Excavations at Eccles 1964

EXCAVATIONS AT ECCLES, 1964 THIRD INTERIM REPORT By A. P. DETSICAS, M.A., P.S.A. INTRODUCTION Excavations were continued,1 in association with the Lower Medway Archseological Research Group, every weekend from late March to early November 1964, at the site of the Romano-British villa at Rowe Place Farm, Eccles, in the parish of Aylesford (N.G.R. TQ/722605; O.S. 6-inch Sheet TQ/76 SW). For permission to continue with this work, I am again indebted to the landowners, Messrs. Associated Portland Cement Manufacturers Limited and the Reed Paper Group Limited, and to the full co-operation of the tenant farmers, Messrs. A. A. and A. C. Southwell, without whose generosity the tasks of the excavation would have been considerably harder. The excavation was financed by generous grants from the Carnegie United Kingdom Trust, the Kent Archaeological Society, the Society of Antiquaries of London, the Craven Fund and the Haverfield Trust of the University of Oxford, and by public contributions. My thanks are due to the members of the Lower Medway Archseological Research Group and the many other volunteers who made this excavation possible, and in particular to the following for their sustained support throughout the season: Misses L. S. Smith and M. Webster, and Messrs. I. J. Bissett, A. C. Harrison, B.A., T. Hetherington, T. Ithell, W. A. Knowles, C. E. J. Martin, and A. M. Rowland, B.A. I am also especially grateful to Messrs. I. J. Bissett and A. C. Harrison, B.A., for much help with field drawings, to Miss A. M. Arnott for preparing for drawing some of my sections, to Mr. R. G. Foord for undertaking the monochrome photography and supplying the prints illustrating this report, to Mrs. K. F. Hartley for reporting on the mortaria, to Dr. J. P. C. Kent, F.S.A., for identifying the coins, and to Mr. E. R. Swain for taking charge of the drawing of the small finds. 1 Arch. Cant., Ixxviii (1963), 125-41; Ixxix (1964), 121-35. I am greatly indebted to the late Professor Sir Ian Richmond, F.B.A., P.S.A., and to Professor S. S. Frere, V-P.S.A., for reading this report in draft form and kindly contributing several valuable suggestions. 69 EXCAVATIONS AT ECCLES, 1964 THE EXCAVATION Work this year was carried out again to north and west of the 1962-3 areas with the aim of (a) completing the plans of the second and third bath buildings, and (b) recovering the plan of the earliest bath building found partly under its successors. As a result, two further periods of occupation of the site have been estabhshed, antedating the earhest baths, and this has necessitated both re-numbering the periods of occupation and a partial revision of the tentative dating suggested in the earher reports. Period I, to c. A.D. 55: The Ditch A short length of ditch (Fig. 3, Sections C-D, Layer 4, and G-H, Layer 4), was exposed below the granary and this constitutes to date the earhest known occupation of the site. This ditch was found to be some 6 feet in width; neither its true width nor its depth are at present known as later building over the ditch has meant some loss of both depth and width. V-shaped in profile, the ditch was cut into the subsoil and was found filled with domestic refuse including many oyster-shells, bones, etc., and a large amount of pottery, both imported and of local manufacture. One or two post-holes were observed at irregular intervals where posts had been driven into the subsoil; if others had existed, the posts had not penetrated beyond the filling of the ditch. The purpose and the full dimensions of this ditch are not yet known, but the most likely interpretation, on the basis of known archaeological evidence, is that it may have enclosed a Belgic type of homestead further to the west and served it as an open boundary ditch, which was eventually filled with domestic rubbish and clay and may have had a fence erected over its hne. Certainly, the pottery contained both in the primary and the secondary filhngs shows marked pre-Roman characteristics, but included some Romano-British material as well, which suggest that the ditch remained open for a few years after the conquest. PeriodII, c. A.D. 55-65: The Granary Excavation below the opus signinum floors of Rooms 47 and 59, the corridor giving access to the laconicum and thence to the rest of the earhest baths, revealed first a short length of foundations and eventually the complete plan of a small barn or granary (11 by 19 feet 3 inches), constructed partly over the filled-in Period I ditch (Plate IIA). Owing to the subsequent demolition for the building of the first baths, very little of the upper courses of the walls of this granary remained, but the method of construction was quite evident. Construction trenches for the five short walls of the granary were cut into the subsoil