Later Prehistoric Settlement on the Hoo Peninsula: excavations at Kingsmead Park, Allhallows
LATER PREHISTORIC SETTLEMENT ON THE HOO PENINSULA: EXCAVATIONS AT KINGSMEAD PARK, ALLHALLOWS.
christopher greatorex
Further evidence for the prehistoric settlement of the Hoo Peninsula has been discovered by Archaeology South-East (a division of the University College London Field Archaeology Unit) at Kingsmead Park, Avery Way, Allhallows (NGR TQ 823 780) during the construction of a nine-hole golf course and associated amenities (Fig. 1). The site lies between the 5m and 10m contours just to the west of the Allhallows marshes. The British Geological Survey shows the local geology as London Clay with Head Gravel. However, it should be noted that a geoarchaeological evaluation of the sediments exposed during the fieldwork suggested the presence of complex fluvial or estuarine deposits, some of which may have formed under temperate climatic conditions. The geoarchaeological aspects of the project have been reported on elsewhere (Bates 1997 and Bates et al. forthcoming) and will not be further considered here as they do not relate to the archaeological deposits described below. Fortunately, the site’s location within striking distance of a number of distinct environmental habitats was clear even before the start of fieldwork. Such ‘marginal’ settings are known to have been favoured by prehistoric populations for resource exploitation and settlement. Indeed, previous archaeological discoveries in the vicinity include Paleaeolithic and Neolithic flintwork from the intertidal zone approximately 800m north-west of Kingsmead Park and two Late Bronze Age metalwork hoards (Ewart Park stage) near Allhallows village (Wickham 1877).
The first phase of archaeological work, conducted over 10 days in November/ December 1997, comprised the excavation of twenty-three targeted 20m-long evaluation trenches across those main areas of intended sub-surface ground disturbance associated with the proposed golf course (Fig. 1c). The objective of this exercise, undertaken by a JCB 3CX mechanical digger fitted with a 1.50m-wide ditching bucket, was the evaluation of both the archaeological and geoarchaeological potential of the site prior to the onset of golf course construction. This report summarises the main findings of both the evaluation (Greatorex 1997) and subsequent watching brief at the site. Full details of all the features, finds and environmental material are housed with the archive, to be deposited in Rochester Museum.
Trial Trenching
The topsoil across the site ranged in thickness from 0.25-0.45m. Within most of the trenches, a natural orange-brown, sometimes sandy silt clay was located immediately below the topsoil; however, a group of trenches (Numbers 7, 8, 17, 18, 19, 20 and 21) did contain an additional 0.10m thick mid grey-brown silty clay subsoil between the topsoil and subsoil. It was soon apparent that eight of the trenches (Numbers 6, 7, 15, 16, 18, 20, 21 and 23) contained a number of significant archaeological features cut into the natural. As the fieldwork brief was simply to establish the general character, date, depth and extent of any identified deposits or structures, only a representative sample of these features was examined by excavation. Even so, the limited manual investigation enabled the division of all revealed features into one of three broad classes for ease of further discussion. Trenches 1-5, 8-14, 17, 19 and 22 were devoid of archaeological features.
Confirmed prehistoric features (Figs 2 and 3: Trenches 15, 16, 20 and 21)
Ten of the features yielded pottery assigned to the Late Bronze Age/Early Iron Age (c.900/800-600 bc), either from the surface of their fills during cleaning or during sectioning. Sherds of this date were recovered from a range of feature types including two probable post-holes found within Trench 16 (Fig. 2: Cut 98/ Fill 99 (Fig. 4, S1) and Cut 100/ Fill 101), suggesting the presence of structures at the site. A third post-hole located within the same trench (Cut 104/ Fill 105), which although artefactually sterile, was virtually identical in plan and profile to Cut 98 and thus may be contemporaneous.
