The Romano-British Buildings in Church Field, Snodland

THE ROMANO-BRITISH BUILDINGS IN CHURCH FIELD, SNODLAND By M. A. OoocK and M. J. E. SYDDELL INTRODUCTION RESCUE excavations on the site of a known Romano-British building at Snodland revealed that, despite the presence of modern industrial buildings, debris, and thick concrete yards, quite substantial foundations still remained. Much was gained from this sadly-neglected site before the only remaining small area was finally obliterated. The work was carried out by the authors on behalf of the Lower Medway Archaeological Research Group. The parts of the building found showed evidence of three periods. Occupation on the site lasted from about the end of the first to the middle of the third century A.D., and perhaps, elsewhere on the site, into the fourth century. The remains excavated comprised part of what may be a bath suite, a range of rooms and a right-angled corridor. It is possible that, if and when the adjacent factory premises are demolished, further remains may come to light. THE SITE The site of the building (N.G.R. TQ70756203) is within the grounds of the Snodland Gas Holder Station and the factory of the Lead Wool Co. Ltd., approximately 650 ft. north of All Saints' Church, on the west bank of the River Medway; the area is known as Church Field. The present village of Snodland lies on a spur of chalky brick.earth and gravelly loam, overlying low terrace gravel, which extends down to the River Medway. The site lies at the edge of this relatively higher ground, with alluvial deposits immediately to the north. In the area of the excavations the immediate subsoil was a sandy loam, in places 2-3 ft. thick, overlying the head deposits. Waste from the old gasworks had in places raised the level of the ground by several inches. EARLIER, DISOOVERIES The existence of a Romano-British building in Church Field has been known for a long time. The first reference to discoveries was made by C. Roach Smith (Arch. J., i (1845), 164) who noted in 1844 that 'tesserae of Roman pavements and fragments of roof and flue tiles and pottery' were scattered about the field. He also recorded the remains of walls, a floor and other features exposed in the river bank. 192 No.2. Church field. (since dtmolish•dl. Gardtn. (.,􀀂 o'􀀃 rr--1· 􀀂, . 11 2:i 􀀂i r-, r- KEY PERIOD 1. INFERRED. PERIOD 2 INFERRED. PERIOD.3. INFERRED. PERIOD UNCERTAIN. INFERRED. PAVIN<:,. POT. [􀀂 JW@l t.CTz] r• L􀀂 [: 􀀅?itfil:1 'P ---, !· ' Pi1-2􀀉f _ J MORTAR FLOOR 􀀑 -􀀒-􀀓 ... .. ,, ' ___--:-_.􀀆_ j ROMAN FOUNDATIONS REVEALED 6V PIT EXCAVATIONS AT NORTH ENO OF HOLDER STATION. SCALE. FEET 10. 5. o. METRES 5. 10. 20. 0 S. PLAN OF THE ROMANO-BRITISH SITE IN CHURCH FIELD,, SNODLAND. ·-·-·.i SCALE. FEET. METRES. tO 5. 5. 0. 􀀧Edg• of concr•t• yard. o. 10, 20. 30. ,o 50. tO Ft(:. I. 􀁟ite plan. Inset, with Romano-British buildings shown solid. ha;,cd on O.S. I : :?.,,00 plan. C'tom, Cop􀁠·right resen·ed. THE ROMANO-BRITISH BUILDINGS, SNODLAND The second recorded observation was made in the latter part of August 1844 by Thomas Wright, and published ten years la.ter in 1854 in Wanderings of an Antiquary. Wright noted that the adjoining field was significantly named 'Stone Grave Field'.1 He also made some small excavations and uncovered 'part of a floor of large tiles' about a foot below the surface. He received information that 'a bath ... (was) discovered in this field about forty years before and .. . (had been) filled up without undergoing any further inquiry'. At the beginning of this century a gasworks was built on part of the site but no discoveries appear to have been reported. However, extensions to these works in 1927 resulted in a number of finds. These were reported in Arch. Cant., xl (1928), 79, by Mr. Norman Cook, B.A., F.S.A. The structural remains comprised 'some pieces of shapeless foundation and a length of wall showing a flint course on the top of a footing made of two layers of chalk boulders'. The two most important finds were a terracotta mask and a bronze buckle-plate of a rare type.2 The buckle-plate is described in Ant. J., vii (1927), 79. An up-to-date assessment is given in Appendix III. The pottery found was dated to the periods late-first, second and fourth centuries A.D. Five coins were found, ranging in date from A.D. 87 to .A.D. 375-83. In 1933-35, during construction of a new factory for the Lead Wool Co. Ltd., several foundations were uncovered but not recorded, and two stone coffins were also found. One of these coffins contained a complete human skeleton. The skeleton was placed in the Royal College of SUl'geons Museum, but the whereabouts of the coffin is unknown (Arch. Gant., :x:lvi (1934), 202; J.R.S., x.xiv (1934), 217). A possible burial found many years ago somewhere close to the present site was reported to Mr. R. F. Jessup, F.S.A., but further details were unobtainable and the pottery passed into private hands. It may be worth mentioning here the presence in the British Museum of two complete samian vessels and a pottery lamp from 'the Roman villa at Snodla.nd' (Appendix II). THE Exo.A.VATION In August 1964, Mr. E. R. Swain received information concerning site clearance operations by the South Eastern Gas Board which had revealed structural remains. Permission to excavate was quickly sought and readily granted by the South Eastern Gas Board and the Lead Wool Co. Ltd., to whom we are greatly indebted. Work began in September 1964. The site was basically divided into three parts. Firstly, there was the Gas Board area (the lower level), 1 In the 1840 Tithe Apportionment map the area is given as 'Stone Quarries'. 2 The whereabouts of the mask are at present unknown. A cast of the mask and an electrotype of the buckle are in Maidstone Museum. 193 THE ROMANO-BRITISH BUILDINGS, SNODLAND which waa approximately 5 ft. below the adjacent land surface because of site clearance operations which had extended a coke storage yard. The storage yard had a thick concrete slab, part of which once formed the floor of the retort house built in 1927. Bulldozing operations in this area revealed several foundations. The second part (the upper level) was a narrow strip ofland between the coke yard and the adjacent factory; this area had for many years been used as garden allotments, but much was later concreted over. A great deal of the stratification had been destroyed by intensive cultivation which in some places had removed all but the floor and wall foundations of the Romano-British building. The third part, at the same level as the second, was the garden of No. 2, Church Field, in the grounds of the Gas Holder Station. As well as thick concrete garden paths, a large brick and concrete underground air-raid shelter occupied much of the area. Due to the sheltered nature of the site, work continued throughout much of the winter and :finished in June 1965. Much of the lower