and partly into the filling of the Period I ditch 70 EXCAVATIONS AT ECCLES, 1964 (Fig. 3, Section C-D, Layer 4); these construction trenches were next filled with loose ragstone chippings (Fig. 3, Sections C-D and G-H, Layer 67), which were probably derived from dressing ragstone, and then the upper courses of the walls were built, to a thickness of 1 foot 6 inches; of ragstone and bright yellow mortar, some of which had run into the loose foundations below the walls. The overall dimensions of the building and its five short walls dividing it into four compartments suggest that its function was that of a grain store. If this was so, the three short internal walls Would probably have supported a floor of wooden planking, which would allow an adequate flow of air below the grain stored in sacks.2 However, the granary was so thoroughly demolished and sealed over by a compact deposit of clayey material (Fig. 3, Section C-D, Layer 2) before the construction of the first baths that no traces were recovered of any sort of floor within its area; on the other hand, evidence was found that the entrance to the granary was about the middle of its south-east wall. Here an area, 5 feet wide, of white mortar some 3 inches thick was exposed, which extended at least as far as the north-west wall of the first bath building; this mortar floor showed the imprints of boards which had been laid down upon it when wet and probably formed a loading platform. To this period probably also belong Rooms 53 and 54 at the extreme south corner of the excavated area and a system of drains probably associated with them. Room 53 was 10 feet 6 inches wide and Room 54 was hardly 3 feet; their dividing wall, again of ragstone and bright yellow mortar, was 2 feet thick but their common north-west wall was only 1 foot 6 inches thick whilst the south-east wall of Room 54, retained in use for the later Room 39, was of intermediate thickness (2 feet 6 inches). Whatever the purpose of these rooms, the presence of burnt material (Fig. 3, Section E-F, Layer 51) suggests that the building to which they belonged, at present beyond the limits of the excavation, may have been destroyed by fire. Period III, c. A.D. 65-120: (a) The Bath Building Work this season was concentrated mainly on the excavation of the earhest bath building^ almost the complete plan of this building has now been recovered, except for the south-west part of Room 52 and the western portion of the suite. Except for a latrine and one or two ancillary rooms, which may await examination in the south-west part of its area, the total accommodation of the baths suite varied, during the five phases of this period, from no less than 16 to a maximum of 19 2 Granaries are a feature of a number of villa-sites, e.g. Ditchley and Lullingstone. 3 Arch. Cant., lxxix (1964), 122-6. 71 EXCAVATIONS AT ECCLES, 1964 rooms, rendered necessary mainly by the disuse of one of the furnaces (Room 46) and the construction of another (Room 65) for the hot plunge-baths, and also by the need to provide an adequate furnaceroom for the laconicum (Room 32). The position of the laconicum, projecting beyond the main hne of the bath building, and the absence, owing to methodical demolition in later periods, of any evidence for a main entrance into the baths, make it difficult to reconcile the plan of the earhest Eccles bath building entirely with either of Krencker's two main types.* I t is clear, however, that the bath suite provided two alternative kinds of bathing practice, dry heat and damp heat and, if the second alternative were to be adopted by the bather, then the plan of the first bath building could be considered as conforming in general with Krencker's Ringtyp.5 In this case, the bather would progress, at any rate in Phase A, from the apodyterium (Room 49) into the frigidarium (Room 30) before entering the tepidarium (Room 28) and the caldarium (Room 39); from these rooms, he would be expected to pass through Rooms 55-57, before the main hot plunge-bath (Room 58), and emerge either directly . through the frigidarium into the apodyterium or, first, by means of the tepidarium and, next, through the frigidarium. From Phase B, however, it is quite clear that, with the commissioning of Room 50, the baths would be entered through the long corridor (Room 59) and the apodyterium (Room 60); and, if the bather should adopt the alternative of dry heat, he would then visit the laconicum before proceeding into the frigidarium and its cold plunge-bath (Room 31) where the colder temperature and immersion into the cold water of the plunge would be expected, by closing the pores of his skin opened in the very hot temperature of the laconicum, to protect him from chilling. The walls of the bath building were built, with a few exceptions such as the wall of the laconicum and others mentioned below, to a standard thickness of 2 feet with ragstone set in a bright yellow mortar which, when dry, is almost the same colour as the mortar used for the construction of the third baths in Period V. PHASE A Room 47 was in this phase 6 feet wide, narrower than in later phases, and is interpreted as a corridor leading towards the main rooms of the baths; there is a slight possibility that at least part of this area was associated with the furnace-room (Room 48), and some soot * D. Krencker, Vergleichende Vntersuchungen romischer Thermen, in D. Krencker and E. Kruger, Die Trierer Kaiserthermen, Augsburg, 1929,117-8. 