Roman and other remains from Chalk near Gravesend

ROMAN AND OTHER REMAINS FROM CHALK NEAR GRAVESEND By A. F. ALLEN A SHORT interim report of Roman finds made in gravel workings at East Chalk near Gravesend was made in 1953.1 The excavation has now ceased and the foUowing is a fuller report of the various finds made. The site was a featureless field known, according to the old tithe maps, as Lower Shorne Field, in the Parish of Chalk, north of the Lower Higham Road and adjoining the parish boundary between Chalk and Shorne.2 The surface of the field is about 25 ft. O.D., just above the level of the Shorne and Chalk marshes which bound it on its western and northern sides. The Lower Higham Road forms the southern boundary of the field, but to the east no physical mark separates it from the neighbouring fields. Before the excavation of the gravel the site was crossed by a footpath from west to east. The presence of Roman remains in the neighbourhood is well known, there having been a number of earlier finds within a radius of a mile from these new discoveries.3 It should be added that from local enquiries it is fairly certain that other unrecorded finds were made in the older gravel workings which adjoin the present workings to the west, and in another gravel pit two or three hundred yards north-east of the present site. The finds consisted of : (a) Kilns. (b) Rubbish pits and trenches containing pottery and daub or baked clay. (c) Thin strata of pottery charcoal and dark humus which for the want of more certain identification must be called occupation levels. (d) Graves. (e) A medieval well. KILNS These were quite invisible on the field surface and were only revealed when the dragline excavator removing the topsoil cut through them. 1 Arch. Cant., LXVI, p. 150. 8 See 6 inch Ordnance Survey Sheet, Kent X, S.E. Grid ref. 690732. A sketch map showing the location of the finds accompanies this report. Fig. 1. 3 See 6 inch Ordnance Survey Sheet, Kent X, N.W. Arch. Cant., XI, p. 113, XXIII, p. 22, and XXVIII, p. xc. V.C.H. Kent, Vol. I l l , pp. 130 and 167. 144 r Kilns etc. Old gravel pit Medieva Well ""tKtifUtli >.*\M Oueen s ,fr.,tT"> Farm E 4st Court Farm FIG. 1. Sketch map of site ROMAN AND OTHER REMAINS FROM CHALK NEAR GRAVESEND For this reason, although a number of supposed kilns were reported, only one was examined in any detail before it was destroyed by the excavator ; this was found at the extreme eastern end of the pit within a few yards of the Shorne boundary. The bucket of the excavator had cut, in its usual slanting fashion, from west to east through the topsoU and had uncovered a mass of black sooty earth and baked clay, in which a damaged pot could be seen. Upon clearing the loose earth the base of a circular kiln having a diameter of approximately 4 ft. 6 in. was revealed at a depth of between 18 in. and 2 ft. below the surface of the field. It was constructed of a grey/green baked clay and its waUs 2 to 3 in. thick. On the north side the walls stUl stood to the height of about 7 in., diminishing in height until they disappeared to the south and west, having been destroyed at this point by the excavator. The floor of the kUn appeared to have consisted of baked clay, and to have coUapsed after the kiln had fallen into disuse—probably under the weight of rubbish which had been dumped inside. Of the rubbish only pottery remained in a mass of dark humus material of organic origin. In the centre of the collapsed floor was found a circular mass of baked clay of a reddish colour which appears to have formed the support of the kiln floor. Beneath the floor was a layer of charcoal and earth which rested upon undisturbed soil consisting of a yellow clay—a subsoil occurring naturally in the gravel in some parts of the pit. At this stage it was too dark to carry the excavation further and by 9.30 the following morning the work of destruction had been completed. The kUn appears to have been of a similar type to that reported as having been found at Hoo Junction (about 1,000 yards away to the north-east) in 1926.x The soil in and around the kiln produced a comparatively complete specimen of coarse Roman pottery of the midsecond century (Fig. 2, No. 4) and a mass of fragments of the same period. All were apparently thrown in with rubbish after the kUn had gone out of use. Of the other kilns reported, one which had been destroyed to below its floor level was examined, but it was too badly damaged to yield any facts beyond that it was of a circular type. Altogether some ten kilns, or suspected kilns, were reported by the workmen, but this number must be treated as unconfirmed, as the only traces actually examined were mere scrambled masses of dark earth, potsherds, and baked clay, which, as was proved in examining the trench mentioned later, may have been merely rubbish pits. RUBBISH PITS AND TRENCHES I t was whUst looking for traces of other kilns that a number of pits and trenches filled with rubbish were observed. These were scattered over the north-eastern part of the excavations in no apparent order. 