The Re-excavation of the Roman 'Villa' at Wingham

THE RE-EXCAVATION OF THE ROMAN 'VILLA' AT WINGHAM Part 1 FRANK JENKINS, M.A., Ph.D., F.S.A. In 1881 and 1882, the late George Dowker excavated the remains of a Roman building in the field called 'The Vineyard' at Wingham Court, Wingham, Kent. (N.G.R. TR 612457). In 1966, Viscount Hawarden kindly invited the present writer to re-excavate the site. This proved rewarding, because a surprising amount of information was obtained, which Dowker had failed to observe. It is now known that the building was not the villa, but a complete bath-house belonging to it. It was also found that the building had undergone three subsequent phases of construction. This report contains an account of the bath-house as initially built. The subsequent phases will appear in a report being prepared for future publication. Dowker's final report was published in the year 1883. Consequently, we felt that it was the appropriate time for the first part of our report to appear, a century later, in this special volume commemorating the publication of the one hundredth volume of Archaeologia Cantiana. All the bibliographical references in the text of this report relate to either Part I or Part II of Dowker's report, published respectively in Arch. Cant., xiv (1882), 134-9, and xv (1883), 351-7. We have included two excellent reproductions of Dowker's previously unpublished photographs, which were taken before he commenced the excavation of the remains of the heated part of the bath-house. The original ones, in a much faded condition, had been stored away and forgotten in the Eastgate House Museum at Rochester, where the author was delighted to find them, about two years after the completion of the re-excavation of the bath-house. PERIOD l (Fig. 1 cf. Dowker's Plan ACXIV fig.) It is necessary to briefly mention, in the following discourse, certain 87 F. JENKINS structural features which belonged to the later phases of the bathhouse. These will be more fully described in the next report which is in course of preparation. Some structural evidence, which Dowker removed but describes in his report, is also included. Alongside the exterior of the north wall of the below-ground structure in which the hypocaustal system was installed, there were the foundations shown on Fig. 1 as Al and Bl. They were both 2 ft. wide, shallow, and consisted of rammed chalk. Foundation Al was parallel with the wall of the below-ground structure, and 5 ft. 6 in. out from it. The foundation was interrupted by a gap 5 ft. 6 in. wide, which evidently indicated the position of the entrance to a north corridor from the open ground outside (no. 1 on Fig. 1). The foundation had evidently extended further westwards from the remaining part of it, but at the west end it was destroyed when a room with a basement was built on to the north wall of the below-ground structure containing the hypocaustal system. Foundation Bl was for a partition wall. To the west there was the corridor no. 1, and to the east a room no. 2. These, with probably another room at the west end, formed the north front of the bath-house. This could not be ascertained because the addition of a later basement room had destroyed the evidence for it. Hence, the question mark on Fig. 1. Due to destruction the length of Room 2 could also not be ascertained, but like the corridor it was 5 ft. 6 in. wide. Traces of the remains of the floors were probably represented by a layer of crumbled opus signinum in the corridor and at the remaining west end of Room 2. The soil underlying the opus signinum in the corridor yielded a few very small Roman potsherds, apparently of first-century date. The reason for the demolition of the corridor and the room became clear when the re-excavation of the bath-house was completed, and it was possible to assemble all of the available evidence. This indicated that the building was eventually extensively rebuilt. The corridor was not rebuilt and Room 2, at the east end of it was replaced by a larger room, i.e. Dowker's Room no. 3. There was a room adjoining the south end of Room 2, which was replaced by Dowker's Room no. 2. The evidence which follows led to this conclusion. First, the earlier room which was replaced by Dowker's Room no. 2. A portion of the level two courses remained at the north end of the west wall. These consisted of yellow tiles and were based on the top of the east wall of the below-ground structure which was built in Period 1 to containe the hypocaustal system. The walls of this structure were faced internally with yellow tiles which were of the same period of construction as the remaining part of the above mentioned west wall of the earlier room (Room 2, Fig. 1), which was 88 THE RE-EXCAVATION OF THE ROMAN 'VILLA' AT WINGHAM 1 I ^ W///////////^^^^^ I • Stoke 6 • D a o a i a D a a apy O D D D / / ?