level of the site ha-s since been completely destroyed, whilst most of the upper level ha.a been covered by a concrete yard and the concrete floor to a new factory extension. The garden area has been levelled and No. 2, Church Field, demolished. The finds have been deposited in Maidstone Museum with the exception of the bones from the infant burial which are in the British Museum (Natural History). AOKNOWLEDGMENTS The excavation could not have been successfully concluded without the willing help of the landowners. For this our grateful thanks are due to Mr. D. Whittaker of the Lead Wool Co. Ltd. and Mr. J.E. Steel and Mr. A. W. Barnett of the South Eastern Gas Board, for permission to excavate and for considerable assistance, which included providing facilities and equipment. We are very grateful to the willing team of helpers, in particular Misses J. Bevan, P. Flood, J. Fox, T. Luton, Mrs. W. Ocock and Mrs. H. Syddall, and Messrs. D. Ayres, R. Chapman, J. Crudgington, C. P. Flint (who also surveyed the site), R. G. Foord (who also took charge of the monochrome photography), M. J. Jessup (who also drew the plans for this report) and C. Martin. We also wish to thank Mr. E. R. Swain for drawing the small finds. The scope of the excavation wa.s greatly increased by the work of Mr. B. Jessup and Mr. L. Filmer who, by operating pneumatic drills, removed large areas of concrete. _Much help and encouragement came from other members of the Lower Medway Archreological Research Group, in particular our 194 TH;E ROMANO-BRITISH BUILDINGS, SNODLAND chairman, Mr. A. C. Harrison, B.A., and also from Mr. D. B. Kelly of Maidstone Museum. We are also greatly indebted to those who have contributed to this report, and to Mr. A. P. Detsicas, M.A., F.S.A., for reading it in draft form and making many valuable comments. THE STRUOTu:RAL REMAINS Lower Level This area was cleared of rubble and the foundations exposed were found to comprise part of a hypocaust, an apsidal structure, and the floor of what could have been a cold plunge-bath. All connecting links between these and other buildings had been destroyed. It appears that the construction of the 'bath' can be dated to the middle or latter half of the second century A.D. It is impossible to date the apse and hypocaust due to lack of associated finds and the fact that their construction was rather unlike all the other walls found. Hypocaust. The top floor of the hypocaust had been removed by the bulldozer, but the under-floor and many of the pilae, together with the filling, remained, to a depth of 1 ft. 3 in. Due to previous destruction, only one corner of the hypocaust remained intact. However, a second corner could just be traced disappearing under modern concrete, giving a room 9 ft. 6 in. in width. The walls were constructed of ragstone, with a herring-bone chalk and ragstone rubble backing; they had been built by sinking a pit into gravel and lining the sides. The bottom courses of the walls, laid on gravel, consisted of a course of tiles overlying a foundation of flints set in clay. The south-east wall tapered from a thickness of 2 ft. 6 in. at the edge of the concrete to I ft. 9 in. at its junction with the south-west wall, which was 2 ft. thick along its whole length. At the junction, just determinable, of the south-west wall with the north-west wall, the latter was 3 ft. thick and appeared to widen considerably, but further investigation of this point was prevented by the modern concrete yard. The mortar here was generally yellow in colour. The under-floor, l½ in. thick, was composed of opus signinum laid directly on the subsoil. The pilae were arranged in a regular pattern, normally at 2-ft. centres. The base of each pila comprised a red bonding tile, usually measuring about 1 ft. X 1 ft. 6 in. The red pila-tiles averaged 8 in. X 8 in. Maximum height to which a pila survived was 1 ft. 2 in. The filling of the hypocaust consisted of a layer of soot, generally 1-2 in. thick but increasing in thickness to about 6 in. at the east end. Over this lay rubble debris, which contained several large red clay tesserae, iron nails, and some small fragments of red and green painted wall-plaster. 195 THE ROMANO-BRITISH BUILDINGS, SNODLAND Apse. The apse was 10 ft. west of the hypocaust. It had a chalk block wall with a. rough outer 'skin' of bonding tiles laid vertically in mortar, probably constructed in the same manner as the hypocaust walls. The thickness of the chalk wall was about 1 ft. 2 in., and the outer skin was about 1 ft. wide. The inside of the apse had two opus signinum floors. The lowest was laid, at about the same level as the under-floor of the hypocaust, on a 3-in. bed of crushed chalk. The floor was 5 in. thick and had a 3-in. quarter-round moulding at its junction with the outer wall. The floor and moulding were later covered with a second opus signinum floor of the same thickness. After the second floor had been laid, the original mortar facing to the wall was thickened out with two thin layers of vertical bonding tiles, 3! in. thick overall, faced internally with 1 in. of mortar. The first floor was smooth and displayed little sign of wear. The upper floor had a, rougher surface and was covered with traces of soot which ran under the tile and mortar internal thickening. The wall remained only for a height of 4 in. above the upper floor. It is possible that this apsidal feature was either part of a bath or the lower floor of a hypocaust, although the presence of the quarterround moulding on the lower floor would seem to point to the former, at least in the earlier phase. In the case of the second floor, the tile and mortar thickening may indicate that either it became necessary to protect the chalk wall from the heat of a hypocaust or to minimize water seepage if the feature was in use as a bath. Oold plunge-bath or storage t,a,nk. The 'bath' was situated approximately 20 ft. to the south-west of the apse, but there was no trace of any connection between the two. The east corner of the 'bath' had been destroyed by the construction of a reinforced concrete slit trench. The remainder of the 'bath' had been almost removed, down to its floor level, by the bulldozer, except at the south corner by the bank. The internal size of the 'bath' was approximately 9 ft. 6 in. by 10 ft. 9 in. The walls were of bonding-tiles laid on a foundation of flints set in clay; they were 1 ft. 6 in. thick and the 'bath' was built in the subsoil in a similar manner to the apse and hypocaust. It appeared that a pit had first been dug into the subsoil, in line with the other parts of the building. However, the 'bath' walls had not been built parallel with the pit sides and, consequently, the structure was not on the same axis as the remainder of the buildings. The 'bath' was surrounded by a back-filling of gault clay. The opus signinum floor was approximately 5 