Quarry Wood Camp Loose A Belgic Oppidum

QUARRY WOOD CAMP, LOOSE: A BELGIC OPPIDVM By D. B. KELLY, B.A., A.M.A. THE SITE QUARRY WOOD Camp (Fig. 1; N.G.R. TQ 765515) is two and a half miles south of Maidstone and lies partly in Loose parish and partly in Boughton Monohelsea. It is situated on the Hythe beds at about 300 ft. above O.D. on level ground. The escarpment of the Lower Greensand, overlooking the Weald, is about three-quarters of a mile to the south. Immediately to the north is the Loose stream, running underground at this point, but reappearing in several springs and emerging immediately to the north-west of the earthwork to join the Medway two miles to the north-west. The subsoil is a light-brown to reddish loam, which can be up to 12 ft. or more in depth in the locality, but on the evidence of the ditch sections is here some 4-5 ft. in thickness. Below this are the Hythe beds, made up of alternate layers of the limestone known as Kentish 'rag' or ragstone and 'hassock', the local name for the compact sand or argillaceous sandstone.1 The single west and east banks of the earthwork remain and part of the ditch on the east side (Fig. 2; Plates IA and B). On the west side the bank stands for a distance of 300 yards, its height above the present ground level ranging from 21 ft. at the south end to 8 ft. at the north. The east bank stands on average at 8-9 ft. and extends for 380 yards. The south side was leveUed by the landowner, Mr. Charlton, at the beginning of the nineteenth century2 and nothing is visible on the ground. The north side has been entirely removed by the quarrying of ragstone. If the earthwork is assumed to have included the area of the disused quarry to the north,3 with any northern defences resting on the high ground to the south of the Une of the Loose stream, then an area of about 30 acres would have been enclosed by the defences. No gateways remain and the opening in the west bank appears to be relatively modern, though existing at the beginning of the last century.* The increase in height of the west bank at its southern end 1 For a detailed account of the Hythe beds in the area see Geology of the Country around Maidstone (Memoirs of the Geological Survey of Great Britain, 1963), 39-41; Observations on Kentish Ragstone as a Building Material, by John VVhichcord, Jr., London, 1846, 3-10. 2 Topography of Maidstone and its Environs, (Anon.), Maidstone, 1839, 85 and PI. vii. 3 See Appendix. * See note 2. 55 I o oo ^ Yards Maidstone linear earthwork Loose Stream Belgic .potter/ 1970 QUARftY WOOD CAMP Coma Loose ^*-*^ * V -f S i r t a i n v earthwork Boughton Green Cock Street Chart Sutton linear earthwork * \ > Linton !FIG. 1. Map of the Area South of Maidstone, sho-wing Position of Quarry Wood Camp and the linear Earthworks. (Based, upon the Ordnance, Survey Map with the sanction of the Controller of H.M. Stationery Office. Grown Copyright reserved.) QUAKBY WOOD CAMP, LOOSK just before the turn might suggest a gateway at the south-west corner. The surrounding area has been much altered by the clearing and levelling of the ground for the making of orchards. THE EXCAVATION (Fig. 2) The excavation of Quarry Wood Camp was undertaken as part of the Kent Archseological Society's programme of investigating the Iron Age earthworks in the county. Work began in the spring of 1963 under the direction of the late Mrs. N. Piercy Fox, F.S.A., assisted by the author. Some trial trenching was done at the south end of the west bank (Site 1), and the ditch on the east side of the camp (Site 2) was sectioned. In the following year, a section was dug outside the west bank to confirm the presence of a ditch (Site 3). Illness prevented Mrs. Piercy Fox from excavating in 1965 and, after her untimely death in March 1966, the author undertook the completion of the excavation. In the autumn of 1966, the Society's resistivity meter was used to locate the position of the levelled southern defences and a section dug to prove its position (Site 4). A final season's work was undertaken in 1967, when a further section (Site 5) was dug by machine across the hne of the south side of the earthwork and a complete section dug through the west bank (Site 1). PREVIOUS HISTORY OF THE SITE The earthwork has attracted httle attention in the past and even today it is httle known, being concealed from the adjacent roads and paths by orchards. It was discussed in a local guidebook, pubhshed in 1839,5 where a plan of sorts was provided, showing that the south and north sides had disappeared by that time. The section of the Victoria County History deahng with earthworks has a small plan of the site.6 Nothing to date the site was found until 1911, when two pits on the north side of the camp were uncovered during the removal of soil preparatory to the quarrying of ragstone. In these were found potsherds of late Belgic type, a stamped amphora handle and animal bones. The pottery is described in the appendix (Figs. 10-12). SUMMARY OF RESULTS A site of about 30 acres was defended by a single bank and ditch, although this cannot be shown for the north side. The sectioning of the bank showed a work of one period only. From the pottery found in the bank section (Site 1) and the east ditch (Site 2) the earthwork dates to the end of the Belgic period, the second quarter of the first century 5 See note 2. 8 Victoria County History, Kent, i, London, 1908, 399. 57 Belgic pottery N ^ 1911 ^ 5 ^ v SS s\* ; W Ot ^3 oo ^; s?-^< * *^Y .^ S - ^ \*.cS >v:*v I^S £# ^ ^!«Kuv ^ »S^ **' «•»*• N X ^ *»** N SSS X JT N. \ *X>:>.\Sv< **£** A«&« m&yv^ 3 SO soo ft I I - 5 = Sites uiirrtnii iiiuin«n' F I G . 2. Quarry "Wood Camp: Plan with excavated Sites marked. {Based upon the Ordnance Survey Map with the sanction of the Controller of H.M. Stationery Office. Crmim Copyright reserved.) QUARRY WOOD CAMP SITE I BANK SECTION NORTH FACE ^ ^ ^ ^ n V^^?>T^5 ^^^^.°^^/'i^^rTTT EXCAVATE ainiiiiiiiiiiiiuiiiiiiim NOT EXCAVAT NATURAL II RAG LIGHT BROWN LOAM HASSOCK SANDY L O A M / HASSOCK NATURAL HASSOCK FIG. 3. Quarry Wood Carapt site 1: Section through West Bank. D.B.K. [face p. 5 QUARRY WOOD CAMP, LOOSE A.D., not long before the Roman invasion, a date supported by the pottery found in 1911. The linear earthwork to the south and traces of similar works to the north and east are likely to have been part of the same defensive complex (see pp. 69 ff.). ACKNOWLEDGEMENTS A major debt of thanks is due to the landowners for their kindness in allowing the Society to excavate the site; to Miss J. Foster Clark, on whose land the major part of the earthwork hes and who allowed us to work over several seasons, and to the late Sir Godfrey Davies, at that time the owner of part of the eastern defences. Miss Foster Clark also kindly gave us the use of a barn to store equipment and we received every help from Mr. Hicks, then of Filmer's Farm, Mr. Holmans and Mr. J. Mattinson, manager of the Merriehills Farm Estate. The Kent Archseological Society made several grants of money to the excavation. In the first season's work, Messrs. Chaloner Evans, E. S. Greenhalf, A. Miles, J . Ray and boys and girls from local