A Prehistoric Burnt Mound' Site at Crabble Paper Mill Near Dover

219 A PREHISTORIC ‘BURNT MOUND’ SITE AT CRABBLE PAPER MILL, NEAR DOVER KEITH PARFITT In connection with plans for a new housing development (now called Mill Race) at the former Crabble Paper Mill near Dover, Canterbury Archaeological Trust was engaged to undertake an investigation of the site during 2002. Initial evaluation trenching produced some significant results, which led to further excavations on the line of the new estate road, followed by an intermittent watching-brief during the groundworks phase of the subsequent building work. Significant archaeological evidence for activity on the site during the Neolithic/Bronze-Age period was recorded, together with the remains of previously unknown medieval masonry structures (Parfitt 2002; Parfitt and Corke 2003) and important information relating to the early post-glacial environmental history of the Dour valley. The present paper is concerned with the Neolithic/Bronze- Age discoveries; the medieval remains and palaeo-environmental data will be detailed in other reports. The nineteenth-century mill buildings, which have been largely retained in the new development, lie within the historic parish of River and occupy a roughly rectangular plot of ground at the foot of Old Park Hill, adjacent to the River Dour, about 2.5km north-west of Dover town centre (Figs 1 and 2). NGR TR 2995 4311, centre. The plot has maximum dimensions of 130 x 160m and is bounded by the Dover-Canterbury railway line on the north-eastern (uphill) side, Crabble Road on the south-east side, the River Dour on the south-west and Kingston Close on the north-west. The elevation ranges from 30m OD along the north-east side to 19m adjacent to the river. The brick mill buildings are largely confined to the lower, southern half of the site, adjacent to the river, leaving the remainder of the plot essentially undisturbed (Fig. 3). The natural geology of the area is complex and variable, consisting of flint river gravels, tufas and peats, with head brickearth on the higher parts of the site. Some study of this geological sequence had been previously undertaken (Barham and Bates 1990; Bates and Barham 1993) and further investigations were conducted by Dr Bates as part of the present project, KEITH PARFITT 220 Fig. 1 Map of the Dour valley showing position of Crabble and other Neolithic and Bronze-Age discoveries in the Dover area, with inset location map. A PREHISTORIC ‘BURNT MOUND’ SITE AT CRABBLE PAPER MILL, NEAR DOVER 221 Fig. 2 Map of the Crabble area showing the site in relation to relief and the River Dour. Scale, 1:10,000. KEITH PARFITT 222 Fig. 3 Overall plan of the Crabble Paper Mill site showing location of excavated trenches and selected features. Prehistoric site, Trenches 7-10. A PREHISTORIC ‘BURNT MOUND’ SITE AT CRABBLE PAPER MILL, NEAR DOVER 223 with significant results (Bates et al. forthcoming). Particular interest attached to this region because it provided an area where it was possible to examine the junction and relationship of the slope-wash deposits with the valley bottom sediments (Barham and Bates 1990, 54). The Neolithic/Bronze-Age Site (Figs 4-6) A series of prehistoric features and deposits, dated to the Neolithic/Bronze- Age period, were located at the eastern corner of the plot, occupying a sloping spur of brickearth overlooking the river at an elevation of between 23.50 and 21.50m above OD. These remains appeared to relate to a larger site that extended uphill, beyond the limits of the investigated area (Fig. 3). The excavated features consisted of a group of fairly shallow, circular pits (Fig. 4, Fs 102, 105, 109, 133, 138, 140, 151 and 153; see Table 1 for details). These had all been cut into the top of the brickearth and lay buried at a depth of between 1.00 and 2.50m below present ground level, sealed by later down-washed soils (Fig. 6). Each pit was filled with deposits of black ashy soil containing much charcoal and large quantities of calcined flint fragments but virtually no other datable finds. Charcoal samples from the fillings of Fs 102 and 133 were submitted for radiocarbon dating and the results obtained indicate a Late Neolithic/Early Bronze-Age date (see below). Their very distinctive ashy fills leave little doubt that these pits are all broadly contemporary. Nevertheless, pit F. 105 was partially cut through F. 109 (Fig. 4) suggesting that they represent a sequence of individual events, rather than a single episode of multiple pit digging. None of the pits showed any evidence of burning on their sides or base to suggest that they had once contained fires (Table 1). On their downhill side, the pits were surrounded by an ‘apron’ of dumped ashy soil (Figs 4-6, contexts 55 and 96). This deposit was up 0.40m thick and enclosed an