1 V.C.H. Kent, Vol. I l l , p. 130. 146 ROMAN AND OTHER REMAINS FROM CHALK NEAR GRAVESEND Most of them consisted of mere masses of blackened earth conspicuous in the lighter coloured gravel, and were normally quite shallow. Of these pits and trenches only two were investigated with any degree of thoroughness. One was an unusually deep pit, and the other a trench which was traced for a distance of about twenty feet. The deep pit was situated some 200 feet to the west of the area in which kilns had been found. It was the most westerly feature found (except for the graves mentioned below) and its depth of approximately 8 ft. was unique. The results of the examination were, however, somewhat disappointing. The mechanical removal of the topsoil had left the top of the pit showing at about 3 ft. below the field level. The first 2 ft. of soU in this remaining pit consisted of clay from the topsoil with occasional potsherds, sufficiently distinct from the surrounding gravel to show that the pit was circular and approximately 5 ft. in diameter. After this the pit was only dug for half its diameter to save time and obtain a profile. Below the first 2 ft. of topsoil darker traces were found of decayed organic material and at a depth of 2 ft. 10 in. from the commencement of digging, a mass of charred black earth 4-6 in. thick was found. This substance was slightly greasy in texture and when wet formed a tenaciously sticky mass. In it were many pieces of broken pottery and an ox tooth. Beneath this layer there was another 5 in. of clay topsoil, followed by what appeared to be a layer of baked clay. The upper surface of this baked clay was a reddish brown, but underneath it changed to a yeUow brick colour. It was so broken that it could not be certain whether it was in situ, or was a mass of kUn material dumped into the pit. Beneath this layer was a deposit of yellow clay, simUar to that found beneath the kiln reported above, except that this clay contained some fragments of pottery. At this tantalizing stage the excavation of the ballast removed everything. The pottery in this pit appears to have been of an earlier type than that in the kiln, particularly the rim of a storage jar of Patchgrove ware, and it seems that this pit belongs to an earlier period of occupation than that of the kiln, probably first century. The rubbish trench was found slanting in a north-westerly direction from the approximate position of one of the reported kUns, though no direct association could be proved. The trench was approximately 1 ft. 10 in. deep, having a curved bottom. It produced a large number of fragments of pottery and much partially baked clay, resembling kiln walls. The pottery seems to have been mostly second century coarse ware. OOOTTPATION LEVELS The trench and rubbish pits were associated with many " traces " observable in the banks left after the topsoil had been removed. From 147 ROMAN AND OTHER REMAINS FROM CHALK NEAR GRAVESEND these traces in the perpendicular bank of topsoil it was interesting to observe a marked layer of soil, approximately 9 in. to a foot below the surface which had at some time, probably the medieval period, been liberally sprinkled with small particles of chalk, which is not a natural part of the soil at this point. Beneath this the traces of Roman pottery were sufficiently continuous to suggest that the general level of the Roman occupation was between 1 ft. and 1 ft. 3 in. below the present field level. Attempts to define more clearly what these traces represented were defeated by the work of gravel digging, and it is not possible to say more than that these traces indicated a period of occupation extending over the first two centuries of the Roman period. At this point mention may be made of two finds made by Mr. Cornwell, the driver of the excavator. One was that of a number of fragments of a large storage jar of Patchgrove ware which were found apparently in this occupation level a few yards from the kUn reported above. There is no evidence to associate these fragments with the kiln, and most of the casual fragments of pottery taken from the other traces were of the normal second century type. The other find was at a point not clearly identified and it consisted of a flagon and Samian ware plate. Both are early second century ware, and the only information avahable about them is that they were found in the loose soil not far from a large collection of potsherds which appears to have been a waste heap from a kUn. GRAVES