An Iron Age and Romano-British Site on Wood Hill Kingsdown Near Deal

393 An Iron Age and Romano-British Si te on Wood Hi ll, Kingsdown, near Deal keith parfitt Wood Hill, on the outskirts of Kingsdown near Deal, occupies a prominent position overlooking the English Channel at the north-east end of a long ridge of Chalk Downland, about 900m from the present shore-line (Fig. 1). Bounded by dry valleys, the summit of the hill rises to just over 65m od and is capped by a localised deposit of Clay-with-Flints. Today, much of this clay area is covered by Kingsdown Wood, which is visible from a considerable distance in most directions. Outside the wood, surface Fig. 1 Map showing location of the site. KEITH PARFITT 394 pottery and prehistoric flintwork has been regularly collected by local researchers for the last forty years. Field-walking and limited trial-trenching (unpubl.) undertaken on the south-eastern side of the hill during the late 1960s and 1970s, produced significant quantities of Iron Age and Roman pottery (Parfitt 1985, 207- 9; Broady 1989, 144). This was scattered across a wide area below the wood, mostly between the 65m and 55m contours; NGR TR 370 476, centred (Fig. 1). The soils here are thin and chalky, in contrast to the heavy clays found on the summit of the hill. A detailed catalogue of the material recovered by local resident Jonathan Joy, amounting to more than 1,700 sherds, was prepared by Nigel Macpherson-Grant in 1980 and this identified pottery ranging in date from the early Iron Age through to the late Roman period. Such a substantial pottery assemblage was clearly suggestive of a sizable habitation site in the area. In 1982, as a follow-up to the previous work, members of the Dover Archaeological Group (DAG) undertook further trial-trenching across part of the area (Fig. 2), in order to investigate the origin of the surface finds (Frere 1983). Some account of the Iron Age material recovered has been previously published (Parfitt 1985, sites 1 and 2) and the present report is concerned with the late Iron Age and Romano-British discoveries. Thirty hand-dug trenches were cut by DAG across land on either side of Victoria Road, an un-metalled trackway running along the south-eastern side of Wood Hill out onto Ringwould Free Down (Trenches 1-30; Figs 1 and 2). The land investigated on the north-western side of Victoria Road consisted of uncultivated grassland and scrub (OS parcel no. 1564), lying adjacent to an in-filled nineteenth-/early twentieth-century chalk quarry. On the south-east side, the land was arable (OS Parcel no. 2351). Overall, the area investigated measured about 130m (ne-sw) by 70 m (nw-se). Sixteen trenches revealed archaeological features and most of these contained Iron Age and/ or Roman pottery, although four failed to produce any datable finds. the excavated features The 1982 trenching revealed a total of twenty-four features of archaeological interest (Fig. 2; Table 1). Sixteen of these were pits and hollows, three were gullies [Fs 12, 19 and 25] and five could have been either sections of ditches, pits or terraces [Fs 1, 2, 6, 7 and 9]. The bulk of the features were located in the south-western half of the area examined (Fig. 2). Most of the datable ones were Roman, with one early/middle Iron Age pit [F. 15, see below]. Pits and Hollows More than half the features located seemed to be pits and hollows of varying shapes and sizes (Table 1 and Figs 3-7). Many were not fully exposed in AN IRON AGE AND ROMANO-BRITISH SITE ON WOOD HILL, KINGSDOWN 395 Fig. 2 Overall site plan showing location of excavated trenches and features. KEITH PARFITT 396 Fig. 3 Plan of Trenches 27 and 28 (Fs 23, 24 and 25). AN IRON AGE AND ROMANO-BRITISH SITE ON WOOD HILL, KINGSDOWN 397 Tab le 1: Wood Hi ll, Kingsdown, excava ted features and Pottery fi nds. Di mensions in metres Type Shape Length Width Axis Depth Sides Base No. sherds [Assemblage] 1 ?Broad ditch/terrace Linear 0.80 (min) 1.80 (min) E-W 0.26 Sloping Flat 13 [7] 2 ?Broad ditch/terrace Linear 1.00 (min) 3.20 (min) N-S 0.46 Sloping Flat 4 3 Pit Oval 1.27 1.05 E-W 0.63 Steep Flat 104 [4] 4 Pit Circular Dia.= 1.68 - 0.44 Steep/sloping Flat 8 5 Pit Oval 1.75 (min) 1.40 N-S 0.60 Steep/vertical Flat 137 [3] 6 Pit/ditch ? 1.05 (min) 0.30 (min) NE -SW 0. 37 Steep/sloping Flat 10 [8] 7 Pit/ditch ? 0.95 (min) 0.65 (min) NE -SW 0. 