Excavations and other Archaeological Investigations on the Thanet Way 1990-95

37 Excavations and other archaeological investigations on the Thanet Way, 1990-95 keith parfitt, tim allen and jon rady with nigel macpherson-grant Early in 1990 Kent County Council finalised plans for road improvements to the A299 Thanet Way (Phases 2-4). This involved the construction of a 6 mile (9.5km) length of dual-carriageway running eastwards from Seasalter to link previously upgraded sections of the trunk road. Canterbury Archaeological Trust was commissioned to undertake a survey of the proposed route in order to document and then evaluate any archaeological remains present. The course of the new road lay south of the existing A299 and north of the Forest of Blean (Fig. 1). The route ran across the North Kent coastal plain extending eastwards from its junction with the existing Thanet Way at Seasalter across Wraik Hill and Clapham Hill. It then descended gradually to pass south of Radfall Corner and north of Shrub Hill before crossing level terrain towards the Share and Coulter public house, onwards to Strode Farm, to rejoin the existing ro ad at Eddington. This open country, north of the wooded uplands of the Blean, consists of a gravel-capped ridge of Eocene London Clay which extends from near Boughton Street in the west, almost as far as the marshlands of the former Wantsum Channel in the east. Most of the route traversed this London Clay formation. A number of small streams (including Swalecliffe Brook, West Brook and Plenty Brook) flow south across the area from the Blean and in these stream valleys there are some overlying deposits of Head or Brickearth. As a consequence of the geology, most of the land is heavy and poorly drained and, unlike the Chalklands to the southeast, not conducive to the formation of cropmarks. Until recently it had been assumed that these heavy clay soils had remained largely wooded and unoccupied into medieval times and that the Blean Forest extended further north across the coastal plain than it does today (Everitt 1986, 29-30). Place-name evidence may call this into question. Dargate, Radfall Gate, Broomfield Gate and Blean Gate along the present day northern edge of the forest suggest that the boundary has not changed a great deal since medieval times1 and early cartographic evidence supports this view (Symondson’s map of Kent dated 1596 designates only a limited area west of the Canterbury-Whitstable road (now the A290) as woodland, whilst the first edition Ordnance Survey of 1819 shows the northern woodland boundary almost identical to that of today). The route ran close to several isolated farms and hamlets of which Eddington (first recorded in 1466), Strode Farm (first recorded 1240), Bullockstone Farm (first recorded 1348) and Chestfield Farm (first recorded 1242), are mentioned in KEITH PARFITT, TIM ALLEN AND JON RADY 38 later medieval documents indicating that the area was occupied from at least the thirteenth century (Glover 1982). Today, much of the land is arable, interspersed with some grassland (including two golf courses) and a few small areas of scrub and woodland. Most of the modern fields are of moderate to large size, usually bounded by drainage ditches, reflecting the ill-drained nature of the land. Systematic field-walking of the route was undertaken in November 1990. Light scatters of archaeological material were present on all the arable land examined. This material mainly consisted of medieval and post-medieval roof tile and pottery with some prehistoric struck and calcined flints. Concentrations of calcined flints were noted in two separate areas near Radfall Corner. These were subsequently excavated as Sites 7 and 8 (see below). Some material of Roman date was also recovered. Towards the eastern end of the route, in the area of the Plenty Brook and Strode Farm, a number of pieces of Roman tile were discovered (Fig. 1, Sites 11, 12 and 13). At Site 11, adjacent to Owl’s Hatch Road, the fragments were associated with a surface scatter of Roman pottery and subsequent excavation uncovered the site of a Romano-British farmstead. A single rim sherd from a Roman mortarium, together with a few Fig. 1 Location of excavated sites: Site 7 (South Street); Sit e 8 (Radfall Corner); Site 11 (Owl’s Hatch Road). Evaluated sites: Site 12 (Plenty Brook), 13 and 14 (north of Strode Farm) EXCAVATIONS AND ARCHAEOLOGICAL INVESTIGATIONS ON THANET WAY 1990-95 39 sherds of Iron Age flint-tempered ware were recovered during field-walking to the north-east of Strode Farm (John Cotter, pers. comm.; Site 14). A few more flinttempered sherds were collected on Herne Bay Golf Course in 1988 (TR 174 669; Wes McLachlan, pers. comm.). Further to the west, near South Street (TR 1230 6480) and just 400m north of the road route, several fragments of Roman tile were recovered from the levelled embankment of the Canterbury-Whitstable railway. These almost certainly derived from a nearby Roman building examined in 1962 which appeared to have formed part of a more extensive building complex, the bulk of which was probably destroyed when the railway was constructed c.1830 (Jenkins 1962). The medieval and post-medieval roof-tile and pottery recorded by field-walking was nowhere sufficiently concentrated to represent structures, though kilns existed in Clowes Wood a short distance to the south (Millard 1968) and others closer to Canterbury at Tyler Hill (Cotter 1997, 63-8; Willson 2003, 48-50 and fig. 15). The collected material was probably the result of manuring of fields, reinforcing the view that at least part of the area was under the plough by the thirteenth century. The desk-top research and field-walking identified fourteen possible archaeological sites along the line of the new road (Parfitt and Allen 1990). After examination by machine-cut evaluation trenches some of these were eliminated (see below). Work at five sites (7, 8, 11, 13 and 14) revealed traces of sub-surface remains and three of these (7, 8 and 11) were subject to full excavation before road construction began. Near Radfall Corner, the adjacent Sites 7 and 8 proved to be of late Bronze Age or earliest Iron Age date, whilst Site 11 off Owl’s Hatch Road formed part of a Romano-British farmstead, with evidence for some medieval occupation. At Sites 13 and 14 located towards the eastern end of the route, trenching unexpectedly produced evidence for Anglo-Saxon activity. Site 7, South Street Site 7 was located some 300m south of South Street at TR 1319 6460. It was identified by an extensive area of scorched flints exposed by ploughing. Subsequent test-trenching revealed a single cut feature which contained pottery, probably of earliest Iron Age date. The site lay over weathered London Clay at an altitude of approximately 35.5m aod. The excavated area covered approximately 3,600m2. An extensive concentration of features located in the north and west of the site was interpreted as the remains of a hilltop settlement, probably part of a farmstead (Fig. 2). Abundant ceramic material dated the settlement to the Earliest Iron Age (c.950/850 bc). A radiocarbon date might intimate a somewhat earlier provenance, but was derived from a fragment of unidentified charcoal (see Discussion below). A smaller concentration of about twenty-five pits in the south of the site was predominantly of the same period, though a few yielded late Iron Age pottery, albeit, apart from one, in much smaller quantities. The site was severely truncated by ploughing and colluvial erosion so that most of the excavated features were very shallow (mean depth 0.15m). As a consequence it was difficult to establish relationships between intercutting features. However, a substantial quantity of diagnostic ceramic material was recovered. The majority of the features had comparable fills, ultimately derived from, and KEITH PARFITT, TIM ALLEN AND JON RADY 40 similar to, the London Clay subsoil. Out of 140 recorded fills, over 70% were a grey-brown clay, some darker or browner than others but generally similar. An additional small percentage had speckles of orange clay within the matrix. The remaining fills had a slightly siltier composition. Any significant variants are described below. Most of the features contained flint inclusions, either angular or rounded (presumably derived from natural gravels in the vicinity) and a significant number contained burnt or fire-crazed flint. Inclusions of small chalk lumps were common to about 40% of the features while just over 10% contained burnt clay fragments. The chalk is unlikely to derive from intrusive, later marling because