area measuring about 32m across (Figs 3 and 4). It followed the natural slope of the ground, falling away to the west, the south and the east (Figs 5 and 6). The composition of the dump deposit was very similar to the filling of the pits and again contained abundant ash and charcoal fragments with very considerable quantities of calcined flint fragments. The deposit also produced some fresh prehistoric struck flints and five very small fragments of flint-tempered prehistoric pottery (see below). On the west side, a charcoal sample from context 55 (Fig. 5), was submitted for radio-carbon dating and this gave a result similar to those obtained from the pits (Table 3). On the south side, the remains of a pit [F. 151] were actually sealed under the ashy dump (Fig. 4). On this side, too, the dumped material (96) also extended over the top of a deposit of natural tufa, clearly indicating that the tufa was of an earlier date (see below). KEITH PARFITT 224 Well beyond the main site, about 50m to the west of F. 153, another deeply buried pit, F. 37, was found cut into the top of the natural tufa deposit in Trench 5 (Fig. 3). The sides and base of this shallow pit showed some evidence of being burnt but the filling of grey clay loam with charcoal flecks produced no datable finds. The stratigraphic position of this isolated westerly feature suggests that it is of prehistoric date, although whether it is contemporary with the main pit group remains unclear. A trench cut by Barham and Bates in 1990 just west of the main site (Fig. 3) had revealed another, much larger pit, perhaps a quarry, cut into the natural tufa (Barham and Bates 1990, 101, fig. 28). This was about TABLE 1. DETAILS OF EXCAVATED PREHISTORIC PITS F. No. Shape Length (m) Width (m) Depth (m) Sides Base Notes 37 circ. dia.= 0.65 0.14 sloping dished Burnt sides and base; undated; Trench 5 102 circ/ oval 1.08 0.99 (min) 0.52 steep dished Not fully exc. C14 date, 3800 ± 40 BP 105 circ. dia. = 0.98 0.32 steep dished Cuts F.109 109 oval 1.35 0.96 (min) 0.33 steep/ sloping dished Cut by F.105 133 oval 1.18 1.00 0.40 (min) steep flat Cut by ditch, F. 97. C14 date, 3960 ± 40 BP 138 circ. dia. = 0.30 0.29 (min) steep point ? Post-hole Top truncated 140 oval 0.40 0.33 0.40 (min) sloping round ? Post-hole Top truncated 151 oval 0.72 0.42 (min) 0.35 sloping dished Under 96. Cut by ditch, F. 146 153 circ. dia. = 0. 50 0.65 steep round ? Post-hole A PREHISTORIC ‘BURNT MOUND’ SITE AT CRABBLE PAPER MILL, NEAR DOVER 225 Fig. 4 General plan of the prehistoric site (Trenches 7-10) showing location of pits and layer of burnt debris (stippled). KEITH PARFITT 226 6m across and more than one metre deep. Its exact purpose and date remain uncertain but its stratigraphic position, again deeply buried below hillwash, suggests that it is prehistoric. Earlier Activity on the Site Examination showed that the natural tufa deposits which occurred on the site sealed the head brickearth. This observation is of some geological interest (see Bates et al. forthcoming). Here, however, we may usefully note the discovery in Trench 9 of two prehistoric flints (an undiagnostic hard hammer flake and a struck fragment) contained within the disturbed upper zone of the brickearth where it was sealed below the tufa. The deposit producing these flints also contained occasional flecks of charcoal, all of which suggests at least casual human occupation in the area prior to the deposition of the tufa, which itself can be shown to predate the Late Neolithic/Early Bronze-Age remains described above (see Bates and Barham 1993 for a general description of tufa formation in the Dour valley). In addition, analysis of environmental samples from the site produced some fragments of marine mussel shell, fairly certainly brought in from the coast by humans (Bates et al. forthcoming). These shells and the two flints from the brickearth, together with several other specific struck flints discovered in later deposits, could all be derived from activity in the area during the Mesolithic period. Fig. 5 Detailed section across the western part of the prehistoric site (Section No. 14). A PREHISTORIC ‘BURNT MOUND’ SITE AT CRABBLE PAPER MILL, NEAR DOVER 227 Fig. 6 Detailed section across the eastern part of the prehistoric site (Section No. 20). KEITH PARFITT 228 Later Features and Deposits (Figs 4-6) The Late Neolithic/Early Bronze-Age site was sealed by a thick sequence of down-washed soils (Figs 5 and 6). Interspersed amongst these hillwash deposits, on various horizons, were a number of later archaeological features and structures. After the abandonment of the prehistoric site, layers of brown clay (Fig. 6, context 147) and dark brown clay loam (Fig. 5, context 54) had developed over the apron of burnt debris and these deposits must effectively represent an