6 ibid., 178. 72 EXCAVATIONS AT ECCLES, 1964 found by the north-west wall of Room 48 and extending some way towards the north-west would support this, but demolition in the later reconstruction has not allowed this to be established beyond doubt. The corridor was floored with a layer of hard white mortar, varying in thickness from 2 to 4 inches; its full length is not yet known as its north-west wall lay beyond the excavated area. Room 48 (6 by 6 feet) was intended, in the original design of the bath-building, to function as the furnace-room serving the large laconicum (Room 32); it was floored with bridging-tiles, which were laid directly on the subsoil and projected into the flue to form the latter's floor as far as the under-floor of the hypocaust in the laconicum (Fig. 3, Section A-B, Layer 31). This floor soon perished, however, and the furnace-room was re-floored with opus signinum, which varied in thickness from 1 to 4 inches (Fig. 3, Section A-B, Layer 28) and was laid down over some dehris layers (Fig. 3, Section A-B, Layers 29 and 30) resulting from the wearing of the flue and also from a partial reconstruction of the flue towards the east in order both to shorten its rather extended length and to avoid the excessive loss of heat supplied by the furnace; no direct evidence for such a reconstruction was found, but that this must be the case is inferred from the complete absence of any cheeks at the higher level in the exposed section. No entrance into Room 48 was found (cf. Room 46, below), and all its enclosing walls were undoubtedly of the same construction and showed no signs of any entrance blocked in a later reconstruction; the logical position for such an entrance would be where shown in the plan (Fig. 2), through the south-west wall of the room, which would allow the furnace to be tended from the open courtyard to the south-west. The area to the south-west of Room 48 was completely exposed to the subsoil, but no evidence was found either of walls, or even construction trenches, to suggest that Room 48 may have projected into the courtyard. There is also the slight possibility, already mentioned in connection with Room 47, that an entrance into Room 48 may have been sited in its north-west wall rather than in its south-west wall. The surprising feature, however, of this room is its small size; for, even if allowance is made for an extension some 2 feet to the north-east at the time of the shortening of the flue into the laconicum and laying down of its opus signinum floor, an area of 8 feet by 6 feet is scarcely a sufficiently large space for the furnace of a laconicum the size of Room 32, a fact which appears even more surprising when one considers the design of the bath building as a whole. In the event, it was found necessary to build, in Phase B, a new and larger furnace-room, Room 50, beyond the north-west wall of Rooms 47 and 48. Room 49 measured 13 feet 6 inches by 8 feet and was floored with a layer of hard yellow mortar to a thickness of some 3 inches (Fig. 3, 73 EXCAVATIONS AT ECCLES, 1964 Section A-B, Layer 12), which was laid partly upon clay and partly upon the pre-existing plough-soil (Fig. 3, Section A-B, Layers 26 and 27); this floor may have been the under-floor of a hypocaust. Its depth, which is about the same as that of the under-floor in Room 32, would suggest this; on the other hand, no direct evidence was found for a flue through its north-west wall, and to supply heat to yet another room would further strain the likely output of Room 48. I t is more probable therefore, that this room was not heated, which would also suit better its presumed function as an apodyterium. No entrance into this room, through its south-west wall from the open courtyard, has been found, nor a doorway into the frigidarium. Room 30, the frigidarium of the suite, has now been confirmed to occupy a total area of very nearly 19 by 29 feet 6 inches and containing its cold-plunge bath, Room 31. Mention was made in the 1963 Report? that both the frigidarium itself and its plunge-bath were floored with mosaic pavements, which had been removed presumably in order that their