Additional Late Bronze Age/Early Iron Age (LBA/EIA) pot sherds and five fragments of a single perforated fired-clay slab were recovered from the top of an oval-shaped feature running beneath the northernmost baulk of Trench 16 (Fig. 2: Cut 102/ Fill 103). The feature probably represents a pit or large post-hole. Five other cuts of confirmed LBA/EIA date can more confidently be interpreted as pits of some description (Figs 2-4: Trench 18, Cut 11/ Fill 12 (S2) and Cut 13/ Fill 14 (S3), Trench 21, Cut 30/ Fill 31, Trench 15, Cut 73/ Fill 74 and Cut 75/ Fill 76). These features were all apparently sub-circular in shape with maximum documented diameters ranging between approximately 0.60m (Cut 13) and 1.60m (Cut 75). On investigation the two pits uncovered within Trench 20 (Cuts 11 and 13) were each discovered to possess rather irregular profiles (Fig. 4) and contain fire-cracked flint and charcoal flecks. The final couple of probable LBA/EIA pits, revealed within Trench 15 (Cut 73/ Fill 74 and Cut 75/ Fill 76), were not sampled by excavation.
Two linear ditches or gullies (Trench 20, Cut 7/ Fill 8 and Trench 21, Cut 40/ Fill 57) of LBA/EIA date were also located. Cut 7 comprised a 0.60m wide, 0.20m deep, e-w feature with a splayed ‘U’ profile (Fig. 4, S4). Cut 40 on the other hand was approximately 0.85m wide by 0.2m deep and ran in a se-nw direction. This ditch was of particular interest in that it provided the sole relative dating evidence for two separate deposits of cremated bone inserted at least partially into the top of Fill 57 (see below).
One prehistoric feature produced pottery of a considerably earlier period. Cut 9, a linear or perhaps lozenge-shaped feature, discernible within Trench 20 (Fig. 3) was found to have a maximum recorded length of 1.65m, width of 1m and depth of 0.12m (Fig. 4, S5). Its only fill (Fill 10) produced a few sherds of pottery dated to the Middle Neolithic (c.3000 cal bc).
Funerary contexts (Fig. 3: Trench 21)
Three small circular cuts located within Trench 21 were found to contain a mixture of cremated human bone, charcoal and soil (Cut 32/ Fill 33, Cut 51/ Fill 52 and Cut 106/ Fill 107). Contexts 51 and 106 both had diameters of approximately 0.40m with steep sides and flat bases (Fig. 4, S6 and S7), while Feature 32 really only survived as a poorly-defined irregular scoop with a maximum depth of 70mm. It should be noted that the shallow depths of these cuts suggested relatively heavy truncation had occurred. However, even accounting for possible truncation, the tiny quantities of calcined bone gleaned from Fills 33 and 52 (0.40 and 0.70g respectively) indicate the token redeposition of pyre-debris rather than the ‘ashes’ of whole skeletons. By way of contrast, an analysis of the much larger bone assemblage retrieved from Fill 107 (543.9g) points towards the buried remains of one or maybe two cremated individuals. The exact date of these three unurned funerary contexts is open to question. The only artefact recovered during their full excavation was a single scrap of Late Bronze Age/Early Iron Age pottery (Fill 33). Of course, such a find, together with the contemporaneous origin of most of the surrounding features, strongly suggests a broadly similar prehistoric date of deposition for the cremation material. Yet this evidence is complicated by Contexts 32 and 106 clearly being cut into the top of a feature which itself yielded Late Bronze Age/Early Iron Age material (Cut 40/ Fill 57). It is thus entirely possible that the sherd attributed to Context 33 was in fact derived from the disturbed fill of Ditch 40. It is thus possible that Cuts 32 and 106 may even represent a distinctly later, perhaps Roman, phase of activity across the site. However, the lack of Roman material from the evaluation, together with the evidence noted above, would strongly suggest the cremated deposits relate to the later prehistoric period.
Undated features (Figs 2 and 3: Trenches 6, 7, 15, 16, 18, 20, 21 and 23)
The date of the thirty-four outstanding documented archaeological features remains uncertain. These contexts included ditches and gullies (Cuts 16, 22, 34, 38, 47, 53, 55, 60, 62, 64, 71, 85, 89 and 95), probable pits (Cuts 20, 24, 41, 43, 49, 69, 78, 81, 83, 108, 110 and 112), three possible post-holes (Cuts 36, 45 and 66) and two almost certain post-holes (Cuts 87 and 104). One amorphously-shaped and indeed poorly-defined cut (18) located within Trench 20 defied any such categorisation. Other features included a sub-circular, 0.27m diameter, patch of burnt natural clay in Trench 15 perhaps indicative of a hearth (Fig. 2, Context 77), while a diffuse spread of charcoal (Context 127) was also exposes in Trench 18 (Fig. 3). Full details of all deposits recorded during the fieldwork are lodged with the Archive. The ascertained physical relationships between undated features, where discernible, has been marked on Figures 2 and 3. These illustrations demonstrate that only one of the contexts of uncertain origin (Trench 21, Cut 41) was found to have a clear stratigraphic relationship with a separate, defined deposit yielding diagnostic artefacts (predating Cut 40).