schools helped with the work, and Messrs. Chaloner Evans and A. Miles were present throughout the second season. In both years, the photography was undertaken by Dr. E. V. Piercy Fox. In 1966, Miss A. Salter and several other volunteers gave help at various times, which is gratefully acknowledged. Mr. T. Ithell, B.Eng., and Mr. A. Miles were present throughout the whole of the 1966 and 1967 seasons, and the author would particularly hke to thank them for their sustained and regular help in the digging and recording and in the tedious task of restoring the bank section in 1968. THE EXCAVATION IN DETAIL (Fig. 2) Site 1. The West Bank (Fig. 3 and PL IA) A site on the west rather than the east bank was chosen since the east bank is planted with trees along its whole length and the west bank covered only with scrub and small bushes. The highest point of the west bank is at its south end, where it rises gradually in height before turning to the east, and a site here was selected because it seemed likely to provide a better section than one to the north where the bank may perhaps have been less complete. The west bank is built on ground sloping gently to the west, so that on the outside the height of the bank is at present 21 ft. and on the inside just over 12 ft. In the first season both faces of the bank were cleaned back and a series of steps cut to ascertain if there was a stone revetment or cresting. Although there was a certain amount of stone used in the construction of the bank it appeared to have been incorporated as the material was excavated, in the body of the bank and occasionally on the face. On the evidence of this single section there was no regular stone cresting 59 D. B. KELLY or revetting as, for instance, at Oldbury I I 7 or the earher earthwork at Squerryes.8 To check this, in 1970, at a point 50 ft. north of Site 1, a stretch of rampart was cleared and examined to see if it had been revetted or crested. Bushes and rabbits had caused considerable disturbance, but apart from two small patches there was no stone present and certainly no regular revetment or cresting. As at Site 1, it appeared that stone had been used on the face of the rampart if it was immediately at hand, but that no attempt was made to revet the slope or crest of the rampart along its entire length. In 1967, the bank was sectioned down to the original Iron Age turf-line. It was of simple, dump construction and of one building period only. The volume of material used in the making of the bank is too great to have come from the ditch alone, but the subsequent levelling and alteration of the surrounding ground has removed any trace of any quarry pits. The base of the bank was carefully constructed: in the front a mass of stiff loam and large pieces of ragstone and, behind this, layers of sandy and argillaceous hassock, bound towards the rear by a turf layer. The rest, and greater part, of the bank was constructed by dumping the material as it was excavated. The tip-lines could be seen quite clearly, and the mass of ragstone and hassock resting on the turfbinding at from 19 to 26 ft. from the front face formed a barrier against which the material could be tipped. Where the sandy hassock was used the face of the bank was revetted with ragstone to secure it. The turflayer at the front of the bank at about 8 ft. from the top may have been placed with the same intention, but it did not appear on the south face of the cutting and was taken to be a local feature. Several small pottery sherds, mostly well abraded, and fragmentary animal bones and teeth were found throughout the bank. These must be regarded as rubbish lying on the ground before the construction of the earthwork and subsequently incorporated in it and thus can only provide a terminus post quern for the date of construction. The pottery (Fig. 9, 1-6) included a sherd of finely combed ware, two sherds with cordons, one of them part of a lid, and a burnished, recurved rim sherd, aU Belgic wares. The cutting was extended into the ditch, and it was apparent that the lowest part of the bank, though only two feet of it, as seen from the outside, was in fact the upper part of the inner side of the ditch. A short stretch of about 8 ft. was dug by hand and, when the lower part of the bank section was restored mechanically, the opportunity was i J. B. Ward Perkins, 'Excavations on the Iron Age Hillfort of Oldbury', Archaeologia, xo (1944), 137-8 and PI. xxxi. 8 N. Piercy Pox, 'Excavation of the Iron Age Camp at Squerryes, Westerham', Arch. Cant. Ixxxv (1970), 31 and figs. 2 and 3. 60 PLATE I A. The West Bank from the South End. Photo: D. B. Kelly Dr. E. V. Pitrm Por B. Tlie Bank and Ditch on the East Side f face p 60 - m Photo: Dr. E. V. Piercy Fox A. Site 3: the excavated Ditch on the West Side. Photo: D. B. Kelly B. Site 4: the Ditch on the South side during Excavation. QUARRY WOOD CAMP, LOOSE taken to use the machine to extend the section across the ditch for a further 9 ft., up to the farm track. At a point 11 ft. from the base of the rampart, the bottom of the ditch was reached, cut down to the natural hassock. It was flat for the 6 ft. that could be uncovered and was 6 ft. below the present ground surface, thus originally being 8 ft. deep. Site 2: The Ditch on the East Side (Fig. 4, PI. IB) The only part of the site where the ditch is now clearly visible is a stretch along the middle part of the east side of the earthwork. A section was cut heie in 1963 and showed that the overall width of the ditch was 29 ft. and its depth 9 ft. The ragstone was here only some 5 ft. below the surface and in the making of the ditch a projecting shelf of rag, 5 ft. wide, had been left on the outer side. The deepest, central part of the ditch had been cut through this ragstone and the succeeding hassock down to the second ragstone layer, which provided a flat bottom, 3 ft. wide, to the ditch. At the bottom of the ditch, above the small amount of immediate silting, was a layer of loam, about 6 in. thick, and then a thin turf-line with a rather stony surface. This turf-line presumably formed during the period of the earthwork's occupation. On it were some twenty small sherds, fairly worn, from a single vessel of thin sandy ware (Fig. 9, 7). After the site fell into disuse, the ditch became filled with ragstone rubble, including large pieces, and loam up to the level of the ragstone shelf, when it remained open at this level long enough for another turfline to form. The remainder of the ditch filling, of loam, hassock and small stones, up to the present ground level, was undifferentiated. Site 3: The Ditch on the West Side (Fig. 5, PI. IIA) This section was cut by machine in 1964 and then enlarged and cleaned by hand. Its purpose was to see if a ditch had existed on the west side of the earthwork, since none was visible on the ground. The ditch as excavated was 28 ft. wide. The precise position of the outer hp was obscured by the levelling that had taken place to make the orchard, but was taken to be where the buried turf-line ended. The bottom of the ditch is 7 ft. below the present ground level. As at Site 2 the ragstone was near the surface, and the ditch was cut through successive layers of ragstone and hassock. The flat bottom, just over 5 ft. across, is formed by a layer of ragstone, and the sides of the ditch are formed by a series of steps cut through the layers of rock. The profile of the ditch is wide and shallow, the inner face rising in shaUow steps for some feet before rising steeply to the foot of the rampart. No pottery was found in the ditch filling. The buried turf-line represents the ground surface up to the time the ground was levelled. 