that would be ubiquitous in the fills. A number of the contexts contained considerable concentrations of carbon or charcoal. Half the excavated contexts yielded ceramics in varying quantities, but other finds were negligible consisting of only a few scraps of animal bone, pieces of slag or vitrified material and ferruginous nodules. Fig. 2 The excavation at South Street (Site 7). EXCAVATIONS AND ARCHAEOLOGICAL INVESTIGATIONS ON THANET WAY 1990-95 41 Phase 1 – Earliest Iron Age settlement: most of the excavated post-holes were clustered in the north-western quadrant of the site in an area relatively devoid of other features although many pits surrounded them, particularly to the east. The post-holes form no clear pattern but truncation has almost certainly removed additional settings. Most of the features were between 0.19m and 0.38m in diameter and no more than 0.24m deep, usually much less. Their presence in an otherwise clear area is highly suggestive of a structure, possibly a round-house of approximately 9m diameter, and contemporaneity with the surrounding pits is suggested by the similar pottery in their fills (although only a few sherds from some) and the presence of burnt flint. A large linear feature [152], probably a ditch, was partially revealed in the extreme west of the site. This was of north-south alignment although only part of its eastern edge was exposed, leaving the possibility that it was a hollow-way. Partial excavation produced a small quantity (10 pieces) of prehistoric pottery suggesting the feature was contemporary with this phase of settlement. This is perhaps confirmed by the alignment of a group of pits to the east which appear to be broadly disposed on a parallel alignment, the area to the east of them being almost completely devoid of activity. The majority of the features were pits of varying shapes, sizes and depths, the largest probably representing intercutting pit complexes that were virtually impossible to separate in the field. They were circular to oval or subrectangular in shape, the smallest (such as pit 8 at about 0.8m diameter) potentially being large post-pits although these were generally scattered and showed no particular disposition. Two however [183 and 185], in the southern group were of very similar size (about 1m diameter) and could have been large post-pits representing a twopost structure. They were 2m apart, centre to centre. The largest of the group [pits 119 and 256] were between 3 and 4.1m wide and potentially represented clay quarries. The unusually shaped pit complex designated pit 150 proved to be an intercutting series of features, including individually isolated pits 288 and 290. Many of the pits were extremely shallow, depths ranging from 0.05m to about 0.6m at the most. There did not appear to be any significant differences between the two pit groups to north and south (apart from the presence of later ceramics in the latter; see below). The pits supplied the bulk of the artefactual evidence (feature 16 in particular yielding over 200 sherds) suggesting that they were primarily for the disposal of domestic waste. Three features [72, 86 and 127] were slightly different being larger (2.5m to 3m long and 1.5m to 2.2m wide) and subrectangular with shallow (<0.2m), flat-based profiles. They contained discrete layers of gravel at their base which included quantities of burnt flint. The features were disposed, about equidistant, in a line on the eastern side of the site. These may have been metalled working hollows as none of the deeper features elsewhere contained the gravel layer. The gravel could have acted as hardstanding, keeping the hollows relatively workable in wet conditions. It is not impossible that they represent sunken-featured structures of some sort, for which there is increasing evidence for this peri od (see below). Only a few other features need further mention. Many large fragments of daub or burnt clay were recovered from a small pit or large post-pit [146], about 0.5m in diameter. These provided strong evidence for a building on the site. One small KEITH PARFITT, TIM ALLEN AND JON RADY 42 pit [281] contained quantities of finely crushed flint comprising approximately 50% of the fill; it was sealed by a pit [138] containing earliest Iron Age pottery. The size of the flint fragments (10-50mm) was similar to that used for tempering the great majority of late Bronze Age/early Iron Age vessels, suggesting pottery manufacture close by. A few pits [88, 187, complex 150] and ditch 152 yielded forms of slag-like material, but in relatively small amounts (eight pieces, weighing 66g). Most of this proved to be vitrified hearth lining or other undiagnostic material and not indicative of metalworking. In addition, three other features [16, 28 and 288] contained a few ferruginous nodules of ironstone. This does not occur naturally in this area and was not observed on nearby excavations. It was originally thought that these nodules were raw material for ironworking, transported from the coastal plain where they are common, but there was no supportive evidence for this. One potential hearth or hearth-pit [105], probably domestic rather than industrial in nature was recorded on the eastern edge of the northern pit complex. This sub-oval shallow scoop, c.0.9m across and only 0.05m deep contained black/grey clay with abundant charcoal. Finally, on the north-western edge of the pit 150 complex, an oval pit [290], about 1m wide and 0.7m deep contained large fragments of an apparently near complete vessel of earliest Iron Age type as well as a significant assemblage of non-structural fired-clay objects or perforated slabs (the remainder of pit 150 complex also yielded a few of these). These objects, occasionally found on later Bronze Age or early Iron Age sites in the South-East, are of unknown function but possibly relate to a light industrial process or salt production (Macpherson-Grant 2007; Champion 2011, 230). Animal bones and charcoal were also recovered, but there was no evidence for burnt bone so the feature is unlikely to have been a cremation; another ritual function or significance might be surmised due to the varied and unusual ceramic content. Unidentified charcoal from this feature was submitted for radiocarbon dating and provided a date of c.1260-920 cal bc at 95.4% confidence (ref: OxA-8033; 2890+/-55 bp). The significance of this is discussed below. Environmental remains consisted of an appreciable sample of charred cereal chaff with a few grains, suggesting that processing of cereals had taken place on site. Phase 2 – Late Iron Age: subsequent occupation is indicated by late Iron Age ceramics, predominantly from features in the southern part of the site. The most significant of these was pit 209 which yielded nearly 60 sherds of this date and no other material. The nearby pit 207 yielded a much smaller quantity. At least five other pits or pit complexes contained late Iron Age pottery [16, 150, 195, 193, 203]. Apart from pit 16, these were all from the southern group. There did not seem to be any significant difference between any of these features and the other pits on the site and in some cases the later pot could be intrusive. Pit 16 for example contained only one later Iron Age sherd with much earlier material. Of the others, late Iron Age sherd counts were still low (generally one or two sherds), but there was often a similar proportion of earlier ceramics, which in some cases could be residual. The pit 150 complex provided the largest quantity of later pottery, but still only a small proportion compared to the Earliest Iron Age ceramics. However, it EXCAVATIONS AND ARCHAEOLOGICAL INVESTIGATIONS ON THANET WAY 1990-95 43 seems unlikely that all the pottery is intrusive, so at least some of these pits might be later in date and, with the more certainly dated pit 209, represent a later phase of occupation probably focussed further to the south. The later prehistoric pottery (Nigel Macpherson-Grant) (Figs 3-5) Some 8.536kg of later prehistoric pottery was categorised as follows: Earliest Iron Age (EIA) vessel ceramic: 1,266 sherds EIA light industrial ceramic: 29 sherds Late Iron Age (LIA) ‘Belgic’-style vessel ceramic: 74 sherds Apart from the contemporary LIA material from Pit 209 all pottery from this phase of activity is intrusive in earlier EIA features. No Roman pottery was recorded and the implication of this small LIA assemblage is that of an essentially pre- Conquest phase of occupation broadly datable to between c.75/50 bc-ad 50. The LIA material is not considered further, with only the EIA pottery discussed Fig. 3 South