in. thick and laid on a foundation of flints and ragstone blocks set upright in clay on the subsoil; the flints and ragstone blocks were laid in lines diagonally across the 'bath'. .Around the perimeter of the 'bath' was a 3-in. coved skirting, partly destroyed on 196 THE RO:M:.ANO-BRITISH BUILDINGS, SNODLAND the north side but intaot on the other three. By a fortunate chance, a piece of internal mortar of the 'bath' was held in its original position by the concrete sides of the slit trench, 2 ft. above the floor of the 'bath'; the 'bath' therefore was probably at least 2 ft. deep. In the south corner of the 'bath', the filling was relatively undisturbed and comprised a very dark topsoil, similar to the modern topsoil in surrounding areas. This suggests that the 'bath' was the one referred to by Thomas Wright in 1854 as having been dug out in the early part of the nineteenth century and subsequently refilled. As there was no evidence to connect the 'bath' with the rest of the building it is not certain that this feature was a cold plunge-bath; its size and construction, together with the fact that the floor level was only a few inches above the under-floors of the apse and hypocaust, show that it was used to contain water. It could have been either a bath or a storage tank, but it is thought more likely to have been the former. The feature is unlikely to have been fed by a stream as the nearest known today is at a lower level and at least a quarter of a mile a.way. Leading directly towards the 'bath' from the upper level of the site, but destroyed. when about 4 ft. from it, was a shallow, straightsided and flat-bottomed ditch. This was generally 3 ft. wide and l ft. deep and sloped down towards the 'bath' from the direction of the factory. The ditch was dug into the soft brown soil which appears elsewhere on the site to be the Roman topsoil. The bottom and the lower 6 in. of the sides of the ditch were covered with a thin hard brown crust, which is generally considered to be caused by the precipitation of constituents of iron compounds from water seeping through organic material. The bottom of the ditch was, at its lowest point, 3 ft. 9 in. below the present ground surface. The primary silting and the bard bottom crust suggei,-t that the ditch was i n use for a considerable period. At a later stage, the ditch was filled in with rubble and a rough pebble paving laid over it and the surrounding area, patched in places with crushed chalk. No doubt due to consolidation of the ditch filling, the paving sank along the line of the ditch. Above this was a thick layer of building debris and soil. It was impossible to follow the line of the ditch southwards, and due to its destruction at the north end it was also impossible to determine whether it was associated with the 'bath'. Upper Level It is proposed to describe the remains on the upper level in two parts as the e􀃦cavations were conducted in two separate areas and it was not possible to investigate the section between. 197 THE ROMANO-BRITISH BUILDINGS, $NODLAND East End Period 1. Period l comprised Walls A and Bl. Wall A consisted only of foundations, l ft. 6 in. wide, of flint with yellow mortar. Wall Bl was of similar construction and in parts existed to a width of I ft. 9 in. Together they formed part of two sides of a room which extended northwards into the area destroyed by the gasworks. The floor of this room had been removed in Period 2 except for the 'feature' which had been only partly destroyed. The 'feature' appeared to have been a step or raised platform, 2 in. high, of unknown use. One edge ran parallel with Wall BI and the other existing edge parallel with Wall A. Only the edges, which were futished with a smooth mortared and well-rounded top, remained, the centre appearing to have been destroyed and replaced with rubble filling for the floor above. The area south of Wall Bl was paved with a chalk floor. A thin layer of burnt clay and a semicircle of tile fragments on the chalk floor abutting Wall Bl may have been a small hearth. Period 2. In Period 2, Walls A and Bl were demolished except for their foundations and the floor taken up. Walls E and F were constructed and a tessellated floor laid on a foundation of chalk blocks. The Period 2 building was built over and beyond the area covered by the Period l structure. Wall E, 2 ft. wide, was represented by one course of bonding tiles laid on a foundation of flints for part of its length and chalk blocks for the length within the area o f the Period 1 building. Wall F formed a right angle with Wall E and enclosed an area originally outside the building during the previous period; it was 2 ft. 3 in. wide and built of flint and ragstone, with yellow mortar. Walls E and F formed parts of the walls of a room with a tessellated floor. The floor was constructed on a base of opus .signinum laid on a foundation of large angular chalk blocks, within the area of the Period 1 building, and rubble filling for the remainder of the area, e;iccept for the parts where the lower courses of Period 1 walls were used as foundations. Only small a.reas remained with the red clay tesserae in situ, as modern cultivation had penetra.ted in many places to the floor base level. Overlying the floors in several places were fragments of painted wall-plaster showing traces of green and red. During the removal of the tessellated floor base, two small white tesserae were found on the chalk block foundation; they may have come from a mosaic elsewhere on the site but no evidence to support this was found during the present excavations. Under the tessellated floor, and abutting against the west face of Wall E, almost in the corner of the room, was a small tile oist containing the bones of an infant. It was constructed of fragments of roofing tile. It was not possible because of later disturbances to determine whether 198 8. J'lww: R. a. Foord A. Tho 'bath'. '['110 l' l'/1(,tu: R. G. Foord ppe,· Lc\'ol h􀀣·pocat1st. I fa,·, p. I u􀀋 THE ROMANO-BRITISH BUILDINGS, SNODLAND this was an intrusive burial or whether it had been deposited at the time the floor was constructed. Careful examination was made of the surrounding area in an attempt to recover as much as possible of the skeleton. Examination of the bones (see Appendix VII) showed only a few of them to be present and this, together with the presence of animal bones, cuuld suggest a reinterment. In the rubble surrounding the cist were found a number of sherds of a late-second-century storage pot (see Dating Evidence, No. 5). The sherds conjoined to form about one-third of the pot. As the area around the burial was thoroughly searched, and no more was found, it seems likely that only this part of the pot had been originally usedperhaps as a crude bowl containing offerings. Period 3. During Period 3 an extension was made to the earlier building, which continued