early ‘topsoil’, formed after activity on the site had ceased. Moderate quantities of burnt flint, presumably derived from the deposits below, were contained within these layers, together with three struck flints and three more undiagnostic sherds of prehistoric flint-tempered pottery from context 54. Probably some considerable time later, a small ditch was dug across the area (Fig. 4, F. 146). This cut through one of the prehistoric pits [F. 151], the burnt debris sealing it (96) and the overlying soil (147). The ditch ran north-west by south-east along the hillside and was traced in the excavation for a minimum distance of 21.50m before being destroyed by later features at the north-western end (Fig. 4). No datable finds were recovered from the lower filling of the ditch but the upper filling yielded some animal bone, fifteen prehistoric struck flints and four small pot-sherds. The sherds consist of two somewhat indeterminate prehistoric flint-tempered pieces, a fragment of late Iron- Age/early Roman grog-tempered ware and a small chip of Roman samian ware. Based on this very limited evidence, a late Iron-Age or early Roman date for the digging of the ditch seems possible. Most probably, it served as some sort of field boundary and a continuation of it may be represented by F. 43 located running along the hillside on a slightly different axis, further to the west in evaluation Trench 6 (Fig. 3). On the eastern side, a layer of sterile down-washed soil, consisting of a cream-brown clay loam subsequently sealed all these earlier deposits and features (Fig. 6, context 95). The deposition of this layer may well have been caused by ploughing further up-slope but the absence of finds within it precludes any precise dating. A Roman or early Anglo-Saxon date might be suggested on general sequential grounds. Sometime after layer 95 had become stabilised, two further pits [Fs 111 and 113] and a broad ditch [F. 97] were cut into it (Fig. 6). These produced no datable finds. After these features had been infilled they were sealed by more deposits of sterile hillwash (Fig. 6, contexts 92 and 94). Next, most probably sometime during the thirteenth century, a substantial masonry building was erected across the area. A series of other medieval structures, features and deposits appear to be broadly contemporary and this complex will form the subject of a future report. A PREHISTORIC ‘BURNT MOUND’ SITE AT CRABBLE PAPER MILL, NEAR DOVER 229 THE PREHISTORIC FINDS All the finds from the site will shortly be deposited at Dover Museum, together with a copy of the site archive. Flintwork (Fig. 7) by Chris Butler A small assemblage of 148 pieces of worked flint weighing 4.9kg was recovered during the work at Crabble (Table 2). About half this material came from the apron of burnt debris (contexts 55 and 96) around the prehistoric pits, although none of the flints are themselves burnt. The only pit to produce any flintwork was F. 109, which yielded an unburnt hard hammer flake and a miscellaneous fragment. The remaining flints were found as small groups in various down-washed soil layers and later feature fills and many are likely to be derived from the higher slopes of Old Park Hill where surface scatters of lithic material have been previously located (Cross and Parfitt 1999). Four different types of flint were noted: Black coloured flint with variations in shade through to a dark grey, occasionally with grey flecking. This flint generally has a rough grey pebble cortex or a smooth buff cortex, suggesting it probably comes from two different original sources. Mid to dark grey flint, sometimes almost an olive colour, with lighter mottled grey patches. Cortex is a smooth buff colour. Light blue-grey mottled flint. This flint type normally has a buff coloured cortex. White to light grey patinated flint with a buff cortex. Internally the flint is either black or dark grey. Chalk Downland flint. The raw material is typical of the flint found on the local Chalk Downs and its associated Clay-with-Flint deposits. The only exceptions are a few examples of pebble flint, which may have come from a local beach or river gravel source. TABLE 2. DETAILS OF PREHISTORIC STRUCK FLINTS FROM CRABBLE PAPER MILL Hard hammer-struck flakes 84 Soft hammer-struck flakes 26 Hard hammer-struck blades 4 Soft hammer-struck blades 2 Soft hammer-struck bladelets 2 Fragments 22 Shattered pieces 2 Two platform flake core 1 Three platform flake core 1 Core fragments 2 Side-and-end scraper 1 Discoidal knife 1 Total 148 KEITH PARFITT 230 The flint assemblage is predominantly debitage (Table 2), with only two complete cores and two recognisable tools present (Fig. 7.1 and 3). The debitage is mostly hard hammer-struck flakes. These tend to be of a large size, and exhibit broad platforms and large bulbs