tesserae might be re-used. By a stroke of very good fortune, it became possible this season to recover a large part of this material in Room 46, obviously surplus to requirements, and to reconstruct, thanks to the painstaking efforts of Mr. David S. Neal, sufficient of these rejected fragments so as to learn something of their patterns (see Appendix, and frontispiece) .7 The following additional points about these mosaics may be made here. The fact that the backing of the gladiatorial mosaic was as much as 3 inches thick is quite consistent with the evidence of excavation; for this mosaic pavement occupied a rectangular portion of the frigidarium, some 15 by 19 feet in size and excluding the area immediately south-west of Room 31,8 and was laid down on a bedding of compact yellow mortar and gravel whereas the thickness of 1 inch for the backing of the 'dolphin' mosaic is due to the fact that this pavement was laid down on the bottom of the plunge-bath, which consisted of 1 foot of opus signinum (see Fig. 3, Section K-L, Layer 25 in the 1963 Report). The difference, both in consistency and in thickness, of the bedding of these mosaics also accounts for the great variation in the size, of the recovered fragments; the gladiatorial fragments are generally much larger than the 'dolphin' ones as the former could be removed much more easily from their relatively less solid foundations. The point made below (see Appendix), that the 'dolphin' mosaic fragments seem to have been subjected to water action, is wholly consistent with the position of the mosaic on the bottom of a plunge-bath. 0 Arch. Cant., lxxix (1964), 124. ' I am indebted to Professor S. S. Frere, V-P.S.A. and to Mr. A. J. Taylor, F.S.A., through whose good offices this work has been undertaken; also, to Dr. D. -I. Smith, F.S.A., for expert advice. 8 Arch. Cant., lxxix (1964), 125. 74 EXCAVATIONS AT ECCLES, 1964 Room 28, which was excavated in 1963, must have obtained its supply of heat through Room 39, though demolition and further building in Period IV have prevented the discovery of a flue connecting these two rooms; it is now very probable that this room, too, was a hot plunge-bath which, less hot than Room 58, would have become the main hot plunge-bath from Phase B onward (see also Room 39, below). The evidence for this is a large drain which, beginning from the west corner of this room, passed underneath the floors of Rooms 58, 55, and 56, and thence to the north-west of the site below the courtyard wall and the tessellated floor of Room 51. This is the main drain of the bath building, fed by the drains serving Rooms 31 and 58. It was very solidly constructed, with walls of ragstone set in bright yellow mortar to a thickness of 1 foot 6 inches; the bottom consisted of roofing-tiles set with their flanges uppermost in bright yellow mortar directly upon the subsoil and overlapped by the sides of the drain. To the south-east of Room 51 and under the courtyard wall, the drain was arched and this would probably have been the case as far as the north-west wall of the main range of rooms; the depth of the drain, from the top of the arch to the top of its tiled floor, was a little over 3 feet, and the arch was constructed of bondingtiles set in opus signinum (Plate IIB). There were no traces of any rendering, either in opus signinum or mortar, on the sides of the drain, and it may be concluded that it contained lead piping.9 Room 39 was completely cleared where not covered by the massive north-west wall of the baths in Period V; the full size of the room was 32 by 12 feet 6 inches, excluding the apse, Room 38, which was fully excavated in 1963. The south-west wall of Room 39, which abutted against the wall of Rooms 53 and 54, was only 1 foot 6 inches thick whereas its north-west wall was the standard thickness of 2 feet; to the south-west, this room terminates against an earher wall, 2 feet 6 inches wide, which may have continued further to the north-west beyond its present termination at the south-west wall of Room 57 where it was found removed. The floor of Room 39 was of opus signinum, 2-3 inches thick, upon which were found slight traces of its hypocaustpilae. Heat for the hypocaust was provided through a flue at the southeast end of the south-west wall of the room from its own furnace-room (Room 52). The function of this room, certainly in this phase and probably throughout the whole period, was that of a caldarium. Room 52, the praefumium of Room 39, was only partly exposed. I t was 11 feet 6 inches wide and had walls of standard construction, except for the south-east wall which was only 1 foot 6 inches thick. 0 A few courses of bonding-tiles found in 1963 in a narrow trench outside the north-west wall of Room 28 and thought then to be part of a drain are now known to belong to the foundation courses of the walls forming the east corner of Room 68. 