Watching Brief
On completion of the trial-trenching it was obvious that the site proposed for golf course construction, contained dense concentrations of significant archaeological remains meritorious of either excavation or preservation in-situ. Perhaps most problematical was a particular concentration of features found in an area initially intended as the location of a small lake (Fig. 1c). However, after discussions between all concerned parties, the following mitigation measure were agreed, outlined in a fresh Project Brief prepared by the Heritage Conservation Group, Kent County Council. Golf tees, greens and bunkers were built up above original ground level rather than being cut through to the natural. All pipework was laid in the top 0.40m of site formation and tree planting and irrigation located to hopefully best avoid identified clusters of archaeological deposits. More fundamentally, the aforementioned lake was moved north-eastwards, away from the important contexts recorded during the evaluation and its excavation made the subject of a continuous watching brief. Indeed, the top soil stripping of all golf course features and the intrusive groundworks required for the planned site access road, car park and clubhouse were also monitored intermittently by Archaeology South-East.
The removal of topsoil over the area of the relocated lake was conducted during January 1998. As requested, the exercise was carried-out using a back-acting mechanical digger fitted with a wide flat-bladed bucket. An approximately 0.25m-thick overburden was rapidly excavated revealing the immediately underlying natural clay dissected by a single sub-circular feature (Cut 131/ Fill 132). Although the edges of Cut 131 were extremely diffuse, this possible pit was estimated to possess a diameter of roughly 0.70m, and depth of just 14mm. Its investigation yielded four worn scraps of Late Bronze Age/Early Iron Age pottery. No other significant deposits or artefacts were observed during the creation of the lake or indeed the subsequent programme of monitoring undertaken across the rest of the site. The largely negative results from the monitoring confirmed that the development had been achieved with only limited impact on the archaeological resource and that the large part of an apparently substantial LBA/ EIA settlement remains below the new golf course.
The Finds
The Neolithic Pottery by Alex Gibson
Two recently broken conjoining sherds plus one other crumb of pottery from Trench 20, Cut 9/ Fill 10) were submitted to the writer for identification and comment. The fabric was examined macroscopically only and identified by form, fabric and decoration to be from the rim of a vessel in the Peterborough tradition of the Middle Neolithic.
The fabric is quite well-fired and, though abraded, averages some 6mm thick. Both calcined flint and crushed shell inclusions, reaching up to 5mm across, could be identified in the breaks as well as erupting through both surfaces. These white-grey inclusions give the generally reddish-brown surface a speckled effect.
The fresh break indicates that the sherd had a short firing. The surfaces are reddish-brown to a maximum depth of 1.5mm while the core is jet black indicating that the organic inclusions in the clay have not been fully burnt out. No other technological indicators are visible.
The rim is expanded and slopes slightly inwards (Fig. 5). The inside of the rim has a marked lip giving to an internally convex neck. The outer rim moulding is rounded. The top of the rim is decorated with short radial impressions of either whipped cord or bird-bone; the degree of abrasion is too great to be precise. Traces of four impressions are visible, each averaging 4mm across and 8mm long. Internally, the neck below the lip of the rim also carries very faint traces of similar short, vertical maggot impressions. Traces of the internal structure in one of the impressions suggest twisted cord, but once more the sherd is too abraded to be certain. The rounded outer rim moulding is decorated with traces of three maggot impressions, sloping slightly from bottom left to top right. In these impressions the internal structure of the maggots is better preserve and traces of cord twists can be detected.
The sherd undoubtedly belongs to the Peterborough Ware or Impressed Ware tradition of the Middle Neolithic. Too little of the vessel survives to allow the sub-type to be recognised with certainty. The relatively simple rim form may suggest that it belongs to the undeveloped Ebbsfleet style but the density of the impressed decoration may make a developed Mortlake Style identification more likely. Present dating evidence for this style or pottery (both Ebbsfleet and Mortlake styles) suggests a currency in the few centuries either side of 3000 cal bc (Gibson & Kinnes 1997).