Early in 1953 Mr. Cornwell reported that he had found a flagon in the spoU tipped by his machine whUst digging at the western end of the pit, some 300 yards from the earher finds. He also reported that small dishes had been observed by the other workmen in the lower ballast. The flagon (Fig. 3, No. 29) was obviously of a later period than the previous finds. Shortly afterwards, whUst removing the topsoU, he observed in the gravel a painted two-handled flagon (Fig. 2, No. 22) and it became apparent that the excavator was cutting through graves, which showed as darker rectangles in the gravel. A close watch was then kept and nine graves in aU were located and examined as far as circumstances permitted. All were orientated on a north-south line, and from the boot studs found in one or two of them, it seems that aU the burials were with the feet to the south, and with the pottery at the feet. No trace of the skeletal parts of the body could be found in any of the graves, the excessive acidity of the soil having presumably destroyed them. AU the graves were approximately 2 ft. 6 in. to 3 ft. wide, 7 ft. long, and at varying depths between 3 and 5 ft., with the exception of Grave No. 8, which was 4 ft. wide, 8 ft. long and some 2 ft. deeper than 148 ROMAN AND OTHER REMAINS FROM CHALK NEAR GRAVESEND the rest. It should be noted that this grave also contained the most pottery. Because of the varying depths of the graves some were practically destroyed by the excavator, but others (Nos. 1, 4, 5, 7 and 8) were dug completely. The pottery found in these graves is catalogued in the appendix, and in addition various pieces of badly corroded metal were found in those which were completely excavated. In Graves Nos. 5 and 8 what might have been circular brooches were found in a central position in each grave, but the corrosion was too advanced for their nature to be determined. Grave No. 7 was notable for providing a number of large iron nails, presumably coffin naUs. Grave No. 8 had two large nails near the foot of the grave, and Grave No. 9 had what appeared to be an iron pin resting upon the plate. In each of the completely dug graves traces of a wooden coffin (represented by dark lines of decayed matter in the lighter gravel) were found. Generally the pottery from the graves is all of fourth century type and is of better class than the fragments found elsewhere in the field. A notable feature of every article found is its worn state ; a condition clearly present before burial. Two or three are without handles, the fracture point being obviously well worn after the break by constant handling. GENERAL OBSERVATIONS The information derived from these finds has been somewhat reduced by two factors : (a) The extreme acidity of the soU in which only pottery and baked clay survived ; all skeletal remains having completely vanished, whUe metaUic substances were represented only by fragments of rust too corroded to conjecture their original shape and purpose.1 (b) The method of working the gravel. Two machines were at work the whole time, one removing topsoU and the other ballast, and both had to work to a rhythm which did not permit any margin of time to follow up discoveries. The discoveries made, however, indicate a long continued Romano- British settlement at this spot. For part of the time the inhabitants were apparently engaged in the making of pottery, but it seems that this ceased at the beginning of the third century, and then the settlement continued on an agricultural basis, as is suggested by the finding of bootstuds in the fourth century graves.2 The settlement was primitive and poor. Only one fragment of roof tile was found and no structural remains have come to light (though some of the traces classed as " occupation levels " may have been hut floors), whUst although 1 An example of the effect of the soil on bone may be mentioned. When digging out the rubbish trench a row of ox teeth was found obviously preserving the teeth as they were set in the jawbone. No trace of the bone structure remained and the teeth crumbled into flakes on being moved. 2 I am indebted to Mr. G. O. Dunning for his observations upon this point. 