61 w I1IIUM1HM / ; / iwnrrrTTTrnTTTT VZ'°/ '/' '•s HASSOCK RAG HASSOCK ° C? „° RAG ^ RED CLAYEY LOAM LIGHT BROWN LOAM NPF MENS OBK DEL PIG. 4. Quarry Wood Camp: site 2: Section across East Ditch. QUARRY WOOD CAMP, LOOSE < z _J 2 a U_ * a. co Z. Q 0 r± 1 O o o I . \ s \ - s S .x\ 0 3 o M 63 D. B. KELLY The large rocks in the fill above the outer face might possibly have come from a counterscarp bank, but no trace of such a bank was found in any of the ditch sections. Site 4: The Ditch on the South Side (Fig. 6, PI. IIB) No trace of rampart or ditch is visible on the south side of the earthwork. The south end of the west rampart turns to the east or south-east, however, and the east rampart at its southern end swings to the west, so that it could be reasonably surmised that any works on the south side would take the form of a straight hne or shallow curve between the ends of the remaining banks. This assumption was strengthened by the rather inaccurate plan of the earthwork in the local guidebook of 1839,9 where the existing southern ends of the two ramparts were joined by a dotted hne with the entry 'Bank here removed by Mr. Charlton'. A short paragraph in the text states that the 'greater part was levelled down and spread over the adjoining land above 20 years ago, by Mr. Charlton, then proprietor of the site'.10 A resistivity survey of the ground where the suspected south rampart and ditch lay was carried out, and a trench was dug at a point near the east rampart, where the meter gave an unusuahy high reading. The section shown in Fig. 6 is of the trench that was dug by hand, but before the trench was filled the machine was used to extend it southwards. The upper part of the ditch was cut through a thick layer of chert, which occurred as a layer in the loam,11 and this formed the upper part of the inner face. The loam below it was revetted with ragstone, a feature for which there was no need in the other ditches which were sectioned. The fill of the ditch above the primary silting contained much stone, some of it from the revetment of the inner face of the ditch, but the large rocks presumably from the body or revetment of the rampart. The buried turf-line represents the ground surface before the rampart was levelled and spread and the top 6|- ft. of fill is the result of this levelhng. As in the other three sections, the material excavated from the ditch alone would not have sufficed to construct the rampart. The modern fill contained sherds of medieval pottery (Fig. 9, 9-13). When the cutting was extended by machine, it was seen that the natural loam rose gradually for a further 12 ft., when it levelled out at a depth of 2 | ft. below the present ground surface. This outer hp of the ditch is a foot higher than the inner hp, as the ground slopes gently northwards here, a feature more pronounced at Site 5, dug 9 See note 2. 10 Mr. Charlton was a considerable local landowner at the time of the Napoleonic wars, as Mr. John Short kindly informs me. 11 For this feature see the Geological Memoir, op. cit. in note 1, 39. 64 05 '.'/'.'S.'/S //J//, ///* / 7 / " / / ' / / - / / / / / ' / ; . *> rrm7T7iTTT]TTrrrrrrnT^ •' *> * > ^ . ' ^ X X ' X f X X - MR.CHARLTON'S / / . - / / / ,' ' /-* / ,' / ' - X. XJ, EVETMENT NATURAL CLAYEY LOAM HASSOCK VMm CLAYEY LOAM •'//'/ /\ LIGHT BROWN LOAM o o o p t-1 o o FIG. 6. Quarry Wood Camp: site 4: Section across South Ditch. D. B. KELLY further west. The width of the ditch here was thus 37 ft. and its depth 7 ft. Site 5: The Ditch on the South Side (Fig. 7) When the trench at Site 4 was filled, the machine was used to excavate a section on the presumed hne of the south ditch at about its centre. The profile shows the undisturbed natural loam rising on the outer face of the ditch to a level almost 3 ft. above that on the inner face, and though the trench dug did not include either hp of the ditch and the inner face may have risen more steeply to the north, the ground here presumably sloped northwards. As at Site 4, the buried turf-line represents the ground surface before the levelling. The width of the ditch here is not known, but must be more than 26 ft. THE DATE AND EOEM OE THE EARTHWORK Of the smah quantity of pottery found, the sherds of the vessel found on the turf hne a little above the bottom of the east ditch must have reached there within a few years of the construction of the camp (Fig. 9, 7). Not enough remains to show its shape, as the sherds are from the base and body only. The thin, hard, sandy ware is quite unlike any of the Belgic pottery from the site or the Maidstone area and the pot may be assigned to the Roman period. The only other finds were the sherds and bead found in the bank (Fig. 9, 1-6, 8). The sherds all come from different vessels, and their worn state and small size show that they were lying on the ground for several years before the construction of the bank, presumably thrown out from some local farmstead with the manure. Several of them at least are late Belgic, as the bronze bead seems to be. The dump construction of the bank, as shown at Site 1, is the usual feature of Belgic banks. At Wheathampstead, Prae Wood, Bagendon and Silchester the banks are all constructed in this way. The noticeable feature of the earthwork is the considerable width of the ditches and their comparative shallowness in proportion to their width: Width Depth Site 2 29 ft. 9 ft. Site 3 28 ft. 7 ft. Site 4 37 ft. 7 ft. Site 5 over 26 ft. c. 6J ft. At Sites 2 and 3, and in the partly excavated ditoh at Site 1, the ditches had flat bottoms, but this feature was probably adopted to save cutting through more ragstone than necessary, and it will be noticed in the sections how the layers of rag caused the sides to be stepped. The flat bottoms of Sites 2 and 3 are only 3 ft. and 5 ft. across respectively and, 66 N $ ^ $ & & Jf _,» \V V. W -it ¥ il y V J/ i i/ * V v * ii 19th. cent, fill FIG. 7. Quarry Wood Camp: site 5: Profile of Ditch. D. B. KELLY moreover, on the south side the ditch as shown at Sites 4 and 5 had a wide V profile. Of those Belgic earthworks known to have been constructed in the first century A.D., Bagendon had defensive ditches that were narrower and deeper in proportion to their width: the cutting across the Eastern Inner Rampart revealed a V-shaped ditch 10| ft. wide and 6 ft. deep.12 This, however, dated to the first two decades of the century. Wide, relatively shallow defensive ditches are a feature of the works constructed not long before the conquest. Oldbury I I had a ditch 15 ft. wide but only about 18 in. deep13 and the Inner Earthwork at Silchester was about 