22 Sloping Dished 10 [8] 8 Pit Circular Dia.= 0.88 - 0.43 Sloping Rounded No finds 9 Pit/ditch ? 1.10 (min) 0.54 (min) NE -SW 0.45 Sloping Dished 10 [8] 10 Natural - - - - - - - No finds 11 Natural - - - - - - - No finds 12 Gully Linear 1.50 (min) 0.60 E-W 0.35 Sloping Dished No finds 13 Pit ?Circular Dia.= 0.65 - 0.33 Steep Dished No finds 14 Quarry pit ??Oval 17.00 (min) 5.90 NE -SW 1.25 Sloping Undulating 121 [6] 15 Pit ?Circular Dia.= 0.97 (min) - 0.27 Steep Flat 38 (IA) 16 Pit Circular. Dia. = 0.70 (min) - 0.15 Sloping Flat 3 17 Pit Sub-rectang. 0.70 0.35 (min,) NE -SW 0.20 Sloping Undulating No finds 18 Large hollow ? 3.60 (min) 1.00 (min) NE -SW 0.95 Sloping Flat 6 19 Gully Linear 1.00 (min) 0.60 NE -SW 0.25 Sloping Rounded 14 20 Large pit ? 2.30 (min) 1.45 (min) NE -SW 0.51 Sloping Unseen 40 [5] 21 Pit ?Oval 1.85 1.30 (min) NE -SW 0.51(max) Sloping Flat 17 22 Pit ?Circular Dia.= 1.80 (min) - 0.40 Sloping Rounded 1 23 Pit Oval 1.27 0.90 (min) NE -SW 0.18 Sloping Dished 13 [1] 24 Pit Circular Dia.= 0.75 - 0.27 Vertical Flat 92 [2] 25 Gully Linear 11.00 (min) 0.35-0.80 NE -SW 0.05-0.15 Steep/sloping Dished 11 26 Hollow (?terrace) ?Rectangular 5.70 (min) 4.25 (min) NW-SE 0.30 (max) Steep Sloping 59 KEITH PARFITT 398 the trenches excavated. The earliest was F. 15 located in Trench 15. This small circular pit produced over thirty flint-tempered pot-sherds broadly datable to the early/middle Iron Age period. Details have been previously published (Parfitt 1985, 209). A shallow pit immediately adjacent [F. 16] (Fig. 6) produced three sherds from a late first/early second-century Roman jar. In Trench 27 two adjacent circular pits [Fs 23 and 24] (Figs 2, 3, 4 and 6; Table 1), both cut through gully F. 25, produced quantities of Belgic-style pottery (see below, Assemblages 1 and 2; Fig. 8.1-8.3) and distinctive chaff-tempered ware, indicating that they were filled sometime during the mid/late first century ad. Pit F. 24 also produced about one third of a rotary quern made from Folkestone Greensand (Fig. 4 and Fig. 9.3). Trench 8, dug at the south-western end of the uncultivated strip northwest of Victoria Road, revealed no less than seven probable pits [Fs 3-9] (Table 1), in two inter-cutting groups [Fs 3, 5, 8 and Fs 6, 7, 9] with F. 4 adjacent (Figs 2, 5 and 7). It was only possible to fully excavate Fs 3, 5 and 8. The remainder of the features extended beyond the excavation limits. Too little of F. 6 was seen for it to be certain that this was not actually a section of ditch. Fs 3, 7 and 8 were cut across by a modern rectangular trench. Fairly certainly, this trench relates to the archaeological investigations that were undertaken here by local enthusiasts between Fig. 4 Pottery and quernstone in pit F. 24 (photo Ben Stocker). AN IRON AGE AND ROMANO-BRITISH SITE ON WOOD HILL, KINGSDOWN 399 Fig. 5 Plan of features in Trench 8 (upper) and Trenches 15 and 16 (lower). KEITH PARFITT 400 1969 and 1971. Its backfilling included a potato crisp packet bearing the closing date for a competition in October 1969. The pits in Trench 8 produced a combined total of just over 250 potsherds mostly of second and third century ad date (Assemblages 3, 4 and 8, see below). Other finds included animal bone, marine shell, burnt daub, iron, and quern fragments. There can be little doubt that these were rubbish pits and they provide clear evidence for habitation in the immediate area during the Roman period. From the pottery dating it seems likely that at least some of them continued in use until the end of the third century (Assemblage 8). Fig. 6 Sections across pits (Fs 15, 16, 21, 23 and 24). AN IRON AGE AND ROMANO-BRITISH SITE ON WOOD HILL, KINGSDOWN 401 Fig. 7 Sections across excavated features (Fs 3-7, 19 and 26). KEITH PARFITT 402 Along the southern margin of the trenched area, on the south-east side of Victoria Road, part of a large, irregular pit [F. 14] was revealed in five trenches (Fig. 2). It seems probable that this represents a chalk quarry of a type that is now becoming increasingly well-known on Iron Age and Roman sites in east Kent. The full extent was not revealed but as recorded it measured at least 17m (ne-sw) by 5.90m (nw-se). Where seen, the sides were sloping and the base undulating. The feature was up to 1.25m deep and was filled by a series of chalky soil deposits [WHK-82-20, 21, 24, 26 and 30] which produced quantities of Iron Age