in use. Walls Q, Rand S were added to form a corridor running south-east away from the ma.in building. Thirty feet from Wall E, internally, the corridor made a right-angled turn to the south-west and presumably connected with other parts of the building now under the adjacent factory. The corner of Walls Q and R was traced, but the corner of the other walls could not be investigated because of the existence of thick concrete. Wall Q was 2 ft. thick and consisted of a foundation of challc blocks and one course of bonding tiles. At its junction with Wall E, Wall Q appears to have been built on the foundations of Walls A and Bl of the first period. Wall R was in a poor condition with only the very lowest courses e.xisting, but appeared to be 2 ft. thick. Wall S was 1 ft. 8 in. thick and of similar construction; however, instead of abutting against the corner of Wall E, it continued to run alongside Wall F, acting as a thickening and giving an overall width of 4 ft. In this area a greater height of wall remained, showing a construction of flint and ragstone with yellow mortar. The fact that Wall F continued in use during Period 3 is shown by the presence of a pot built into the inside face of Wall S, partly above floor level. The pot, which may have held a votive offering was found containing only earth. It was placed in a 1-ft. square recess and packed around with small flints. The pot is a second- or early-third-century storage jar (see Dating Evidence, No. 16). The floor of the corridor had a foundation of broken tiles laid at a slope in herring-bone fashion. Above this was a layer of opus signinum which acted as a base for what appears to have been a tessellated floor of the same type as that found in Period 2. None of its tesserae were found in situ but the construction of the base and the presence of large numbers of red tessetae in the soil above seem to support this suggestion. Around the eastern sides of the corridor, a paving of compacted pebbles and chalk was exposed. This was laid on a surface which 199 10 THE ROMANO-BRITISH BUILDINGS, SNODLAND sloped away from the building, in some places quite steeply and close to walls. Part of a quern stone was found in the building debris overlying the paving (see .Appendix IV). On the other side of the conidor there were no traces of this pebble paving, only the original topsoil. Fragments of painted plaster, mainly red in colour, and many iron nails were found over the whole of thls east end. West End The work in this area was limited by the existence of modern features. Because of this it was not possible to prove any connection between the east and west ends of the structural remains, although certain assumptions can be made. The remains cannot be definitely associated with any of the periods identified at the east end. What appeared to be the end of Wall Bl of the east end was traced for about 7 ft., the intervening length having been destroyed. This wall, B2, was 1 ft. 9 in. thick (as was BI in places). The construction of the two walls was exactly the same and they were on the same alignment. The south, or inside face, of the wall showed patches of plaster in situ. There was no trace of colouring on this plaster. Six feet three inches to the south of B2, was Wall H. This was of flint and mortar construction also, but showed no traces of any plaster facing. The south-east side of the room was formed by Wall G, which could only be exposed in the concreted face of the bank. Wall G was 1 ft. 6 in. wide and only the foundation of chalk blocks remained. It should be mentioned here that on the south-east side of Wall G traces of the tessellated floor make-up similar to that of Period 2 in the east end were found. The fourth side of the room thus enclosed was formed by Wa.11 C which continued to run beyond it in a north-east direction. The room was 11 ft. X 7 ft. and had a 2-in. thick floor of rammed chalk. Wall C continued beyond the room for 11 ft. and then made a right-angled turn to the north-west. The soil to the south-east side of Wall C was of the soft brown type, indicating that Wall Chad been an external wall. On the south-west side was a chalk floor laid directly on the soft brown soil. The presence, directly on the floor, of a neat fall of tiles from the collapsed roof of the building showed that the floor had been the only one laid during the existence of this part of the building. The right-angle to Wall C was formed by Wall D. The construction remaining was of chalk blocks which were most probably the foundation courses. The inside of Wall D also revealed the chalk floor. Outside Wall D, running parallel with it and only 1 ft. away, was a flat-bottomed ditch about 4 ft. wide; it had sloping sides and a total depth of about 2 ft. There was no evidence of silting in the 200 THE ROMANO-BRITISH BUILDINGS, SNODLAND ditch which suggests that it had been either re-cut immediately before its filling or refilled almost immediately after digging. The fill of the ditch was of gravel, with some cha.Th:, rising to and slightly above the surrounding Romano-British ground level. This initial filling was covered in places by dark soil. Over this, to a depth of 1 ft. 6 in. to 2 ft. was a mixture of gravel and cha.Th:, finally covered by the present topsoil. Another short length of ditch, of similar width and depth, was uncovered in a small area available for excavation between the house and the bank. The ditch ran roughly north-east, south-west on a course that would connect with a continuation of the other ditch. The ditch had some primary silting and wi;ts filled with general building rubble, including a large amount of chalk blocks and mortar. Returning to Walls C and D, it is not possible to delimit any room of which they may have formed part, because of the intrusion of a large air-raid shelter. However, opposite Wall D was another wall, J, which was 2 ft. 2 in. thick. It was of flint and ragstone coruitruction with yellow mortar. To the north-east of Wall J were traces of a mortar floor, laid on a bed of rubble; the level of this floor was the same as the chaTh:floor enclosed by Walls C and D. The difference in composition of these two floors indicates that there must have existed a partition wall, now destroyed by the shelter. Wall J returned as Wall K in a north-east direction to form an 8-ft. wide room of unknown length. South-west of Wall J, enclosed by Walls L, M and N, was another hypocaust. Wall L had been extensively robbed but enough remained to show a width of 3 ft.; its construction was of ragstone with yellow mortar. Wall M is mainly conjectural. Also an external wall, it formed the return of L. This point could not be investigated further due to an existing structure. N was an internal wall, 2 ft. thick, of bonding tiles and mortar. It formed the south-east side of the hypocaust area, with a flue 2 ft. wide. This wall