The Allhallows sherd brings the Peterborough Ware findspots in Kent to 15. Ebbsfleet style pottery has been found at nine of these sites and sherds in the Mortlake style at five. Both styles are represented at Baston Manor (Philp 1973). Close parallels for the Allhallows vessel, particularly with its less-developed rim form, come from the Ebbsfleet type site itself (Piggott & Burchell 1939), Baston Manor, Cheriton (Smith & Philp 1975), Castle Hill Folkestone, Grovehurst and Tankerton (Greenfield 1960).
The distribution of Peterborough Ware in Kent has a markedly eastern and coastal pattern though one suspects that, since many of the finds are recent, this reflects more current development trends than archaeological reality. Most discoveries of Peterborough Ware still comprise individual stray finds and, other than the Ebbsfleet and Baston Manor sites, well-stratified assemblages with associated lithics are few.
The Late Bronze Age/ Early Iron Age and later pottery
by Nigel MacPherson-Grant
Apart from the very small unstratified medieval assemblage, which prob-ably represents settlement fringe field manuring (details archived), and the Neolithic sherds discussed above all the pottery from the site is of the Late Bronze Age/Early Iron Age (c.900/800-600 bc).
Despite small sherd sizes and low count of diagnostic elements there is sufficient to confirm the LBA/EIA date for the main period of activity represented. The thin-walled and generally profusely flint-tempered nature of the sherds recovered is typical of the period; one fineware sherd has traces of an inner rim bevel which is equally consistent with formal trends for this period. If there was any doubt it is removed by the presence of scrappy sherds (from Trench 16, Context 103) from two perforated slabs, diagnostic type fossils for this period; one is flint-tempered, the other organic-tempered (both recognised fabric variations for this light industrial artefact type). In addition there is a sherd from an oxidised organic-tempered object which may well be, though not necessarily, associated with salt-production.
Rather more intriguing are sherds from Trench 15, Contexts 74 and 76 (which are probably from the same object). Surface curvature suggests the sherds are not from a loomweight but from a rounded element; use-sooting on this surface suggests use in either food or some light industrial preparatory process. Original form and exact function are not determinable.
The Cremated Bone by Jacqueline McKinley
Cremated bone from three possible burial contexts was received for analysis; although undated, the deposits are believed to be Late Bronze Age/Early Iron Age. Osteological analysis followed the writer’s standard procedure (McKinley 1994a). Age was assessed from the stage of skeletal development (McMinn and Hutchings 1985). All three deposits had been substantially truncated, the maximum surviving depth of the cuts being between 0.07m and 0.10m; consequently, it is likely that some cremated bone may have been lost. All three deposits comprised a mix of charcoal and cremated bone, with no clear concentrations of the latter in any instance.
The very low level of bone recovery from two of the contexts – 0.4g from 33 and 0.7g from 52 – together with the mixed archaeological components within the deposits, suggests they may represent redeposited pyre debris rather than the remains of cremation burials. The 543.9g of bone from Context 107 represents the remains of an adult of 18-45 yr., possibly female; evidence to suggest the presence of remains from an infant/juvenile were also recovered, but this was inconclusive.
The small maximum fragment size (40.5mm) and the recovery of the majority of the bone in Context 107 from within the 5-10mm range (52%), is probably more indicative of poor bone preservation than of deliberate fragmentation prior to burial, though the latter cannot be fully ruled-out (McKinley 1994b). All the bone is slightly worn with a chalky surface appearance, and this, together with the absence of spongy bone elements, is indicative of adverse conditions (acid) in the burial environment (McKinley 1997, p. 245). The bone was buff-white in colour (oxidised; Holden et al 1995). Elements from each skeletal area were recovered from 107 and there is no evidence to suggest any deliberate selection of specific bones for burial.
discussion
The investigations at Kingsmead Park uncovered a number of prehistoric archaeological features, the majority being found on the higher ground towards the centre of the southern half and south-western corner of the golf course development. Although these deposits were recorded in relatively high densities, the absence of any formal ‘open area excavations’ focused on the two main zones of identified interest pre-cludes detailed interpretation of settlement function and morphology. Particularly problematic is the impossibility of directly associating individual features exposed within separate trenches (i.e. continuations of ditches/gullies or post-hole distributions across more than one trench). Consequently we are left with a series of tantalising, but rather isolated, insights into the character of an obviously intensively utilised settlement. Nevertheless, the broad phases of activity have been ascertained.
