A Late Iron-Age Early Roman Site at Bredgar near Sittingbourne

A LATE IRON-AGE / EARLY ROMAN SITE AT BREDGAR, NEAR SITTINGBOURNE

damien c. boden

In 2004 an area of land lying to the west of Bredgar Primary School was subject to archaeological excavation in advance of the construction of a new classroom block. This was undertaken by a field team from the Canterbury Archaeological Trust under the direction of the writer and followed a watching brief on two geotechnical test pits (Willson 2003) and a strip and map evaluation which had identified the presence of a number of ditches, post-holes and the badly disturbed and truncated remains of a flint-built wall foundation (Helm 2003). A large corpus of pottery was recovered during this work, analysis of which pointed to activity and settlement of the area during the later Iron Age or ‘Belgic’ period, and the first century of the Roman occupation.

The village of Bredgar is situated 4.8km south-west of Sittingbourne, 3.8km south of Roman Watling Street and some 4.5km north of the Pilgrims’ Way on the dip slope of the North Downs. The historical centre of the village lies at the staggered junction of Bexon Lane, Gore Road and the B2163 (known as the High Street through the village) which traverses the Downs between Sittingbourne and Hollingbourne. The area investigated lies to the east of the historic centre of the village (Fig. 1) on Bexon Lane (centred on NGR TQ 8805 6026), adjacent to the medieval church of St John the Baptist.

Little previous archaeological work has been done in the general area of the village, although during the construction of a bungalow in Gore Road during the summer of 1957, a very important Roman coin hoard was uncovered, consisting of thirty-four gold aurei, the latest of which were four issues of Claudius of ad 41 in near mint condition (Carson 1959, 17-22). Much has been written on the circumstances surrounding this hoard; deposition by an officer before an engagement with the native Britons during the crossing of the Medway is one popular theory (Frere 1974, 81).

An archaeological watching brief during alterations to the Old Post Office which lies some 150m west of the site (TQ 8748 6012), identified a linear feature with a roughly e-w alignment (Willson 2002). This ditch was recorded as being at least 1m deep and although not securely dated may be associated with the ditches identified during this excavation. An evaluation undertaken by CAT in 1996 in advance of residential development at Travers Gardens (TQ 8785 6037), on the suggested site of a Roman marching camp, located some 250m west of the school and only 90m south of the coin hoard find-spot, proved fruitless (Ward 1996).

Excavation in advance of a housing development at Wises Lane, Borden (TQ 8870 6370), which lies some 3km north of Bredgar revealed three phases of field system and a small cremation cemetery, both dating to the first century ad. Little evidence for domestic occupation was found although a settlement spanning the decades either side of the Roman conquest must lie nearby (Hammond et al. 2003).

Various soil and crop marks have been recorded in the area around Bredgar including possible ring-ditches at TQ 8837 5920 and TQ 8931 5915; a curvilinear soil mark at TQ 8915 6106; circular crop-marks at TQ 8787 6155 and TQ 8700 6070 and circular crop marks and soil marks at TQ 8847 5906 which have been interpreted as either a prehistoric enclosure or post-medieval windmills. The fields to the east of school buildings beyond the village hall were quarried for brickearth during the late nineteenth/early twentieth century.

the excavation

The site was bounded to the south by Bexon Lane, to the west by the church of St John the Baptist, and to the north by private gardens. The modern ground surface gently slopes down towards the north-east, from 88.49m to 88.0m od. The underlying geology is Upper Chalk overlain by Clay-with-Flints.

The excavation area was 25m long with a width of 10m at its northern end and roughly 15m at the south and lay immediately to the north-west of the Victorian school building (Fig. 2). It was formerly known as the Headmaster’s garden and consisted of a lawn, flower beds and the hardcore base of a small garage. The turf, topsoil and other modern overburden [110] had been removed during the strip and map evaluation and was found to form a relatively even layer some 0.40m thick over the entire area. Although minimal excavation had been carried out during the evaluation it had been possible to suggest at least three distinct phases of activity encompassing the later Iron Age and early Roman periods. The work carried out by CAT during the spring of 2004 identified the presence of many more features and deposit sequences which has augmented the evidence gathered from the original work. At least two phases of prehistoric activity, and five phases of early Roman occupation or activity can now be suggested.

Phase 1: Undated prehistoric features (Fig. 3)

Toward the south of the site two shallow features [F63 and F64] were identified. F63 was located on the western side of the excavation to the north of, and cut by, curvilinear ditch line [F103] (described below). This consisted of an irregular subcircular cut c. 2m long and 0.36m deep, containing a single fill of light greyish brown clay. F64 lay some 3m further north along the western limit of the excavation and formed a sub-circular cut 0.80m long with a maximum depth of 0.17m and contained a clay fill very similar to that of F63. No dating evidence or cultural material was recovered from either feature and they were therefore interpreted as either depressions in the underlying natural subsoil or more likely tree-throw scars.

Phase 2: Late Iron Age( ?) enclosure ditch F103 (Figs 3 and 4)

The earliest activity on the site was represented by a curvilinear ditch [F103] which attained a maximum depth at its western end of 0.85m which traversed the southern end of the site in a roughly wnw to ese direction. Although later activity had removed the inner or southern side, enough of the feature remained to suggest an original width of c.1.50m at its western end, widening to c.2.64m towards its eastern end where it turned abruptly toward the south-east. The outer or northern side of the ditch was steeply convex in profile and although its southern side had been truncated, a more gently sloping profile is suggested by the sections shown in Fig. 4. Four fills of greyish yellow, silty clay were identified which produced Late Iron Age ‘Belgic’ pottery, struck flint flakes, a flint scraper and fragments of calcined flint.

Phase 3: Late Iron Age ?enclosure ditch F102 (Fig. 3)

The Phase 2 ditch F103 was superseded by a more substantial ditch [F102], some 2.50m wide at its western end and gradually widening to c.3.80m, where it was lost below the eastern limit of the excavation. F102 was slightly deeper than its predecessor, attaining a maximum excavated depth at its eastern end of c.1m. This had removed the southern or inner edge of F103 and completely truncated the upper fills of the earlier feature along its eastern end. At least six deposits were seen to infill the lower extent of the cut which probably represent a period of natural silting toward the end of its lifetime. Sherds of flint-tempered and other pre-conquest ‘Belgic’ pottery in association with a fragment of unidentified metal slag and calcined flint were recovered from these deposits. The remainder of the ditch was filled by deposits of clayey silt which contained a dozen sherds of late first-century Roman pottery, and probably represents a deliberate infilling contemporary with post-conquest occupation of the site.

Phase 4: Mid first-century Roman occupation (Fig. 4)

This phase of occupation is represented by the deliberate backfilling of ditch F102, the cutting of a number of ditches, pits and post-holes and the construction of a postulated timber structure in the south-west corner of the site.