and Roman pottery (Assemblage 6, see below), animal bone, marine shell, pieces of lavastone quern, two iron nails and a prehistoric struck flint. The latest pottery recovered is of fourth-century date, implying that the pit was being used for casual dumping of refuse until the end of the Roman period. From the nature of the filling and the date of the pottery it contained, it would seem that this feature had silted-up gradually over many years, rather than being deliberately back-filled at any specific time. A broad, shallow hollow [F. 26] located in Trench 16 may represent part of a terrace for a timber building [Figs 2, 5 and 7]. It cut through an earlier gully [F. 19] and two pits [F. 17 and un-numbered]. Neither pit contained any datable finds but the gully yielded a quantity of material, including fourteen sherds of pottery, the latest of which are datable to the period c.ad 170-250. From this, it would seem that the hollow could not have been excavated before the late second century. The chalky soil filling the hollow, itself (Fig. 7, WHK-82-18) yielded more than fifty Roman sherds, the latest of which seem to be of fourth-century date. A short distance to the south-west, large hollow F. 18, located in Trench 18 might represent part of another such terrace. It produced six potsherds of second-century date. Trenches further to the south-west revealed portions of three more pits (Fig. 2) [Fs 20-22]. The largest of these was F. 20 which produced forty sherds of pottery datable to the third century (Assemblage 5). Pit F. 21 produced seventeen sherds of first/second-century ad date, whilst F. 22 yielded a single piece of late Roman grog-tempered ware. Possible Ditches [Fs 1 and 2] Sections of two features located on the western side of the area investigated (Fig. 2) [Fs 1 and 2], were interpreted during the excavation as parts of very broad, shallow ditches. The full width of neither was exposed, however, so their identification cannot be certain. It is not impossible that they actually represent shallow terraces cut into the slope, similar to F. 26 (see above). F. 1 produced a small collection of late Roman pottery (Assemblage 7), whilst F. 2 yielded a few sherds no later than the second century. Both features contained some animal bone. AN IRON AGE AND ROMANO-BRITISH SITE ON WOOD HILL, KINGSDOWN 403 Gullies [Fs 12, 19 and 25] Portions of three gullies [Fs 12, 19 and 25] were located (Fig. 2; Table 1). Too little was seen to reveal any clear overall pattern but Fs 19 and 25 were roughly aligned north-east by south-west, whilst F. 12 ran east-west. It seems unlikely that all these features are contemporary and only F. 25 was traced for any significant distance (Fig. 3). Excavated for some 11m, gully F. 25 was cut by the two pits containing Belgic pottery [Fs 23 and 24, see above] and itself yielded eleven sherds of similar material. A piece of Romanised Upchurch ware also recovered suggests, however, that the gully was finally in-filled sometime after ad 43 (see below). If so, the two subsequent pits must also be post-Conquest. Gully F. 19 yielded fourteen potsherds datable to the period c.ad 170-250 but the filling of F. 12 and a later pit [F. 13] which cut it were both sterile. the finds The archive for the project includes thirty-six recorded deposits, twentysix recorded features (Table 1), two site plans, twenty-four measured sections and five boxes of finds. All the finds from the site have been placed in store at Dover Museum, together with a copy of the field records. Prehistoric Struck Flints (not illustrated) A total of just eleven prehistoric struck flints was recovered during the excavations. These were contained either in the topsoil deposits or as residual material within the fillings of cut features. There are five unworked waste flakes and blades, with another three which show signs of edge damage, suggesting that they may have been utilised. There is also a flake with a worked notch, a bifacially worked pointed ‘rod’ and an end-scraper. This small assemblage must be considered in relation to the much larger collections of lithic material recovered over many years from the surface of Wood Hill (see above). Detailed analysis indicates these collections include extensive Lower Palaeolithic, Neolithic and Bronze Age industries (Halliwell and Parfitt 1993; Scott-Jackson 2000). Based on the blue-white