45 ft. wide and only 8 ft. deep.14 Both these sites have been compared to the Fecamp series of north French hill-forts,15 but the difference of a century in date remains a problem16 and the profiles of the Fecamp type oppida17 show ditches far wider and flatter than all the English examples adduced, except for the Silchester Inner Ditch and the non-Belgic Caburn. Of the non-Belgic fortifications of the period, the Caburn, refortified on the eve of the conquest18 has a wide, flat-bottomed outer ditch on the north side, about 40 ft. across and only some 4-5 ft. in depth.19 The re-fortification of High Rocks, not earher than the second quarter of the first century A.D. and perhaps just before the conquest,20 added in places a ditch shallower and wider than in the early phase, but not of any great width—about 11 ft. across at Site J, for instance. The excavator noted that the ditches were wide and shallow when cut in the rock but U-shaped when dug in the clay.21 It seems, then, that there was a fashion for wide, shallow ditches in defensive earthworks in south-east England in the two decades before the Roman invasion, perhaps, as Dr. Wilson suggested in 193922, as 'an attempt to frustrate the Roman methods of attack by filling up the ditches with earth or brushwood'. These ditches were not usually 12 Elsie M. Clifford, Bagendon: a Belgic Oppidum, Cambridge, 1961, 8 and fig. 3. 13 Archaeologia, xc (1944), PI. xxxi. 14 G. C. Boon, 'Belgic and Roman Silchester: the Excavations of 1954-1958', Archaeologia, cii (1969), 14 and PI. vii. 16 Archaeologia, xc (1944), 139-41; cii (1969), 14. 16 Mrs. M. A. Cotton's suggestions in Problems of the Iron Age in Southern Britain (Ed. S. S. Frere, London, 1960), 66, involve a rather arbitrary re-dating of both phases at Oldbury. 17 Sir Mortimer Wheeler and Miss K. M. Richardson, Hill-Forts of Northern France, Oxford, 1957,10, fig. 2. 18 A. E. Wilson, 'Excavations at The Caburn, 1938', Sx. A.C, Ixxx (1939), 194. 10 A. E. Wilson, 'Excavations in the Ramparts and Gateway of The Cabum . . . 1937', Sx. A.C, Ixxix (1938), 181 and PI. i. 20 J. H. Money, 'Excavations in the Iron Age Hill-Fort at High Rocks, near Tunbridge Wells, 1967-1961', Sx. A.C, cvi (1968), 184. 21 Ibid., 179. 22 Sx. A.C, Ixxx (1939), 194-6. 68 QUARRY WOOD CAMP, LOOSE 'flat-bottomed' in the sense in which the term is used of the Fecamp series of a century earher. Among this group of earthworks Quarry Wood Camp may fittingly be placed. THE LINEAR EARTHWORKS (Fig. 1) In 1965 Messrs. A. Clarke, A. S. Phillips and C. F. Wardale, of the Archseology Division of the Ordnance Survey, discovered and traced a linear earthwork, consisting of a bank and ditch, running from Linton Park to Chart Sutton, a distance of about 2-f miles.23 The earthwork is set well back from the escarpment and runs from east to west, except for a stretch immediately south of Cock Street, where it makes a sharp turn to the north. Both the west and east ends terminate at about the 400 ft. contour line, the earthwork closing the gap between these areas of shghtly higher ground (Fig. 8). The ditch is on the south side of the bank. Quarry Wood Camp is about half a mile north of the linear earthwork, towards its west end, and about a mile east of the camp, concealed by the trees and undergrowth which make up Hogtrough Shaw, there is a short section of a massive bank and ditch, running north-south, the ditch being on the east side. From the top of the bank to the bottom of the ditch is 31 ft. and from the top of the bank to the outer lip of the ditch is 48 ft. This short stretch, if extended southwards, would join the Linton-Chart Sutton earthwork at a point about 750 yds. eastsouth- east of Cock Street, where there is a slight break in the earthwork, and the Ordnance Survey noticed traces of a double bank and ditch. In the Topography of Maidstone2i the author describes a bank and ditch in Loose parish, facing north, about 200 yds. long, in a slip of chestnut plantation at the back of Pickering Street, and another small portion 'near Mr. Cull's Oasts'. Until a few years previously the earthwork had extended from the Loose turnpike road as far as Boughton Mount, and an old inhabitant told the author that it was thought that it had continued through the whole of Langley parish. The part then existing was 'about 30 ft. wide at the base, 10 ft. high on the north side, where the hollow is, and about 5 ft. on the south, above the level of the ground.' In the accompanying plan the earthwork is shown running from the Loose Road to Boughton Mount and the lengths on either side of the existmg stretch are described as 'removed by Mr. Charlton', the landowner who demohshed the south side of the Camp. The present writer has visited the site and found that the earthwork described behind Pickering Street still exists, though with a profile much flattened and eroded. The short stretch near Boughton Mount has disappeared. 23 Arch. Cant., Ixxx (1966), 277. 24 Op. cit. in note 2; 104, 118-9 and PI. vi. 69 9 D. B. KELLY The Linton-Chart Sutton earthwork varies considerably in its state of preservation, ranging from the only slightly discernible bank and ditch to fairly well preserved sections. The best preserved section at the time of the Ordnance Survey investigation, at about N.G.R. TQ 771504, has since been leveUed for the planting of fruit trees. From this point eastwards, and then northwards until it again turns east and runs along the north side of the road to Amber Green and Chart Corner, the earthwork provides the boundary between orchard and woodland. Immediately west of Amber Green, the Ordnance Survey noticed that there was an apparent break in the earthwork at the approximate point where the Roman road from Rochester to Hastings is joined by the road from Lympne,26 though just over a hundred yards west of the line given by Margary. Margary shows an ahgnment angle immediately north of the earthwork, his line for the road turning east-south-east before the junction. At the time of its discovery this part of the earthwork had aheady been almost ploughed away and was to be ploughed again after the orchard through which it ran had been grubbed. It was hoped that if the Roman road did cross the earthwork it could be shown whether or not the earthwork pre-dated the road. In November, 1965, Messrs. F. G. Aldsworth, of the Ordnance Survey, A. Miles and the author carried out a small excavation north of the earthwork in an attempt to find the hne of the Roman road, which seemed essential before an investigation of the point where it was thought to cross the earthwork.26 We were unsuccessful in this and can confirm Margary's remark that the road here 'seems to have been thoroughly obhterated.' Further investigation at this point might weU prove worthwhile, and it is hoped to undertake this at some future date; elsewhere along the line excavation would reveal the profile of bank and ditch, but is unlikely to provide anything to date the earthwork. In the absence of direct dating evidence alternatives to a Belgic date for the earthwork must be briefly considered, the two possibilities, however slight, being that the work is the remains of an Anglo-Saxon linear defence or a medieval park enclosure. The siting of the linear earthworks around a known Belgic camp is a very strong argument against both alternative attributions, especially the former, though, of course, the contemporaneity of these stretches of linear work has not been proved. The Linton- Chart Sutton section of the earthwork is constructed on the head overlying the Hythe beds, which extends to the west and east of it. Despite a careful search by the Ordnance Survey, no trace has been found of the earthwork beyond the limits discovered and published by them. The ends of the work stop approximately on the 400 ft. contour 2 6 1 . D. Margary, Roman Ways in the Weald, London, 1965, 214 and map 218. 