75 EXCAVATIONS AT ECCLES, 1964 The furnace-room was floored with a double layer of bonding-tiles which were found covered with a thick deposit of soot and ashes; the entrance into this room is not definitely known, though a gap in its south-east wall may represent it. Room 57 measured 15 by 10 feet 6 inches and was built with walls of the standard construction and thickness. It s function is not certain as its floor had been completely removed in later periods; it could have been heated from Room 52 through a flue, probably at the north corner of Room 52, but no evidence was found for this. Similarly uncertain is the function of Rooms 55 and 56 (5 feet 6 inches by 10 feet 6 inches, and 7 feet 6 inches by 10 feet 6 inches, respectively). But Room 56 at least appears to have been heated from Room 58, since there was evidence for a flue in its north-west wall. However, both these rooms, and especially Room 55, were thoroughly demohshed and levelled in later periods, in particular to allow the construction of Room 43 in Period V; there is a possibility that both rooms had floors of opus signinum laid down upon a solid foundation of mortared ragstone as elsewhere in the baths. Room 58, the hot plunge-bath of the suite, measured 14 feet 6 inches by 23 feet 6 inches and had walls of standard construction, except for the wall dividing it from its furnace-room, Room 46, which was 2 feet 6 inches thick and built entirely of coursed bonding-tiles set in bright yellow mortar on a foundation of ragstone. Room 58 was floored with opus signinum to a thickness of 6 inches, laid down on a solid foundation of mortared ragstone resting on the subsoil; this foundation layer was some 13 inches thick and identical with that found in 1963 below the floor of Room 28.10 Little doubt remains that this room was hypocausted in this phase as it certainly was in all subsequent ones; but, owing to the several major reconstructions in this area during the life of the baths, no direct evidence was found for a hypocaust on the opus signinum floor in the form either of pilae, their mortar pads or even masons' markings. But that one existed in this phase is clear from the function of the room as a hot plunge-bath and by surviving details bearing on the floor level. I ts identity as a hot plunge-bath is proved by a very unusual feature in the form of evidence for a testudo in the arch connecting Room 58 with its furnace in Room 46. A testudo was a semi-cylindrical metal tank with its flat side set over the furnace and its end open to the bath below the surface of the water; by this means a circulation was set up which maintained the heat of the bath water.11 In Britain structural evidence for testudines is rarely found because walls do not often survive to sufficient height to preserve it. In the 10 Arch. Cant., lxxix (1964), 125. 11 J. B. Ward Perkins and J. M. C. Toynbee, The Hunting Baths at Lepais Magna, in Archaeologia, xoiii (1949), 176-7. 76 '•TTTnTTTTTTTTTTTTTHTTTT iiiiiiniiiiiiiiiiiiui / . , ' ' ' ' * „ - " * Periods iv-v ECCLES 19 6 4 ^tPtnod ft Pwriod II Ptnods HlQ-t Ptriod lila-e l^p*riod Ulb-tj ^Ptriod lllc-d Pirtod Hid-9 '%]P*rio ><. .'.ti Photo: R. G. Fooril A. Room 46, showing Flue and blocked Testudo Arch. "1 />Ao(o: R. O. Foortl B. Testudo Arch, Chimneys, Water Supply and Outflow Drain before Reconstruction in Phases B and C. PLATE IV Photo: R. 0. Fooril A. Room 63, Hypocaust and Partition Walls. • - J ^ J M U J ? - - • • "••• Photo: R. G. Foord B. Room 43, the Pmefurnium of Period VB-C. EXCAVATIONS AT ECCLES, 1964 height cannot be accurately deduced but, from the measurements of its emplacement, its diameter could not have been much more than 6 feet, a size commensurate with the volume of hot water needed in a plunge-bath the size of Room 58. Room 46 did not continue in use after the reconstruction of part of the bath building which marked the end of this phase but, though it was neither destroyed, demolished nor entirely abandoned, it is impossible to suggest what use it could have been put to; it remained intact, however, until the very end of Period III as the mass of mosaic fragments from the destroyed frigidarium pavements and much building debris, containing two coins of Nerva, lay directly upon the soot and ash inside the furnace-room. At the close of this phase, the boiler and the testudo tank were removed, and the arch of the testudo was blocked with bonding-tiles in the usual bright yellow mortar (Plate IIIA); this helped to preserve the south-east wall of the furnace-room in such good condition that it was retained and incorporated into the bath building in Period IV; it survived to a total height of some 7 feet. At some time within this period, the dating of which