of percussion. Although many have prominent ripples on the ventral face, there are proportionately few hinge fractures. Apart from a few exceptions, the soft hammer-struck flakes also have large platforms, but generally have diffuse bulbs of percussion and lips. This could indicate that these were struck with a soft stone hammer rather than one of antler. Two hard hammer-struck flakes and six soft hammer-struck flakes, together with two of the hard hammer-struck blades have evidence of platform preparation. One soft hammer-struck flake may be an axe-thinning flake. Two cores and two core fragments were also found, all coming from the burnt debris deposit, context 55. The cores are both flake cores, and have two and three platforms respectively (Table 2). The two-platform core has also been utilised as a hammerstone, while the three-platform core has some evidence of platform preparation. There is some evidence in the assemblage that care was taken during the Fig. 7 Prehistoric worked flints from the site (scale 1:2) A PREHISTORIC ‘BURNT MOUND’ SITE AT CRABBLE PAPER MILL, NEAR DOVER 231 knapping process. This is suggested by the platform preparation on one core and some of the flakes and blades, together with the presence amongst the debitage of two flakes that could be classified as core rejuvenation flakes. The first is a flanc de nucleus recovered from a thick hillwash layer above the main prehistoric levels (Fig. 6, context 90). This flake has been struck from the side of a core at 90 degrees to the original platform, and has removed the flaked face of the core and the edge of the original platform. The second ‘rejuvenation’ flake is from another sealing hillwash layer (context 91, not shown on section), and has resulted from a sharp blow set some way back from the edge of the core platform to remove the platform edge, where the previous removals had undercut the platform. In both cases these removals would have allowed flaking of the core to continue, and it is unlikely that they are accidental removals. It seems fairly certain from the excavated contexts that both these pieces must be derived, washed down from further up the hill. The first tool is also derived from a hillwash layer (context 130, not shown on section). It consists of a single side-and-end scraper manufactured on a flake fragment (Fig. 7.3). Abrupt retouch has modified one lateral edge and the proximal end of a flake or blade fragment, after removal of the bulb and platform. The opposite lateral edge could not have functioned as a knife, so the retouch cannot have been intended as backing. It is therefore concluded that this was intended for use as a side-and-end scraper, and was most probably used by a left handed person. The second tool came from the burnt apron deposit on the west side, context 55, and must be directly associated with the main site under consideration here. It is unburnt and has been manufactured on a disc-shaped flake that has been heavily abraded or ground around the proximal end, along one lateral edge and partly around the distal end (Fig. 7.1). A small number of semi-abrupt or invasive retouch flakes have then been removed bifacially, probably as a result of this preparation rather than as secondary working. The opposite lateral edge has some semi-abrupt retouch forming a shallow concave area, together with possible utilisation damage along this edge. The abrasion or grinding may have been intended as preparation of the flake edge for subsequent bifacial invasive removals, although if this was the case it was either abandoned or utilised without this additional retouch taking place. This tool does not fit any current standard implement type, but is most closely related to a discoidal knife, possibly unfinished, or utilised in its partly made state. In addition to the above tools, there was also a hard hammer-struck blade from context 55, with possible utilisation damage along one lateral edge (Fig. 7.2), together with a retouched flake and a retouched fragment from later contexts (not illustrated). The largest sub-assemblage of flints, comprising 72 pieces, was recovered from the apron of burnt debris around the prehistoric pits (contexts 55 and 96). This includes 54 flakes and blades, 11 flake fragments, both of the cores and the core fragments, together with the discoidal knife and utilised blade (Fig. 7). The flakes are predominantly hard hammer-struck, whilst most of the ‘soft hammer-struck’ pieces were probably removed with a soft stone hammer. Apart from a single residual, probable Mesolithic bladelet, all this material would fit a Later Neolithic KEITH PARFITT 232 or Early Bronze-Age date (see below). Interestingly, none of the pieces had been burnt and they do not appear to have been directly involved with the recorded flint burning process. The