patination, and in the absence of any particularly diagnostic tool-types, a broad Neolithic/Bronze Age date-range may be tentatively suggested for the present material. Objects of iron (not illustrated) The excavated features produced a total of eleven pieces of iron-work. These are all simple hand-made nails of small to medium size, with square sectioned shafts. Six came from late second/early third century KEITH PARFITT 404 pit F. 5 and two more from the quarry pit F. 14. Single examples were recovered from the fillings of Fs 2, 3 and 19. Pottery (Fig. 8; Table 2) by Malcolm Lyne The thirty trenches put down on the site yielded a total of 594 sherds (4,901g) of late Iron Age to late Roman pottery from twenty-two separate contexts. A further 354 sherds (3,055g) of similarly-dated, medieval and post-medieval pottery were retrieved from the topsoil in the various trenches and 250 more sherds (2,499g) came from field-walking over the excavated site. All the assemblages were quantified by numbers of sherds and their weights per fabric. These fabrics were classified using a x8 magnification lens with built-in metric scale for determining the natures, sizes, forms and frequencies of added inclusions. Finer fabrics were further examined using a x30 magnification pocket microscope with artificial light source. None of the assemblages are large enough for more meaningful quantification by Estimated Vessel Equivalents (EVEs) based on rim sherds (Orton 1975). Tab le 2. Detai ls of Potter y in Assemb lage 3 from Pit , F. 5 Fabric No. of sherds % Weight (g) % R1 71 51.8 552 49.9 R14 20 14.6 190 17.2 R16 30 21.9 198 17.9 R43 4 2.9 68 6.1 R71 1 0.7 1 0.1 R75 3 2.2 14 1.2 LR2.1 3 2.2 42 3.8 LR2.2 4 2.9 26 2.3 LR2.3 1 0.8 16 1.5 Total 137 (100) 1,107 (100) Fabrics were classified using Canterbury Archaeological Trust codings (Macpherson-Grant et al 1995) as follows: B2 ‘Belgic’ coarse grog-tempered ware B2/R1 Transitional ‘Belgic’ grog-tempered/Native Coarseware B3 ‘Belgic’ grog-tempered with sparse flint B5 ‘Belgic’ grog-tempered with sand B8 ‘Belgic’ fine-sanded ware BER16 ‘Thanet Dry’ ware R1 N ative Coarseware AN IRON AGE AND ROMANO-BRITISH SITE ON WOOD HILL, KINGSDOWN 405 R5 Canterbury coarse grey sand-tempered ware R6.3 Canterbury coarse buff sand-tempered ware R13 Black-burnished ware, BB1 R14 Black-burnished ware, BB2 R16 Fine grey Upchurch ware R17 Fine orange Upchurch (Hoo) ware R36 M oselkeramik R42 South Gaulish Samian R43 Central Gaulish Samian R50 Baetican Dressel 20 amphora fabric R65 Verulamium Region Whiteware R71 M iscellaneous pink-buff wares R73 M iscellaneous coarse grey sand-tempered wares R75 Miscellaneous white/cream wares R98 Miscellaneous amphora fabrics R99 M iscellaneous mortaria fabrics LR1 L ate Roman grog-tempered ware LR1.1 L ate Roman grog-tempered ware with white siltstone grog LR2.1 Thameside fine-sanded grey ware LR2.2 Thameside fine-sanded grey ware with superficial reddening LR2.3 Thameside coarse-sanded grey ware LR3 H arrold shell-tempered ware LR5 Alice Holt/Farnham grey wares LR6 Overwey/Portchester D buff sandy wares LR10 O xfordshire Red-Colourcoat LR13 H adham Oxidised wares LR19 M ayen ware Late Iron Age to c.ad 100: the features of this date were concentrated in Trench 27. The earliest feature here was gully F. 25 which produced 11 sherds (68g) consisting almost entirely of ‘Belgic’ grog-tempered and sand-tempered wares. No rims or other diagnostic sherds are present but a fragment from a Romanised closed form in Upchurch Fabric R16 indicates that the feature was still receiving rubbish after ad 43: it could conceivably be entirely post-Conquest in date. Assemblage 1, from the fill of Pit F. 23 cut into Gully F. 25 (WHK- 82-22) The thirteen sherds from this feature are all in ‘Belgic’ grog-tempered Fabrics B2 and B2.1 and include large fresh sherds from the following two vessels:- Fig. 8.1 Rim from bead-rim jar with corrugated shoulder of Thompson (1982) Type B2-2 in Fabric B2.1 fired brown-black. Ext. rim diameter 210mm. c. 50 bc-ad 70. KEITH PARFITT 406 Fig. 8 Pottery from excavated features (1:4). AN IRON AGE AND ROMANO-BRITISH SITE ON WOOD HILL, KINGSDOWN 407 Fig. 8.2 Round, cordoned jar of Thompson Type B3-5 in Fabric B2 fired buff-brown. Ext. rim diameter 160mm. c. 50 bc-ad 70. Assemblage 2, from the fill of Pit F. 24 cut into Gully F. 25 (WHK-82-34) The fill of this feature produced 