Street (Site 7): earliest Iron Age pottery. No 1 (Pit 150), Nos 2, 4, 6-7 (Pit 290), Nos 3 and 5 (Pit 288). Scale 1:4. KEITH PARFITT, TIM ALLEN AND JON RADY 44 here. The majority of the latter assemblage stems from shallow plough-reduced features and the material has suffered accordingly, both in terms of condition and size. Despite this, the value of the assemblage lies in its essential single-period purity and the presence from Pit 290 of both vessel and light industrial ceramic (perforated fired clay objects) accompanied by a radiocarbon date. In addition Pit 138 produced quantities of unfired but ready-to-use flint-tempered potting clay. The EIA assemblage’s recovered set of manufacturing trends has been analysed according to fabric, form, finishing and decorative characteristics. This data is held in the site archive together with a catalogue of available form and decoration parallels. Fig. 4 South Street (Site 7): earliest Iron Age pottery from other features. Scale 1:4. EXCAVATIONS AND ARCHAEOLOGICAL INVESTIGATIONS ON THANET WAY 1990-95 45 Eight vessel fabrics were recorded representing four or five obviously different clay sources. The latter are represented by fabrics with only silt-grade matrices (108 sherds), those with varying quantities of fine quartzsand (1121 sherds) or those with naturally-occurring organic and iron oxide inclusions (30 and 7 sherds respectively). In terms of deliberately added filler types, the assemblage is dominated by the use of moderate-fairly profuse crushed flint fillers with smaller quantities of mixed-temper fabrics, flint and grog (106), flint and organic (30) and solely quartzsand (1). Most finewares are visually dominated by filler grades >1-2mm, though some (e.g. 1) contain predominantly coarser grades >3mm. Coarsewares contain grades mostly averaging >2mm. The low count or absence of grades above 3mm indicates a preferred upper limit. Thin-walled vessels dominate both vessel classes and there is a distinct preference for 2-5mm thick walls; any thicker-walled examples fall close to 10mm. A single fineware sherd with a 10mm plus body wall is exceptional. Most fineware finishes are confined to rather poor-quality light-moderate horizontal burnishing externally with variably even surfaces. Open bowl forms usually have a lighter internal burnish, closed forms with burnished exteriors and inner surfaces simply wiped smooth. Coarseware finishes are typically fairly rough and characterised by surfaces with irregular light horizontal wiping. Some lower body sherds exhibit marked vertical or diagonal grit-drag scoriation. A review of firing trends indicates a predominance of reduced vessels, particularly amongst the finewares where there appears to be a preference for vessels with light grey through to black surfaces. Amongst the coarsewares there is greater variety suggesting a less rigid control of desired end-appearance, with a mixture of reduced, partially oxidised (predominantly brown or buff surfaces) and oxidised vessels. Of the finewares, 1, 2, 8 and 9 are bipartite bowls (Figs 3, 4). All are highshouldered although there is less certainty with 1 (8 is a larger and thicker-walled variant of the others). No. 10 is a simple hemispherical bowl. Bowls 1 and 8 both have simple above-shoulder linear decoration, the former with groups of diagonal grooves probably spaced around the body, the other with a band of probably three (possibly more) thin horizontal lines. The latter are incised – another undrawn example from a smaller bowl has multiple combed lines. Amongst the coarsewares, 3, 5 and 13-16 are high-shouldered bipartite jars. Most are from a range of mediumsmall diameter cooking-pots or jars, but the elegantly flaring thin-walled 3 and the angle-rimmed 16 are both from large-diameter storage jars. The two bases 7and 17 are probably from large storage vessels, and the skin of profuse basal grits on the latter is a typical, but not ubiquitous, manufacturing trait for the period. The marked inner-rim bevels of 14 and 16 (and less obviously on 5 and 15) are also characteristic. No 4 is probably from a slack-shouldered deep basin, 11-12 are hemispherical bowls and 6 is a straight-sided tub. The neatly cabled or incised decoration on 3, 13 and 16 (and the finger-tip impressions on a scrap from an applied horizontal cordon), are typical of regional EIA coarseware decoration. Perforated slabs: between seven and eight non-structural fired clay objects were recovered (Fig. 5), the majority from pit complex 150/290 (1-5) and one fragment from Pit 63 (6). No 1 is conventionally flint-tempered as are a number of other KEITH PARFITT, TIM ALLEN AND JON RADY 46 fragments, 2-5 are in a mixed-temper fabric (organic and sparse flint) and 6 is profusely organic-tempered. The structures of 2-5, particularly 5, are distinctly laminar. This trend is partly due to their organic content but also because some items (e.g. 5) appear to be made from several clay strips being tightly compressed together. All were fired in oxidising conditions – drab pink-buff or light buffbrown. The thicknesses of 1, 2 and 6, and the grooved edge of 2 are similar to many perforated slabs from contemporary South-East England assemblages. No. 4 however is different and its greater thickness, footed base and single perforation suggest a different function. The close-set holes of 3 and the tapering section of 5 are again atypical. A single C-14 date of 1260-920 cal bc (OxA-8033 at 95.4% confidence) was obtained from a sample of charcoal from the mid fill (289) of pit 290. This fill contained pots 2, 4, 6 and 7 and perforated slab 4. Since there appear to be no radical chronology-related differences in the material from the pit complex 150/288/290 Fig. 5 South Street (Site 7); earliest Iron Age perforated slabs. No 1 (Pit 288), Nos 3 and 5 (Pit 150), No 4 (Pit 290), No 6 (Pit 63). Scale 1:4. EXCAVATIONS AND ARCHAEOLOGICAL INVESTIGATIONS ON THANET WAY 1990-95 47 sequence this date can, initially, be applied to pots 3, 5 and perforated slab 1 from Pit 288 and pot 1 and perforated slabs 3 and 5 from Pit 150. For the same reason and in the same manner, it can be extended to the pottery from the rest of the site. However, there is an important qualification attached to this sample (s ee below). Typological comparisons indicated that the fineware bowls 9 and 10, the coarseware jars 14 and 16 with their internally bevelled rims and, to a lesser extent, the curving everted-rim coarseware jar 3, or minor variations of these, were relatively frequent occurrences in other broadly contemporary regional EIA assemblages. Similarly for the decorated fineware bowls 1 and 8. The number of these equations, particularly with the large assemblages from Highstead (Couldrey 2007, 118-21, figs 57-81) and Monkton Court Farm, Thanet (Macpherson-Grant 1994, 248-88, figs 5-22), initially indicated a c.900 or 850-600 bc date for South Street. However, the general similarity between the South Street bowl 9 and bowl 1 from the lower ditch fill of the Highstead enclosure B70 suggested that 9 might be closer to c.900 bc than radically later. B70 was typologically dated to Highstead’s Period 2 (c.900-600 bc). Its construction and first use was considered to be probably fairly early within that range, partly on the basis of the bowl, which had a fairly close parallel in Belgium dated to Halstatt B2/B3, c.900-700 bc (Couldrey 2007, 118 and fig 57, 1). At face value this dating is at variance with the above C-14 result. In addition, the sample obtained was from associated charcoal rather than the ideal sample type of burnt food residue from a vessel interior. Despite the caveat that the wood was old (although the sample suggests not radically older than the associated assemblage) the upper end-date is close to the typological dates quoted above. Equally, seen from the other end, South Street’s assemblage is totally unlike regional mid-late Bronze Age ceramic of 1350-1150 cal bc date, and only marginally similar to its currently recognised range of LBA forms, 1150-800 cal bc (Morris 2006, figs 3.2, 3.4a-b, 3.5a-b). It is worth stressing that the latter’s late end-date was qualified by the presence in the radiocarbon-dated assemblage of a few decorated elements of EIA character. These are atypical of the traditionally recognised material of LBA date so that, in radiocarbon years, the true end-date of the LBA is more likely to be around 1000 cal bc. Coincidentally, another EIA assemblage, from Tothill Road, Minster in Thanet has produced a single C-14 date of 1190-920 cal bc for soot deposited on the exterior of a large cooking-bowl. The latter