upper filling of later ditch F. 146 produced 15 residual pieces of worked flint probably mainly derived from the earlier apron but these are undiagnostic and could fit either a Neolithic or Bronze-Age date range. The remaining contexts either have too few pieces or the flintwork is too undiagnostic to allow dating. This small flint assemblage has a few residual Mesolithic pieces, suggesting that there was some activity during the Mesolithic period in the vicinity of the site (see above). The majority of the flintwork, however, appears to belong to the Neolithic or Bronze-Age period, with the ashy dump layer (contexts 55 and 96) most probably dating to the Later Neolithic or Early Bronze-Age period. The evidence for a careful knapping process is interesting, especially as some of these pieces are present in both contexts 55 and 96. However, the two rejuvenation flakes occur in layers above these, and it is possible that all the pieces represent earlier residual activity, possibly during the earlier Neolithic. The assemblage associated with the main prehistoric site is too small, and has too few tools to be able to suggest any specific site function. Calcined Flint by Keith Parfitt A sample of 86 pieces (2.44 kg) of calcined flint was retained from the apron of dumped burnt debris surrounding the pits (context 55). The vast majority is thoroughly calcined and clearly has been subjected to intense heating. Most fragments are between 40 and 100mm across but much larger quantities of fragmented material, of 20mm or less, was also contained within the deposit. The burnt flints recovered from the fillings of the pits were identical to those from the apron and none was retained. Pottery by Keith Parfitt A total of fifteen small flint-tempered, prehistoric pot-sherds (not illustrated) was recovered from the excavations (56g). Of these, just five pieces (10g) came from the apron of dumped burnt debris surrounding the pits (contexts 55 and 96), with a further three pieces from the early ‘topsoil’ formed above (Fig. 5, context 54). The remaining seven fragments were found as residual material in later features and deposits and need not be directly related to the prehistoric site under consideration here. None of these sherds is closely datable. Other Finds by Keith Parfitt Pit F. 105 produced a single small fragment of burnt red clay, whilst F. 109 contained a piece of ironstone, foreign to the immediate area. A PREHISTORIC ‘BURNT MOUND’ SITE AT CRABBLE PAPER MILL, NEAR DOVER 233 TABLE 3. RADIO-CARBON DATES FROM PREHISTORIC FEATURES AT CRABBLE PAPER MILL CAT Sample Beta Lab Ref. Conventional R’carbon Age (BP) Calibrated* 1 sigma (Cal. BC) Calibrated* 2 sigma (Cal. BC) CPMD-55 (apron dump) 181796 3870±40 2445–2285 2465–2205 CPMD-101 (Pit, F. 102) 181797 3800±40 2295–2190 2165–2150 2340–2130 CPMD-132 (Pit, F. 133) 181798 3960±40 2485–2455 2570–2340 *using Pretoria Calibration Procedure Overall, datable prehistoric artefacts associated with the site were few and three radio-carbon dates (see Table 3) provide the main evidence for the age of the recorded remains. Samples taken from pits F. 102 and F. 133 and apron dump layer 55 have each provided results which fall within the Late Neolithic/Early Bronze-Age period. Analysis of the associated flintwork also implies such a date for the site (see above) and the prehistoric pottery recovered is consistent with this. In addition, a few struck flints that seem to belong to the Mesolithic period are present but only two undiagnostic pieces, sealed below the natural tufa, could represent in situ finds related to habitation during this earlier period. A possible Mesolithic occupation site previously identified at Beresford Road, some 500m north-west of the Crabble site, has now been disproved (Parfitt 1982). DISCUSSION OF THE NEOLITHIC/BRONZE-AGE SITE There seems little doubt that the sheltered valley of the River Dour, with its abundant supply of fresh running water, was an important focus for settlement throughout the prehistoric period (Fig. 1). Discoveries include the internationally important Middle Bronze-Age boat located close to the river mouth in 1992 (Clark 2004) but other finds are fragmentary and imprecisely located (see Parfitt in Clark 2004, 266-9 for a summary). The prehistoric remains found at Crabble clearly represent part of a significant new site, dating to the Late Neolithic/Early Bronze-Age period, with the suggestion of some earlier, Mesolithic activity close by. Although perhaps unrepresentative of the complete site, the remains excavated in 2002 do not appear to readily equate with what might be typically associated with a prehistoric settlement. Thus, burnt debris and calcined flints abound but there are only about seventy struck flints KEITH PARFITT 234 (unburnt) with very few tools; there