. 3). The pit was sealed by a rammed chalk floor (2013) that was laid right up to the slightly sloping chalk and clay pad on which the cobble cellar lining was constructed. Interleaved patches of soot and scorching as well as varying depths of material indicated that the floor was composed of numerous spreads and dumps of chalk deposited at different times. Five small kilns and other pits were cut into the chalk floor of the cellar although it is likely that no more than two of these were in operation at any one time and the floors had been repaired each time a kiln was replaced or a pit ceased to be used. Kilns 1 and 2 were the earliest in the IRON AGE & ROM ANO -BRITISH SETTLEMENT AT BISHOP ’S AV., BRO ADST AIRS 205 sequence. Kiln 1 (Cut 2015) had a very shallow platform at the south end dipping to a flat based bowl at the north. The upper edges were reddened by the heat of firing. Kiln 2 (Cut 2030) was a narrower, irregular, elliptical shape with a flue extending to the south; the upper edge of the structure was also heat-reddened. Kiln 3 was composed of two roughly elliptical elements, a steep-sided body with a curved base and at the southern end a shallower flat-based stoke chamber. The upper edge of the kiln showed distinct red scorching. The sooty fill was sealed by a distinct band of fine ash with a patch of scorched red clay above it and in turn was sealed with a chalk dump restoring the level of the floor surface. The latest date for pottery from the body of the kiln indicated it had been abandoned late in the second century ad. A sub circular pit (Cut 2041) cut through the chalk floor was backfilled with two further deposits of chalk (2026 and 2027), restoring the original floor surface level. Pottery from these deposits could date as late as the third century. A second pit (Cut 2023) cutting the western side of 2026 may have been the body of another kiln as the fills contained large amounts of charred wood, seeds, grain and chaff. Pottery from the upper deposit and the stratigraphic association with the earlier pit and kilns suggest a slightly later, late second-century date. Charred grains and weed seeds preserved within the sooty fills of Kilns 1-3 (2014, 2016 and 2034 respectively) indicated that crop waste from the cultivation of spelt wheat and hulled barley had been used as fuel. The kilns may have been used to parch grains prior to milling. Both the kilns were sealed with dumps of chalk replacing the original floor surface. Kiln 4 (Fig. 3, Inset 1) was slightly different to the earlier kilns within the cellared building. A rough elliptical platform of broken Roman roof tiles and combed box flue tile fragments was set within a compact deposit of clay surrounding the platform which had been reddened and baked hard (2008, 2018 and 2019). Large flat sandstone slabs were set within the scorched fringe. The platform may have been more extensive and was perhaps originally almost circular in plan. A thick soot deposit (2010) overlying the platform produced cereal grains and charred fragments of wood. Many metal small finds including iron nails, rings and fittings, copper alloy fragments and a Roman openwork military belt buckle (Fig. 4, 1) and a copper alloy spoon (Fig. 4, 2) were recovered from the deposit suggesting the burning of structural timbers and even clothing as fuel within the kiln. The belt buckle is paralleled at Richborough (Bushe-Fox 1928, 55) and from military sites in Dacia and Osterburken, and Pfunz in Germany (Moody 2005b). Fragments of burnt daub and clumps of degraded lava quern were also present in the deposit. Among other pottery associated with the structure of Kiln 4 and the soot deposit were GER ALD MOOD Y 206 Fig. 4. The military belt buckle, copper alloy spoon and Dorset BB1 pottery. IRON AGE & ROM ANO -BRITISH SETTLEMENT AT BISHOP ’S AV., BRO ADST AIRS 207 a number of sherds of a Dorset Black Burnished ware (BB1) beaded-and-flanged bowl with burnished arcading and six fresh joining sherds from a straight-sided dish in black BB1 fabric (Fig. 4, 1-2). Kiln 4 was sealed by a mixed deposit of clay interleaved with soot residue formed by trampling irregular spreads of clay and chalk over the sooty debris below. Pottery dated to the late second to early third century and two coins (Gordian III AD 243-244; Otacillia Severa AD 244-249) date the floor deposit to the mid-late third century ad. Kiln 5 (Fig. 3, Inset 2) consisted of an L-shaped flue (Cut 2037) formed in the chalk floor of the north-western corner of the building. The sides of the flue and the body of the kiln were formed of stiff brown clay strengthened with occasional chalk blocks and large fragments of tegula, set against the angle of the corner of the flint cobble lining of the cellar (2011). The kiln body also contained fragments of painted wall plaster. The surviving structure of Building 1 had no evidence of fine finishes such as plastered walls and these and the tiles possibly derived from the demolition of the superstructure of the cellar or from another high quality Roman building nearby. A sandstone platform (2044) was set at the entrance to the flue that was constricted by chalk blocks and tiles set on edge. The internal chamber was parallel sided and roughly 0.36m wide throughout its length until it finally petered out with a narrow channel. The exit from the flue was blocked with a large fragment of millstone and a cobble, perhaps a mechanism for regulating the flow of heat. The flue was filled with a fine soot deposit (2012) that also included a combed box flue tile fragment, pieces of lava quern, fragments of painted plaster, some daub and small mortar fragments where the upper part of the structure had collapsed into the flue chamber. The end