92 sherds (978g) in ‘Belgic’ grogtempered Fabric B2 of which all but two were fresh joining sherds making up the greater part of the following pot:- Fig. 8.3 Bead-rim jar of Thompson (1982) Type C1-1 fired patchy brown/ black with combing all over its body. Ext. rim diameter 180mm. Late Iron Age to c.ad 70: the only other feature of this date on the site was pit F. 4 in Trench 8, immediately to the north-west of Trench 27 on the other side of the track. This produced eight sherds (86g) comprising seven fragments in grog-tempered Fabric B2 and a sherd from a jar in silt-tempered ‘Thanet Dry’ Fabric BER16 fired grey-black (c.ad 40-80). c.ad 100-270: there are only two assemblages which could possibly be dated to the period c.ad 100-150: their small sizes, however, makes such precise dating uncertain. The upper fill of ditch F. 2 in Trench 6 produced four sherds (24g) including a fragment from a possible early-secondcentury jar in sandy grey Canterbury Fabric R5: the fill of hollow, F. 18 in Trench 18 yielded a further six sherds including the rim from a lidseated jar in similar fabric. This fragment and that from ditch F. 2 could, however, equally well be dated to the period c.ad 150-175. Examination of the pottery from the field-walking and topsoil also failed to reveal any specifically early second-century forms other than Central Gaulish Samian Dr. 27 cup fragments (c.ad 120-150) and a Dressel 20 amphora rim of Martin-Kilcher Type 15 (1983, c.ad 120-150): a small number of residual early second-century sherds did, however, come from pit F. 3 (Assemblage 4). Trench 8 revealed a complex of inter-cutting pits of late-second to early-third-century date: Assemblage 3, from the fill of Pit F. 5 (WHK-82-16) The 137 sherds (1,107g) from this feature are well mixed and lacking the quantification-distorting factor of large numbers of fresh sherds from just a few vessels. The assemblage was therefore quantified by numbers of sherds and their weights per fabric (Table 2). The biggest single suppliers of pottery to the site at the time that this assemblage was deposited appear to have been the Native Coarseware producers who, from the percentage distribution of their wares, are thought to have been based along the Wantsum Channel west of Richborough and may also have been engaged in sea-salt manufacture. Half the pottery in this assemblage by both sherd count and weight KEITH PARFITT 408 comes from this source, consists entirely of high-fired handmade cooking-pots and can be dated to the period c.ad 170-250/300: this dates the back-filling of the pit to later than ad 170. Another 42 per cent of the pottery comes from the Thameside kilns around the Medway estuary, on the Isle of Sheppey and elsewhere along the south shore of the Thames estuary and comprises open forms and cooking-pots in BB2 Fabric R14, fineware jars, beakers and flasks in Upchurch Fabric R16 and much smaller numbers of jar sherds in Thameside Fabrics LR2.1, 2.2 and 2.3 of post ad 150 date. The BB2 ‘pie-dishes’ consist almost entirely of plain examples of Monaghan (1987) Class 5C, dated c.ad 150/170-250 and include examples of Form 5C1-5 (c.150/170-240) and 5C7-1(c.170-230): examples of the earlier ‘pie-dish’ Class 5D with lattice decoration (c.110/120-180) are completely absent: this further indicates a 170- 230 date for the filling of the pit. The four sherds of samian include large fragments from a Central Gaulish Dr. 31 platter (c.ad 150-200). Assemblage 4, from the fill of Pit F. 3 cutting Pits F. 5 and F. 8 (WHK- 82-14) Pit F. 8 cutting into pit F. 5 was totally lacking in pottery but pit F. 3 which cut both it and F. 5 produced 104 sherds (800g), of which some at least was probably derived from F. 5. There are, however, a greater number of residual late first- to early second-century sherds in this feature, including a Monaghan (1987) Type 2H1 beaker rim (c.ad 90- 130), a BB2 pie-dish of Type 5D4.1 with vertical burnished lines on its body (c.ad 120-180), a South Gaulish samian Dr. 18 platter (c.ad 70-90) and a Martres-de-Veyre samian Dr. 27 cup (c.ad 90-120). Assemblage 5, from the fill of Pit F. 20 (WHK-82-28 and 29) The forty sherds (277g) of pottery from this feature in Trench 20 include eleven joining fragments from the lower part of a flagon in black-cored orange Hoo Fabric R17. In the absence of rim sherds, close dating of this vessel is