20 Thanks are due to Hernden Farms Ltd. for allowing us to undertake the work. 70 tf »• 10 j o o M 10°'' (/t i « ° s 4oo •**> _y Quarry Wood Camp Ccmct try Linear earthwork + B unaluJ Xoc Pottery, b r o o c h o . e tc Aylesford Coins (inscribed) M Maidstone I OOO Yds r l M I c I_I_I i F I G . 8. Belgic Sites of t h e first Century A.D. in t h e Maidstone Area. (Based upon the Ordnance Survey Map with the sanction of the Controller of H.M. Stationery Office. Crown Copyright reserved.) [fact p. 71 QUARRY WOOD CAMP, LOOSE lines (Kg. 8), though this is apparent on the map rather than in the field, since the rise in level is scarcely noticeable on the ground. This earthwork, then, would hardly have been constructed to demarcate a very large tract of land, nor do its ends rest on natural features which would themselves have constituted a barrier. In any case, the Weald itself would provide a fairly formidable natural barrier to the south. One possibihty is that the earthwork, like so many dykes of supposedly Anglo-Saxon date, was bunt to block the Roman roads from Hastings and Lympne. In the Roman period, however, the Hastings road is assumed to have been used in connection with the iron-working sites in the Weald and, apart from these, there are no other Romano- British sites known there; after these sites feU into disuse the road itself would have been httle used though a stretch below the escarpment is foUowed by the present main road. Lympne was presumably part of the Kingdom of Kent from the early years of the Anglo-Saxon settlement, but the road, running through the Weald just south of the Greensand ridge, may have faUen into disuse quite early, as its line now runs across open country rarely foUowed by tracks, let alone roads.27 If, in the sixth century, the Medway was the frontier of the Kentish Kingdom28 defensive earthworks, if they existed, are more likely to have faced west, blocking the land between the westward bend of the river above Maidstone and the Greensand escarpment, as did the later dyke on the Surrey border just west of Westerham.29 FinaUy, with the gradual settlement of the Weald from the eighth century onwards no such defensive earthwork would be likely. There is no documentary evidence for a medieval park in the area, though records exist in most cases of such enclosures.30 Indeed, one would not expect such a document, since Loose belonged to Christchurch, Canterbury, until the Dissolution, and Boughton Monohelsea, Langley and Chart Sutton aU had different owners, except for a short period after the Conquest, when they aU belonged to Bishop Odo.31 A park so constructed, too, would have enclosed the ragstone quarries at Boughton Monohelsea, worked in the medieval period, and several farms and hamlets of some antiquity. In 1841 a Roman bath buUding was discovered and excavated near Brishing Court, in a field known as The Slade, by the Loose stream and about 500 yds. west of Hogtrough Shaw.32 The whereabouts of the 27 Margary, op. cit. in note 25, 228-43. 28 S. C. Hawkes, Arch. Journ., cxxvi (1969), 189. 20 A. Clark, Sy.A.C, lvii (1960), 72-4. 30 Field Archaeology, Ordnance Survey Professional Papers no. 13, 1963, 131; 0. G. S. Crawford, Archaeology in the Field, London, 1964,197. 31 Edward Hasted, History and Topographical Survey of the County of Kent, 2nd edition, 1797-1801, Vols, iv and v, under parishes mentioned. 32 Archaeologia, xxix (1842), 414-20. 71 D. B. KELLY pottery found there is not known, but among the small finds preserved in Maidstone Museum is the perforated catchplate of a large, bronze La Tene I I I fibula. More significant are the six Belgic coins from the site, of Dubnovellaunus (2), EppUlus, Cunobehn (2) and Amminius.33 Another coin of Cunobehn was found at Cox Heath, about a league west of our site.34 In the autumn of 1970, a large quantity of late-Belgic pottery was brought to Maidstone Museum,35 found at Furfield Hole, about 300 yds. north-west of Hogtrough Shaw (Fig. 1). This pottery was not in situ and was found along a ledge well above the quarry floor and about 12 ft. below the level of the field to the north. I t included a small amount of Romano-British pottery of varying date up to the third century A.D., but over 90 per cent, of the sherds were Belgic, many of them simUar to the pottery found in 1911 in Quarry Wood Camp. They presumably came from pits destroyed when the overburden was removed before the stone was quarried. A complete biconical jar, in the Kent Archseological Society's coUection, and presumably from a burial, was found at Chart Sutton in the last century, though outside the linear earthwork. I t seems reasonable, then, to accept these earthworks as Belgic. They provided defensive outworks for the Quarry Wood Camp earthwork and embraced a large tract of land and a stretch of the Loose stream. Similar earthworks of Belgic date are too well-known to require more than a mention here. At Colchester, Wheathampstead, Prae Wood, Bagendon, Chichester and, in part at least, at Silchester they have been shown to be Belgic. DISCUSSION At Colchester, by the time of Cunobehn, a site of some twelve square miles between the Colne and Roman Rivers had been marked off and defended by a series of dykes; within, the area with the strongest natural defences was fortified with the Sheepen Dyke and became what the excavators describe as the focus of British Colchester.36 The oppidum at Bagendon, constructed in the first two decades of the first century A.D., was about 200 acres in area, defended by earthworks on three sides and relying on thick woodland for its defence on the remaining side.37 Here there was no central stronghold. The Wheathampstead oppidum, of the previous century, was about 100 acres in size and 33 Listed in D. P. Allen, 'The Origins of Coinage in Britain: a Reappraisal', in Problems of the Iron Age in Southern Britain, Ed. S. S. Frere, London, 1960, 215,213,232,238. 31 Ibid., 232. 36 By our member Mr. D. Attlee and boys from Senacre School. 30 C. P. C. Hawkes and M. R. Hull, Camulodunum, Oxford, 1947, 8-16. 37 Clifford, op. cit. in note 12, 4-5. 