of occupation in Building 1 Building 1 appears to have been abandoned in the late third century. A compact deposit of gritty silt with a high proportion of chalk fragments and small pieces of painted wall plaster overlaid the upper structure of Kiln 5 and extended over much of the area of the northern corner of the cellar. A more extensive demolition deposit of large flints derived from the walls of the structure in a matrix of yellowish brown soil with profuse chalk inclusions and fragments of painted wall plaster was concentrated in the better preserved north-western corner of the cellar and extended patchily over most of the excavated area of the structure. The upper course of walling in the north-western part of the cellar had tumbled east into the building, slipping off in coursed layers (Fig. 3). At the southern end of Building 1 a roughly rectangular robber cut (Cut 2007) had truncated the trampled floor deposit (2005). The cut had GER ALD MOOD Y 208 been backfilled with a spread of large cobbles in thick silty clay (2006), which appeared to be contemporary with the demolition material filling the northern part of the structure. T he archaeological deposits on the BAB site were covered with cultivated garden soil from the former gardens. All the archaeological deposits on the SNB site were sealed by a deep overburden of fine colluvium and imported garden soil (2000). the fiFInds R oman Pottery (comments by Dr Malcolm Lyne) Analysis of the 85 contemporary pottery sherds from the c.ad 150-300 dated occupation within Building 1 indicate a site where most of the pottery in use consisted of high-temperature fired handmade jars from local ‘Native Coarse Ware’ producers and imported Thameside industry products from production sites around the Medway estuary, which accounted for about two-thirds of the occupational pottery. The third significant source of pottery is more surprising and may have something to say about the nature of occupation within the Stone Road buildings. Dorset Black-Burnished ware (BB1) is quite rare in Kent because of the proximity of the Thameside kilns producing its wheel-turned BB2 equivalent. What little BB1 that there is tends to occur in mid-to-late third-century contexts; with its appearance coinciding with the final occupation on several rural sites. Unusually large quantities are also present in constructional and earliest occupational contexts within the stone shore forts at Richborough and thought to have been constructed during the period c.286-296 at the time of the secession of the British provinces under the usurpers Carausius and Allectus. It may be that military units transferred from the northern frontiers to man the new forts against the threat from the Continent brought their existing BB1 pottery supply arrangements with them. The presence of both BB1 vessels and an openwork belt buckle within the Stone Road buildings may, in turn, indicate late third-century military activity on the site. P ainted Plaster, Querns, Millstones and Coin N inety-four painted plaster fragments were recovered from deposits within Building 1. Although some derived from demolition deposits, others were incorporated into the final clay floor and into the body of Kiln 5 suggesting they had already been removed from their original context and deposited while the building was still occupied. Fragments of worn millstones were found within the soot deposit of Kiln 4 (2010) and within the structure of Kiln 5 (2011 and 2012) and IRON AGE & ROM ANO -BRITISH SETTLEMENT AT BISHOP ’S AV., BRO ADST AIRS 209 along with fragments of Lava quern within demolition rubble filling the building. Most of this material – like the conglomerate rotary quern possibly used as a door pivot – had been casually reused for building material. P otin coin of the Kentish Primary series, BAB 04, Context 192, Small Find 3, comparable with Von Ardsell 1428 (D. Holman, pers. comm.). general discussion T he features from both the Iron Age and the Roman period on both sites have been heavily eroded by later use of the land particularly for agricultural and later residential development. In common with many archaeological sites in Thanet survival of features is restricted to deep cut pits and sunken elements of buildings such as cellars. It is possible that the remaining elements of Buildings 1 and 2 at SNsnB represent relicts of a single multiple roomed structure such as a villa. Preservation of the floor surface of the cellar has encapsulated a sequence of deposits giving a brief glimpse of the daily events carried out within the building and some clues to its status in the form of the painted plaster and building materials. The artefact assemblage has allowed a chronology to be constructed and associations to be made with other archaeological and historical events. T he building’s location within the valley affords an impressive view of the sea at Stone Bay and the building may have originated as a private house with the cellar reflecting the service quarters. Alternatively the sunken element could represent part of a smaller purpose-built structure forming part of a complex of buildings and may have served a semi-commercial or service function within a larger community. The association with the kilns utilising waste from cereal crop production as fuel and the charred cereal grains recovered from the kilns in addition to several fragments of mill stone, lava quern and hand querns suggest continuous use over a period of time as part of a domestic bakery. The two later kilns were fundamentally different in structure and character and contained the pottery and belt buckle with military associations. These were the last activities carried out within the building before its