149 ROMAN AND OTHER REMAINS FROM CHALK NEAR GRAVESEND some allowance must be made for the acidity of the soil and the difficulties of examination, the absence of coin finds or of any other objects common in Romano-British settlements, suggests the poverty of the inhabitants. MEDIEVAL WELL The well discovered at the extreme western end of the gravel pit belongs to another period of occupation on this site, and the following short note is given. After the topsoil had been removed a number of flints were observed in the greensand underlying the baUast. The ground was cleared and what was first thought to be a flint cist filled with topsoil was discovered. The dimensions at the top were 1 ft. 7 in. by 2 ft. 6 in., and the aperture was roughly " D "-shaped. WhUst clearing out the filling a small medieval schist hone (now in Maidstone Museum) was found at a depth of approximately 6 ft. from the land surface (see Fig. 3, No. 40). Further digging showed that the shaft was roughly square at its lower levels (about 2 ft. 6 in.) and constructed of coursed flints bedded in what appeared to be a mixture of clay and lime. At a depth of about 6 ft. 6 in. from the original surface a water level was reached, and although excavation was continued against a steady flow from springs on the east and west sides of the shaft, the bottom was not reached. The flow of water made it reasonably certain that the shaft was that of an ancient well, and further excavation had as its main object the determination of the age of the well. Several small fragments of coarse medieval earthenware were found in the sediment at the bottom, and when the flint walls were demolished, another fragment of a similar type was found in the filling between the flints. (Fig. 3, No. 38). At the extreme bottom of the digging (but not apparently the bottom of the well) wooden beams were found on the west and north sides of the well, in an advanced state of decay. The sections, however, were not sufficient for dendrochronology to determine their date. Although the bottom was not reached, the pottery found shows that this well formed no part of the Romano-British settlement, but to some otherwise unrecorded medieval settlement at this spot.1 A survey of what is known of this area in medieval times suggests that it was probably part of the manor of East Court at Chalk, but there were two other places loosely called '' Manors," Tumberwood and Ockenden, in this district, the sites of which are now lost, which may be represented by this old weU. I should like to take this opportunity of expressing my great 1 A fragment of medieval pottery was found in the topsoil covering one of the Roman graves at a depth of about a foot (see Fig. 3, No. 39). 150 ROMAN AND OTHER REMAINS FROM CHALK NEAR GRAVESEND obligation to Mr. A. Warhurst of Maidstone Museum, who has not only contributed the appendix upon the pottery but helped me considerably with the excavating and technical side of this report. I should also like to express my thanks to Messrs. Holman, Ford and Butler, who have dug so hard and so often with me, and finally to Mr. Cornwell, the driver of one of the excavators, without whose constant interest and reports much of the foregoing would have never been recorded. THE POTTERY By A. WARHURST, B.A., A.M.A. (A) POTTERY EROM THE KILN SITE AND SURROUNDING AREA Fig. 2, Nos. 1-16 Many hundreds of coarse ware sherds were recovered from the kiln site and surrounding area and it is clear that the mechanical excavators were cutting through the waste heaps of the kilns and occupation debris. Due to the exigencies of excavation in the gravel pit, which was continuously worked at the time, none of the pottery can be regarded as securely stratified. The most that can be said is that it gives evidence of occupation on the site during the first and second centuries and shows a period of activity for the kilns about the middle of the latter. It was at this time that the demand of the comparatively dense population of the Thames and Medway estuarine area for pottery would be at its highest. An attempt has been made, on the evidence of numbers of rims found, to assess the main types which were produced by the kilns. The pottery is notable for its consistently hard, thin, brittle fabric, which is often sandy in texture. Its monotonous drabness is relieved occasionaUy by scored lines, lattice patterning and slip coating, although this last feature is now less prominent than it may have been originally, due to the acidic action of the gravel subsoil upon the surface of the pottery. 1-4. Wheel-made ollae with outbent rims in hard thin fabric varying from red to grey in colour. The rim sections vary from a simple recurve to a sharp eversion and although it would be possible to draw up an evolutionary series of rim forms, no significance can be attached to this variation under the present circumstances. In all cases the diameter of the rim exceeds that of the girth and occasionally the sides of the pots may be scored with single lines or acute-angled lattice patterning. Many of the examples show slip coating on the rim 151 ROMAN AND OTHER REMAINS FROM CHALK NEAR GRAVESEND wawiwwm FIG. 2. Pottery from the kiln site and Graves 1-3 (\) 152 ROMAN AND OTHER REMAINS FROM CHALK NEAR GRAVESEND and shoulder which are frequently burnished smooth.1 This type of jar seems to have been the most common product of