is a more sharply-profiled version of the South Street coarseware bowl 4. Again, there is a possibility of old wood being used. The range of forms from Tothill is closely matched amongst those from South Street; a bevel-rimmed jar virtually identical to coarseware jar 14 came from the same pit group as the dated bowl (Macpherson-Grant forthcoming). Set alongside South Street, the scientific end-dates from both assemblages strongly enhance the likelihood of a date close to the beginning of the first millennium bc for the start of the EIA in the eastern part of the county. Again coincidentally, a report on excavations at Cliffs End in Thanet discusses a suite of C-14 dates obtained for a sequence of EIA pottery, this time more desirably from burnt food residues. This contains a useful set of vessel profiles that can be placed into 100-year phases (McInley et al. 2013). Of these only the first two apply here with, between 1000-900 cal bc predominantly plain vessels with just a KEITH PARFITT, TIM ALLEN AND JON RADY 48 few decorated coarsewares and, between 900-800 cal bc, both decorated fine and coarseware elements typical of regional EIA assemblages (Alistair Barclay, Matt Leivers, pers. comm.). Two aspects of the South Street assemblage are pertinent to this sequence. First, the fineware bowl 9 is closely similar to one from the first Cliffs End phase. Second, although the recovered range of decorated fineware bowls from South Street is rather sparse (only 1 and 5) its set of coarseware forms is mirrored at Tothill, which had a higher quantity of decorated finewares. This suggests that the currency of both settlements continued into the second Cliffs End phase. Summarising, both South Street and Tothill have C-14 end-dates of 920 cal bc, and within the first Cliffs End phase 1000-900 cal bc, implying that the currency of both settlements began during that phase. For South Street, this likelihood is reinforced by the equation with Cliffs End for fineware bowl 9. The close typological equations between South Street and Tothill strongly indicate that the occupation of both sites continued into the second Cliffs End phase, 900-800 cal bc. Estimates of settlement longevity for South Street are hindered by the small area sampled. However, an unquantified scan of the overall EIA assemblage suggests that the likely vessel number was quite high. More specifically, the cluster of intercutting pits represented by pit complex 150/288/290 is a notable exception within a zone of single pits. This suggests same-function re-use of a particular location over a period of time which in this case, since there are a number of perforated slab fragments from this complex, might be associated with ‘light industrial’ activity. These two points, potentially high vessel numbers and the repeated use of the same location for the discard of broken ‘industrial’ material, together indicate settlement stability and occupation over at least one to two, possibly three, generations, a period of c. 50-75 years (if estimated at approximately 25 years per generation). On the basis of the above evidence, all reasonable caveats accepted, this period is likely to have occurred between c.950 and 850 bc. Site 8, Radfall Corner This site was identified during field-walking when a scatter of calcined flints was recorded in a field a short distance south-east of Radfall Corner, Chestfield (TR 1340 6480). Located on London Clay at an elevation of about 25m aod on gently sloping ground at the foot of Shrub Hill, three evaluation trenches were cut which revealed traces of a metalled trackway apparently dating to the prehistoric period. Consequently, an area measuring some 900m2 was fully excavated before roadconstruction began. Interestingly there was no evidence of colluvial downwash on this low-lying site, which might suggest that Shrub Hill, whose summit is still partly wooded, was not tilled in antiquity, though severe truncation, probably a result of plough action, was evident on site. The mean depth of the excavated features was approximately 0.15m. The main excavated features were two intercutting prehistoric trackways and an adjacent group of possibly contemporary post-holes (Fig. 6). The associated ceramic assemblage all dated to the late Bronze to early Iron Age, c.1000/900-600 bc.A forked linear feature [27 and 43] of possibly two phases extended across most EXCAVATIONS AND ARCHAEOLOGICAL INVESTIGATIONS ON THANET WAY 1990-95 49 of the site on a north-west/south-east alignment. Although the relationship between the two features was difficult to ascertain, the main near-straight alignment [43] was possibly the later. Linear feature 27, curved away to the north-west. These relatively wide and shallow features with broad U-shaped profiles were between 1.75m and 2.6m wide, 0.08 to 0.15m deep and only partly excavated. Much of the base of feature 43 was sealed by a 20-50mm thick spread of gravel (57), presumed to extend along the entire length of the cut, which produced some forty-two sherds of pottery. Marks consistent with cartwheel ruts, evident cutting into the gravel near the junction of the two features, suggested that feature 43 was the later. An axle width of approximately 1.75m was suggested. These features have been interpreted as prehistoric trackways, broad hollow ways leading to the higher ground of the Blean. The gravel spread appears to represent deliberate consolidation and perhaps Fig. 6 The excavation at Radfall Corner (Site 8). KEITH PARFITT, TIM ALLEN AND JON RADY 50 suggests the trackway was an important route. Similar examples have since been recorded elsewhere (see main discussion below). The gravel and wheels ruts were sealed by a discrete dump of what appeared to be occupation waste (54). This was composed of lumps of burnt clay, admixed with large amounts of charcoal and scorched flints and a relatively large pottery assemblage (nearly thirty sherds). The remainder of the cuts were filled with clays of various hue, usually a greyish-yellow or brown, undoubtedly naturally eroded material, though a substantial amount of pottery was recovered from these fills. Adjacent to the spread of occupation material on the north side of the track was a cluster of twenty-two post-holes or post-pits. These were of varying size and shape, generally circular or oval in plan with similar fills of yellowish or greyish brown clay often containing fragments of burnt flint and charcoal or burnt clay. Size varied between 0.2m and 0.5m across, though some were larger. Although none showed any evidence for post-pipes, two of the larger examples had narrow and deep pointed sections at the base, suggesting that the posts had been pushed into the base of the post-pits. Three of the features yielded a few sherds of late Bronze Age/early Iron Age ceramics. At least some of these features can be interpreted as the incomplete remains of a circular structure, about 8m in diameter. That a domestic structure is represented by the post-holes is suggested by its position adjacent to the trackway and the detritus within the trackway at this point, this being composed of domestic rubbish possibly mixed with demolition debris. Three other slightly larger and very shallow features recorded in the vicinity might represent contemporary pits. They contained burnt flint and small quantities of similar pottery. The 204 sherds from this small site were derived from only ten contexts, of which two from hollow way 43 produced reasonably large assemblages. Despite this, the whole assemblage is represented by small and heavily abraded material, mostly bodysherds with one scrappy coarseware rim. The latter and the associated fabric characteristics are broadly similar to the assemblage from South Street (above); however its overall appearance is marginally coarser and superficially closer to elements from some regional late Bronze Age Deverel-Rimbury assemblages. A claim for occupation of this date is not made; instead a fairly broad early first millennium bc bracket is suggested of c.1000/900-600 bc. Site 11, Owl’s Hatch Road Roman farmstead This site lay on the outskirts of Herne Bay to the south of Greenhill, some 2km inland from the present coastline (TR 1633 6610). It was first discovered during field-walking when a surface spread of Roman pottery, together with a few pieces of Roman tile, was located. These were scattered over an area of about 70m within a large arable field on the southern side of Owl’s Hatch Road, opposite the Abbotswood Estate (Parfitt and Allen 1990, 11). Following evaluation trenching during 1991, which produced more positive evidence for Roman