A substantial ditch [F101] was excavated which had an e-w alignment and ran centrally across the full width of the excavation area. This possessed a V-shaped profile with a fairly constant width of c.3m although this did undulate, particularly along its northern edge. This erosion of the edges was probably caused by a period of cultivation of the area described below. The depth of this feature ranged from just over 1m at its western end to 0.85m at its eastern end. These measurements represent the depth of the cut in relation to the contemporary topography which sloped away to the east. When measured from the horizontal the ditch cut possessed a slight fall from west to east. As a result of the re-cutting or re-establishment of this feature (Phase 6, below) only the basal fill deposits had survived. No pottery was recovered from these fills, although a fragment of Roman tile was present in the upper extent of fill [168] located at the western end of the feature.

A rectilinear ditch system [F104] was identified in the south-eastern corner of the site, partially overlying the eastern extent of curvilinear ditch lines F102 and F103. This consisted of a narrow, V-shaped, n-s aligned ditch c.10m in length which extended from the southern limit of the excavation to the southern edge of the e-w aligned ditch F101. Truncation from the re-cutting and re-establishment of F101 (Phase 6, below) had removed any evidence of the true relationship between it and F104, although given that F104 was not traced any further northwards beyond F101 it is possible that F104 had originally drained into it. Two further elements of this probable drainage system, an e-w aligned ditch and a parallel, shorter ditch line were seen to extend eastwards, at right angles to the n-s ditch. The three ditches together enclosed a subrectangular area 4.80m wide by c.7m long. This area had unfortunately been very badly disturbed by later activity including the laying of drains and other below-ground services.

This ditch system may suggest that a structure had originally occupied this area, although no definite evidence in the form of post-holes or other structural elements for this was forthcoming. Very little cultural material was recovered from the ditch fills with three sherds of flint-tempered Late Iron-Age pottery retrieved from fill [263] of the northern e-w aligned ditch and a number of post-conquest sherds from the basal fill [311] of the n-s aligned ditch.

Seven subcircular post-holes with diameters of between 0.30-0.60m and depths of between 0.25 and 0.45m can also be included in this phase. Four of these [F11], [F13], [F59] and [F60], were located to the north of e-w ditch F101, may represent an irregular timber post structure, further elements of which may have been removed with the cutting of F101. Fragments of Roman brick together with small fragments of calcined flint were recovered from deposits [127] and [131], the fills of F11 and F13 respectively. F48 located on the western edge of the excavation, [F70] on the eastern side and [F39] toward the southern limit of the site represent single posts with no definite function. Immediately to the south of F48 four stake-holes were identified [F46, F47, F50 and F51]. No meaningful arrangement or structure was definable, although further elements may lie to the west, beyond the limits of the excavation. A further stake-hole [F61] located some 9m south of the main cluster, was also identified although this appeared to be an isolated feature.

Twelve features of varying shape and dimensions were investigated to the north of ditch F101 and included shallow, subcircular pits, rectilinear ‘troughs’ or gullies and a possible square-sectioned ditch or drain. Two of these features [F45 and F49] which lay on the periphery of the group, were discrete in nature, while the remaining eight possessed relat-ionships with other features. F49 was located on the eastern side of the excavation and consisted of an irregular elongated cut 1.49m long, 0.90m wide and a maximum depth at the eastern limit of the excavation of 0.40m. F45 was located on the western side of the site and consisted of a subcircular cut 1.20m long, 0.94m wide and 0.39m deep, the fill of which [202], produced a sherd of flint-tempered pottery and a small fragment of Roman tile.

The earliest of the eight features representing a sequence of pits and gullies in the north-eastern corner of the site [F62] was a small pit roughly circular in shape with a diameter of 0.86m and a depth of 0.75m. This feature contained a single stoney fill [250] which produced no cultural material. F62 was cut on its eastern side by a similar although slightly wider pit [F57] which possessed a concave profiled cut with a diameter of c.1m and a depth of 0.58m, and contained a single fill [242] again devoid of cultural material. Pit F57 was in turn truncated on its south-eastern side by a straight, vertically-sided, flat-bottomed cut [F53] and on its northern side by a similar although badly truncated feature [F56]. Both features may represent gullies or drains although truncation from later activity and the limited sample available for investigation precluded any secure interpretation. Cut F53 which possessed a sw-ne alignment, had a length of 2.50m, a width of 0.54m and a maximum depth at its eastern end of 0.20m and contained a stoney, clayey silt fill [223] which produced three fragments of Roman brick or tile. F56 which cut across the north-eastern corner of the excavation consisted of a straight, steep sided cut with a distinct step or ledge running down the base on either side. This had a length of c.1.46m, a width of 0.70m and a depth of 0.49m and contained a primary silty clay fill [238] and an upper stoney clay fill [239], which produced fragments of Roman tile and a scrap of undiagnostic Roman pottery. Much of the visible extent of this feature had been removed by a large circular pit [F58] which consisted of a steep-sided, concave cut with a diameter of c.2.30m and a maximum depth of 0.55m. This contained two fills [274] and [275] of stoney, yellowish brown clay, the upper fill [274] produced three small fragments of Roman tile.

To the west of gully F53 a rectilinear ‘trough’ or elongated pit [F54] was excavated. This consisted of a straight, vertically-sided and flat-bottomed cut with a roughly n-s alignment. This had a surviving length of 3.80m, a width of 0.71m and a maximum depth at its northern end of 0.53m. Two fills were recorded, an upper fill [230] and a primary fill [237] both of which consisted of clayey silt and produced small fragments of Roman tile. The basal fill [237] was subject to environmental analysis which detected small deposits of cess-like material, the presence of which suggests that this feature and possibly all of the rectilinear gullies or ‘troughs’ in this area, can be interpreted as latrines.

F54 was cut along its eastern side by a similar feature [F52] and on its western side by pit [F55]. F52 consisted of a linear, steep-sided flat-bottomed cut [220] which contained three fills [217, 218 and 219]. This feature was 3.68m long, a maximum of 0.78m wide at its northern end and 0.58m deep. Fragments of Roman brick and pottery were recovered from fill [218]. F55 was subcircular in plan with a shallow, concave pro-file, 0.74m long, 0.33m wide and 0.12m deep. Two fragments of Roman tile were recovered from a single fill [235] of stoney, greyish brown clay.

A large circular, although somewhat irregular feature [F35] was also investigated. This had steep, concave sides and a concave base with a diameter of c.2m and a depth of 0.49m. No cultural material was recovered from either of its two silty clay fills [170] and [256]. Pit F35 was cut by a smaller feature [F31]. This was subrectangular in plan with shallow sloping concave sides and a concave base, 1.11m long, 0.60m wide and 0.09m deep, and contained a single fill [162] which produced no cultural material.