is no food debris in the form of animal bone or marine shell, and pottery is confined to a few very small fragments. The excavated contexts actually appear to contain virtually no household rubbish and almost everything seems to be derived from the heating of flints in fires, apparently on a large scale. In fact, the investigated area appears to be concerned with a single, very specific activity or industrial process, involving some sort of hot stone technology. The close proximity of the River Dour is probably important and implies that quantities of fresh water were also required by this activity. The very distinctive, apparently non-domestic, character of the excavated site, situated no more than 60m from fresh running water, strongly suggests that Crabble should be linked to a somewhat enigmatic group of prehistoric features known as ‘burnt mounds’. These are characterised by low, oval, crescentic or kidney-shaped heaps, or spreads, of burnt stones and charcoal, and are often associated with pits. They are invariably located adjacent to streams, rivers or other water sources. Such sites are scattered across Britain and are usually dated to the Bronze-Age period. Although several sites have recently been discovered in London and the Thames valley (Moore et al. 2003, 184), no definite examples are yet published from Kent. Crabble, however, seems to fall within a general pattern which is now emerging across the country. Various purposes for burnt mounds have been suggested and it is probable that they actually represent several different types of activity. The two most frequently suggested interpretations for them are as specialised (?ritual) cooking sites, or (based on ethnographic parallels) prehistoric saunas/sweat-houses (Barfield and Hodder 1981). The pits excavated at Crabble had all eventually been filled with burnt waste but their original purpose is less apparent. Table 1 (above) shows that these pits fall into two basic sizes:- large, about one metre across [Fs 102, 105, 109, 133 and 151], and small, 0.50m or less, across [Fs 138, 140 and 153]. It is possible that they were all originally dug as simple rubbish pits but the three smallest, located towards the western end of the excavated area [Fs 138, 140 and 153], seem rather inappropriately sized for such a use (Table 1). On other sites, such small pits might well have been identified as post-holes. At Crabble any posts could have been withdrawn and the resulting holes back-filled with some of the burnt debris that abounded in the area. If these are to be seen as post-holes, however, they probably do not form part of a single structure (see below) with Fs 138 and 140 lying side-by-side but F. 153 positioned 8m further north-west (Fig. 4). The larger pits found at Crabble are of a size more appropriate for use in the disposal of rubbish (Fig. 6; Table 1) but they could equally well be interpreted as cooking pits, in which raw fish or meat was prepared using heated flints. However, the amount of burnt material dumped on the A PREHISTORIC ‘BURNT MOUND’ SITE AT CRABBLE PAPER MILL, NEAR DOVER 235 slopes beyond the pits seems rather too great to have been generated by a succession of single pit-cooking events and the absence of food bones also seems to be very significant in this context. Moreover, the fact that one of the pits [F. 151] was actually sealed beneath the apron of dumped material might suggest a sequence, in which unwanted burnt debris had initially been dumped into specifically dug pits but, as time went on, was simply tipped down the hillside without any attempt at burial. As an alternative to a cooking site, the possible use of Crabble as the location of a prehistoric sauna or sweat-house must be considered. In this interpretation (Barfield and Hodder 1981), it is envisaged that heated flints would be contained within some sort of temporary building or tent. Water would be poured over these hot stones to produce steam for the cleansing (spiritual or otherwise) of individuals inside the structure. After steaming, the adjacent river would have provided them with a place for an invigorating, outdoor cold plunge. At Crabble, the apron of burnt debris cannot be seen as the undisturbed base of a steaming structure, although it might be explained if the complete sweat-house had been dismantled after each period of use and the waste thrown downhill. If so, it could imply that the associated structures were erected upon the more level ground enclosed by the apron. The three small western pits (Fig. 4, Fs 138, 140 and 153), conceivably back-filled post-holes (see above), might now take on more significance, and could perhaps be suggested as marking the positions of central poles for successive tent structures. What function the larger pits could have served, however, is rather less