may have served either as a support to the flimsily-constructed hypocaust or as a foundation, probably to a partition above. Whether there was a continuation of the hypocaust on the other side of Wall N was not proved but it is strongly suspected (see below). The under-floor of the hypocaust was of clay baked hard. Its thickness varied from a thin 'skin' to about 2 in., depending on the contours of the ground beneath. The pi"lae were roughly built of a mixture of complete pi"lae tiles and broken fragments of roofing and bonding tiles, bonded with clay. The pilae tiles were the same size as those of the other hypocaust. There were no base tiles, the pilae standing rather precariously directly on the uneven underfloor. The pilae were arranged at about 1 ft. 6 in. centres in a roughly symmetrical pattern. Maximum height to which a pila existed was 1 ft. 4 in. 201 THE ROMANO-BRITISH BUILDINGS, SNODLAND In a northwards extension to the hypocaust room was a tile feature, which partly filled the space between Wall Land the corner of J and K. This formed at least one flue, I ft. 6 in. wide, the east side of the flue being formed by the one side of Wall K. The traceable part of the feature consisted of a 9-in. wide course of tiles laid on a thick bed of clay. The angle formed by the feature contained a filling of decayed mortar. Further investigation was precluded by the air-raid shelter. On the south-east side of Wall N, which was built entirely of bonding tiles, most of the area available for excavation had been badly disturbed, but enough could be examined to show that the hypocaust under-floor continued, burnt a dark red on both sides of Wall N. A layer of ash, 1-2 in. thick, lay over the floor, but towards Wall L it disappeared perhaps because of robbing and other disturbances. In view of the concentration of soot at this point, the burning of the clay floor, and the bonding tile construction of Wall N, it seems likely that the firing pit was located somewhere near this end of the hypocaust. The function of the hypocaust was not clear, but it may well have served as a corn-drying floor. To the west of Wall L, outside the building, was a large area of the pebble and chalk paving, forming perhaps a courtyard. Its level was 9 in. above that of the hypocaust under-floor. The 'courtyard' showed two major layers of pebble paving, both of which had been extensively repaired where they had worn, with crushed chalk and mortar; both layers were littered with trodden-in tile fragments, potsherds, animal bones and shells. The paving appeared to end about 20 ft. from the building. Sealed under the lower surface, and embedded in the soft brown soil, was a small pot (Dating Evidence, No. 24). Other pottery is dated from the late-first century to the third century A.D. In the chalk and pebble external paving was a fragment of blue window glass, 5 mm. thick, with one bevelled edge and with traces of pink mortar adhering to one side. LATER DrsooVERIEs As a result of civil engineering work carried out after the end of the excavation, several more structural remains were exposed in pits and trenches and recorded as far as possible under difficult conditions by Mr. M. Jessup. The parts of the site where these later discoveries were made had not previously been available for archaeological excavation. During the digging of a narrow trench to the south-ea.at of the lower hypocaust, two walls forming the corner of a building were exposed. The south-east wa.11 was 1 ft. 6 in. thick and built of chalk blocks set in mortar. The other wall was of similar construction but 2 ft. thick. Traces of a chalk and mortar floor were observed between them. 202 THE ROMANO-BRITISH BUILDINGS, SNODLAND The soil on the outside of the walls was the typical soft brown of the original topsoil. .Although the walls are on the 'upper level', the level here was much lower than elsewhere as the ground sloped towards the river, and it is very likely that the structure was associated with the hypocaust and 'apse' at the 'lower level'. Approximately 400 ft. north of the excavated remains, close to the river, the construction of two large machine bases uncovered traces of other structures. The ground where the machine bases were dug is today at about the sa,me level as that of the lower site. Originally, however, as the excavations revealed, the ground had sloped away from the south end of the site, and the top 4 ft. of the layers exposed was made-up ground. The excavations had out 2 ft. 6 in. into the original ground surface removing most of the remains and leaving only a few traces exposed at the sides. Pit I revealed two chalk-block walls, each about 2 ft. thick, and running parallel 2 ft. 9 in. apart; they were bonded with pink mortar. Three feet six inches north of one wall and 4 ft. south of the other, ,vere traces of hard packed yellow mortar, possibly indicating the existence of mortar floors. If these were floors then the two walls would each be the external wall to a building. Covering the whole area was a layer of Roman building rubble nearly I ft. thick. Pit 2 was 20 ft. to the north of Pit I and was dug to a similar depth. On the north side of this pit was a wall of chalk blocks and flint, running parallel with the walls in Pit I. The section on the west face showed a yellow mortar floor, 3 in. thick, overlying the original RomanoBritish topsoil. A small pit a foot deep had been dug into it, and was lined with baked clay and filled with charcoal and burnt material. The mortar floor sealed the pit. No building debris was apparent. If the unexoavated area betwe􀄠n the two pits did not conceal other walls, the remains found suggest the existence of at least two buildings, both with mortar floors and one some 31 ft. wide. DATING EVIDENCE (Figs. 2 and 3) . Arch. Gant. Ant.J Areh.J. J.R.S. Gamulodunum Ganterbury, I-IT .ABBREVIATIONS AND REFERENCES Arohceowgia Oantiana Antiquaries Journal Arehreologieal Journal Journal of Roman Studies 0. F. C. Hawkes and M. R. Hull, Gamuwdunwrn, Oxford, 1947. Audrey Williams and Sheppard Frere, 'Canterbury Excavations', {I) Arch. Oant., Ix (1947), 68-100; (IT) Arek. Gant., lxi (1948), 1-45. 203 THE ROMANO-BRITISH BUILDINGS, SNODLAND , (􀀂 2 ( 􀀁 􀀄6 4{" 4 XL/ 5 8 ,􀀉 9􀀊􀀋\ 1-----01 ----,-1----1 -..,..,----( 1 II L. ..........\ I ' 12 I 7 ) 7 '),. ( 1-, 15- - 􀀌' ,17 ! ) ,. 16 ' FIG. 2. Coarse pottery (t). 204 THE ROMANO-BRITISH BUILDINGS, SNODLAND l ,.􀀒'.:-·.,·.:;.•.·.􀀓1 􀀂7 J (.. I , I ' K\\\\\l 25 􀀁 26 􀀂 I 7 27 (,. 26 "\ l 29 ( 􀀁 30 \ 31 Cobham Oolches-ter Dover Greenhithe Jewry Wall -) 7 ) 􀀱, FIG. 3. Coarse pottery(¼ except No. 24 [½]). P. J. Tester, 'The Roman Villa in Cobham Park, near Rochester', Arch. Gant., lxxvi (1961), 88-109. M. R. Hull, The Roman Pot-ters' KiV/1,8 of Colchester, Oxford, 1963. L. Murray Threipland and K. A. Steer, 'Excavations at Dover 1945-47', Arch. Gant., lxiv (1951), 130-49. A. P. Detsicas, '.An Iron Age and Romano-British Site at Stone Castle Quarry, Greenhithe', Arch. Gant., lxx.xi (1966), 136-90. K. M. Kenyon, Excavations at the Jewry Wall Site, Leicest,er, Oxford, 1948. 