Three pits [F76, F81 and F82] were excavated to the south of ditch F101. F76 was located at the southern end of site and cut through the backfill deposits of Phase 3 curvilinear feature F102. F76 was only recognised in section during the excavation of F102 and its complete extent was not visible. This feature possessed a visible width of c.1m and a depth of 0.66m. Two fills were identified [281] and [282] which produced sherds of post-conquest Roman pottery and fragments of calcined flint. F82 was located immediately to the east of and cut by pit F81, and also extended under the southern limit of the excavation. The surviving, visible part of the feature was an elongated quadrant with a concave profile, 0.47m long, 0.21m wide and 0.12m deep. No cultural material was recovered from its single fill [298]. F81 was located in the far south-eastern corner of the site and was seen to extend under the southern limit of the excavation. The visible part was subcircular in plan with shallow sloping, concave sides and a concave base. This had a width of 1.08m and a depth of 0.45m and contained a single fill [296]. A single sherd of late third-century Roman pottery was recovered from the upper surface of this deposit although this should be regarded as intrusive.

Phase 5: temporary abandonment and deposition of plough soil deposit

The very early Roman occupation and activity on the site appears to have ceased toward the later part of the first century when the area was used for agriculture purposes. All features and deposits in the preceding phases were sealed by a layer of stoney ploughsoil. This deposit was hand excavated in various locations, with the remaining material removed by machine. These deposits were recorded as three groups or deposit sequences; [G117] which formed a general layer present over most of the site; [G118] which formed a basal fill of the Phase 4, e-w aligned ditch F101 and deposits (S102) which formed the lower fills of the Phase 4, rectilinear ditch system F104. G117 formed a layer 0.20-0.25m thick and consisted of light yellowish-brown slightly sandy clay with frequent inclusions of small, rounded and sub-angular flint pebbles and pebble fragments. A number of very sparse areas or lenses of small, dark rounded flint pebbles were also observed within this deposit. This material does not occur naturally in the underlying clayey subsoil and therefore it is possible that these pebbles were imported onto the site and may represent very badly disturbed surfaces. Apart from a concentration of pottery and tile dated to the later first century, present in the south-east corner of the excavation and recorded as deposit [155], little cultural material was recovered from this deposit.

Phase 6: later first- to early second-century Roman occupation (Fig. 5)

A second phase of domestic occupation was represented by the construction of rectangular building [F114] which was located in the south-eastern corner of the site immediately to the north of, and partially truncated by, the existing school building. Further disturbance from the installation of several modern services including inspection pits had destroyed all of the eastern side and internal elements although the northern and western sides including the north-western and north-eastern corners had survived.

From the surviving elements it is possible to suggest a square or rectangular building 8.12m long and at least 4m wide which consisted of flint nodule filled trenches. These were c.0.60m wide and c.0.25m deep and probably formed the foundations for either a timber or half-timbered structure. No mortar or other bonding material was identified, although where excavated along its northern side, a single course of smaller more angular flints was seen to overlie the larger more rounded flint nodule fill of the foundation trench. A small projection was present extending westwards from the north-west corner which may represent a small buttress or strengthening to this corner of the building which straddled the n-s arm of the earlier Phase 4 ditches F104. No internal structural elements or above ground wall construction survived although large fragments of tufa (travertine) together with imbrex, tegula and other ceramic building materials were present in the considerable deposits of demolition rubble which were found as spreads and the upper fills of the ditch system F104 located to the west and north of the surviving foundations.

The re-cutting of the Phase 4 e-w ditch line F101 also appears to have been carried out at this time. This consisted of a slightly convex sided V-shaped cut [F100], c.10m long running centrally along the partially silted ditch F101. This possessed a maximum depth of 0.64m along its central and eastern extent although it was somewhat shallower at its western end where it reached a depth of 0.45m. This contained a single lower fill recorded as deposits [167], [165] and [176], one of which [165], produced a substantial pottery assemblage dated to the later first century in association with a few fragments of Roman tile and daub. The upper fills of this feature probably represent a period of deliberate backfilling or disuse and are described in Phase 7, below.

Phase 7: early to mid second-century abandonment (Fig. 6)

This phase is represented by demolition deposits of flint nodules and large pebbles together with fragments of ceramic building material overlying and surrounding the building foundations F114. These deposits were also seen to infill the partially silted rectilinear ditch system F104. Little cultural material was recovered from these deposits although layer [330], which was located to the north of and partially covered the north-west corner of the flint foundations, contained over forty sherds of a single-handled flagon and other Roman fabrics dated to the mid second century

Two deposits [123] and [124] of dark greyish, brown clayey silt were identified at the southern end of the site to the west of structure F114 and infilling a slight depression created by the underlying curvilinear features F102 and F103. Deposit [124] was c.0.12m thick and covered an area some 6m north to south and 4m east to west. This produced over 130 sherds of pottery which can be dated to the mid second century, together with fragments of Roman brick and tile and two fragments of unidentified metal slag. This was overlain by a less extensive deposit [123] which covered an area of 2.20m n-s by 1.60m e-w. This was slightly darker than [124] and produced eighty-seven sherds of pottery of the same date as deposit [124]. These deposits probably represent a remnant of buried soil contemporary with the Roman occupation, which has been removed or disturbed in other areas by later ploughing, although a ‘kitchen garden’ associated with building F114 is also possible.

Similar deposits were found to infill ditch F100 and probably represent deliberate infilling or disuse of the feature. These deposits consisted of dark greyish brown clayey silts which produced a considerable quantity of pottery dating from the later first to the mid second century. A single sherd of mid third-century pottery was recovered from deposit [156] located at the eastern end of the feature although this is probably intrusive.

Phase 7a: later Roman through to the construction of the school buildings

Although no definite features or deposits can be securely dated to the later Roman period, a number of pottery sherds retrieved from the upper fill of ditch F100 and intrusive material from pit F81 which are considerably later than the bulk of the assemblage, does suggest later Roman activity in the area of the investigation. Similarly, the paucity of cultural material dating from any post-Roman period suggests very little activity in the area until the construction of the school buildings in the late nineteenth century.

Phase 8: nineteenth and twentieth centuries

This phase of activity is associated with the construction of the school building in the mid 1860s and subsequent use of the area as the school garden.

the finds

The ‘Belgic’ and Roman Pottery (Figs 7-10) by Andrew Savage

The modest assemblage of pottery recovered during the course of the excavation comprises a relatively wide range of types, which are principally of ‘Belgic’ and early Roman date. Relatively little of the material would appear to post-date the first century ad. Of particular interest is the relatively large and distinctive corpus of flint-tempered ware.

All of the pottery was examined by eye and with the aid of a X20 hand lens. The fabrics in each layer were quantified by sherd-count, weight, and EVEs (estimated vessel equivalents), using rims (Orton 1975, 30-35). A quantified list of fabric types from the site, giving CAT fabric code and common name is presented below. In the absence of a suitable reference, a brief fabric description is included, if appropriate. Generally, in this report, fabric codes only are used. There were 1,044 sherds of ‘Belgic’ and Roman pottery, weighing approximately 7,507 grams, with a total EVE’s value of 877.64 per cent.