The recovery of the Mid Neolithic pottery demonstrates the exploitation of the region during the few centuries either side of 3000 cal. bc. Unfortunately, due to its abraded nature it is uncertain if this is a residual piece in a later feature or is associated with actual archaeological features at the site. Even so, the Allhallows find does at least bring the confirmed Peterborough Ware findspots in Kent to fifteen and certainly comprise the earliest discoveries made throughout the current archaeological fieldwork.
Fortunately, the evidence for the Late Bronze Age/Early Iron Age (900/800-600 bc) utilisation of the site is far less ambiguous. It is apparent from the range of feature types (pits, ditches, gullies, post-holes, a possible hearth) and artefacts (pottery, probable light industrial or kitchen items) that a potentially low-status, yet fairly extensive prehistoric settlement was situated at this site. Unsurprisingly, given the ‘keyhole’ nature of the evaluation trenches, no structures, fence-lines, land plots or coherent ditch systems could be identified amongst the contexts dated to this period. Thus although sited on what was almost certainly prime agricultural land within easy reach of marine/estuarine resources (even accounting for coastal change), the precise function of the individual features found at Kingsmead Park is unclear. Consequently, many wider issues inescapably grounded in such fundamental interpretations, such as the nature, size and duration of the occupation remain a matter of conjecture until further, more extensive, archaeological work is undertaken at the site.
The three unurned deposits of cremated human bone and charcoal recorded suggests the site, at some point, had a ritual/funerary function. The location of these cuts within the confines of a confirmed Late Bronze Age/Early Iron Age settlement and the recovery of a contemporaneous pottery sherd from one of the features does suggest a similar prehistoric origin for this funerary activity. Certainly it is true that during the Later Bronze Age most people were probably ‘cremated with their ashes scattered or buried in shallow pits without any pots’ (Parker Pearson 1996). However, as two of the Kingsmead Park cremations dissected the top of a known Late Bronze Age/Early Iron Age ditch, they may relate to the final use of the site for burial just before, or as part of, its abandonment. The possibility of these cremated remains being of Roman date, based on current evidence, is thought unlikely.
It is often stated that the distribution of metalwork and documented Later Bronze Age settlements in Kent, including the regularly quoted sites of Mill Hill (Stebbing 1934), Minnis Bay (Worsfold 1943), Richborough (Bushe-Fox 1949) and Highstead (Tatton-Brown 1976) illustrate the extensive exploitation of the county’s river valleys and north/north-eastern extremities. Clearly the discoveries at Kingsmead Park conform with this established pattern. Indeed, even more recent fieldwork undertaken in the area has revealed a small Later Bronze Age (c. 1500-1000 bc) linear feature and pit at Middle Stoke (James 1998 and forthcoming), a series of superimposed ditches of similar date near Malmayes Farm (James 1999), possible Bronze Age droveway and settlement at Kingsnorth Power Station (Johnson 1999 and Griffin and Griffin in prep.). The evidence when taken from all of these sites confirms the intensive prehistoric exploitation of the Hoo Peninsula perhaps from as early as the Mid Neolithic, but certainly throughout the Bronze Age and into the Early Iron Age. It would appear that after the Late Bronze Age/Early Iron Age the site at Kingsmead Park was not utilised intensively, particularly for occupation. It is probable the land was used primarily for pasture and, at times, arable cultivation until the development of the new golf course.
acknowledgements
The author would like to thank Lis Dyson of Kent County Council and Luke Barber of Archaeology South-East. Thanks is also given to the developer, Kingsmead Parks Ltd, for funding the work and Bill Hales (Laurel Consultants) and Ted Awcock (Artington Golf) for practical help throughout the fieldwork. The specialist contributions of Nigel Macpherson Grant, Alex Gibson, Jaqueline McKinley, Martin Bales and Frances Wenban-Smith are also gratefully acknowledged. The programme of trial trenching was undertaken by Christopher Greatorex and Mike Seager-Thomas, while the drawings were prepared by Justin and Jane Russell.
bibliography
Bates, M., 1997, ‘A Geoarchaeological Evaluation of Test Pits from a site at Allhallows, Hoo Peninsula, North Kent’ (unpublished Developer Report).
Bates, M., Keen, D., Whittaker, J., Merry, J. and Wenban-Smith, F., forthcoming, ‘Middle Pleistocene Molluscan and Ostrocod Faunas from Allhallows, Kent’, Proceedings of the Geologist’s Association.