/////////////////^^ CD Chalk foundations left • _ » - • _ • UDo or7 J . .SKIS Fig. 1. Wingham: The Bath-house. based on the top of the east wall of the below-ground structure. Hence, this wall was also built in Period 1. Overlying this there was the remaining lowest portion of a wall built of flint stones and mortar, which was the west wall of Dowker's Room no. 2, the other walls of which were built up with the same materials from the foundations. As the lowest remaining portion of the west wall was built of yellow tiles in Period 1, it appeared that it was left standing two courses when the earlier room was demolished. The other walls were destroyed down to the foundations which were used for the corresponding walls of Dowker's Room no. 2 to be built upon them with flint stones and mortar, and the west wall of it was built up with the same materials on the remaining stump of the earlier tile-built wall. Turning to Dowker's Room no. 3 which replaced the earlier Room no. 2. The east wall was in line with that of his Room no. 2 and the remaining lowest part of it was also built of flint stones and mortar. The two walls were fully bonded to show that they were of the same period of construction, but there was a straight joint where the ends of the two foundations met. This indicated that the foundation of the wall of Dowker's Room no. 3 was later than that of the wall of his Room no. 2. In view of this, and the fact that the flint-built walls on them were fully in bond, and therefore of the same build, it is logical that the foundations of an earlier room were utilized for the three 89 F. JENKINS walls of the later room, i.e. Dowker's Room no. 2. Furthermore, it is equally logical that the latter and Dowker's Room no. 3 were built as part of the re-building of the bath-house. At the same time tessellated floors were laid in these rooms. Turning now to Dowker's Room no. 4. He had removed all of the debris from its hypocaust, and in so-doing some important information was destroyed, but fortunately he records it (Arch. Cant., xiv (1882), 138). He tells us that the floor of Room no. 4 had fallen in, and its debris was lying at the bottom of the flues of the hypocaust. He also also makes a significant remark. It appeared as if the tessellated floor of his Room no. 2 had combined with Room no. 4. We found that this was correct because its remaining end, still paved with the tesserae, slightly projected into the east end of the room. This confirmed that the debris Dowker removed from the hypocaust was from this tessellated floor, but it is now clear that it was not the Period 1 floor which was destroyed in the course of rebuilding the bath-house. Dowker had failed to observe that the hypocaust he describes, was installed at that time (Arch. Cant., xiv (1882), 138-9). Consequently, he did not know that the rebuilding work had entailed the destruction of the Period 1 floor and that the hypocaust below it was replaced by the new one which was of an entirely different pattern. The earlier hypocaust was of the type in which pilae were placed to support the floor of the room. These pilae were removed for the new hypocaust to be installed, but the imprints of a number of them were in the concrete underfloor to indicate where they had stood. From these it was clear that the pilae were built up with tiles 8 in. sq. Having reached this stage, it can be stated that Dowker's Rooms nos. 2, 3 and 4 with their tessellated floors were built when the rebuilding of the bath-house took place in Period 2; also, that his Rooms no. 3 and no. 2, not heated by hypocausts, replaced the earlier rooms which were also not heated, shown respectively as Rooms 1 and 2 on Fig. 1. Turning to Dowker's Room no. 1 which contained a rectangular cold bath. We examined the exterior of the west wall, and the inner face of the north wall of the substructure in which the bath was installed. The north wall consisted of yellow tiles and evidently the west wall which was coated with mortar on the exterior. The water in the bath was probably not much more than 1 ft. 6 in. deep. There was a drain at the bottom which passed through the west wall at the south-west corner. No traces were found either of an external main drain into which the water was discharged, or of the trench for it. The evidence was probably destroyed when Dowker's trench alongside the exterior of the west wall was dug. It should be noted his 90 THE RE-EXCAVATION OF THE ROMAN VILLA' AT WINGHAM PLATE I View from the North, showing the sunken cold bath (Dowker's Room no. 1) when first uncovered in 1881. photograph of the bath was taken before his men dug the trench (Plate I). The south wall, at its highest remaining part stood 2 ft. 6 in. high from the bottom of the bath, i.e. c. 1 ft. above the water line of the bath. As the latter equated with the floor level of Dowker's Room no. 2, the 1 ft.-high upper part of the wall evidently stood above the contemporary ground level. This could not be confirmed because, 91 PLATE II t o H m Z g General View of the Remains of the Rooms at the east End of the Bath-house in 1881. THE RE-EXCAVATION OF THE ROMAN'VILLA'AT WINGHAM although stratigraphical evidence outside the wall had not been removed by Dowker, it could not be examined, because some of the trees of mature growth forming the southern boundary of the field were standing there. Owing to these trees, the thickness of the wall which Dowker states was apparently 2 ft. could not be checked. The tessellated interior of the bath yielded some interesting information. It was found that Dowker was incorrect in stating that the tesserae were embedded in the opus signinum forming the c. 2 in. thick primary lining of the bath (Arch. Cant., xiv (1882), 136). In one place, a patch of the tesserae was slightly detached from the opus signinum to