Most of the pottery fabrics are either common locally, or have a more widespread distribution and have therefore been previously discussed and described. No fabric descriptions are given here for these wares. References are given, where appropriate, to published descriptions or discussions; they are limited to those considered by the writer to be reasonably accessible and most relevant. In many cases the reference(s) chosen will act as links to further, more extensive bibliographies. The fabric codes are as follows:

B1: ‘Belgic’ grog-tempered ware (fine) (Thompson 1982, 20-25; Pollard 1995, 589-90). 20 sherds. c.140g.

B2: ‘Belgic’ grog-tempered ware (coarse) (Thompson 1982, 20-25; Pollard 1995, 589-90; Tomber and Dore 1998, 214). 51 sherds. c.705g.

B3: ‘Belgic’ grog-tempered ware (coarse), with flint. As B2 with sparse flint (Thompson 1982, 20-25; Pollard 1995, Tomber and Dore 1998). 21 sherds. c.185g.

B6: ‘Belgic’ shell-tempered ware (coarse) (Pollard 1987, 208; Monaghan 1987, 249). 26 sherds. c.362g.

B8: ‘Belgic’ sand-tempered ware (coarse) (Pollard 1995, 595). 13 sherds. c.152g.

B9: ‘Belgic’ sand-tempered ware (fine) (Pollard 1995, 595). 15 sherds. c.225g.

BER7: Early Gaulish white beaker ware, Rigby fabric 1B (Rigby 1995, 648). 2 sherds. c.11g.

BER11: Early Gaulish white flagon ware, Rigby fabric WW1 (Tomber and Dore 1998, 22). 3 sherds. c.20g.

B/ER16.1: ‘Belgic’ flint-tempered ware, cf. ‘Thanet Dry’ silty/sandware (fine/coarse), Monaghan fabric F2/1 (Monaghan 1987, 251). 129 sherds. c.1,140g.

Colour: mid or dark grey. Feel: soft and smooth where burnished, otherwise slightly rough. Inclusions: sparse, fine, sub-angular clear and colourless quartz, and sparse fine to medium (occasionally coarse) angular and sub-angular flint, set in a fine silty matrix; sparse to moderate fine white mica; occasional rounded ferruginous grains, up to c.1mm in diameter; occasional burnt-out organic inclusions; rare fine to very coarse grains of ?siltstone up to c.1.5mm. Surface treatment: exterior surfaces sometimes finely smoothed, but weathering at this site has rendered many sherds matt or rough. Manufacture: hand-made. Fracture: finely irregular.

IB2: ‘Belgic’ flint and sand-tempered ware (coarse). 31 sherds. c.331g.

R7: Reduced, probable Canterbury sandyware (fine) (Pollard 1995, 599). 1 sherd. c.1g.

R8: oxidised (orange) probable Canterbury sandyware (coarse) (Pollard 1995, 599). 2 sherds. c.16g.

R9.1: Oxidised (pink-buff) Canterbury sandyware (coarse) (Pollard 1995, 599). 6 sherds. c.86g.

R14: Black-burnished ware, fabric 2 (BB2), mostly of north Kent (Cooling and Cliffe) manufacture (Monaghan 1987, 246; Tomber and Dore 1998, 165-66). 8 sherds, c.199g.

R16: Fine Upchurch-type ware (reduced) (Tomber and Dore 1998, 168; see Monaghan, 1987, for discussion of all fine Upchurch-types). 231 sherds. c.843g.

R17.1: Fine Upchurch-type ware (orange). 91 sherds. c.484g.

R18.1: Fine Upchurch-type ware (purple-grey with white or cream slip). 53 sherds. c.282g.

R20: Lyon Ware. (Tomber and Dore 1998, 59). 28 sherds. c.23g.

R42: South Gaulish samian (Webster 1996, 13; Tomber and Dore 1998, 28). 13 sherds. c.69g.

R50: South Spanish Dressel 20 amphora = Peacock and Williams class 25 (Peacock and Williams 1986, 140, Tomber and Dore 1998, 84). 1 sherd. c.10g.

R62: ?Kent fabric 2 mortarium (Hartley 1982, 150-58). 1 sherd. c.330g.

R71: Misc. pink-buff fabrics (Pollard 1995, 601). 4 sherds. c.18g.

R73: Misc. reduced sand-tempered ware (coarse) (Pollard 1995, 702). 275 sherds. c.1,813g.

R74.1: Misc. oxidised (orange) sand-tempered ware (coarse); distinguished from R73 on the basis of colour. 6 sherds. c.20g.

R81: Eggshell terra nigra (Tomber and Dore 1998, 16). 9 sherds. c.26g.

R87: Gaulish white flagon ware, Rigby fabric WW2 (Tomber and Dore 1998, 23). 2 sherds. c.12g.

LR11: Nene-valley type colour-coated ware (Howe et al. 1980; Tomber and Dore 1998, 117–18). 1 sherd. c.4g.

The physical condition of the pottery is generally poor. Most sherds are small and heavily weathered. In many cases this weathering has resulted in the complete loss of the original surface finish, and as a consequence, it was difficult to assign a date to some of the smaller groups and individual sherds.

The assemblage contains a significant proportion of typically ‘Belgic’ fabric types, tempered with flint, grog, sand and shell (29.8 per cent by sherd count). Although there is also a significant quantity of romanized, wheel-thrown, reduced, R73 (26.2 per cent), recognizable forms in this fabric are largely limited to bead-rim and everted-rim jars.

A similar chronological trend is observable among the finewares. By far the most abundant fabrics are fine Upchurch-types, R16, 17, and 18, which together total 36 per cent of the assemblage by sherd count (21.4 per cent by weight). All of the identifiable forms date to either the second half of the first and possibly (but not necessarily) the early second century. The commonest forms are dishes, flagons and carinated beakers of Monaghan classes 7A, 1E1, and 2G. Girth beakers of class 2F are also present. There is a conspicuous absence of barbotine-dot decorated ‘poppyhead’ beakers of Monaghan class 2A. (Monaghan 1987).

Although there are only thirteen sherds of samian, they are all of Southern Gaulish type (R42), probably belonging to the second half of the first century. With the exception of a single sherd of LR11 (see below), all of the other imported fineware sherds, in fabrics BER 7 and 11, R20 and 81 also belong to the first century. The presence of such a diversity of early fineware types is striking, and particularly that of a roughcast-decorated cornice-rimmed beaker in fabric R20. This ware, dated to between c. ad 40-70, is not common in Kent, occurring principally on high status urban and villa sites (Pollard 1988, 37). Another unusual find is a stamped R81 base (Fig. 8, no. 18). Although the form is uncertain, it is likely to be a necked and carinated beaker, Holwerda forms 25-27 (Holwerda 1941). The ware is usually thought to post-date ad 50 in Britain (Rigby 1995, 647). The detail of the stamp itself is, unfortunately, entirely unintelligible and not illustrated.