72 QUARRY WOOD CAMP, LOOSE defended partly by massive dykes and partly by the dense forest of the site.38 To the south-west was the formidable Beech Bottom Dyke, providing a boundary between the rivers Ver and Lea, but no outworks surrounding the oppidum have been traced. At the succeeding Belgic site at Prae Wood the dyke was not defensive and served rather as a boundary.39 The Quarry Woodoppidum differs from these sites in that it consists of a strong central fortification with outworks surrounding it, at least partially, on three sides. Except on the west side, where the site is protected by the Loose vaUey, there seem to be no natural features which would govern the siting of an oppidum at this place, whereas the natural features which led to the choice of site were apparent at Colchester, Bagendon and Wheathampstead. All the Belgic oppida mentioned above—Camulodunum, Prae Wood, Bagendon, Chichester (Selsey) and Galleva—were succeeded in the Roman period by important towns, situated near at hand, yet no such town appeared on the site of what is now Maidstone. Here, in Roman times, where the Rochester-Hastings road ran near the Medway, have been found no more than four buildings, two of them substantial villas, some cemeteries and scattered finds.40 In fact, aU the Medway valley below and for a short distance above Maidstone was equally well populated then.41 The bath-house at Boughton Monohelsea, some nearby burials and pottery from White Lodge Farm at Cock Street are the only Romano-British remains from within the area of the oppidum and are presumed to be connected with the quarrying of ragstone here. The nearest Roman town was Rochester, where, in 1961, Belgic coin moulds were found.42 Yet the Maidstone area can show an appreciable number of Belgic remains (Fig. 8). The important cemetery at Aylesford43 continued in use into the first century A.D., and the nearby Allington cemetery,44 with native butt-beakers, a platter and dishes and an imported two-handled jug of Camulodunum type 161 Ab, is mostly ofthis century, too. In Maidstone, four sites, two of them burials, 38 R. E. M. and T. V. Wheeler, Verulamium: a Belgic and two Roman Cities, Oxford, 1936,19. 30 Ibid., 11. 40 With minor additions the account in the Victoria County History of Kent, iii, London, 1932, 98-101, stands today. 41 V.CH. Kent, iii: Ordnance Survey: Map of Roman Britain, 3rd Edition, 1956. 42 Arch. Oant., Ixxvii (1962), li. 13 A. J. Evans, Archaeologia, Iii (1890), 317-88; Ann Birchall, P.P.S., xxxi (1965), 241-367. 44 An unpublished find of about 1860, from Tassel's Quarry; pottery in Maidstone Museum. 73 D. B. KELLY have produced Belgic pottery of the first century A.D.45 and there are finds of late-Belgic date from Hermitage Farm, Allington—a burial, Barming,46 Detling47 and the site of the Roman villa at Eccles.48 Two coins, of Vosenios and EppiUus, were found at the site of the alleged Roman temple on Blue BeU HiU49 and three coins of Cunobehn at the Eccles viUa site.50 There was, then, a substantial Belgic population in the Maidstone area for whom the Quarry Wood oppidum may have provided a refuge. They probably hved in open settlements or farmsteads for which, until Quarry Wood Camp was built, they saw no need for defences. East Kent, as far as the Medway, was an area of primary Belgic settlement, remaining firmly Belgic despite dynastic changes. West Kent, it is thought, populated by Wealden peoples, was not annexed by the Belgae until the time of Cunobehn, though coins of Tasciovanus have been found on seven sites.51 The nearest hill-fort, Oldbury, was shown by excavation to have been captured by the Belgae soon after its construction by Wealden people in the first century A.D. and to have had its defences strengthened by the Belgae not long before the Roman invasion.52 Of the other excavated hill-forts in West Kent, Holwood53 and Squerryes54 provided no Belgic pottery, and at High Rocks the excavator found that the period II re-fortification could not be attributed to the Belgae.55 It might be argued that Quarry Wood Camp was constructed on the western boundary of Cunobelin's Kentish kingdom, but, if so, it is surprising that the river crossings at Maidstone and Aylesford with their Belgic settlements should have been left undefended. Even in the time of Tasciovanus, Belgic influence in West Kent was strong and during Cunobelin's reign the Wealden people of West Kent were threatened by the Belgae, not threatening them. Allen has shown, on 45 (a) A cremation burial with a pedestal urn, at Haynes' Garage, Ashford Road (4rc/i.Can«., Ixxviii (1963), 194-6; (b) Cherry Orchard Way, 1920; (c)Scrubb's Lane, 1889; (d) Northborough School, 1939 (Arch. Oant., Iiii (1940), xliii). The pottery from all these sites is in Maidstone Museum. 40 A small jar found in Arnold's Quarry in 1931 and further pots found in 1963 and 1966, one containing a cremation. In Maidstone Museum. 17 A large La Tdne I I I brooch found in building Detling vicarage in 1831, and a site with pottery found by J. W. F. Edwards in 1968 (Kent Arch. Rev., 11, 7 and 15,12-13). In Maidstone Museum. 48 A. P. Detsicas, 'Excavations at Eccles, 1964', Arch. Cant., Ixxx (1966), 70; 1966, Ixxxi (1966), 45. 49 Allen, op. cit. in note 33, 213, 214; V.O.H., Kent, iii, 104. 50 Arch. Cant., Ixxxi (1966), 45. 6X Allen, loc. cit. in note 33, 222; Ordnance Survey: Map of Southern Britain in the Iron Age, 1962. 62 Archaeologia, xo (1944), 153-4. 63 N. Piercy Fox, Arch. Oant., Ixxxiv (1969), 185 if. 64 Arch. Oant., lxxxv (1970), 31-2. 66 J. H. Money, loo. cit. in note 20, 184. 74 QUARRY WOOD CAMP, LOOSE the coin evidence,66 that by A.D. 25 Cunobehn was in control of East Kent, where he succeeded EppiUus. West of the Medway fewer of his coins have been found than those of Tasciovanus, but by the time of his death in about A.D. 40 the Belgae had occupied most of Kent west of the Medway as far as the south side of the Greensand ridge. If the Quarry Wood oppidum is dated not long before A.D. 43, its construction is unhkely to have been undertaken in connection with expansion into West Kent and the Weald. It is, moreover, too far removed from Verica's kingdom to have had any connection with inter-dynastic conflicts. The reason for its construction must have been the fear of a Roman invasion, either in A.D. 40 or 43. The construction of the oppidum at this time explains the choice of a site not especially well protected by natural features, compared with other Belgic oppida, nor near enough to the Medway to control the important river crossings. An oppidum built as a dehberate act of pohcy rather than as an emergency measure must surely have been sited with due regard to one or other of these considerations. In fact, a secluded and rather remote site was chosen and its short-lived nature is one reason why the Belgic settlement—or, perhaps, oppidum—at Rochester became the site of a town after the conquest, whereas the Maidstone area remained well but not heavUy populated, as in Belgic times. Professor Frere67 foUows Collingwood in taking the Roman advance from East Kent along the hne of the later Watling Street, its immediate objective a possible oppidum at Rochester. A. R. Burn58 suggests an advance along the North Downs, though the main trackway along the top59 seems a more likely route than his suggestion of the Pilgrims' Way. AU these authors place the site of the Medway battle at Rochester. Even so, the Belgic leaders, though evidently not strategists of the highest class, may have foreseen the possibihty of a Roman crossing of the Medway in the Aylesford area and the Quarry Wood oppidum would have provided not only a refuge for the local population but also a stronghold in the event of defeat. There is no evidence that the Roman army attacked it, but the reduction of Oldbury60 shows that such forts were not by-passed. THE FINDS (Fig. 9) Pottery Site 1: In the Rampart. 