clear. Another recently suggested interpretation for some burnt mound sites is as places where steam bending of wood occurred for producing curved timbers, such as might be required in prehistoric boat building (Gifford and Gifford 2004, 502). The discovery in Dover, some 2.7km further downstream (Fig. 1), of a Middle Bronze-Age boat (Clark 2004) makes this a particularly interesting idea for Crabble. However, the present site appears to be located too far up an essentially un-navigable river for this function to have been likely. The exact interpretation of the burnt flint site at Crabble must presently remain uncertain. Yet the site is not without some local parallels. Other prehistoric pits filled with distinctive deposits of calcined flint and charcoal, and broadly comparable to those found at Crabble, have been discovered at several other sites around Dover (Fig. 1). An example at Church Hougham, some 3.5km south-west of Crabble gave a radiocarbon date of 3790 ± 50 BP (2340-2035 Cal. BC; 2 sigma; Beta 130968; Parfitt forthcoming), whilst another at Whitfield, 1.7km north-east, was dated to 3730 ± 40 BP (2270-2260 or 2220-2020 Cal BC; Beta 179755; Parfitt 2003) and yet another, at Eythorne 6km to the north of Crabble (not shown on Fig. 1), has provided a date of 3360 ± 40 BP (1740-1530 KEITH PARFITT 236 Cal. BC; Beta 147436, Geoff Halliwell pers. comm.). Thus, a series of pits filled with calcined flint and dating to the Late Neolithic/Early Bronze- Age period is beginning to be recognised as occurring in the Dover area. However, these other pit sites differ from that at Crabble in two important respects; firstly, they are all located high on the Chalk Downs well away from any water sources and secondly, they are all represented by single isolated features, rather than forming parts of a more extensive site. It presently seems most likely that they relate to one-off pit-cooking events, which took place away from any established settlement, perhaps at some hunting party’s kill site. ACKNOWLEDGEMENTS Thanks are due to a number of people who assisted with the project in a variety of ways. First and foremost, the contribution of the developers, Crabble Mill Developments Ltd, who allowed access and funded both the fieldwork and post-excavation analysis must be gratefully recorded. On site, every assistance was given by manager, Richards Howells and the ground-working team. Simon Mason of Kent County Council’s Heritage Conservation Group monitored progress of the work throughout and provided much assistance and encouragement. Barry Corke has prepared the site drawings for publication and Chris Butler analysed the flint assemblage. Pete Clark and Geoff Halliwell kindly read earlier drafts of this text and suggested a number of improvements. BIBLIOGRAPHY Barfield, L. and Hodder, M., 1981, ‘Birmingham’s Bronze Age’, Current Archaeology, 78, 198-200. Barham, A.J. and Bates, M.R., 1990, The Holocene prehistory and palaeoenvironment of the Dour Valley catchment. A geoarchaeological assessment report for Dover District Council. (Geoarchaeological Service Facility Technical Report 90/04. University of London Institute of Archaeology, London). Bates, M.R. and Barham, A.J., 1993, ‘Recent Observations of Tufa in the Dour Valley, Kent’, Quaternary Newsletter, 71, 11-25. Bates, M.R., Barham, A.J., Jones, S., Parfitt, K., Parfitt, S., Pedley, M., Preece, R.C., Walker, M.J.C. and Whittaker, J.E., forthcoming, ‘Holocene sequences and their regional significance from the Crabble Paper Mill site, Dover, Kent, UK’. Clark, P. (ed.), 2004, The Dover Bronze Age Boat (English Heritage). Cross, R. and Parfitt, K., 1999, ‘Former Old Park Barracks, Whitfield, Dover: Proposed Redevelopment – Archaeological Desk Study and Impact Assessment’ (CAT Archive Report for Dover Harbour Board, May 1999). Gifford, E. and Gifford, J., 2004, ‘The Ferriby ship experiment’, Current Archaeology, 191, 498-505. A PREHISTORIC ‘BURNT MOUND’ SITE AT CRABBLE PAPER MILL, NEAR DOVER 237 Moore, P., Bradley, T. and Bishop, B., 2003, ‘A late Bronze Age burnt mound site at The Phillimores, Campden Hill Road, Kensington’, London Archaeologist, 10, No. 7. Parfitt, K., 1982, ‘A Trial Excavation at River, near Dover’, Kent Archaeol. Rev., 69, 293. Parfitt, K., 2002, Crabble Paper Mill, River, Dover: an Assessment of the Archaeological and Geoarchaeological Discoveries (unpublished CAT report submitted to KCC, August 2002). Parfitt, K., 2003, A Watching-Brief at White Cliffs Business Park, Honeywood Parkway, Whitfield, February 2003 (unpublished CAT report submitted to KCC, May 2003). Parfitt, K., forthcoming, Archaeology along the Folkestone-Dover Transfer Pipeline, 1998 (CAT report). Parfitt, K. and Corke, B., 2003, ‘Crabble Paper Mill, Dover’, Canterbury’s Archaeology 2001-2002 (CAT 26th Annual Report).