205 THE ROMANO-BRITISH BUILDINGS, SNODLAND Joyden's Wood P. J. Tester and J.E. L. Caiger, 'Excavations o n the Site of a Romano-British Settlement in Joyden's Wood near Bexley', Arch. Gant., lxvili (1954), Lulling stone Richhorough, I-IV PER.TOD 1 167-83. Lieut.-Colonel G. W. Meates, E. Greenfield and Edwyn Birchenough, 'The Lullingstone Roman Villa, Second Interim Report', Arch. Gant., lxv (1952), 26-78. J. P. Bushe-Fox, Excavations of the Roman Fort al, Richhorough, Kent, Reports I-IV, 0.xford, 1926-49. No stratified datable finds were discovered in association with the two lengths of wall and chalk floor which constituted the only remains of this period. PERIOD 2-Mm-THmD CENTUitY A.D. The overwhelming majority of potsherds are of a late-second/earlythird century A.D. date. However, it is impossible to ignore the presence of two sherds of flanged bowls (Noa. 4 and 6) generally considered to be mid-third century at the earliest. From under tessellated floor mainly over remains of Period 1 structures but partly on light brown loamy soil. Samian Ware Rim band of Form 31 or 37. Hadria.nio--.Antonine. Coarse Pottery 1. Jar or cooking pot in grey fabric and paste, burnished externally above the shoulder to just inside the rim. 2. Cordoned jar in grey fabric and paste (Richhorough, IV, 467, Antonine). 3. Rim in brown fabric with light brown paste; traces of burnishing. Diameter 6 in. (Canterbury, I, 97, Fig. 15, No. 5, late-second century). 4. Rim of flanged bowl in dark grey-brown fabric with partly worn burnishing. Diameter 7 in. (Colchester, 305A, c. A.D. 250-400). From rubble .filling surrounding tile cist of infant ·burial 5. Jar with everted rim in reddish-brown fabric and light brown paste, partly burnt, decorated with burnished lines and burnished above the decoration. Sherds conjoined to form about one-third of the complete vessel (Greenhithe, 71, c . .A..D. 120-50; Colchester, 278, into third century). 206 THE ROMANO-BRITISH BUILDINGS, SNODLAND 6. Rim of flanged bowl, dark grey-brown fabric, diameter 7 in. (see No. 4 above). 7. Base of rough-cast beaker (of. Richborough, III, 302, c . .A.D. 90-140). PERIOD 3-AFTER Mm-THIRD CENTURY .A.D. From und,er tessellated floor t,o corrid,or Samian Ware Sherds of Form 38, and two decorated sherds of Form 37 (see Appendix I). Coarse Pottery 8. Bead rim jar in orange-brown fabric with soapy paste containing shell grit. 9. Jar in grey fabric and sandy paste (Canterbury, I, 96, No. 5). 10. Jar in brown fabric with grey rather gritty paste similar in form to No. 5 above (cf. Lullingstone, Nos. 68, 69, late-Antonine). ll. Rough-cast beaker in cream paste with reddish-brown colour coating. Similar types occur up to c . .A.D. 300 at Colchester and elsewhere. 12. Imitation of samian Form Curle 15 in light red fabric with partly twisted rim (Greenhithe, 134). 13. Dish in black fabric with dark grey sandy paste ( of. Lullingstone, 51, 52, late-Antonine). 14. Jar in light grey fabric (Camulodunum, 234A, late-first century). 15. Dish in dark grey-brown fabric and grey paste burnished internally and externally, decorated with burnished lines and with a horizontal line heavily scored on the inner surface about ¾ in. below the rim (Greenliithe, 181). From Period 3 thickening t,o Wall F 16. Jar of reddish-brown fabric and grey paste, part of rim missing, probably late-second century/early-third century. MlSOELLANEOUS From clay backfilling to sides of 'bath' 17. Bead-rim jar in black sandy fabric and paste. 18. Dish of grey fabric with sandy grey paste (Greenhithe, 178, .A.D. 120-150; Cobham, Fig. 3, No. 14, A.D. 70-180). 19. Jar in dark grey fabric with light brown gritty paste, the inner surface pitted and uneven. Sherds which were most probably of the same jar showed burnished lattice decoration externally below the shoulder (Greenhithe, 146, .A.D. 120-50). 207 THE ROMANO-BRITISH BUILDINGS, SNODLAND From rubble filling to Upper Level hypocaust 20. Dish of dark grey fabric and sandy paste, groove under rim (Dover, 136, No. 19, second century). 21. Mortarium in light-red fabric and paste, with flint grit (of: Camulodunum, 191, late-first century). 22. Mortarium in buff fabric with polished surface. Generally fourth century but the form has occurred in rare cases as early as the late second century (of. Jewry Wall, Type J). From in or on the pebble paving at the West End of the Upper Level unless otherwise indicated. 23. Comb-ware jar in grey paste with light brown fa.bric burnt black in patches. Native tradition (Rwhborough, II, 97-99, pl. xxix, Fig. 2, No. 1, up to end of Flavian period). Under paving. 24. Small beaker in grey paste, burnished externally with groove around middle. Evidently coarse ware copy of a rough cast beaker form. Sealed under chalk and pebble paving. 25. Bowl with flat lip in hard buff sandy fabric; burnt black around rim (of. Greenhithe, 228; Ricliborough, m, 216, .A..D. 50-75). 26. Dish in dark brown fabric, burnished internally and externally. Diagonal hatched pattern (cf. Greenhithe, 59, A.D. 120-50). 27. Small jar with outbent rim, in hard grey fabric (Joyden's Wood, 176, No. 15, second century). 28. Bead-rim jar of red-grey fabric and hard sandy grey paste. Heavy bead rim (of. Greenhithe, 94, c. A.D. 90-120). 29. Jar in Patch Grove type ware. Reddish-brown leathery fabric and grey paste (Greenhithe, 146, A.D. 120-50). 30. Straight-sided dish with a deep groove, in brown fa.bric (Greenhithe, 189, .A..D. 120-50). 31. Platter of grey-black fabric and grey-pink paste. Coarse ware copy of terra nigra form (Canterbury, II, 28, No. 32, first century). 32. Rough-cast beaker with cornice rim. Light brown colour coating over cream paste. Fine sand particles (Greenhithe, 200, 201, .A..D. 120-50). 33. Jar in brown sandy paste and fabric, burnished around base, and black slip applied over top. Decorated with quadruple and triple oblique lines (of. Colchester, 163, Fig. 93, Nos. 14 and 17). 34. Not illustrated. Fragments of three colour coated folded beakers with rubrfoated scales on ridges (Colchester, 169, No. 15, A.D. 210-300). In topsoil over pebble and chalk paving. 35. Not illustrated. Poppy-head beaker in light grey paste and dark grey fine fabric. Cordon at base of neck. Late-first to late-second century. 