is still to be established, the bath building underwent one major addition (Room 50) and several reconstructions of its existing rooms in the course of the subsequent Phases B-E of Period III; and, though the reasons for the addition of Room 50 are fairly obvious, it is not possible to say what caused the major reconstructions affecting the middle range of the bath building to the south-east of the disused Room 46. No evidence was found for any destruction by fire or decay in the structure of the walls; quite to the contrary, the overall impression gained is one of methodical reconstruction rather than piecemeal repair. It could be that the size of Room 58 proved too large and uneconomic, and it was decided to convert Room 28 into a hot plunge-bath replacing Room 58; all this, however plausible, can only be conjectured since excavation has provided the facts only but no clues to their causes. PHASES B AND C During these two phases, steps were taken to provide an adequate supply of heat for the laconicum, and Room 48 was then abandoned. The flue into the laconicum was blocked with mortared ragstone. The north-west wall of Room 48 and the north-east wall of Room 47 were demolished and levelled down before the laying'of an opus signinum floor, some 4-5 inches thick (Pig. 3, Section A-B, Layer 12), at the same level as the floor of Room 60; this floor was extended to the south-east as far as the south-east wall of the old Room 48 and, to the north-east, up to a newly-built north-east wall, of standard thickness and construction, continuing the alignment of the north-east wall of the baths. 79 EXCAVATIONS AT ECCLES, 1964 Room 59 was a new corridor superseding Room 47 at a greater width (8 feet) and extending over the demolished Room 48; though its full length is not yet excavated, it probably functioned as a corridor leading into the bath building through Room 60. Room 60 is to all intents and purposes the old Room 49 and, though it has undergone no structural change, its function as an apodyterium, before entering either the laconicum or the frigidarium, cannot now be doubted: there is still no evidence for an entrance into this room but, logically, it should have been in its north-west wall as there seems to be no reason why the bather should approach it through the open courtyard rather than through Room 59. Similarly, no traces remained of the doorway leading from Room 60 into the frigidarium. Room 50, the new pmefurnium for the laconicum, was built with ragstone walls only 1 foot 6 inches thick; these were trench-built against the subsoil externally and dressed only internally, and the south-west wall was bonded with the north-east wall of Room 59. The furnaceroom measured 8 by 10-12 feet.15 Again as in the case of the other rooms, no direct evidence was found for an entrance, but the north-east wall was much destroyed, and it is likely that the entrance was through this; it certainly was not through the north-west wall, which survived without any signs of robbing or demolition to any depth. The main reconstruction carried out during these phases affected the area previously occupied by Rooms 58 and 56. The rest of the accommodation remained unchanged, though in some cases (e.g. Room 28) the function of some rooms may have altered. Room 28 remained unchanged, but it may have become the main hot plunge-bath since, quite clearly, not one of Rooms 61-64 could have served that purpose; they all lack the drainage that would have been necessary. Likewise, Rooms 38 and 39 with their own furnace-room, Room 52, remained structurally unaltered, the only addition in this area being the length of wall closing the mouth of the apsidal Room 38.16 This may mean that Room 39 was converted into the main hot plunge-bath in replacement of Room 58, and this would overcome the difficulty of having Room 28 supplied with hot water from a furnaceroom, Rooms 65 and 66, at some distance from the plunge-bath. Room 39, served by its own adjacent furnace-room, Room 52, provided some evidence for a drain (shown on the main site plan, Fig. 1) which, probably built in Period H, may have continued into use as the outflow of Room 39.17 In the order of construction, first, the north-west wall of the old 16 Arch. Cant., lxxix (1964), 123. 10 ibid., 126. 