The almost complete absence of products of the Canterbury pottery industry provides further evidence for the limited distribution of these wares in this area of Kent (Pollard 1987, 69-9).

Three groups of material have been selected for discussion. In terms of the quality and quantity of their ceramic content, the only significant pre-conquest or conquest-period deposits are ditch groups F102 and F103, and the most significant early Roman deposits are the two principal fills of ditch group F100, deposits [164] and [165]. Soil layers [123] and [124] are of particular stratigraphic importance, as they accumulated over F102 and F103. Representative forms from these deposits have been illustrated, and are supplemented by a number of types from context [156].

Ditches F102 and F103, contexts [180, 181, 184, 194, 226, 228, 232 and 233] Table 1; Fig. 7, nos. 1-3

These deposits yielded pottery which would appear to be entirely of pre-conquest or conquest-period date. There is a complete absence of pottery displaying distinctively ‘Romanized’ features, or characteristics such as wheel-throwing. Most of this material is tempered principally with flint or grog.

Most of the flint-tempered sherds have a very fine, grey, silty clay matrix. This distinctive fabric, BER16.1, is tempered with sparse fine to medium flint, but is otherwise, macroscopically, remarkably similar to CAT fabric BER16 (colloquially known as ‘Thanet Dry’). The distribution of the latter is almost entirely confined to the extreme east of the county, being commonly found in Thanet in contexts dating to the mid first to early second century (Nigel Macpherson-Grant, pers. comm.). It is only very occasionally found in Canterbury. BER16.1 constitutes approximately 8.1 per cent of the total site assemblage at Bredgar, suggesting the probability of local production, possibly in the area of the Upchurch Marshes, whose southernmost edge lies little more than 7km to the north of the village.

Monaghan (1987, 215-16) has discussed the production of flint-tempered wares on the Upchurch Marshes, which began in the early first century ad, and declined after the Roman conquest, until its final demise around ad 70-80. The principal forms produced included bead-rim jars of Monaghan class 3E and facet-rimmed jars of class 3G, both of which are represented in BER16.1 at Bredgar (see for example Fig 10, nos. 1-3). BER16.1 would appear, on the basis of his published fabric description, to be very similar to Monaghan fabric F2/1 (Monaghan 1987, 251).

Pollard has also discussed the production of flint-tempered wares on the Marshes. Its distribution appears to be highly localised, occurring no more than 15km away (Pollard 1988, 46). Similar forms and fabrics were recovered from first-century deposits at Wises Lane, Borden, which lies little more than 4km to the north of Bredgar (M. Lyne, in Hammond et al. 2003, 72-77).

Ditch F100, contexts [164, 165] Table 2; Fig. 8, nos. 4-18

In addition to small quantities of a range of ‘Belgic’ fabrics, this group yielded much larger amounts of post-conquest material. The principal coarseware fabric is wheel-thrown R73 (19.93 per cent by sherd count). Forms present in this ware are almost all closed, the rim-sherds representing mainly out-turned or flange-rim vessels (Fig. 8, nos. 16-17). Fragments of two lids were also identified. There is a complete absence of R14, which occurs widely in Kent from the Hadrianic period onwards (Pollard 1988, 88-90). The commonest fineware is R16 (38.85 per cent by sherd count, 29.76 per cent by weight). The most prominent forms present in this ware are beakers of cordoned or carinated type (Fig. 8, no. 13) and shallow dishes of Monaghan type 7A. There are also sherds from flaring ring-necked flagons of class1E1. These types date to the later first or early second century ad (Monaghan 1987). All of the other finewares, however, fabrics R20, 42, 81 and 87, are usually assigned a first-century date. Most of these first-century fineware sherds occur in the lower fill [165], but the significance of this is difficult to assess, given the small vessel quantities involved, and there are a number of vessel links between [164] and [165]. On balance the evidence would seem to suggest that very little, if any, of the assemblage need be any later than first century in date.

Soil layers 123 and 124 Table 3; Fig. 9, nos. 19-28

The pottery in these layers is significantly less abundant than in Ditch F100. It is broadly similar in character, but lacks several of the early finewares found in the latter, and contains proportionately rather less B2. The strongest dating evidence is provided by the range of late first- or early second-century R16. The rim of an R9 flanged-rim mortarium in fabric R9 might be later, but not necessarily. It is one of only two mortarium sherds recovered from the site.

There is little ceramic evidence for occupation of the site after the first century, consisting of only eight sherds of R14 (0.76 per cent of the site assemblage). Six of these, representing two simple-rimmed dog-dishes and a bead and flange-rim pie-dish (Fig. 10, no. 29), came from context [156], the upper fill of ditch F100. R14 is commonly found in Kent in deposits dating to the mid second century and later.

Disregarding a small number of deposits yielding only a few sherds of chron-ologically undiagnostic sandyware sherds, only one other feature produced pottery which could be confidently assigned a later Roman date. Context [296], a fill of pit 297, contained a sherd of LR11. It probably represents a flask or flagon and dates from the mid third to fourth century, although it might be as early as the late second.

Ceramic Building Material by Louise Harrison

A total of 371 fragments weighing c.28kg were recovered during the excavation, the majority of which was found in association with later material, or was considered too small or fragmentary to warrant further analysis. The remaining assemblage available for analysis, although fragmentary, is well fired, fresh and generally in an unabraded and good condition. This has been sorted by tile type and by fabric type and consists of 42 fragments of brick and tile weighing 8.395kg (Table ⁴).

The assemblage has been studied with the use of a 10x binocular microscope; the fabric types are as follows:

Fabric 1: a fine sandy fabric with a scatter of small sized (under 0.5mm) quartz grains. There are few other inclusions apart from very occasional larger sized quartz grains and calcareous fragments. This fabric is commonly found in Canterbury and surrounding areas and is very similar to material excavated from tile kilns at Whitehall Gardens (Jenkins 1956) and St Stephen’s (Jenkins 1960). However, considering the location of the site and the fabric’s condition, the material is much more likely to have been produced locally.

Fabric 8: is recognised mainly by its colour which can vary from a pale orange to a white/cream colour. The fabric’s matrix contains a moderate quantity of small sized quartz grains (up to 0.5mm), some iron oxides, and occasionally lenses or pellets of red clay. Additionally, the moulding sand present on the back of the tile usually consists of rose coloured quartz grains (up to 1mm). This fabric is thought to have been produced at Eccles and dates in London from the early first century through to the later second century (Betts 1992).

Fabric 11: varied in colour from a pale orange to a red colour. It has a fine, sandy consistency; its distinguishing feature is that of common silty lenses and ‘swirls’ visible in the fabric’s matrix. This fabric is dated in London to ad 100-120 and is presumably an import into Kent.