impossible but it is unlikely to be later than ad 250. Eight featureless pieces from both closed and open BB2 forms and seven equally-featureless sherds from beakers in Upchurch fine ware Fabric R16 are also present and are equally unlikely to be later than 250 in date. A sherd from a straight-sided dish in Dorset BB1 fabric is, however, also present and dates to after 220. A date during the second quarter of the third-century is therefore suggested for the assemblage from this pit. c.ad 270-400: all the late Roman assemblages from the site are small and scrappy with the exception of that from the fills of Quarry F. 14: AN IRON AGE AND ROMANO-BRITISH SITE ON WOOD HILL, KINGSDOWN 409 Assemblage 6, from the fills of Quarry F. 14 (WHK-82-20, 21, 24, 26 and 30) This feature yielded 121 sherds (757g) of badly broken up, mainly late first- to early third-century pottery, nearly all of which came from the upper fills and is probably residual in its contexts. Later material is also present and includes fragments from a developed-beaded-andflanged BB1 bowl (c.ad 270-400), handmade cooking-pots in Late Roman grog-tempered ware fabrics LR 1 and LR 1.1 (c.ad 270-400), an Oxfordshire Red Colourcoat Dr. 38 copy (c.ad 240-400) and a hook-rimmed Class 3C jar in coarse Alice Holt/Farnham ware (c.ad 330-400). The presence of the latter fragment indicates that the rubbish was dumped in the quarry during the late fourth century or even later. The following sherd is of particular interest:- Fig. 8.4 Fragment from shoulder of jar in grog-tempered grey Fabric B2.1 fired black with painted diagonal lines. Painted decoration is uncharacteristic of the ‘Belgic’ grog-tempered wares of east Kent but is found on Late Iron Age East Sussex Ware vessels in similar fabric from Wealden east Kent and East Sussex (Green 1980, 74). Such painted East Sussex Ware vessels have been found in south-east Kent at the Dollands Moor site, near Folkestone (Lyne forthcoming), and at the Harville Farm villa, Wye (late J. Bradshaw pers. comm.). WHK-82-30. Other, much smaller, assemblages came from other features: Assemblage 7, from the fill of ditch F. 1 (WHK-82-13). The thirteen sherds (202g) of pottery from this feature include a rim fragment from a BB1 cooking-pot of post ad 280 date, five abraded sherds in ‘Belgic’ grog-tempered and Native Coarseware and four handmade jar fragments in Late Roman grog-tempered ware Fabric LR 1.1 (c.ad 270-420): large, fresh sherds from a Dr 38 bowl copy in Oxfordshire Red Colourcoated ware (c.ad 240-400) and an evertedrim jar in late BB2 fabric with thick pink margins (c.ad 270-350) are also present in an assemblage which was probably deposited c.ad 270-400. Assemblage 8, from the combined fill of Fs 6, 7 and 9 (WHK-82-17) The ten sherds (58g) of pottery from this combined fill include five fragments from Native Coarseware closed forms (c.ad 170-250/300) and a developed-beaded-and-flanged bowl rim sherd in late Thameside BB2 fabric (c.ad 270-370). This suggests a date of c.ad 270-300 for one at least of these features, if not all of them. KEITH PARFITT 410 Miscellaneous: more than half of all the pottery from the site is unstratified and comes from field-walking and the topsoil of the various trenches. Most of this material is abraded but it does include the following interesting pieces: Fig. 8.5 Rim sherd from Mayen ware bowl of Gose (1976) Type 492, c.ad 350-400. Small but persistent numbers of coarseware jars, bowls and dishes from this Rhineland source were traded across South-East Britain during the mid/late fourth century (Fulford and Bird 1975). Topsoil, Trench 1 (WHK-82-4). Fig. 8.6 Four joining sherds in fine grey Upchurch Fabric R16 from the base of a jar or other closed form. There are little dimples in the interior of the pot around the junction of the sides with the base, the purpose of which is uncertain. Subsoil, Trench 16 (WHK-82- 18). Chaff-tempered Ware (not illustrated) by Keith Parfitt and Geoff Halliwell Some 196 fragments (460g) derived from small, crude ceramic vessels of chaff-tempered ware (Macpherson-Grant 1980) were recovered from the excavated features. The bulk of this material came from two pits, F. 23 (88 pieces) and F. 24 (96 pieces). The date-range of this ceramic-type, now well-known in east Kent, seems to be entirely confined within the first century bc to first century ad. The overall size, shape