1 71 owe much in this discussion to the advice kindly given to me by the late Professor Sir Ian Richmond, F.B.A., P.S.A. 80 EXCAVATIONS AT ECCLES, 1964 Room 58 was reduced in thickness by cutting it back 6 inches and thus to the standard width of the baths, and the new south-east face of this wall was rendered with two superimposed coatings of opus signinum— one of these was applied immediately in order to make good the rough surface (Plate IIIB), and the second at the same time as the building of the floor of Rooms 61 and 63 with which it forms one unit. Next, the level of the floor was raised some 10 inches, by a packing of much rubble and mortar dehris, and a new opus signinum floor was constructed; this floor, 3 inches thick close to the new north-west wall and 5 inches thick further away to the south-east, was carried across the whole area and below the north-west wall of Rooms 62 and 64. The partition walls between Rooms 61 and 63 and 63 and 65 were built before the final application of opus signinum making up the required thickness of the floor, as their tiles were bedded through the latter (Plate IVA); this final coating of opus signinum was the second layer applied to render the new north-west wall of Rooms 61 and 63. Lastly, the wall dividing the whole area, standard in thickness and construction, was built on top of the opus signinum floor. Room 61 (7 by 9 feet) is a small room heated by a hypocaust whose pilae had been removed in later reconstruction, though enough traces remained to show that the pilae were built of tiles of the same dimension (1 foot square); there was, however, no means of deducing the height of suspension of the floor. Room 63 (7 by 7 feet 6 inches), with the same type of hypocaust as Room 61, was divided from the latter by a slight partition wall, 1 foot thick, constructed of bonding-tiles set in bright yellow mortar and rendered with opus signinum (Plate IVA); this wall had three flues (1 foot wide) through it for the supply of heat for Room 61. Room 62 (6 by 10 feet) was entered through a door, some 3 feet in width, in its north-west wall but, owing to the poor state of preservation in this area, there were no signs of an entrance into Room 64 (6 by 11 feet 6 inches) through their partition wall. Both these rooms were floored with a compacted layer of hard bright yellow mortar and do not appear to have been heated. If they were hypocausted, then one would expect the same floor as in Rooms 61 and 63 to have been carried up to the north-west wall of Room 39; similarly, there should have been flues both in the north-west and north-east walls of Room 64. More conclusively, the doorway in the north-west wall of Room 62 and the floor of both rooms were built at a higher level than the under-floor in the hypocausted Rooms 61 and 63. Room 65 (7 by 8 feet 6 inches) was floored with the same opus signinum floor as Rooms 61 and 63 to a thickness of some 4 inches and was divided from Room 63 by a slight (1 foot) wall of bonding-tiles faced with opus signinum, which had a single flue (2 feet wide) through 81 6 EXCAVATIONS AT ECCLES, 1964 it (Plate IVA). Cheeks constructed of heavily-burnt tiles continued this flue right across Room 65 to the same 2-foot width up to the wall dividing it from Room 66 (Pig. 2). I t is very probable that a boiler was placed over this long flue for the supply of hot water, through piping, possibly into Room 28; but for the absence of a drainage system, there would be a strong possibihty that Room 65 supplied the hot water for a new plunge-bath, Room 63. No signs of such drainage were found, and it is beyond doubt that the drain serving the old Room 58 was not retained in use. Constructed on the opus signinum floor of the flue and for its whole length, was a curious channel of clay-bonded imbrices, which had been burnt red and enclosed much soot; this was no doubt intended to drain away condensation. Room 66 (5 by 7 feet 6 inches), once again rather on the small size, was the furnace-room for the new hypocausts; its floor, which would have probably been of baked clay, had completely perished, but the whole room was found quite filled with soot and ash, part of which at least would have derived from its later incorporation into Room 68. PHASES D AND E If fragmentation of the middle range of the bath building into small rooms had been the hall-mark of Phases B and C, simplification and a partial reversion to the plan of PhaseAin this area seem to have been the aims of Phases D and E for reasons that can only be hypothetical; and one of these is once more probably the miscalculation (surprising as it is for the architects of such an elaborate bath building) of the size of the furnace-room heating the new hypocausts. Some reconstruction, too, took place in the corridor, Room 59, obviously the inevitable result of the hard wear to which its floor had been subjected. The length of Room 59 was increased by the reconstruction of the wall separating it from Room 60 some 2 feet further south-east to align with the doorway into Room 32, during Phase E. Before this happened, however, the floor level was raised on average some 6 inches by the deposition of much rubble, and a new opus signinum floor, 4-5 inches thick, was laid down over the combined area of Rooms 69 and 60 (Pig. 3, Section A-B, Layers 11 