and some medieval occupation here, two substantial areas along the line of the new carriageway (Areas 1 and 2) were cleared of topsoil in the spring of 1995 in order to further examine the nature and extent of the buried remains before road construction began. The site occupied EXCAVATIONS AND ARCHAEOLOGICAL INVESTIGATIONS ON THANET WAY 1990-95 51 almost the highest point on a slight rise in essentially flat land, some 380m to the west of the Plenty Brook and stood at an elevation of 16-17m aod on heavy, illdrained London Clay. Over sixty individual features of archaeological interest were revealed in the excavations. Most of these appeared to be of Roman date and at least three were medieval. Cut into the natural clay, all the excavated features were found to be very shallow, having been previously truncated by the plough. Of the two areas stripped, the most extensive was Area 2 measuring some 90m (e-w) by 30m (n-s). Extending across the full width of the excavation, more than fifty individual archaeological features, mostly pits of varying shapes and sizes were exposed (Fig. 7). Area 1 lay 50m from Area 2 and measured 30m square. Apart from a series of modern field-ditches and drainage trenches, only three features of archaeological significance were revealed, all small pits located on the western side of the cleared area. Their dense carbon fillings suggested that they could have been hearth-pits but they produced no datable material. Located between the two main area excavations, evaluation Trench 2A revealed two adjacent oval pits of medieval date. One produced ten sherds of shell-filled sandy ware datable to the period c.ad 1150-1200/25, mostly derived from a single vessel. The other yielded six sherds of a similar date, together with three worn middle-late Anglo-Saxon sherds, two datable to c.775-850. In Area 2 the features were concentrated in the central sector, quite densely packed into a zone about 23m across. The bulk of these appeared to be pits, broadly similar although not all were excavated. The fills, generally of grey, brown or greyishbrown clay, derived predominantly from the underlying subsoil, some with the occasional speck of carbon in the matrix. Only variations are d escribed below. Apart from field-drainage ditches of recent date, just one gully [87] and two intercutting ditches were revealed (Fig. 7). The gully was aligned roughly eastwest and was traced for a minimum distance of 7m from its eastern terminal on the western side of the complex. It was 0.20m deep with steeply sloping sides and a slightly rounded base and produced significant quantities of oyster shell, some animal bone, about fifty sherds of pottery datable to the second century ad and a few small fragments of basalt lava-stone quern. The two short intercutting ditches [59/85 and 59/83] were located near the central part of the excavated complex. Feature 59/85 was the earliest and comprised a straight ditch about 11.7m in length, with rounded terminals, sloping sides and a slightly rounded base. With dimensions ranging from 1.10m to 2.10m wide, between 0.26m and 0.36m deep, it was filled with a series of clay deposits, which produced about 270 sherds of heavily broken up Roman pottery, all datable between the late first and early third century, together with some animal bone and oyster shell, Roman tile and part of a bronze pin. It would seem that this ditch had been completely infilled by the middle of the third century and then replaced by another ditch (83). Ditch 83, which represented a replacement or re-cut of the original, was curved in plan with a total length of about 9m. From its rounded east terminal it extended down the southern side of the earlier feature and then curved to the northwest, cutting across the width of 59/85 ending in a rounded terminal on its northern side. The feature was between 1.25m and 2.15m wide and 0.37m to 0.47m deep with sloping sides and a rounded base. More than 300 Roman pot-sherds, together with a few pieces of Roman tile and animal bone were recovered. Overall, the KEITH PARFITT, TIM ALLEN AND JON RADY 52 Fig. 7 The excavation at Owls Hatch Road Roman farmstead (Site 11); Area 2 features with inset showing positions of both Areas). EXCAVATIONS AND ARCHAEOLOGICAL INVESTIGATIONS ON THANET WAY 1990-95 53 pottery ranged in date from the late first to perhaps the early to mid fourth century and clearly included a significant proportion of residual sherds, derived from the earlier ditch. A few larger and fresher sherds were of late third- to early fourthcentury date. The available evidence thus suggests that this ditch was filled during the early fourth century, having been dug to replace ditch 59/85 around the middle of the third century. The purpose of these two successive ditches is not clear. Due to their short length it seems unlikely that they were for drainage, nor do they appear to delimit any particular group of features. The quantity of pottery they produced, much of it residual, suggests that they lay immediately adjacent to a long-established habitation area and the fact that the original ditch (59/85) was subsequently replaced by another in roughly the same position (83) indicates that the feature continued to be of significance for some considerable length of time. Several pits examined covered a sizeable area but one large hollow [112] located on the northern side of the area seemed too large to be regarded as a normal pit. Its full extent was not revealed and perhaps half lay beyond the limits of the excavation. As recorded, the feature appeared to be oval or circular in shape and measured a minimum of 4.50m (e-w) by at least 2.40m (n-s). It was cut by a later pit [61] on the eastern side. The hollow was 0.34m to 0.40m deep, with sloping sides and an undulating base and produced about 260 pot sherds, together with a corroded and illegible coin, probably of first- or second-century date and two pieces of Roman tile. The bulk of the pottery, including a few sherds of worn southern and central Gaulish samian ware, ranged in date from the late first to the early third century and clearly included a significant proportion of residual material. A few sherds were datable to the late third to early fourth century. From this it would seem that this feature was being filled over a long period, between the late first and late third century but must have been full by the end of the third century ad, or shortly after. The purpose of the large hollow [112] is not certain but it could perhaps represent a quarry-pit for clay or potentially the truncated remnant of a pond or dew pond. The carbon-rich filling of pit 61 cutting the hollow produced a further quantity of pottery, the latest of which was of the late third to early fourth cen tury. The bulk of the features investigated consisted of shallow pits no more than 0.40m deep. These were of varying shapes and sizes but most were oval, ranging in length from 0.80m to 3.88m. Fills were similar but a few, such as 61, contained denser deposits of carbon. Most produced at least a few pottery sherds, the bulk of which were of second- or third-century date. Pit 97 contained a single sherd of Iron Age flint-tempered ware which was probably residual, while pits 75 and 105 yielded small undistinguished groups of late third- to early fourth-century pottery. Pit 93 however appeared to be of post-Roman origin as in addition to residual Roman material, produced ten worn sherds of late Anglo-Saxon pottery, perhaps of tenth-century date, together with a somewhat fresher piece of early medieval ware, possibly of the eleventh century. A few fragments of plain medieval roofing tile were also recovered, suggesting that the pit was filled during the twelfth to thirteenth century. It seems likely that the majority of the Roman pits were originally dug for refuse. Their generally shallow depths tend to preclude use as storage pits, unless subsequent plough damage has been so severe as to have removed all but the bases. KEITH PARFITT, TIM ALLEN AND JON RADY 54 A few of the larger pits [e.g. 81 and 114] could perhaps represent quarries for clay and may thus be more closely related to the large hollow, 112. Six small features may be regarded as shallow post-holes. These were found scattered across Area 2 and did not appear to relate to any clearly defined structure. Features 119, 39, and 103 do form an east-west row and an imaginary line extended from 103 to 101, located to the south, creates a right-angle. Careful examination failed to reveal any other post-holes in this area, however, and the observed arrangement is probably fortuitous. The