208 THE ROMANO-BRITISH BUILDINGS, SNODLAND DISOUSSION It has not been possible to date the construction of Period I, but some sherds of late.first century forms were found, suggesting that there was occupation at this time somewhere close at hand. Although the present excavations disclosed no evidence of actual occupation after the third century A.D. the date of some of the previous finds and the presence of two unstrati:fiedfourth-century coins (see Appendix VI) suggests the possibility of later occupation on another part of the site-­ presumably now obscured or destroyed. However, it is also possible that these later finds were stray items, dropped when the buildings were already in ruins. Taking earlier discoveries into account as well as the parts of the building established by the present excavations, it becomes clear that the site was of some size and stretched for several hundred feet along the west bank of the River Medway. The position by the river would have given it easy communications with near-by Rochester and the remainder of the valley. No other sizeable Romano-British buildings have so far been discovered in the vicinity. There is a conspicuous absence of remains from this period on the fertile, well-drained lower slopes of the North Downs, behind Snodland. As only few parts of the building have remained relatively intact, very little can be said about the nature of the occupation, its purpose, its economic life or its wider implications. The latest phase does correspond with the date of a burial in the Holborough Barrow,8 which was only three-quarters of a mile to the north-west on a hillside overlooking the valley. It has been suggested4 that the barrow was erected for a local dignitary who was probably quite wealthy, but if any claim to his residence is made the imposing villa at Eccles5 must surely have the advantage over the Snodland building. APPENDIX I SAMIANWARE (Fig. 4) By A. P. DETSIOAS, M.A., F.S.A. (i) Pl ain Forms. Most of the material submitted was very fragmentary but otherwise in a fairly good state of preservation. An examination of the forms present shows that Form 33, represented by sixteen vessels,6 was the most common cup, followed by Form 27, with sherds 3 Arch. Oant., Ixviii (1954), 1-61. ' Ibid., 28. 6 Two miles south-east of the barrow on the eaat bank of the River Medway. 0 One of these vessels bears a graffito (X) on the underside of the foot-ring. 209 THE ROMANO-BRITISH BUILDINGS, SNODLAND from five cups; the early Form 24 is totally absent. Form 31, including both 31R and 3l(Sa), was the most common plate, with pieces from eleven different plates; three plates of Form 18 and five belonging to Form 18/31 complete the total of plates found. The presence of a few sherds of Form 18, some of which could be of South Gaulish manufacture, could suggest occupation rather earlier than the bulk of the samian, but the amount of this material is so small that, unless such settlement can be supported by the evidence of the associated coarse pottery, it would be safer to ignore these sherds. On the other hand, the presence of sherds from vessels of Forms 38 and 45 clearly points to continued occupation into the third century A.D. The plain samian found is almost entirely of Central Gaulish provenance, with very few sherds from South or East Gaulish factories. s A/fo Fm. 4. Samian ware(½)- (ii) Decorated Forms. Except for one small fragment from a Form 30 vessel mentioned below, all the decorated sherds belong to Form 37; Form 29 is not at all represented. There is one small fragment close to the base of the bowl which may be of East Gaulish manufacture, otherwise all the material is of Central Gaulish origin. (Not illustrated.) Form 30, in good condition but with glaze completely worn. A small piece from the lowest part of the vessel, with part of the small column (detail 12) used by BVTRI0.7 Date: c. A.D. 120-145. Form 37. In poor condition, probably of East Gaulish origin; from the lowest part of the bowl, the decoration is almost completely worn off. Date: c . .A..D. 160. I. Form 37. Central Gaulish. In good condition, but with very poor relief owing to blurring in manufacture; in the style of the problematic potter DONNAVOVS. The remnant of the decoration consists of an arcade (detail 23)8 terminating upon large beaded rings (detail 1)8 which are placed partly upon thls potter's vine (detail 27);8 7 J. A. Stanfield and Grace Simpson, Oentral Gauli8h Potters, Oxford, 1958, 55, Fig. 13 (hereafter abbrevia.t􀁒d to OGP). 8 OGP, fig. 11, p. 42. 210 THE ROMANO-BRITISH BUILDINGS, SNODLAND the space below the arcade contains one figure-type, Crane to right (0.2196),9 which has been recorded on a bowl from Caerleon stamped DONNAVCI and on sherds from London.10 Another vine ornament is impressed below the crane, with seven-beaded rosettes masking the junctions of the vines and the bead-row border which encloses the whole decorative scheme. Date: c. A.D. 100-120. 2. Form 37. Central Gaulish. Good condition and glaze. A small fragment from the upper part of a, small, thin-walled bowl. The ovolo is very badly blurred and makes positive identification virtually impossible; it is a fairly small, double-bordered ovolo with a thin tongue which is attached to the right of the ovolo and may end in a small rosette. A bead-row border encloses the ovolo and a similar vertical border, ending at the upper terminal on a plain ring, is used for the probable panel decoration. To left of this border, remnants of a, medallion, and to right, the upper part of a figure-type which is probably a, Caryatid (D.656=O. 1196), used by several Central Gaulish potters. Date: c. A.D. 100-150. 3. Form 37. Central Gaulish. Good condition and glaze. A small sherd from the upper part of the decoration, with a double-bordered ovolo whose tongue is composed of rounded beads and ha,s a tip slightly bent to left. Insufficient remains to attempt even a tentative attribution, but it is not unlikely that this bowl dates to early Antonine times. 4. Form 37. Central Gaulish. Good condition. A remnant of decoration suggests panels divided by wavy-line borders. The rounded contours of this border recall the early style of CASVRIVS, but positive attribution is impossible. To right of the border, Appolo (D.54A=O.91A), used by several potters; to left, a remnant of an animal figure-type and part of probably a column which cannot be restored with confidence. Date: c. A.D. 120-160. 5. Form 37. Central Gaulish. Good condition and glaze. The ovolo is very similar to CINNAMVS's No. 3, and is enclosed by a mediumsize bead-row border.11 Date: c. A.D. 145-190. APPENDIX II SAMIAN w ARE AND POTTERY LAMP IN THE BRITISH MUSEUM IN the British Museum publication Antiquities of Roman Britain a 􀃑ention is made of a samian vessel from the Roman Villa, Snodland {p. 30, No. 7, acquisition No. 1948, 4-2.3). 9 D. followed by a number=J. Dechelette, Les Vases cbamiques ornes de la G􀁱ule romaine, Paris, 1904; O. followed by a number=F. Oswald, Index of Figure-Types on Terra Sigilla 10 ta, Liverpool, 1936-37. OGP, pls. 45/519 and 47/549. 11 CGP, fig. 47, p. 267. 