the Shorne kilns and the series may be dated with confidence to the middle of the second century.2 1. Olla with simple recurved rim of hard red-grey fabric. Shp coated on the outer rim and neck and on the inner lip of the rim. Diameter of rim 9 | in. 2. Olla with more sharply recurved rim in hard red-grey fabric. Slip coated as above, but burnished smooth. Diameter of rim 5 in. 3. Olla with everted rim in hard red-grey fabric. Shp coated on the exterior rim and neck and on the inner lip of the rim. Diameter of rim 9£ in. 4. Olla with sharply everted rim in hard sandy red-buff fabric, the side scored with single lines. Found in the kiln. 5-10. Wheel-made bead-rim jars of hard sandy fabric varying in colour from red to grey. Bead-rim jars are normaUy dated early in the Roman period,3 but second century examples occur at Richborough4 and at Lullingstone.5 The fabric would seem to be a more reliable guide than the form in assigning the Shorne bead-rims to the second century. 5. Bead-rim jar of hard sandy red-grey fabric with flecks of flint in the paste. Diameter of rim 1\ in. 6. Bead-rim jar of sandy buff-grey fabric and smooth paste. Diameter of rim 9 in. 7. Bead-rim jar, fabric as 6. Diameter of rim S\ in. 8. Bead-rim jar with lid ledge in hard grey-red porridgy fabric. Diameter of rim 10 in. 9. Bead-rim jar, fabric as 6. Diameter of rim 1\ in. 10. Bead-rim jar, fabric as 5. Diameter of rim 6£ in. 11-13. Straight-sided dishes with roll-over rims and chamfered, or rounded, bases. The fabric is grey and hard and the sides are often scored with acute angled lattice patterning. These dishes are sometimes slip-covered and burnished. The type is well-known in the Antonine period.6 1 Cf. A. J. Clark, " The Fourth-Century Romano-British Pottery Kilns at Overwey, Tilford," Surrey Arch. Coll., LI, p. 48. The same phenomenon was noted here on similar jars of a later date. 2 Cf. Richborough, types 320 and 464. 3 A. W. G. Lowther, "Excavations at Ashstead," Surrey Arch. Coll., XXXVIII, p. 17 and Fig. 7. 1 Richborough, type 251. 6 " The Lullingstone Roman Villa," Arch. Cant., LXIII, p. 17 and Fig. 5, Nos. 4 and 5. 0 Richborough, type 46. 153 ROMAN AND OTHER REMAINS FROM CHALK NEAR GRAVESEND 11. Side of a dish with roll-over rim in hard red-grey fabric, slipcovered and burnished on the inside and much of the outside. Diameter of rim 9 in. 12. Side of a dish with roll-over rim in hard buff-red fabric with traces of salmon coloured slip. Chamfered base. Diameter of rim 9 in. 13. Side of a dish with roll-over rim protruding internally, in hard grey fabric with a smooth surface. The junction of the side and base is rounded and the side bears traces of acute angled lattice patterning. Diameter of rim 9 in. 14 and 15. Straight-sided dishes in hard sandy fabric varying in colour from grey to red. The type is normally late second century.1 14. Side of a dish with chamfered base in hard buff fabric, fumed black. Diameter of rim 8 in. 15. Side of a dish in hard red-grey fabric with salmon coloured slip. Burnished on the outside and decorated with a scored zig-zag pattern. Diameter of rim 8|- in. 16. Bowls of hard sandy red or grey fabric with a cordon at the neck just above a zone of scored hatching. The example illustrated is coated with a pinky-white slip and burnished on the outside and the inside lip of the rim. A similar bowl at Richborough is dated to the end of the first century,2 but the fabric and pottery associations of the Shorne examples suggest a second century date. (B) POTTERY FROM THE GRAVES (Figs. 2 and 3, Nos. 17-36) Despite the interference of the mechanical excavator with the investigation of the graves, it has been possible to reconstruct the grave groups satisfactorily. There is little reason to doubt that all nine graves were roughly contemporary but the burials cannot, unfortunately, be dated by any other method than a consideration of the pottery types. Much .of the pottery shows a superficial resemblance to New Forest ware but it is unlhkely that such ware exists in Kent other than as isolated imports, and the pottery was probably made near to Shorne itself. Many pots show a considerable amount of wear and tear before their consignment to the grave. Indeed, vessels, Nos. 20 and 29, were obviously much used after their handles had been broken off and the Samian dish, No. 36, must have been a real family heirloom. On the other hand, vessels, Nos. 20, 21 and 23, which are all hand-made, can hardly have been manufactured before the very end of the fourth century and one of these, with worn handle fractures, was old at the time of the burial. It seems hkely, then, that the Shorne burials took 1 Richborough, type 166. 2 Ibid., type 386. 154 ROMAN AND OTHER REMAINS FROM CHALK NEAR GRAVESEND HKKK 28 K \U\W\V u'.uiYu 38 37 39