post-holes were all either circular or oval in shape and ranged from 0.20m to 0.59m wide and from 0.04m to 0.11m deep. Post-hole 103 contained a large lump of sandstone, perhaps a quern fragment, re-used as a packing stone. Three other post-holes produced small quantities of Roman pottery datable to the second to early fourth century, the bulk from 101 which yielded fourteen sherds. Two pits on the eastern side of Area 2 [41 and 45] were filled with dense carbon and ash deposits sealing traces of a burnt surface to the natural clay in their base suggesting that these features were hearth-pits. Neither was associated with metalworking debris and a domestic function seems more probable. Both produced a few sherds of pottery broadly of late first- to early third-century date and large quantities of burnt red clay came from pit 45, too fragmentary for any identification of its original form or purpose. The three pits located in Area 1 also contained dense carbon deposits suggesting that these too could have been hearth pits, but no trace of burning was noted on their bases and there was no dati ng evidence. The pottery:2 the site yielded 1,683 sherds (10.103kg) of late first- to early fourth-century pottery from thirty-six contexts. Most of the sherd assemblages were small and rather comminuted but somewhat larger pot groups were retrieved from Ditches 59/85 and 83, Hollow 112 and Pit 61. None of the assemblages were large enough for quantification by EVEs (Orton 1975) but all had weights and numbers of sherds per fabric recorded for archival purposes. The small size of most of the assemblages makes it difficult to draw conclusions regarding pottery supply to the site. However, the size of the assemblage from Ditch 59/85 is large enough to suggest that nearly all of the coarse pottery in use during the late second and early third century came from three suppliers in roughly equal proportions. Native potteries probably centred along the Wantsum Channel just east of the site produced Belgic/Native Coarse ‘transitional’ ware cooking-pots and storage jars, the Thameside and Upchurch potteries in the Medway marshes to the west supplied cooking pots, bowls, dishes and finewares and the Canterbury kilns to the south supplied slightly smaller amounts of jars and lid-seated carinated bowls. The poor showing of finewares suggests a low status settlement: the few pieces include Central Gaulish Samian and colour-coated wares and Moselkeramik beakers from the area around Trier. The twenty small finds from this site include a fragmentary copper alloy pin, one coin (illegible but of early Roman date), several pieces of lead and numerous small fragments of querns of both basalt lava and millstone grit. All of the finds are of common and familiar Roman types, and although they provide a little supplementary dating evidence, there are too few to be able to make any broader statements about the nature of activity on the site. Most of the other find assemblages, which included worked flint, tile, burnt clay or daub, glass and animal bone were in small quantity or too abraded for clear identification. Details remain in the archive. EXCAVATIONS AND ARCHAEOLOGICAL INVESTIGATIONS ON THANET WAY 1990-95 55 Other sites Of the fourteen areas of potential interest identified, only those three described above were examined to any great extent; others could not be evaluated due to difficulties of access and the condition of the farmland, while cropmark Sites 2 and 6 were evaluated with negative results. Two sites (Sites 13 and 14; TR 1722 6620 and TR 177 664) north of Strode Farm were evaluated, but the low density of features did not merit further work. The sites lay some 350m apart near Strode Farm, Lower Herne. Site 13: was situated on flat land (at about 11m aod) a short distance to the west of the Plenty Brook and was identified from several fragments of Roman tile picked up during field-walking about 150m north of Strode Farm. Site 14 lay adjacent to Herne Bay Golf Course, about 500m north-east of Strode Farm and was represented by surface finds of flint-tempered prehistoric pottery. Evaluation trenching was undertaken across both sites in October 1991 when a few sub-surface features of archaeological interest were located. At Site 13, although eighteen trenches, totalling some 178m, were excavated to the top of the natural brickearth subsoil, only one feature of archaeological interest was located on the south-western side of the investigated area, although three early Roman sherds of grog-tempered ware were recovered from ploughsoil. The feature consisted of a broad, shallow negative feature, the full extent of which was not revealed. It was over 6m across but only 0.25m deep, with a steep side and undulating base. It could represent a large pit, or conceivably a sunken trackway. No traces of it were located in any of the other trenches, however. The feature produced fragments of animal bone, Roman tile and burnt flint, together with seven pot-sherds. Four of these were of early Roman date and somewhat worn, the others fresher and of Anglo-Saxon organic-tempered and chalk-filled wares datable to the period c.ad 575-650. This could suggest a seventh-century date for the in-filling of the feature and raises the possibility that it might have represented part of a sunken-featured building. It would thus appear that there was both Roman and Anglo-Saxon activity here, although the main focus of occupation must have lain beyond the area investigated, unless the general absence of associated features is due to subsequent plough erosion. Site 14: twenty-five trenches, totalling some 258m in length, were excavated at, but only two separate features were located, both sub-rectangular pits or ditch terminals. One of these yielded two fresh pot-sherds datable to c.ad 825-875. A ninth-century date for the filling of this feature is thus suggested. No other features were discovered but a piece of Roman tile came from the plough-soil, together with eight small fragments of flint-tempered pottery, perhaps broadly datable to c.850-550 bc and probably contemporary with the sherds previously found in this area. Two more mid to late Anglo-Saxon pot-sherds were also recovered. discussion At the time of these excavations, virtually nothing was known of the archaeological potential of this general area north of the Blean between the Wantsum to the east KEITH PARFITT, TIM ALLEN AND JON RADY 56 and Seasalter to the west. As mentioned above, it was often assumed that much was forested in the past and that because of the heavy, unworkable and ill-drained nature of the land, settlement prior to the medieval period would have been sparse; it was even considered desolate in more recent times (Allen 2009, 189). The excavations on the Thanet Way were probably the first to challenge these views, but since then have been superseded by a number of sometimes much more extensive archaeological investigations (see, for example, Allen 2009). Excavation on the London Clay is always difficult and the material remains and associated artefactual assemblages not always of great quality as the harsh, acidic environment of the clays is not conducive to preservation of bone or organic material. Such was the case with the sites described above. However, they form an integral part of an emerging wider picture of the area which suggests that occupation, settlement and agriculture were more common and widespread than previously supposed.3 Significant forest clearance probably started in the middle Bronze Age increasing later in the period. There is much less evidence for early prehistoric occupation, possibly reflected by the scarcity of late Neolithic/early Bronze Age barrows in this region, although some possible examples have been recorded (Allen 2009, 199). There is a caveat to the area’s lack of cropmarks; clays are generally not conducive to their formation and struck flints are not uncommon in the region. In the later Bronze/earliest Iron Age, when Sites 7 and 8 were probably occupied, settlement densities seem to rise. Excavations suggest the existence of smaller settlements and farmsteads, perhaps relatively short-lived (as suggested by Macpherson-Grant at South Street, above), some with associated ditched field systems. Ditches were probably necessary for drainage (Allen 2009, 203). There was limited evidence for crops at South Street (below); a pastoral economy was probably equally important for these settlements. Since only a part of the probable settlement at South Street was exposed, it is difficult to comment on its overall form, whether, for example, it was enclosed