1. Sherd, dark grey fabric, orange-red surface on both sides; decorated with fine horizontal combing; abraded. 50 Archaeologia, xc (1944), 1-35, 158-9. 67 S. S. Frere, Britannia, London, 1967, 64. 68 A. R. Burn, 'The Battle of the Medway, A.D. 43', History, 39 (1953), 105-15. 59 I. D. Margary, Arch. Journ., cix (1952), 47-9 and Arch. Oant., lxiv (1951), 20—3. 60 Archaeologia, xc (1944), 139. 75 D. B. KELLY TH==f f 7 7 Fia. 9. Quarry Wood Camp: Finds from the Excavation. (£; no. 8, }) 2. Sherd from the rim of a lid; reddish-brown ware with remains of black-burnished surface; very worn. So far as can be judged from a single sherd the hd appears to be a late type. Early lids at Colchester are 'tall and conical or boldly domed'. The nearest paraUels there are nos. 9 and 10, both Roman (Camulodunum, PI. lxxxv). A lid from Silchester, found in a layer of c. A.D. 60 at the latest, is also close in form (Archceologia, cii (1969), 65, fig. 13, 78). In a pre-Roman layer from outside the Roman Theatre at Canterbury was a similar hd (Britannia, i (1970), 100 and fig. 9, 6). 3. Sherd with a cordon, ware as no. 2. Perhaps from a native butt-beaker. Very worn. 76 QUARRY WOOD CAMP, LOOSE 4. Sherd from a jar with recurved rim; hard, sandy, grey ware with surface black-burnished on both sides. (Cf. jars from Greenhithe: Arch. Cant., Ixxxi (1966), 149-52 and figs. 3 and 4. Mr. A. P. Detsicas informs me that the ware is identical.) 5. Rim sherd with flattened top, from a dish with incurved side; ware as no. 4; outside and top burnished. At Colchester the dish or platter with incurved side is rare (Camulodunum form 30) and dateable to the earliest Roman period. Form 19 at Bagendon (Bagendon, 219 and fig. 48, 15) is reminiscent of our sherd, but has a more pronounced internal offset. There it is from period IIA—the second quarter of the first century A.D. 6. Basal sherd with vestigial foot-ring; ware as nos 4 and 5. Site 2: Ditch on east Side; resting on lowest turf-hne. 7. Sherds from the base and body of a jar of hard, sandy, buff ware. One body sherd decorated with lattice pattern. The ware is quite unlike any of the local Belgic wares and is Romano- British. Bronze Site 1: In the Rampart. 8. Bead, of biconical shape. The only parallel I can find to this is from Bridge Hill, near Canterbury, where a similar bead was found in a rubbish pit with late Belgic pottery. (Arch. Cant., Ixxviii (1963), 185-6 and fig. 14,11). Medieval pottery Site 4: South Ditch. Unstratified; from the Nineteenth Century Ditchfilling. Four different wares were present: (i) A hard, sandy, grey ware with grey, buff or pink-buff surface. The recognizable sherds include two from sagging bases and a jug rim (no. 9). Presumably a jug ware. (ii) A hard, light-grey ware with a hght sheU-fiUing61 and a lightgrey or buff surface. Sherds include a cooking-pot rim (no. 10) and one with part of an apphed vertical strip. (iii) A grey ware, coarser than (ii), but with only a hght shellfilling and a pink-buff surface. Only example was a cooking-pot (no. 11). (iv) A coarse, grey ware with a medium sheU-filling and dark-grey surface. Among the sherds are rims from a cooking-pot (no. 12) and a bowl (no. 13). Although it is unstratified a date towards the end of the thirteenth century may be assigned to all the pottery. 61 Following S. E. Rigold's definition in Arch. Journ., cxxii (1965), 127. 77 D. B. KELLY 9. Jug; flattened rim and grooved neck. Ware i. (For form, cf. B. Philp, Excavations at Faversham, 1965, 56 and fig. 17, 115-6; S. E. Rigold, 'Temple Manor, Strood', Arch. Journ., cxxh (1965), 129 and fig. 14, 6—just after 1300 here.) 10. Cooking-pot; down-turned, flanged rim with shght inner bead. Ware ii. (Cf. example from Pivington: Arch. Cant., Ixxvii (1962), 40 and fig. 4, hi, with shght sheU-filling, though a different ware.) 11. Cooking-pot; flanged rim with upturned bevel on underside. Ware hi. (Cf., e.g. Pivington, ibid., fig. 4, h and v.) 12. Cooking pot; horizontal flanged rim, undercut. Ware iv. 13. Bowl; flanged rim with single groove on top. Ware iv. APPENDIX The Pottery Find of 1911. In July, 1911, pottery sherds were found while the soil and loam overlying the ragstone were being removed before quarrying (see Fig. 2). These were correctly identified as 'Late Keltic' by the Rev. W. Gardner-Waterman, vicar of Loose, who inspected the site with Mr. Hubert Elgar, of Maidstone Museum. The pottery (Figs. 10-12) is in Maidstone Museum, with a short account of the discovery by Elgar and a watercolour by C. Walter showing the site of the find, but neither the account nor the pottery have been pubhshed previously. In his account Elgar describes the section, which rested on the uncovered ragstone, many feet above the quarry floor: 'The section showed 6 ft. of undisturbed brick-earth, about 3 ft. of loam with brick-earth debris and 1 ft. of surface soil. For some distance along the face of the excavation, and just above the undisturbed brick-earth, fragments of urns and pieces of bones could be seen in the loam, evidently marking the site of the bottom of a ditch or trench. On the west of the excavation, where the deposit had not been removed quite so rapidly, a cross-section was displayed, which showed plainly that at one time a ditch existed to a depth of about 3 ft. in the brickearth, and was filled with loam and gravels from the surrounding rocks. The lower portion of the infilling . . . contained numerous fragments of pottery, interspersed with pieces of bones belonging to domestic animals,62 thin layers of carbonaceous matter and small pieces of burnt wood.' GeologicaUy, the section differs from those observed during the excavation and in the existmg quarry face. There are, however, extensive sheets of head brick-earth north of the site in the dip-slope of the Hythe beds, and it apparently occurs quite commonly in guUs in the 02 The few bones still preserved are of pig, sheep or goat and cow. 78 QUARRY WOOD CAMP, LOOSE Hythe beds.63 Alternatively, 'hassock' may be meant by 'brick-earth'. Unfortunately, any hopes that the 'ditches' might be the remains of defences on the north side of the camp are dispelled by the watercolour drawing. This is intended to be a sketch rather than an accurately measured section, since the depths of the various deposits shown are very different from those described in the written account, but, even so, the main features must be correct. I t shows two sections, at right angles to each other, both cut through what are clearly adjacent pits. Any northern defences had disappeared by the early nineteenth century, as the account and plan of that time show64. The Pottery (Figs. 