reveal that its face was remarkably smooth. This indicated that enough time had elapsed for it to set hard before it was tessellated. The length of time is not known, but the tesserae on the sides of the bath were identical in every respect with those of the tessellated floors in Dowker's Rooms no. 2 and no. 3. He states that some identical tesserae remained on the south wall of his Room no. 2. This must mean that the sides of the bath were also tessellated in, Period 2 when the bath-house was rebuilt. As the bath and the substructure enclosing it were apparently part of the first bath-house, it is reasonable to think that originally it was lined with the opus signinum which was not tessellated until the time came to rebuild the bath-house. As it was the usual Roman practice to provide a cold bath, we have shown it as part of the Period 1 building (8 on Fig. 1). It should be noted that as the area southwards from the bath has not been excavated, we do not know if the building extended further in that direction, or that the room containing the cold bath was actually a rudimentary south-east wing. The description of the bath-house which follows is divided into Section A, the unheated rooms, and Section B, the rooms of graded temperatures and the heating system, the Period 1 arrangement of which was easily identified. PERIOD l (Fig. 1) A. The unheated rooms. (1) The north corridor The corridor was entered from outside through a doorway in the north wall. This was probably the main entrance, the corridor was 5 ft. 6 in. wide, but of unknown length. It probably had a floor of opus signinum. (2) The apodyterium As this was the first room entered, and in view of its position in the 93 F. JENKINS layout of the bath-house, it is logical that it was the disrobing room or apodyterium. Most of it was destroyed. It was 5 ft. 6 in. wide north to south. It appeared equally logical to restore it on plan with a length of 16 ft. 6 in. east to west. It is possible that this room, and the corridor, were half-timbered structures and the sill beams resting on the shallow chalk foundations. (3) The frigidarium As it was the normal Roman practice for the room next to the apodyterium to be the frigidarium, it is reasonable to think that Room 3 served that purpose. There was evidence to suggest that the foundations of this room remained, with the walls of a later room on them. It is restored on plan with walls 2 ft. thick, and measuring internally 11 ft. 4 in. by 11 ft. 10 in. The walls were evidently built of yellow tiles laid in level courses, bonded with mortar. A small channel of a similar cross-section in the mortar covering the top of the foundation of the south wall in the south-west corner was evidently the cast of a drain-pipe. It is not known whether the waste water flowed out from it into a main drain because Dowker's deeptrenching had destroyed the evidence. B. The hot rooms and the hypocaustal system The hypocaustal system, the furnace, and the stoking-room were placed in a large rectangular below-ground structure. The walls, 2 ft. thick, were built of flint stones and mortar, faced internally with yellow tiles, 11 in. long, laid in level courses bonded with mortar. The walls were c. 2 ft. 4 in. high, and enclosed an area c. 48 ft. long east to west, and c. 11 ft. 4 in. wide. Its bottom was floored with a white concrete. It appeared to Dowker that the north and south walls were subsequently extended further westwards (Arch. Cant., xv (1883), 352). This is incorrect because we found that the walls were of the same build throughout. It is now certain that the hypocaustal system described by Dowker had been installed to replace the first one when the bath-house was extensively rebuilt (Arch. Cant., xiv (1882), 138-9; xv (1883), 352). (4) The tepidarium Room 4 was evidently the tepidarium, because it was entered from the adjoining frigidarium, and the hypocaust which heated it was furthest from the furnace. Its doorway was in the east wall. The damaged lowest portions of the sides of a doorway remained, but it does not necessarily follow that it was the original one. Because of this doubt the original doorway is shown tentatively on Fig. 1, 94 THE RE-EXCAVATION OF THE ROMAN 'VILLA' AT WINGHAM centrally placed in the wall, until the later re-building work is described and discussed more fully in our next report. The walls of the tepidarium were almost destroyed, but small portions of the lowest few courses of the north and east ones remained, where they were in bond, to form the north-east corner of the room. These consisted of yellow tiles bonded with mortar, like the corresponding walls of the hypocaust on which they were based. Nothing was found to suggest that the walls of the room were not built up simultaneously with the walls of the hypocaust. The west wall of the hypocaust was destroyed down to the underfloor. The position it had formerly occupied was indicated by the broken tiles where it had been cut away from the lateral walls, and a scar 2 ft. wide across the underfloor from north to south. This proved that the hypocaust had measured 11 ft. 4 in. sq. The destroyed west wall must have contained an arched opening to allow the heat from the furnace to pass into the hypocaust. The hypocaust was of the type which contained tile-built pilae to support the floor of the