Although the assemblage is small, it is clear from Table 4 that the most common fabric type is fabric 8, representing 61.91 per cent of the assemblage by quantity.

Brick: only one brick fragment with one corner 38mm thick, presumably a bessalis was present in the assemblage. No diagnostic marks such as signature marks or tally marks were present.

Flue Tile: these were all fragmentary but all exhibit keying consisting of wavy vertical combing made with a comb with 5-9 teeth, the width of the stroke varying from 15-28mm.

Imbrex: these were all fragmentary with no complete or virtually complete tiles surviving.

Tegulae: these were the most commonly represented type of ceramic building material in the assemblage. Although they are all fragmentary, they could virtually all (except one) be assigned a flange profile type (see Table 4).

Miscellaneous tile: tile that bore no characteristic features such as flanges (in the case of tegulae) were classed as miscellaneous tile. All three of these fragments bore incomplete signature marks which appear to be signature type 1, which consists of a one grooved semicircle.

Although the size of the assemblage from Bredgar School is small and consists of fragmentary brick and tile, it is well fired, and of a fresh, unabraded condition. This would suggest little post-depositional movement, probably deriving from building F114 or other buildings nearby. Over half of the assemblage consists of roofing tile (52.4 per cent by quantity), while all the other tile types (grouped together) represent 47.6 per cent of the assemblage (by quantity). The presence of the roofing tile, along with flue tile and tesserae suggests that the assemblage was from a substantial building containing a hypocaust system and tessellated floors. The majority of the dateable brick and tile consists of fabric 8, which can be dated to the earlier first century ad.

The Small Finds by Lynne Bevan

Prehistoric Worked Flint: thirty-five items of humanly-worked flint were recovered, weighing a total of 462g. These comprised three retouched flakes (SF 247, Context 232 x 1 and SF 250, Context 199 x 2), a core rejuvenation flake from a blade core (SF 245, Context 153, Phase 8) and 31 un-retouched flakes (Table ⁵). Although it had a fresh, sharp appearance, the flint was of a poor quality, light to medium brown and grey in colour, and was derived from secondary sources, probably from local boulder clays or river gravels. A high incidence of cortical inclusions and hinge fractures among the flakes attests to the difficulty of working this unpredictable material. Although certain observations can be made about the assemblage, the small size of the collection precludes fuller analysis.

The flint appears to be residual in nature, given its association with Late Iron Age, Roman and modern cultural material (Boden 2004). However, some of the flint appears to have been derived from the same small nodules, implying that prehistoric core reduction processes took place on, or in the vicinity of, the site and that the flint had not moved very far from its place of original deposition. The small collections of flint from Contexts 155, 193 and 227 appeared to have been derived from the same small nodules, although refitting of any of the nodules was not possible.

Close dating of the flints was not possible, given the paucity of chronologically-diagnostic items, and for this reason, as well as its circumstances of discovery, the flint cannot be regarded as a cohesive assemblage. However, a broadly Neolithic date is probable for one core rejuvenation flake from a core with small blade detachments. The general morphology of the flakes is suggestive of a later prehistoric date for the collection (Pitts 1978, 17-37) which was probably produced at various times during the later Neolithic to Bronze Age periods. An Iron Age date may also be possible for some of the collection, based upon recent work on later prehistoric flint assemblages (e.g. Young and Humphrey 1999, 231-242).

Later Finds: there were no small finds relating to the Late Iron Age or Early Roman occupation of the site. Amongst the post-medieval finds were two fragments from slate writing boards and a fragment of slate pencil.

The use of slate pencils may date to before 1600, examples of which date have been recovered from archaeological contexts in Norwich (Margeson 1993, fig. 38:442-444, 70-71), although they were not in common use until the eighteenth century (Rhodes 1984, Nos. 116a, b). Slate writing equipment was being mass-produced and was in general use in Victorian schools by the late eighteenth to early nineteenth centuries (Sturt 1967; Rhodes 1984, 120-123; May 1994, 24).

The slate pencil from Bredgar Primary School has a machine-turned point, rather than a facetted end, unlike the slate pencils from Norwich (Margeson 1993, Fig. 38: 442-444). The Bredgar pencil, together with the slate boards was probably contemporary with the earlier use of the school during the latter half of the nineteenth century, before slate pencils were replaced by chalks.

Environmental Samples by Enid P. Allison

Six bulk samples of soil from the fills of selected ditches and pits were processed to recover plant and animal remains. Flotation was carried out to recover a ‘washover’ on 0.5mm mesh, and residues on nested 2mm and 1mm meshes. The soils at Bredgar appear to have been unsuitable for the preservation of organic material unless it was carbonized (converted to elemental carbon by burning in a reducing atmosphere) or mineralized (which most commonly occurs in deposits containing cess). Preservation of bones and shell was also poor.

Sparse amounts of material derived from human occupation of the site were recovered. The fill of ditch segment S41 (ditch F102) yielded a small assemblage of rather poorly preserved cereal remains and a charred fruit pip. Uncharred blackberry seeds (Rubus) were common but may represent brambles growing beside the ditch rather than the remains of food. The same sample also produced poorly preserved fragments of mussel shell (Mytilus edulis) suggesting an unknown degree of exploitation of rocky coastal locations as a source of food.

The fill of pit F54 contained poorly preserved mineralized fragments and a small mineralized fruit. A comparison of these remains with mineralized material from other from sites with similar soil types, suggests that this deposit is likely to have contained cess. Two other pits sampled produced between them a small quantity of charcoal and a single charred cereal grain. There were no indications as to their use.

discussion

The siting of the excavation was fortuitous in that although only a very small area was investigated, being limited to the footprint of the proposed development, the exposed features and deposits illustrate a fairly rapid sequence of events dating to the later Iron Age, Belgic and early Roman periods. The size and shape of the enclosure represented by ditches F102 and F103 remains unknown, although farmsteads and settlements of this later Iron-Age period typically, but by no means universally, consisted of an enclosed group of curvilinear or circular timber buildings (round-houses), four-post granaries, corn drying racks and other post-built structures and were usually associated with field systems (Dark and Dark 1997, 13).

Although in some areas of Britain the Roman invasion and occupation was systematically resisted, the establishment of Roman ways of life was surprisingly rapid. This ‘Romanisation’ included improvements to agricultural practices, intensification of cultivation and, in a few instances, the development of the existing native farmsteads and villages into villa estates. Recent excavations at Broad Oak to the west of Canterbury (Boden 2003, 24-26) have identified the site of a late Iron Age ‘Belgic’ settlement which clearly shows continuity of occupation from the late first century bc through to at least the fourth century. At Broad Oak the remains of post-built native structures, drainage ditches, small enclosures and pits were overlain by a succession of Roman features, which included metalled surfaces, drainage ditches and masonry structures including a small apse-ended bath-house. We see a native settlement at Broad Oak, with probably earlier origins, developing into a farm or villa estate of some size.