by ditches or of a dispersed, open type. Ditch 152 on the west, which appeared to be contemporary (other features appear to respect its position and the ditch cut none of them), may have formed part of an enclosure. Without the benefit of cropmark evidence however, its function remains uncertain. Nevertheless, the features of the early and later phases probably represent elements of a small settlement or farmstead of one or a few family units although difficulties in judging the overall character of the late Bronze/earliest Iron Age activity are compounded in Kent, where settlements of the period seem to vary considerably (Champion 2007, 102- 6). Of interest here are the gravel-based or metalled working hollows. Similar features have been recorded at other sites and have been seen as possible early examples of sunken-featured structures, otherwise evident in Kent from the Iron Age, Roman and later periods (Hicks 2008; Clark and Lane 2013). The debate on these earlier examples remains open. Macpherson-Grant discusses the dating in detail above, but the site can be fairly confidently assigned to the Earliest Iron Age, with a single radiocarbon date broadly in agreement. Taking all this into account, Macpherson-Grant concludes that Phase 1 activity at South Street is likely to have occurred between c.950 and 850 bc. The site could therefore be slightly later than Site 8 (Radfall Corner) to the east (see below). EXCAVATIONS AND ARCHAEOLOGICAL INVESTIGATIONS ON THANET WAY 1990-95 57 The presence of slag and ironstone in some of the features at South Street was initially considered significant because evidence for ironworking is rarely found on sites of this period in Kent. However, there is some doubt as to the integrity of this evidence due to late Iron Age activity on the site, which has probably introduced intrusive material into some of the early features, though for a proportion of the slag-containing pits there is no inherent reason to suggest that such contamination has occurred. In any event, the material did not conclusively indicate that ironworking or smithing took place here, comprising as it did undiagnostic pieces including for the most part, vitrified hearth lining or cinders. Like the ironstone nodules they could equally well relate to ironworking close by, perhaps on the Phase 2 settlement likely to be slightly to the south (below). As for the remainder of the material assemblages, such as the animal bone and worked flints, these were too small, abraded or undiagnostic for any firm conclusions to be made, although cow, horse and sheep were represented in the bone assemblage, providing clear evidence that the economy was based in part on stock breeding. Environmental sampling suggested cereals were also grown and processed. Turning to the trackway at Radfall Corner, routes such as this, sometimes forming hollow-ways, would have connected settlements to both the resources of the coast and the woodlands of the Blean. Until recently there was little evidence for trackways of such an early date in Kent, particularly not metalled, but at least one other of a similar period has since been excavated at Ramsgate (Rady et al. forthcoming) and an early Bronze Age trackway is recorded at Holywell Coombe, Folkestone (Bennett et al. 1998, 273). At Radfall Corner, the remains of occupation close to the trackways were heavily truncated and therefore difficult to assess, but may have been similar in form, although slightly earlier than the activity at South Street. The siting of both settlements may have been influenced by ease of movement along the coastal zone and Thames estuary, a corridor which it has been postulated may have provided an important trade route during this period (Allen 2009). As elsewhere in Kent and the South-East, evidence for settlement declines during the early Iron Age, with relatively few early to mid Iron Age sites known. Again many factors were probably involved partly perhaps connected to the adjustment from a bronze-related economy to that of iron; environmental factors may also have played a part. Whatever the causes it is possible that occupation aggregated into fewer and somewhat larger but more discrete settlement sites each perhaps occupied by a number of co-operating family units and which must have represented safer, more stable economic entities than the smaller late Bronze Age settlements and farmsteads. Such early to middle Iron Age sites may be represented at Sunset Caravan Park, Whitstable and Underdown Lane, Eddington near Herne Bay in the immediate locale (Allen and Willson 2001; Jarman 2005). The evidence suggests that smaller settlements again proliferated in the late Iron Age (Allen 2009, 202-3), possibly representing an increase in population, or some other economic or climatic imperative. The second phase of activity at South Street dates to this period, the century or so prior to the Roman Conquest, but as at Radfall Corner, not enough was exposed to inform further discussion. Neither is there evidence to suggest occupation into the Roman period, whereas either KEITH PARFITT, TIM ALLEN AND JON RADY 58 shifting occupation or the establishment of new settlements is represented by the activity at Site 11, which may have been occupied from the late first through to the earlier fourth century. A light scatter of struck flints located in the area of Site 11 was part of a background spread recorded along much of the road corridor (Parfitt and Allen 1990, 11). The flints demonstrate some early prehistoric activity in the region, although the numbers recovered suggest that such activity may never have been very intense. Additionally, a small quantity of residual, flint-tempered Iron Age pot-sherds was recovered, implying that there was pre-Roman occupation somewhere nearby but no associated features were recognised. Other sherds of Iron Age pottery were also found a short distance to the north-east, close to Strode Farm (Site 14) and on Herne Bay Golf Course, tending to confirm that there was pre-Roman settlement in the region. The excavated remains suggest that the main phase of occupation at Site 11 was of Roman date and of two separate phases. The ceramics however, suggest a long sequence of occupation extending from the late first through to the earlier fourth century, with much of the earlier pottery found as residual material in later features (see above). It is likely therefore that the full extent of the Roman site has not been revealed and that further parts of it, perhaps the main settlement area, extends under Owl’s Hatch Road, onto the Abbotswood Estate to the north and across the field to the south. A few pieces of Roman tile were retrieved from the surface between the main excavated areas and the Plenty Brook (see above), and further Roman material has been found in the field beyond the brook, near Strode Farm (Site 13, see below). Trenching in both these areas however, failed to reveal any definitely associated features and the present site currently appears to represent the main settlement of the immediate area. The general character of the excavated remains together with the associated finds, indicate that Site 11 represents a Romano-British native farmstead of relatively low status. The scatter of pits, ditches and a few post-holes is largely typical of such rural settlements across most of Kent, although the absence of well-defined ditched enclosures is a little unusual, especially in view of the necessity for good drainage on these clay soils. No clear evidence for buildings was revealed but the general density of features suggests that these stood close-by. Fairly certainly they would have been of timber. The general absence of building debris largely confirms this; the occasional fragments of Roman tile recovered were not necessarily derived directly from buildings and could have arrived at the site in imported rubble or brought in for use in small structures such as hearths or ovens. The few postholes recorded during the excavation must relate to timber structures of some kind, though these could have been fences, drying frames, animal pens or the like. Evidence for the economy of the site is minimal; just a few pieces of poorly preserved animal bone were recovered due to the acidic nature of the subsoil. Evidence for grain processing is suggested by fragments from two quern-stones. In view of the difficult nature of the subsoil, a mostly pastoral and/or woodland based economy may perhaps seem most appropriate for the site. Shellfish, in the form of oysters, were certainly brought to the settlement, no doubt from local beds off the north Kent