tepidarium. This was indicated by their imprints on the surface of the concrete underfloor. These showed where twenty-four of the original thirty-six 8 in. sq. pilae had stood. From the positions of these imprints there were six rows of pilae, with six of them to each row. With this type of hypocaust the walls of the room were lined with vertical flues formed of box-tiles, faced with wall plaster. Taking this into account the room internally was about 10 ft. 4 in. sq. Its floor was destroyed, but its upper surface was probably c. 2 ft. 10 in. above the underfloor. (5) The caldarium As Room no. 5 was supplied with an atmosphere of dry heat at a higher temperature than that of the tepidarium, it was the caldarium. The imprints of twenty-three 8 in. sq. tile-built pilae on the surface of the underfloor indicated that there had been thirty-six standing in the same foundation as those in the hypocaust of the tepidarium. At the west end there was the remaining lowest tile of each of four pilae, which were placed in a row and spaced c. 1 ft. 6 in. apart to allow the heat from the furnace to pass through into the hypocaust. These pilae divided off the hypocaust from the furnace. The hypocaust was therefore 11 ft. 4 in. sq. The walls of the caldarium were destroyed but allowing for the vertical flues which must have lined them, the room was probably c. 10 ft. 4 in. sq. (6) The hot tank The furnace measured 4 ft. long west to east, and 11 ft. 4 in. wide. 95 F. JENKINS The walls and the floor above it were destroyed. In the north-west corner there was the imprint of an 8-in. pila in the surface of the underfloor. It seems logical that a water-tank was set above the furnace to supply the hot water used in the bath-house. (7) The stoking-room This room was below ground and was 16 ft. long west to east and 11 ft. 4 in. wide. This suggested that there was sufficient space for the stokers to work and for the storage of the wood fuel undercover. Much of the east wall containing the furnace arch was destroyed. If traces of the latter remained they were underlying the sides of a later main flue which was connected to a different furnace. This later flue was not removed; thus, the width of the Period 1 furnace arch could not be ascertained, but it seems logical that it was c. 1 ft. 6 in. like the spaces between the large pilae at the rear of the furnace. The concrete floor terminated with a ragged edge c. 7 ft. to the west of the furnace arch. Dowker thought that this indicated that the area was not paved between the floor and the west wall of the room (Arch. Cant., xv (1883), 352). There are reasons, however, that this was not the case. Dowker failed to observe that a new east wall containing a furnace arch had been built during later alterations, with the result that it was 7 ft. to the west of the former position. Hence, he examined the bottom of the later and smaller stoking-room, which we found was covered with a hard layer of trodden ash. When this was removed two isolated patches of concrete were underlying it. These appeared to be vestiges of the destroyed floor. It seems from Dowker's account that he was unable to make a thorough examination (Arch. Cant., xv (1883), 352). He tells us that the room was filled with debris up to the height of the lateral walls, and on cleaning this out he soon reached the water level. Since that time this has been lowered because after we reached the floor of the room we left it exposed for a year, and throughout this time the room was never flooded. Dowker also mentions that at the level of the top of the lateral walls the debris which was dumped in the room was covered by a rough concrete (Arch. Cant., xv (1883), 352). Although he destroyed this, it is now certain that it was a floor which was laid at the same level as the floors of the other rooms when the west end of the bath-house underwent alterations. This work was carried out when it was decided to abandon the below-ground part of the furnace room and the associated hypocaustal system. As the isolated patches of concrete appeared to be vestiges of the floor of the stoking-room, it seems reasonable to think that the floor was destroyed in the initial stage of the later alterations, and then the room was filled up with the 96 THE RE-EXCAVATION OF THE ROMAN 'VILLA' AT WINGHAM debris. This work, which marked the penultimate structural phase, will be fully described in a later report. (8) The cold bath For reasons given in the foregoing discourse (p. 00), the cold bath, i.e. Dowker's Room no. 1, is shown on Fig. 1 as part of the Period 1 bath-house. The room is shown tentatively as a rudimentary southeast wing until future excavation of the area extending south of it proves otherwise. It is certain that the bath would have been lined internally with some kind of material, which we have suggested was the opus signinum which remained. The bath measured internally 8 ft. 4 in. east to west, and 6 ft. 5 in. north to south. It was entered from the frigidarium (no. 2) through an opening in the north wall of the bath-room to steps which led down to the bottom of the bath. A drain at the bottom in the south-west corner went through the west wall. It is virtually certain that the remaining traces of the external main drain were destroyed when Dowker's men dug to a deeper level alongside the outer face of