The presence of the very early post-conquest pottery associated with the possible timber structure and the slightly later and more substantial flint built building F114, suggests that here, too, in Bredgar a native farm or farmstead continued and developed although probably along Roman lines with the occupants enjoying a higher standard of living. The possibility that further masonry remains may be located in the immediate area has been suggested, in the most part by verbal reports of ‘wide flint wall foundations’ being uncovered during drainage work to the north of the nearby village hall, and in the rear garden of the neighbouring property, during the summer of 1976. No records exist of these discoveries, although a large pile of flint nodules still lies to the rear of the village hall, examination of which by the writer proved inconclusive. The location of the Iron-Age and Roman remains in this area may suggest that the present High Street through the village of Bredgar and possibly Bexon Lane to the east and Gore Road to the west represents a ‘crossroads’ of ancient trackways traversing the North Downs.

acknowledgements

The writer would like to thank Diocesan Architectural Services Ltd who commissioned the work; the staff and pupils of Bredgar C. of E. (Aided) Primary School whose enthusiasm and interest was always welcome; the excavation team and volunteer Mr Don Rudd, who often put up with terrible weather and site conditions.

bibliography

Betts, I.M., 1992, ‘Roman Tile from Eccles, Kent, found at Colchester’, in Excav-ations at Culver Street, The Gilberd School and other sites in Colchester 1971-85.

Boden, D., 2003, ‘Shelford Farm Estate, Broad Oak’, Canterbury’s Archaeology 2001-2002, 24-26.

Boden, D., 2004, Archaeological Excavation at Bredgar Primary School, Kent, unpublished CAT stratigraphic report.

Carson, R.A.G., 1959, ‘The Bredgar treasure of Roman coins’, Numismatic Chronicle, 6th ser., 19, 17-22.

Dark, K. and Dark, P., 1997, The Landscape of Roman Britain, Stroud, Gloucs.

Frere, S.S., 1974, Britannia, London.

Hammond, S., Preston, S. and Taylor, A., 2003, A First Century AD Field System and Cremation Cemetery at Wises Lane, Borden, Sittingbourne, Kent, Thames Valley Archaeological Services; Monograph 4.

Hartley, K.F., 1982, ‘The Mortaria’, in P. Bennett, S.S. Frere and S. Stow, Excav-ations at Canterbury Castle, The Archaeology of Canterbury, Vol. I, 150-8.

Helm, R., 2003, Interim report on Archaeological strip and map excavation on land at Bredgar Primary School, unpublished CAT client report number 2003/105.

Holwerda, J.H., 1941, De Belgische Waar in Nijmegen, Nijmegen.

Howe, M.D., Perrin, J.R. and Mackreth, D.F., 1980, Roman Pottery from the Nene Valley: A Guide, Peterborough City Museum Occ. Paper no. 2, Peterborough.

Jenkins, F., 1956, ‘A Roman Tilery and Two Pottery Kilns at Durovernum (Canterbury)’, Antiquaries Journal, xxxvi, 40-56.

Jenkins, F., 1960, ‘Two pottery kilns and a tilery of the Roman period at Cant-erbury (Durovernum Cantiacorum)’, Archaeologia Cantiana, lxxiv, 151-61.

Margeson, S., 1993, Norwich Households: Medieval and Post-Medieval Finds from Norwich Survey Excavations 1971-78, East Anglian Archaeology 58, The Norwich Survey/Norfolk Museums Service, Norwich.

May, T., 1994, The Victorian Schoolroom, Shire Publications Ltd, Princes Risborough.

Monaghan, J., 1987, Upchurch and Thameside Roman Pottery, British Archaeological Reports (British Series), No. 173, Oxford.

Orton, C.J., 1975, ‘Quantitative pottery studies: some progress, problems and prospects’, Science and Archaeology, 16.

Peacock, D.P.S. and Williams, D.F., 1986, Amphorae and the Roman economy: an introductory guide, London.

Pitts, M., 1978, ‘On the Shape of Waste Flakes as an Index of Technological Change in Lithic Industries’, Journal of Archaeological Sciences 5, 17-37.

Pollard, R.J., 1987, ‘The Other Pottery’, in G.W. Meates, The Lullingstone Roman Villa, Volume II – The Wall Paintings and Finds, Maidstone, Kent Archaeological Society Monograph, 3, 164-305.

Pollard, R.J., 1988, The Roman Pottery of Kent (Maidstone).

Pollard, R.J., 1995, ‘Pottery from the Augustan to the Vespasianic years’, in K. Blockley et al., Excavations in the Marlowe Car Park and Surrounding Areas, The Archaeology of Canterbury, Vol. V, 585-624.

Rhodes, M. 1984, ‘Slate Writing Implements’, in Thompson, A., Grew, F. and Schofield, J., ‘Excavations at Aldgate’, Post-Medieval Archaeology, 18, 120-122.

Rigby, V., 1995, ‘Early Gaulish and Rhenish Imports’, in K. Blockley et al., Excavations in the Marlowe Car Park and Surrounding Areas, The Archaeology of Canterbury V, Canterbury, 640-70.

Sturt, M., 1967, The Education of the People: A History of Primary Education in England and Wales, Routledge and Kegan Paul, London.

Thompson, I., 1982, Grog-tempered ‘Belgic’ pottery of South-Eastern England, British Archaeological Reports (British Series), No. 108, Oxford.

Tomber, M. and Dore, J., 1998, The National Roman Fabric Reference Collection, MoLAS Monograph 2.

Ward, A., 1996, ‘An Archaeological Evaluation at Travers Gardens, Bredgar’, unpublished CAT client report.

Webster, P., with contributions by Dannell, G.B., 1996, Roman samian pottery in Britain, CBA Practical handbook in Archaeology, 13.

Willson, J., 2002, ‘An Archaeological Watching Brief at The Old Post Office, The Street, Bredgar, nr. Sittingbourne, Kent’, unpublished CAT client report (arch. no. 1637).

Willson, J. 2003, ‘A Brief Interim Report on the results of an Archaeological Watching Brief of Two Geotechnical Test Pits at Bredgar Primary School’, unpublished CAT client report.

Young, R. and Humphrey, J., 1999, ‘Flint Use in England after the Bronze Age: Time for a Re-Evaluation?’, Proceedings of the Prehistoric Society, 65, 231-242.

Fig. 1 Site Location Plan (1:2500).

Fig. 2 Site Plan (1:400).

Fig. 3 Prehistoric Phases 1-3 (1:150).

Fig. 4 Phase 4, Initial Roman Occupation (1:150).

Fig. 5 Phase 6, Later Roman Occupation (1:150).

Fig. 6 Phase 7 Features and Deposits.

Fig. 7 ‘Belgic’ Pottery from Ditches F102 and F103, nos. 1-3 : BER 16.1 (1:4).