Bushe Fox, J.P., 1949, ‘Fourth report on the excavations of the Roman fort at Richborough’, Society of Antiq. London.
Gibson, A.M. and Kinnes, I.H., 1997, ‘On the urns of a dilemma: radiocarbon and the Peterborough problem’, Oxford Journal of Archaeology, 16 (1), 65-72.
Greatorex, C., 1997, ‘An Archaeological Field Evaluation of Land Proposed for Development at Kingsmead Park, Avery Way, Allhallows, Kent’, Archaeology South-East Archive Report No. 791.
Greenfield, E., 1960, ‘A Neolithic pit and other finds from Wingham, East Kent, Archaeologia Cantiana, 74, 58-72.
Griffin, N. and Griffin, F. (in prep.), ‘Excavation of a Late Prehistoric Settlement at Kingsnorth Power Station’, Archaeology South-East Archive Report No. 1129.
Holden, J.L., Phakey, P.P. and Clement, J.G., 1995, ‘Scanning electron micro-scope observation of incinerated human femoral bone: a case study’, Forensic Science International, 74, 17-28.
James, R., 1998, ‘A Post-excavation Assessment on the Archaeological Invest-igations on the A228 Stoke Bends Improvement, Middle Stoke, Kent’, Archaeology South-East Archive Report No. 853.
James, R., forthcoming. ‘Archaeological investigations at Middle Stoke, 1995 and 1998’, Archaeologia Cantiana (Archaeology South-East Report No. 853).
James, R., 1999. ‘An Archaeological Evaluation at Malmaynes Hall Farm (Kingsnorth Pipeline), Stoke, Kent.’ Archaeology South-East Archive Report No. 1001b.
Johnson, C., 1999, ‘An Archaeological Post-Excavation Assessment of Damhead Creek Kingsnorth Power Station and Associated Works, Hoo St Werburgh, Kent.’ Archaeology South-East Archive Report No. 1001.
McKinley, J.I., 1994a, The Anglo-Saxon cemetery at Spong Hill, North Elmham Part VIII: The Cremations, East Anglian Archaeology No. 69.
McKinley, J.I., 1994b, ‘Bone fragment size in British cremation burials and its implications for pyre technology and ritual’, JAS, 21, 339-342.
McKinley, J.I., 1997, ‘Cremated human bone from burial and pyre-related contexts 244-252’, in Fitzpatrick, A.P., Archaeological excavations on the route of the A27 Westhampnett bypass, West Sussex, 1992. Volume 2: the cemeteries, Wessex Archaeology Report No. 12.
McMinn, R.M.H. and Hutchings, R.T., 1985, A colour atlas of human anatomy, Wolfe Medical Publications.
Parker Pearson, M., 1996, Bronze Age Britain, Batsford Ltd/English Heritage, 50.
Philp, B.J., 1973, ‘Site I: A Neolithic Site near Baston Manor, Hayes, Kent’, Excavations in West Kent 1960-1970, Second Research Report in the Kent Series.
Piggott, S. and Burchell, J.P.T., 1939, ‘Decorated Neolithic pottery from the bed of the Ebbsfleet, Northfleet, Kent’, Antiquaries Journal, 19, 405-420.
Smith, I.F. and Philp, B.J., 1975, ‘A new Neolithic site at Cheriton, Folkestone’, Kent Archaeological Review, 42, 42-45.
Stebbing, W.P.D., 1934, ‘An early Iron Age site at Deal’, Archaeologia Cantiana, 46, 207-209.
Tatton-Brown, T., 1976, ‘Excavations in 1976 by the Canterbury Archaeological Trust’, Archaeologia Cantiana, 92, 235-244.
Wickham, H., 1877, ‘Celtic remains found in the hundred of Hoo’, Archaeologia Cantiana, 11, 123-125.
Worsfold, F.H., 1943, ‘A report on the late Bronze Age site excavated at Minnis Bay, Birchington, Kent, 1938-1940’, Proceedings of the Prehistoric Society, 9, 28-47.
Fig. 1 Site and Trench location plans.
Fig. 2 Plans of Trenches 6, 7, 15 and 16.
Fig. 3 Plans of Trenches 18, 20, 21 and 23.
Fig. 4 Selected sections.
Fig. 5 The Neolithic Pottery.