the wall. (9) Apsidal hot bath (not shown on Fig. 1) Finally, it must be mentioned that possibly the Period 1 bath-house had a heated apsidal bath, which was entered through the south wall of the caldarium to which it was attached. In order to prove or disprove this, it is necessary to introduce full details of certain alterations which followed in Periods 2 to 5 inclusive. These will be fully discussed in our next report. Hence, to avoid undue repetition the correct order in which the apse was built will be determined in the forthcoming report, and for this reason it is not included in the Period 1 plan on Fig. 1. SUMMARY The re-excavation of the remains of the building was very worthwhile because it produced a surprising amount of new information. It is now known that the building was not the main one in the villa complex, but it was a complete bath-house. It is also now known that it stood on the south side of a large courtyard, at the west end of which the remains of a building await excavation. On the north side of the courtyard we have almost completely excavated the remains of a large aisled building (publication pending). The east end of the courtyard apparently was open, and a trackway probably connected it with a Roman road which according to the late Mr. I.D. Margary was on the line now taken by the modern road B2046, and went south and 97 F. JENKINS west from Wingham over Neavy Downs, and Cooting Downs, to its junction with the Canterbury to Dover Roman road on Barham Downs (Arch. Cant., lxi (1948), 130). The bath-house stood on rising ground, with its west end a few yards east of a small stream. The source of this is at Wingham Well about a quarter of a mile to the south. This stream, a feeder of the Lesser Stour, was probably wider and deeper in Roman times. The bath-house was normal for a domestic bath-house of its size. The north front of it comprised a corridor Room 1, with an adjoining Room 2 at the east end. It is now impossible to prove that there was a similar room at the west end of the corridor. Room 2 was evidently the apodyterium. Room 3 was the frigidarium from which the suite of rooms of graded temperatures extended westwards. Room 4 was the tepidarium, and Room 5 the caldarium. Room 6, at the west end, over the furnace, was probably a tank which supplied the hot water used in the bath-house. The hypocaustal system was of the type in which 8 in. sq. tile-built pilae were placed to support the floors of the hot rooms. Room 7, a basement area, was the stoking-room which may have been roofed over at a lower level than the hot rooms. This room was large enough to provide ample space for the stokers to work and for the storage of the wood fuel under cover. Room 8 contained a cold-bath. The date when the bath-house was built could not be firmly fixed. The few small pieces of Roman pottery which pre-dated the north corridor were apparently of first-century date. These may indicate that the bath-house was built either at the end of the first century, or in the first decade of the second. The length of time the bath-house was in use until it was rebuilt, and the reason why, are questions which cannot be answered with certainty. The large-scale rebuilding work suggested that the building may have been accidentally destroyed by fire, which was always a hazard with this type of structure. The villa-estate was situated 6 miles from Canterbury to the west, and about the same distance from Richborough to the east. Hence, the estate had easy access to the local markets. It is perhaps worthy of mention that before Dowker excavated the remains of the bath-house, a depth of from 2 to 3 ft. of soil had accumulated over them. This was due to soil creep which was the combined effect of rainwash, the activities of worms and burrowing creatures, and ploughing since the time that the bath-house became a convenient source of building materials for use elsewhere. 98 THE RE-EXCAVATION OF THE ROMAN 'VILLA' AT WINGHAM ACKNOWLEDGEMENTS Throughout the work we received much kindness and encouragement from the Viscount and Lady Hawarden who financed the project. With the permission of the then Ministry of Works a small sum was transferred from a grant to the Canterbury Archaeological Society for archaeological rescue work in Canterbury. This was facilitated with the kind co-operation of the Inspector, the late Mr. S.E. Rigold, M.A., F.S.A., who was always available for consultation on the site. Our thanks also to the following volunteer diggers. Dr. P.B. Stones, the Misses R. Maidment, J. Fowler, and S. Martin; also Messrs. R. Hawkins, P. Taylor, H. Howe, K. Elks, N.A. Spurring, C. Gray, W. Gray, S. St. Leger, R. Bull, M.G. Stewart, and C. Willis, and some of his fellow scholars of King's School, Canterbury. Our thanks also to the late Major W.F.J. Harvey, M.B.E., D.F.C., T.D., who was so helpful in many capacities other than digging. We record with gratitude that the late Mr. J.C. Taylor, curator of the Eastgate Museum at Rochester kindly permitted us to have Dowker's photographs copied. Finally, our congratulatons to Messrs. Sunbeam Photo Ltd. of Margate for the excellent reproductions of the original much-faded photographs. We are indebted to Mr. John Rady, of the Canterbury Archaeological Trust, for drawing the copy of our plan of the bath-house for publication. 99