10-12). PracticaUy aU the pottery appears to the eye to be of the same fabric: a fairly hard, sandy, grey ware, slightly gritted. Occasionally differences in firing have altered the surface colour, but with the exception of a few sherds—nos. 15, 17, 19, 32 and 33—which are of a different fabric, it has not been thought necessary to describe the ware for each individual vessel, though surface colour and treatment are described. Of the thirty-three vessels illustrated, ten are bead-rim jars and two of the body sherds probably come from this type of vessel. The remainder are nearly aU jars, ranging in size from very large storage-jars with heavy rims to a smaU cordoned bowl. Combing or furrowing is a frequent form of decoration and occurs on about half of the sherds which are not illustrated. Twelve vessels have some form of cordoning or grooving and five have tooled decoration. The assemblage is typical of the pottery found on late Belgic domestic sites in East Kent, such as those at Canterbury, Snargate and Faversham.65 Bead-rim jars, large storage-jars and smaller jars and bowls, with cordoning and combed and tooled decoration, are well represented on these sites. When, for lack of Kentish examples, reference has had to be made to roughly comparable pottery from Camulodunum,66 it is noticeable that this pottery is, or can be, late. 1. Large storage jar; rim burnished inside and outside. 2. Large storage jar; rim black-burnished inside and outside. 3. Bead-rim jar, corrugated below rim; brown-grey burnishing outside and on inside of rim. 4. Bead-rim jar, the shoulder decorated with grooves; black burnishing on outside. 03 Geological Memoir, loc. cit. in note 1, 106. 64 See note 2. 05 Canterbury, Rose Lane, Arch. Cant., lxviii (1954), 104-14; Snargate, Arch. Cant., Ixxxiii (1968), 265-6; Faversham, B. Philp, Excavations at Faversham, 1965,1968,76-81. 06 Op. cit. in note 36. 79 D. B. KELLY 1 1 f I ( f / Fia. 10. Quarry Wood Camp: Belgio Pottery found in 1911. (i) 80 QUARRY WOOD CAMP, LOOSE 5. Bead-rim jar, corrugated on shoulder; blaok burnishing outside. 6. Bead-rim jar, shoulder shghtly corrugated; brown burnishing outside. 7. Bead-rim jar, shoulder decorated with grooves, body furrowed; black burnishing on outside. 8. Large bead-rim jar, single line of stamped decoration below rim, body furrowed or striated; dark grey surface, black burnishing on rim and neck. 9. Large bead-rim jar, two cordons below rim, body (or its upper part) decorated with panels of horizontal combing separated by vertical combing; rim and cordons burnished. 10, 11. Large bead-rim jars, with combed decoration on body; rims and necks black-burnished outside. 12. Large wide-mouthed jar, buff surface inside and outside; exterior and inside of rim and neck have hght brown burnishing. 13. Jar, with single cordon at junction of neck and shoulder; black to dark grey burnishing outside. 14. Corrugated bowl with single line of tooled lattice decoration; fired inside and out to a reddish brown; black to dark grey burnishing on exterior and inside of rim and neck. 15. Cordoned jar; hard, reddish-brown ware, exterior black-burnished. (Cf. Rose Lane, Canterbury, no. 39 (Arch. Cant., lxviii (1954), 110)). 16. Sherd from shoulder of jar, with cordons and tooled chevron decoration; brown-grey exterior and traces of burnishing below neck. 17. Jar with short, upright neck; hand-made; coarse, dark grey ware, heavily gritted; exterior brown-grey. Base of jar of similar ware from Scrubb's Lane, Maidstone (Maidstone Museum, unpublished). 18. Jar with short, upright neck; hand-made; exterior black-burnished and striated. 19. Jar with everted rim; hard, reddish-brown ware, brown-grey, surface, with traces of burnishing. 20. Jar with horizontally out-turned and flattened rim; dark grey burnished exterior. Rim hke Camulodunum 267B, but this is Roman. 21. Large cyhndrical cup or beaker, with noticeable grooving on outside; dark grey burnishing. Possibly the top of a pedestal beaker, something like Camulodunum 79, but larger and without the cordons. • 22. Small globular bowl, with bead-rim; black burnishing. 23. Sherd of a large vessel of nodular rusticated ware; black-grey surface. Camulodunum 97, a globular beaker, has similar studded decoration, though applied more regularly. Our example is a 81 D. B. KELLY ; / ; f 7 1 ; ;JU^S FIG. 11. Quarry Wood Camp: Belgio Pottery found in 1911. (i) 82 QUARRY WOOD CAMP, LOOSE *L > \ 28 / lH 33 :l 29 V 1 / • I . a 30 34 V • 1 WSBHtm 35 31 BRK FIG. 12. Quarry Wood Camp: Pottery and Brooch found in 1911. (i; no. 34, %; nos. 33 and 35, J) coarser and thicker ware. The only stratified example at Colchester was from period IV (A.D. 49-61). This type of decoration is fairly common in the Roman period, though not in Kent. (See F. H. Thompson in Ant. Journ., xxxviii (1958), 15-51.) 24. Sherd of vessel of hnear rusticated ware; ware and surface as no. 23. Camulodunum 99 is another globular beaker with similar hnear rustication and occurs in periods IV-VI there (A.D. 49-65). (See F. H. Thompson, ibid.) 25. Sherd of vessel with decoration of close and continuous stabbing; ware and finish as nos. 23-4. The various forms of Camulodunum 108, a globular or ovoid beaker of finer ware, have wide zones of stabbed decoration, though not as close as on our sherd. At Colchester these occur in periods I-VI (A.D. 10-65). A sherd from a large jar from Rose Lane, Canterbury (Arch. Cant., lxviii (1954), 106 and fig. 4, 13) has simUar stabbing. 26. Sherd from neck and shoulder of, perhaps, a biconical jar; orangered surface, the neck burnished; decorated on shoulder with tooled obhque hnes. 83 D. B. KELLY 27. Sherd from body of large jar; brown surface with tooled decoration of concentric loops. 28. Sherd from body of jar, with tooled lattice decoration. 29. Sherd from shoulder of a large storage jar with single zone of stabbed herring-bone decoration below the neck groove; brown, smoothed surface. Fragments of a large storage-jar from Northborough School, Maidstone, have similar decoration, which also occurs on large storage-jars found at Bexley (Arch. Cant., lxxn (1958), 187, fig. 1, 1-2). The type is common in the early Roman period. 30. Sherd from a large storage-jar of furrowed ware; one of a number of sherds from large jars decorated in this fashion. 31. Base of jar with decoration of vertical combing; reddish-brown surface. 32. Butt-beaker; thin, creamy-white ware. Camulodunum 113, occurring there in periods I-IV (A.D. 10-61). 33. Sherd from base of amphora handle; orange-red ware, with stamp EB. A stamp ED (? D or B) occurs on an amphora found at Nages. (M. H. Callender, Boman Amphorae, London, 1965, no. 566; <7J.L.,xii,82.) 34. Spindle whorl, made from sherd of pot with tooled decoration. Bronze. 35. Bow and part of spring of a La Tene I I I brooch, with knob representing the junction on a La Tene I I brooch. (Cf. R. E. M. Wheeler: Maiden Castle, Oxford, 1943,258 and fig. 83,9, found with Romano- Belgic pottery of c. A.D. 25-70; Borough Green: Arch. Cant., lxvi (1953), 159; Deal, in J. P. Bushe Fox, Excavation of the Late-Celtic Urnfield at Swarling, Oxford, 1925, PL xv, 16. 84