abandonment. In 2001 Dr David Perkins summarised the current state of knowledge of Roman Archaeology in Thanet suggesting that ‘a new picture of Roman Thanet is emerging as a populous island landscape liberally sprinkled with Villas and farming settlements’. The distribution map of known and possible Roman buildings (Perkins 2001, fig. 2) showed a gap in the North Foreland area and the article postulated the presence of an undiscovered building as a focus for the three Thanet SMRsmr references recorded for the area at the time. An ongoing research project carried out by the Trust for Thanet Archaeology and Thanet Archaeological Society has identified further references to Roman sites in this area. GER ALD MOOD Y 210 The distribution pattern of sites identified by Dr Perkins (2001) reveals other apparent gaps which suggest a significant further number of Roman buildings once existed on Thanet. Excavations since 2001 have indeed continued to produce evidence of Roman settlement and structures. Little in the way of plans, stratigraphic sequences or dating for Roman sites on Thanet have been available until recently. It is difficult without this data to propose any chronological patterns or to comment on the status of the Roman community on Thanet. In proposing a sequence of events for the Thanet area in the Roman period the excavations of the Roman settlements and forts at Richborough and Reculver provide the most significant published evidence (Bushe-Fox 1926, 1928, 1932, 1949; Cunliffe 1968; Philp 2005). Recent work has provided such information for three sites; Abbey Farm Villa (Parfitt 2006), now the most extensively excavated Roman structure in Thanet, and the excavations at BAB and SNsnB. T he pottery from the Abbey Farm Villa spans the period from the late Iron Age to the third century ad (Lyne 2006). In common with BAB and SNsnB, Abbey Farm Villa was probably a successor to an extensive Iron Age community and developed into a building complex in classical style. The latest phase was also associated with the supply to the villa of BB1 vessels before the villa buildings were abandoned and extensively robbed. This pattern seems to be reflected in the development of the building at SNsnB from its origins in an Iron Age community to its abandonment in the third century. The small amount of pottery and coins reflecting settlement of third- to possibly fifth-century date at Abbey Farm were associated with the two flues of a malting kiln set within a post-built barn within the enclosing perimeter wall of the villa and robber trenches for the villa buildings. The great villa with its amenities, mosaics and fine painted walls were replaced by timber sheds. There was no evidence of a later occupation phase at SNsnB or BAB implying that the focus of settlement had moved elsewhere. In view of Dr Lyne’s comments on the association between BB1 vessels and the events of the late third century, it seems that we are now able to go further than simply demonstrating the presence of Roman settlement in Thanet and attempt to construct a chronological sequence pinned to the events surrounding the usurpation of Caurausius and Alectus after which the settlement pattern of the first two and a half centuries of Roman Thanet was fundamentally changed. More evidence is required to prove this but the increasing volume of data will certainly contribute to focussing the picture of Roman Thanet proposed by Dr Perkins. Fully illustrated archive reports with specialist artefact analysis and photographic archives for both sites are held by the Trust for Thanet Archaeology and are available on request. IRON AGE & ROM ANO -BRITISH SETTLEMENT AT BISHOP ’S AV., BRO ADST AIRS 211 acknowledgements T he author would like to thank HMEhme Developments Ltd for funding the BAB excavations and post-excavation and Mr Jonathan Kelly for funding the SNsnB excavation and post-excavation. Thanks are due to the volunteers who assisted the excavation team at both sites namely Pip Fisk, Chris Pout and Maggy Redmond. The site plans were produced by Stephen Clifton and Susan Deacon and the artefacts illustrations were produced by Maggy Redmond. Chris Tucker provided parallels for the military belt buckle. T he specialists who contributed to the archive reports were: Dr Malcolm Lyne (Pottery); Kate Roberts (Botanical remains); and Louisa Gidney (Animal bone). David Holman identified the Iron Age coin. The late Trevor Anderson carried out the analysis of the human bone: he will be greatly missed. bibliography Bushe-Fox, J.P., 1926; 1928; 1932; 1946, First/Second/Third/Fourth Report on the Excavation of the Roman Fort at Richborough, Kent, Reports of the Research Committee of the Society of Antiquaries of London, Nos vi, vii, x, xvi. Cunliffe, B.W. (ed.), 1968, Fifth Report on the Excavations of the Roman Fort at Richborough, Kent, Reports of the Research Committee of the Society of Antiquaries of London, No. XXIII. H ogwood, P., 1995, ‘Investigations at North Foreland Hill’, Archaeologia Cantiana, cxv, 475. L yne, M., 2006, The Pottery from Minster Roman Villa (AFM 1996-2004) (Archive report commissioned by the KAS). M oody, G.A., 2005a, Hamilton Lodge, Bishops Avenue, Broadstairs. Archaeological Report, unpubl. Trust for Thanet Archaeology client report. M oody, G.A., 2005b, Land to the rear of 103 Stone Road. Archaeological Report, unpubl. Trust for Thanet Archaeology client report. Parfitt, K., 2006, ‘The Roman Villa at Minster in Thanet. Part 3: The Corridor House, Building 4’, Archaeologia Cantiana, cxxvi, 115-133. P erkins, D.R.J., 2001, ‘The Roman Archaeology of the Isle of Thanet’, Archaeologia Cantiana, cxxi, 43-60. P hilp, B., 2005, The Excavation of the Roman Fort at Reculver, Kent (KARUru, Dover).