TABLE 1. QUANTIFICATION OF POTTERY FABRICS

FROM DITCHES F102 AND F103 (PHASES 2 AND 3)

Fabric	F102	F103	Fabric Totals
IB2	9/116/0	3/35/0	12/151/0
B1	1/4/0	2/7/0	3/11/0
B2	2/13/0		2/13/0
B3	2/3/0	2/19/0	4/22/0
BER16.1	16/134/8.4	10/115/14.88	17/164/23.28
Totals	31/270/8.4	8/144/14.88	38/361/23.28

For each fabric, the following information is given: sherd count/weight in grams/EVEs value.

Fig. 8 First-century Roman pottery from Ditch F100, No. 4: B1. Nos. 5-7: B2. No. 8: B8. Nos. 9-13: R16. No. 14: R17. No. 15: R42. Nos. 16, 17: R73. No. 18: R81. (1:4).

TABLE 2. QUANTIFICATION OF POTTERY FABRICS

FROM THE PRINCIPAL FILLS OF DITCH F100 (PHASE 6), CONTEXTS [164, 165]

Fabric	Context 164	Context 165	Fabric Totals
B1	1/27/11.76%		1/27/11.76%
B2	14/130/29.68%	5/31/16.24%	19/161/45.92%
B6	2/37/0		2/37/0
B8	1/16/6.44%		1/16/6.44%
BER11	2/7/0		2/7/0
BER16.1	17/206/0	3/51/0	20/257/0
R9	2/9/0	2/12/0	4/21/0
R16	57/339/71.12%	58/178/24.36%	115/517/95.48%
R17	2/25/39.20%		2/25/39.20%
R18	14/59/0	4/17/0	18/76/0
R20		28/23/0	28/23/0
R42	5/39/0	5/23/29.68%	10/62/29.68%
R50	1/10/0		1/10/0
R73	46/379/6.16%	13/48/13.76%	59/427/19.60%
R74	3/12/0		3/12/0
R81		9/26/0	9/26/0
R87		2/12/0	2/12/0
Totals	167/1295/164.36%	129/421/84.2%

For each fabric, the following information is given: sherd count/weight in grams/EVEs value.

Fig. 9 First-century ‘Belgic’ and Roman pottery from Ditch F100, soil layers 123 and 124: No. 19: B6. No. 20: B8. Nos. 21-23: BER16.1. No. 24: R9. Nos. 25–27: R16. No. 28: R73. (1:4).

TABLE 3. QUANTIFICATION OF POTTERY FABRICS

FROM SOIL LAYERS 123 AND 124 (PHASE 7)

Fabric	Layer 123	Layer 124	Fabric Totals
IB2		3/37/0	3/37/0
B1		2/16/0	2/16/0
B2		5/251/0	2/251/0
B3		1/5/0	1/5/0
B6	4/7/0	9/107/0	13/178/0
B8	1/14/12.88%	9/105/40.88%	10/119/53.76%
BER11		1/13/0	1/13/0
BER16.1	13/85/19.32%	19/184/26.88%	32/269/46.20%
R8	1/3/0	8/36/35.12%	9/39/35.12%
R9	1/64/13.44%		1/64/13.44%
R16	17/38/0	40/123/41.20%	57/16/41.20%
R17	1/2/0	5/27/0	6/29/0
R73	45/217/15.68%	34/221/9.80%	79/238/25.48%
Totals	83/494/61.32%	136/1125/153.88%

For each fabric, the following information is given: sherd count/weight in grams/EVEs value.

Fig. 10 Second-century Roman pottery from Ditch F100, context 156: this material has been selected to supplement the range of forms illustrated from the Groups discussed above. No. 29: R14. Nos. 30-31: R73. (1:4).

TABLE 4: BRICK AND TILE

Cxt. No.	Fab. No.	Form	Flange	Weight (gm)	Tile No.	Comments
102	8	tessera		10	36
103	1	flue		395	26	combed with wavy stoke, teeth 5, width 15mm, worn and abraded
110	8	tessera		15	34
131	1	brick		1,400	8
156	8	imbrex		240	5
156	8	tegula	5	495	10
156	1	imbrex		240	18
156	1	imbrex		185	19
156	1	flue		130	30	with corner, combed with wavy vertical comb, 9 teeth, width 28mm.
156	8	tessera		110	31	x 7
164	1	tegula	2	300	1
164	1	tegula	12	230	2
164	11	tegula	7	700	3
164	8	tegula	10	185	4
164	8	imbrex		320	9
164	8	imbrex		250	13
164	8	tegula	11	540	17
164	8	tessera		20	33
173	1	imbrex		180	15
173	1	tile		415	20	with signature mark
173	1	tegula	1	130	27
173	8	tegula	11	195	28
173	8	tegula	11	70	29
179	8	tegula	1	120	6
179	8	tegula	16	110	7
179	8	tile		60	21	with signature mark
179	1	tile		285	25	with signature mark
218	1	flue		115	16	combed with wavy stoke, 6? teeth, width 17mm, faint and unclear
218	8	tessera		10	35
257	8	tessera		10	32
294	8	tegula		220	23	flange unclear
294	8	tegula		175	24
Cxt. No.	Fab. No	Form	Flange	Weight (gm)	Tile No.	Comments
330	8	tegula		75	11
330	1	imbrex		185	22
u/s	1	imbrex		175	12
u/s	1	flue		100	14	unclear as to combed/roller stamped?

TABLE 4 (cont): BRICK AND TILE

TABLE 5. THE FLINT ASSEMBLAGE

Cxt	Description	No.	Wt(g)	Comments
105	Fill of ditch F102 (Recov-ered during the evaluation).	9	199	8 flakes plus 1 possible core
110	Topsoil and overburden	1	27	Un-retouched flake
124	Occupation deposit.	1	6	Un-retouched flake
153	Fill of modern cultivation feature F24	1	9	Core rejuvenation flake
155	Plough soil deposit	3	22	Un-retouched flakes
164	Upper fill of ditch F100.	1	11	Un-retouched flake
179	Fill of ditch F100.	1	7	Un-retouched flake
180	Upper fill of ditch F102.	1	10	Un-retouched flake
181	Primary fill of ditch F103.	3	44	Un-retouched flakes
183	Primary fill of ditch F102.	1	7	Un-retouched flake
193	Plough soil deposit part of G117	4	32	Un-retouched flakes
198	Fill of ditch F102.	2	15	Un-retouched flakes
199	Fill of ditch F103.	2	23	Retouched flakes
221	Fill of ditch F102.	1	15	Un-retouched flake
225	Fill of ditch F102.	4	75	Un-retouched flake
227	Fill of ditch F102.	6	107	Un-retouched flakes
232	Fill of ditch F103.	2	77	Retouched flake; flake
263	Fill of ditch F104.	1	3	Un-retouched flake