211 THE ROMANO-BRITISH BUILDINGS, SNODLAND It appears that the complete acquisition (Noa. 1948, 4-2.1-3) comprised two samian vessels and a pottery lamp. One of the samian vessels was a complete Form 18 shallow platter bearing the stampOF. OER. The other was a complete Form 33 cup, with the stamp, FELIX.F, with an initial retrograde F. Mr. A. P. Detsioas, M.A., F.S.A., has kindly informed us that. OF .CER is usually considered a mis-reacling of the fairly common stamp of the South Gaulish potter GERMANVS (A.D. 70-100). FELIX is a South Gaulish potter of that name. There are several instances of an initial retrograde F for his name. FELIX's period of activity was second-half of the first century A.D., probably the last quarter. The pottery lamp has been identified by the British Museum as most probably early second century A.D. in date. It is of a type introduced in the second half of the first century and could conceivably be dated to the very end of that century. It is almost certainly of local manufacture. It is not known whether the lamp and samian vessels were found together, or indeed when and exactly where any of them were discovered. All three items were presented to the British Museum by J. P. T. Burchell, M.C., F.S.A., who acquired them in 1944 from the collection of the late W. M. Cook. APPENDIX Ill THE SNODLAND BUCKLE By M:Rs. L. WEBSTER Department of British and Medieval Antiquities, The British Museum THE Snodland buckle fragment, found in 1927 and now in the British Museum (reg. No. 1928, 5-11, 1), originally formed half of a one-piece buckle plate of Hawkes class IVA ,12 showing the central recess for the buckle loop. It is bronze, decorated with two medallions containing incised human heads on a niello background, separated by chip-carved scrollery and flanked by chip-carved foliate patterns. Round the edge runs a border of bronze squares reserved in niello, each decorated. with a circle of pricked dots. Its greatest width is 2 • 85 in. The Snodland buckle plate belongs to a class of late Roman metal• work which has received particular attention during recent -years, 12 S. C. Hawkes and G. O. Dunning, 'Soldiers and Settlers in Britain, fourth to fifth century', Meil. Arch., v (1961), l ff. · 212 THE ROMANO-BRITISH BUILDINGS, SNODLAND both from Continental and English scholars.18 This genre comprises late-fourth-century bronze belt equipage including buckles, strap ends and other fitments, which are found all along the frontiers of the Roman Empire. These are now generally recognized as the mass-produced military equipment of Germanic laeti andfoederati serving in the Roman army in the later part of the fourth and fifth century. Mrs. Hawkes has analysed the English material in great detail, and it is to her class IV A that the Snodland piece belongs, by virtue of its rectangular plate structure and rich ornamentation. On others of this class, those from Richborough and Smithfield for example,14 we see the typical adaptation of such Roman techniques as chip-carving, and motifs such as vine scroll and rosettes to a Germanic taste for the abstract, becoming the characteristic florid all-over embellishment, that we see on native Germanic metalwork. But the Snodland buckle, uniquely among the English finds, shows, in its little portrait medallions and precise workmanship, a still classical temper. Its nearest parallels in this are a lavish buckle set formerly in Castellani collection at Rome and another from Hungary.16 The Snodland buckle was certainly, like others of this class, made abroad, most probably in the Rhineland or Northern Gaul. The fact that continental buckles of Class IVA have been found only in South-East England strongly supports the suggestion that such metalwork was first brought over by Germanic troops sent in 368/9 from Northern Gaul to man the Saxon shore forts, as part of Count Theodosius's army. This ,vould put the Snodland buckle's date somewhere in the second half of the fourth century, and probably before 383, when Magnus Maximus withdrew troops, and Britain once more became a precarious province. APPENDIX IV FRAGMENTARY QUERN STONE By W. S. PENN, B.Sc. Tmi quern is of Mayen lava. Only a one-sixth segment survives but this is particularly well preserved with little flaking. There appear to 13 E.g. J. Werner, 'Spii.tromisohe Giirtelgarnituren in Keilsohnitt Teohnik" Jahreahefte dea Osterreichischen Archaologischen Institute in Wien, xxvi ( 1930), 53.ff. J. Werner, 'Zur Entstehung der Reihengrl!.ber Civilisation', Archa!ologia Geographica, l (1950), 23.ff. S. J. De Laet, J. Dhondt, J. Nenquin, 'Les Laeti du Namurois et l'origine de la civilisation merovingienne'. Etudea d'histoire et d'arcMologie dediees a Ferdinand Oourtoy (1952), 149.ff. S.C. Hawkes and G. C. Dunning, op. cit., and with full bibliography. 11 S. C. Ra.wkes and G. C. Dunning, op. cit., Fig. 21, and pl. IIA. . 16 A. Riegel, Sp!itromische Kunstindustrie (1927), Fig. 86. J. Forssa.nder, 'Provinzial Romisches and Germanisohes', Meddelanden fr&n Lurui.8 Universitetete Hi-storiska Museum, vii (1937), Fig. 19. 213 THE ROMANO-BRITISH BUILDINGS, SNODLAND be traces of mortar whioh would suggest re-use of the quern for building purposes which may account for its preservation. It is an upper stone with a plain upper surface and a grooved, concave lower surface. There is a trace of one recess for a rynd on the lower surface which does not penetrate right through and although the shape is not clear it is probably dovetail. No hole for a handle socket survives on the existing fragment and there is no sign of a hopper in the upper surface of the stone. Dimensions are as follows: External diameter c. 29 inches Thickness at outer edge 2½ inches Thickness at inner edge 2 inches Angle of grinding surface c. 10° Angle of grooves. . c. 10° Dating of this stone cannot be precise. It is certainly not very early since it is not thick enough and it has too large a diameter. It cannot be very late as it is too thick and the angle of the grinding surface is too great. The stone thus probably dates from between .A.D. 150 and 250. In the absence of the lower stone it is not possible to be certain how the stone was driven but it could have been mechanical. APPENDIX V THE SM.ALL Fnms (Fig. 5) Reported on by E. W. Tn.LEY, Esq. Description 1. Fragment of shale bracelet. of. Arch. Oant., lxxi (1957), p. 87, Fig. 18, 3 (Springhead). 2. Fragment of bone pin with spherical head. Not illustrated. 3. Fragment of bone pin (2 ·7 in.). Not illustrated. 4. Small glass bead. 214 Position Disturbed building debris near 'Apse'. Foundation trench to 'Apse' wall. Under Period 2 tessellated floor. Under Period 3 tessellated floor, in corridor. a 􀀌 ? p.. ., 􀀍 􀀎 􀀏\\\\\\\\\\\') i ...i I If>,. I f I .......... ,, ... , ' , ... __ --------· .,,. ,,,. 0) CTNV'ICTONS 'SONICT'II.O.S: HSLLnrn·O.NV:w:0'8: :nLL THE ROMANO-BRITISH BUILDINGS, $NODLAND 5. Iron key-handle with ornamental bronze head. of. J. P. Bushe-Fox, Excavations

Previous
Previous

The Ritual Shaft at Warbank, Keston

Next
Next

The Skeleton of Halling Man