and function of the vessels represented, however, continues to be uncertain and little new information can be obtained from the present assemblage. The great bulk of the fragments recovered are very small, but portions of the crudely applied collars with protrusions or lugs as recognised at Green Lane, Whitfield (Parfitt 2002, 390-2) again seem to be present. Almost one third of the pieces in the Wood Hill assemblage appear to represent fragments of such collar or lug. There is one largish piece of collar which still encases a rim (now loose), together with the broken, rounded end of a quite neatly shaped projecting lug. Quernstones The excavations produced a total of 11 pieces of broken quernstone. Seven of these are made from local Folkestone Greensand (Fig. 9); three are of Millstone Grit-type sandstone, probably obtained from the Pennines in northern England and there is part of an imported Mayen lava stone quern from the German Rhineland (now represented by eight shapeless fragments). Querns made from all three of these rock-types are wellknown on many other Romano-British settlements across east Kent. The greater frequency of querns made from the local Folkestone Greensand AN IRON AGE AND ROMANO-BRITISH SITE ON WOOD HILL, KINGSDOWN 411 is to be expected. Investigations at East Wear Bay, Folkestone, about 16 km to the south-west of Wood Hill, have demonstrated that Greensand quernstones were being manufactured there throughout the late Iron Age and earlier Roman periods (Keller 1989). Quern fragments were recovered from Fs 4, 5, 18 and 24, together with several topsoil and general soil deposits. Found protruding through the surface of Victoria Road was about one quarter of a top stone in Greensand (Fig. 9.1). This would have had an overall diameter of around 347mm and a thickness of 80mm. In Trench 16 an upper soil deposit produced a small fragment of another top stone in Greensand. This would have had an overall diameter of around 370mm and is 53mm thick (Fig. 9.2). The largest stratified fragment was recovered from the filling of Belgic pit, F. 24. This consists of about one third of an upper stone, again of Greensand (Figs 4 and 9.3). It would have had an overall diameter of around 400mm and is 85mm thick, with a side handle-hole about 25mm across. There Fig. 9 Quernstones from the site (1:4). KEITH PARFITT 412 is a wide and deep central hopper with traces of the central spindle hole remaining. Worn traces of a pecked grinding surface survive. Animal bone (not illustrated) Sixteen of the excavated deposits yielded a total of 102 pieces of domestic animal bone (751g), of which about one quarter are teeth. Most of the material is fragmentary and there are no significant groups. The fills of quarry pit, F. 14 produced almost a quarter of the overall assemblage. Marine shell (not illustrated) A total of 75 marine shells (506g) was recovered from twelve different contexts. The largest groups came from pits, Fs 3 and 5, and quarry pit F. 14. Most common are oyster shells, which totalled 27. Also present are winkle (17), mussel (14 fragments), limpet (12) and whelk (5). These finds indicate that the resources of the nearby seashore were being exploited by the inhabitants of the site, at least on an occasional basis. discussion There can be no doubt that the surface pottery scatter and excavated features relate to an extensive settlement that once occupied the southeastern slopes of Wood Hill at Kingsdown. Habitation here began well before the Roman period. Pottery from a ditch excavated in 1971 indicates some occupation during the late Bronze Age/early Iron Age period (Parfitt 1985, 207, 215-7) and pit F. 15 contained pottery broadly dated to the early/middle Iron Age (Parfitt 1985, 209). The bulk of the features excavated in 1982, however, are datable to between the later first and mid-third century ad. Moreover, sufficient late Roman pottery, including Oxfordshire, Alice Holt/Farnham and Mayen wares, is present to indicate that, unlike many rural sites in east Kent, occupation here continued well into the fourth century. The trenching revealed a significant number of features but no overall plan is apparent from the relatively limited work possible. The full extent of the site was not established although its north-eastern boundary may have been defined. The bulk of the features located were pits and a number of