coast, then perhaps rather more than 2km away. Despite the presence of at least two substantial hearth-pits, there is no evidence for industrial activity EXCAVATIONS AND ARCHAEOLOGICAL INVESTIGATIONS ON THANET WAY 1990-95 59 such as ferrous metalworking. The nearby Plenty Brook no doubt provided a ready source of fresh water for the settlement. The site at Owl’s Hatch Road is part of growing evidence for spread and concentration of settlement in this zone in the Roman period. Work on a new housing estate at Chestfield, some 3km to the west, revealed traces of a Roman field system (Blockley 1987), while traces of another were recorded off Bogshole Lane, some 4km to the north-east (Parfitt and Hutcheson 1995); these were presumably associated with other settlements. The structure partly destroyed by the Canterbury-Whitstable Railway (described above), just over 4km to the south-west, implies the presence of another, perhaps major, settlement here, quite possibly a villa. Further to the west, work on an earlier phase of the Thanet Way road improvement scheme near Seasalter Lane located a small quantity of Roman pottery and tile, but no associated features (TR 091 636; Parfitt and Allen 1990, Appendix) while other sites of the period are indicated near Eddington (Shand 2002) and Beltinge (Parfitt and Hutcheson 1995). Sites 13 and 14 suggested Iron Age, Roman and Anglo-Saxon activity at the eastern end of the road route, though again the main centres of occupation probably lay beyond the areas investigated. Of the three features located in evaluation trenching here, at least two appeared to be Anglo-Saxon. No definite Iron Age or Roman features were present, although finds of the date were recovered from the plough-soil and as residual material in one of the features. The evidence for regular, if not continuous, occupation extending back to at least the Iron Age period along this eastern part of the road corridor is perhaps primarily due to features of the local topography. The gentle terrain and the presence of the Plenty Brook would have been important, but the local soils were perhaps of greater significance. The extensive tracts of heavy, ill-drained London Clay partially give way here to slightly lighter, brickearth-based soils which are better drained and more easily cultivated. The evidence for an Anglo-Saxon presence is of particular interest since this is still rarely encountered in rural east Kent. The discoveries must be considered in conjunction with the mid-late Anglo-Saxon pottery finds made at Site 11 (Owl’s Hatch Road), about a kilometre to the south-west. Here, a few residual pieces of middle-late Anglo-Saxon pottery were recovered from two medieval features, although no Anglo-Saxon features could be identified. Taken together, these finds suggest at least small-scale occupation across this region throughout the Anglo- Saxon period from perhaps as early as the end of the sixth cent ury. Evidence for medieval activity at Site 11 contributes in a small way to a recently much increased body of information on rural settlement patterns and associated agriculture such as revealed at Thanet Earth, near Monkton, and on road and rail schemes south of Gravesend (Rady 2010, 11-16; Allen et al. 2012, 485-583), to name a few. The above mentioned sites were extensive, on different subsoils and in topographical settings dissimilar to those at Thanet Way, but a comparably complex landscape organization might be represented, perhaps hinted at by agricultural features encountered in other excavations in the area (near Chestfield and Greenhill; Allen 2002; 2004; Helm 2008), all providing evidence for this previously somewhat neglected period. KEITH PARFITT, TIM ALLEN AND JON RADY 60 bibliography Allen, T., 2002, ‘Churchwood Drive, Chestfield’, Canterbury’s Archaeology 1999-2000, 23-7. Allen, T., 2004, ‘Swine, Salt and Seafood – a Case Study of Anglo-Saxon and Early Medieval Settlement in North-East Kent’, Archaeologia Cantiana, cxxiv, 117-35. Allen, T., 2009, ‘Prehistoric settlement patterns on the North Kent coast between Seasalter and the Wantsum’, Archaeologia Cantiana, cxxix, 189-207. Allen, T., Donnelly, M., Hardy, A., Hayden, C. and Powell, K., 2012, A Road Through the Past: Archaeological discoveries on the A2 Pepperhill to Cobham road-scheme in Kent, Oxford Archaeology Monograph 16. Allen, T. and Willson, J., 2001, ‘Sunset Caravan Park and Church Lane East, Whitstable’, Canterbury’s Archaeology 1998-1999, 10-11. Bennett, P., Ouditt, S. and Rady, J., 1998, ‘The Prehistory of Holywell Coombe’, in R.C. Reece and D.R. Bridgland (eds), Late Quaternary Environmental change in North-west Europe, London, 261-314. Bennett, P., Couldrey, P. and Macpherson-Grant, N., 2007, Highstead near Chislet, Kent, Excavations 1975-1977, Archaeology of Canterbury, New Series, IV, Canterbury. Blockley, P., 1987, ‘Observations and watching briefs’, in Canterbury’s Archaeology 1986- 1987, 22. Champion, T., 2007, ‘Prehistoric Kent’, in J.H. Williams (ed.), The Archaeology of Kent to AD 800, Kent County Council, 67-132. Champion, T., 2011, ‘Later Prehistory’, in P. Booth et al., On Track – The Archaeology of High Speed 1 Section 1 in Kent, Oxford Wessex Archaeology Monograph No. 4, 151- 241. Clark, P. and Lane R., 2014, ‘St Edmund’s School, Canterbury’, Canterbury’s Archaeology 2012-2013, 8-10. Couldrey, P., 2007, ‘The Pottery’, in P. Bennett et al., 101-71. Cotter, J., 1997, A twelfth-century pottery kiln at Pound Lane, Canterbury, CAT Occas. Paper No. 1, Canterbury. Everitt, A., 1986, Continuity and colonization: the evolution of Kentish settlement, Leicester University Press. Glover, J., 1982, The Place Names of Kent, Meresborough Books. Helm, R., 2008, ‘The Grange, Greenhill Road, Herne Bay’, Canterbury’s Archaeology 2006-2007, 14-15. Hicks, A., 2008, ‘The Roman Settlement’, in P. Bennett, P. Clark, A. Hicks, J. Rady and I. Riddler, At the Great Crossroads: prehistoric, Roman and medieval discoveries on the Isle of Thanet 1994-95, CAT Occas. Paper No. 4, 101-278. Jarman, C., 2005, ‘Underdown Lane, Eddington’, Canterbury’s Archaeology 2003-2004, 14-16. Jenkins, F., 1962, ‘A Roman Building on South Street, near Whitstable’, Journal of Roman Studies, lii, 190. Macpherson-Grant, N., 1994, ‘The Pottery’, in D.R.J. Perkins et al., ‘Monkton Court Farm Evaluation 1992’, Archaeologia Cantiana, cxiv, 248-88. Macpherson-Grant, N., 2007, ‘Perforated pottery slabs’, in P. Bennett et al, 267-8. Macpherson-Grant, N., forthcoming, ‘The Prehistoric Pottery’ in J. Cotton et al., report on the MoLA 2010 excavation at Tothill Road, Minster, Thanet. McInley, J., Leivers, M., Schuster, J., Marshall, P., Barclay, A.J. and Stoodley, N., 2013, Cliffs End Farm, Isle of Thanet, Kent: a mortuary and ritual site of the Bronze Age, Iron Age and Anglo-Saxon period with evidence for long distance maritime mobility, Wessex Archaeology Monograph 31. Millard, L., 1968, ‘Notes from Canterbury Museum’, Archaeologia Cantiana, lxxxiii, 267-8. EXCAVATIONS AND ARCHAEOLOGICAL INVESTIGATIONS ON THANET WAY 1990-95 61 Morris, E., 2006, ‘The Later Prehistoric Pottery’, in P. Booth (ed.), Ceramics from Section 1 of the Channel Tunnel Rail Link, Kent, CTRL Specialist Report Series (2006), 34-121. Orton, C.R., 1975, ‘Quantitative pottery studies: some progress, problems and prospects’, Science and Archaeology, 16, 30-5. Parfitt, K. and Allen, T., 1990, ‘An archaeological survey of the Thanet Way (Phases 2-4)’, unpubl. report. Parfitt, K. and Hutcheson, A., 1995, ‘The Herne Bay Water Treatment Project: assessment of the archaeological discoveries’, unpubl. CAT report. Rady, J., 2010, ‘Thanet Earth, Monkton’, Canterbury’s Archaeology 2008-2009, 1-16. Rady, J., Boden, D. and Wilson, T., forthcoming, ‘Evidence for a Neolithic midden, later prehistoric and Anglo-Saxon settlement at the site of the new Ellington and Hereson School, Ramsgate’. Shand, G., 2002, ‘Eddington Farm, Herne Bay’, Canterbury’s Archaeology 1999-2000, 18-23. Willson, J., 2003, ‘The University of Kent. An Archaeological and Historical Desk Based Assessment and Survey’, CAT unpubl. report. endnotes 1 A ‘gate’ is defined as an ‘entry onto the downs or into the Blean’ (Everitt 1986, 160). ‘Blean’ signifies an area of common scrubland or heath, not necessarily woodland. 2 Malcolm Lyne’s assessment of the Roman pottery is held with the site archive . 3 Allen (2009) lists thirty-two separate sites. An additional five sites (mostly earlier and later prehistoric/Roman in date) are detailed in the same volume (‘Notes and summaries’, 359-89), giving some indication of the burgeoning amount of potential information.