these are likely to have been for rubbish disposal. The extensive pit F. 14 may have been a chalk quarry, whilst the broad shallow hollow, F. 26, could represent a terrace for a timber structure. The general absence of building debris in the form of stone, mortar and roofing tile, indicates that there were no substantial Roman masonry buildings on the site and a native settlement provided with simple timber buildings may be inferred. This is typical of many habitation sites known AN IRON AGE AND ROMANO-BRITISH SITE ON WOOD HILL, KINGSDOWN 413 on the Downlands of east Kent, where farmsteads rather than villas predominate. The comparatively small number of villas which are known in the region tend to be found in the river valleys and at the foot of the Downs. Nevertheless, some of the settlements on the Downland seem to have been quite large, occasionally approaching village proportions. Evidence for Roman activity in the immediate area of Kingsdown is fairly limited. Traces of another probable native farmstead site occupied between the late first and mid third century ad has been previously recorded at the Scout Camp above Oldstairs Bay, some 850m to the north-east (Fig. 1; Parfitt 1995). About 2 km north-west of Wood Hill, on the outskirts of Walmer, the recent discovery of part of a third-century aisled barn, could perhaps indicate the presence of a previously unknown villa complex there (Jarman 2010). acknowledgements Thanks are due firstly to Kingsdown Parish Council and Mr Jim Pearson- Wood, the respective owners for allowing access to the land in their possession. Jonathan Joy and Nigel Macpherson-Grant readily provided details of the surface pottery recovered. Of the Group members who undertook the work thanks are due to Mrs Joan Briggs, Mrs Shirley Creasey, Miss Tina Avann (now Mrs Parfitt), the late Ben Stocker (Group Chairman), and Messrs Geoff Halliwell, Mike Halliwell and Howard Jones, for their hard work over many weekends. The site plans and sections have kindly been prepared for publication by Barry Corke. Dr Malcolm Lyne produced the report on the excavated pottery. bibliography Broady, J., 1989, Deal in Old Photographs (Gloucester). Frere, S.S., 1983, ‘Roman Britain in 1982. Kingsdown’, Britannia XIV, 335. Fulford, M.G. and Bird, J., 1975, ‘Imported Pottery from Germany in Late Roman Britain’, Britannia VI, 171-81. Gose, E., 1976, Gefasstypen der Romischen Keramik im Rheinland, Beihefte der Bonner Jahrbucher Band 1. Green, C.M., 1980, ‘Handmade pottery and Society in Late Iron Age and Roman East Sussex’, Sussex Archaeol Collect., 118, 69-86. Halliwell, G. and Parfitt, K., 1993, ‘Non-River Gravel Lower and Middle Palaeolithic Discoveries in East Kent’, Kent Arch. Rev., 114, 80-89. Jarman, C., 2010, Excavations at Downlands, Walmer, Kent (CAT Occ. Paper no. 7). Keller, P.T., 1989, ‘Quern Production at Folkestone, South-East Kent: an Interim Note’, Britannia 20, 193-200. Lyne, M.A.B., forthcoming, ‘The Late Iron Age and Roman Pottery from the Channel Tunnel Sites’. KEITH PARFITT 414 Macpherson Grant, N., 1980, ‘Chaff Tempered Ware’, Kent Arch. Rev., 61, 2-4. Macpherson-Grant, N., Savage, A., Cotter, J., Davey, M. and Riddler, I., 1995, Canterbury Ceramics 2. The Processing and Study of Excavated Pottery. Martin-Kilcher, S., 1983, ‘Les amphores romaines a huile de Betique (Dressel 20 et 23) D’Augst (Colonia Augusta Rauricorum) et Kaiseraugst (Castrum Rauracense). Un rapport preliminaire’, in Blazquez, J.M. and Remesal, J. (eds), Prod. y Com. del Aceite en la Antiguedad. II Congresso, Madrid, 337- 47. Monaghan, J., 1987, Upchurch and Thameside Roman Pottery. A ceramic typology for northern Kent, first to third centuries A.D., B.A.R. Brit. Series 173 (Oxford). Orton, C.J., 1975, ‘Quantitative Pottery Studies; Some Progress, Problems and Prospects’, Science and Archaeology 16, 30-5. Parfitt, K., 1985, ‘Some Iron Age Sites in the Deal Area’, Kent Arch. Rev., 79, 206-219. Parfitt, K., 1995, ‘A Roman Site above Oldstairs Bay, Kingsdown’, Kent Arch. Rev. 119, 202-4. Parfitt, K., 2002, ‘A Prehistoric site off Green Lane, Whitfield, near Dover’, Archaeologia Cantiana, cxxii, 373-396. Scott-Jackson, J., 2000, Lower and Middle Palaeolithic artefacts from deposits mapped as Clay-with-flints (Oxbow Books). Thompson, I., 1982, Grog-tempered ‘Belgic’ Pottery of South-eastern England, B.A.R. Brit. Series 108 (Oxford).