and 10); a tessellation may also have been constructed on this floor as several loose tesserae, cut from tiles were found on it; but, owing to disturbance by ploughing, this cannot be certain. The walls of the corridor were plastered with bright' yellow mortar which had been painted with various polychrome designs, seemingly mostly leaves, flowers and plants. It was not possible to establish where the partition wall between Rooms 59 and 60 would have been located in Phase D, as no traces of it remained anywhere; it would probably have been a sleeper-wall 82 EXCAVATIONS AT ECCLES, 1964 S.W.MJ 1 1 8 1 83 EXCAVATIONS AT ECCLES, 1964 built upon the floor, which was unbroken along its whole length, and some 2 feet north-west of its successor in Phase E. Por this latter wall, only 1 foot 6 inches wide and of standard construction, was not contemporary with the construction of the floor on which it rested; this is proved by the fact that it abutted against the painted wallplaster surface of the south-west wall of Rooms 59 and 60 which it preserved in situ; moreover, many of the large number of fragments of painted wall-plaster found in this corridor show unmistakable signs of re-painting. The only other reconstructed rooms in these phases are Rooms 67 and 68, the former largely replacing Rooms 61 and 63 and the latter becoming the new and rather larger furnace-room for the new hypocaust. Room 67 measured 23 by 7 feet and was heated by a hypocaust; the under-floor of this hypocaust was of opus signinum, 5 inches thick, raised 6 inches over the level of the earlier under-floor by the deposition of a make-up layer consisting of mortar rubble and debris undoubtedly originating from the walls demolished in the reconstruction of this area. The upper floor of the hypocaust was suspended 19 inches on pilae, which had been almost totally removed, except for a few base tiles (1 by 1 by I in.) remaining along the north-west wall of the room; the height of suspension is definitely known in this instance as a remnant of the suspended floor, 3 inches thick and built of opus signinum and bridging-tiles, was found in situ at the junction of floor and northwest wall. Room 68, the furnace-room serving Room 67, measured 7 feet 6 inches by 10 feet 6 inches and reverted to the size of the earher Room 56; it contained a floor of bonding-tiles laid over a layer of opus signinum and tile debris, and served Room 67 through a flue (2 feet wide) in its north-east wall. The amount of soot and ashes filling the lower levels of this room makes its function as a furnace-room quite certain. A further range of rooms, of unknown number as yet, was added in Phase E, with walls of standard thickness and construction, to the south-west of Room 57, but too httle has so far been exposed to establish their plan and probable purpose. Associated with these rooms was a length of drain, with ragstone walls and tiled bottom, likely to have been intended for the disposal of rainwater from the roof of these rooms, which it carried to its junction with the main drain to the south-east of Room 51. The whole area north-west of the main range of rooms and Room 46 was enclosed throughout Period III by a long wall of standard construction, projecting south-west from its junction with Room 59 and used later as the foundation for the courtyard wall of the baths in 84 E C C L E S 1 9 6 4 S E C T I O N rn firrTTTTTTT nnmTmTTTnmnTTWfTi i n ( i l i l i i i i i l i i i i i in^l i i i i i i i i i i i i i inini i i i ^m\\\i s ^ ^ ^ ^ ^ g p ^ ^ ^ ^ ^ » ^ » 'mm. rrTmm 777777777711,•m/lll/IBlZlB/SSZi ?f>l,l/in,„ - -M£*ZZ*r£i Period Ilia Period I lib -d III a-d l i i i i i i ^ n i i inf V'l l < y mmM A-E Period II. Granary Walls . WMWWmwmwmv i i i i i i i i i M i i u i i i i i I I I I I I i M i II I I i~rn-rrna?nT i i i I . i I i TTLLI i ' ' i ITI i M rrrrTTTrTTTTTTTTTTTTTPTr1 ' ' i ' i i I i I T I I i i i i i i i TTTTTT ^ i . ' ' 1 ' - ' ; , 1 ; ' - ' : 1 , ' !'.!' h u" ' ' & i I " -Veriod Hid. . . • • ' : i i ' I I I i I I I I I I I I reel 'smZ22W&M7WJoW>%wk Period Ilia kxux^Z^^^^^^f--- - ,.j Metres S £ C T / O N Period tlla-c m irrnTTTT -yj^v^A-VLxx- i i i i r-t i ' OlMJJMJ-Wl^^4Mt ^A -£UJ4 l f f l f f i - n - 1 is* i i i i i i-r- rT-m-T- Cheek mmw\m SPXTTTrlTrTT^^ rrrnrmTn i , i . 1 1 . 1962 Trench Trr#rtrfff Rendering 'Hi-llllnllii'!^ l l l l l N l l t l l i i i i J I I M I i I I I I I T I 1111 • 111 r 11 111& 1111i J i l l Period Vac j ! 11 I UB.-ACH mens |i|i||JI|IJJJIllllMll.lM I HUl |llTtlllii|l|iUI I" Dark Brown Soil §M Graf el Period ^Mortar t C, Rubbish filling Tile Debris Period It J]\ Subsoil Hard Core ^Clayey Filling \Sott Brown Soil || | || Yellow Mortar Rubble ||§| Soot F I G . 3. Brown Soil t Charcoal ^Mortar S Gravel W^ftobber Trench ^^Rogstone Chipping* Burnt Mortar/Clay \fc^Cholk Lumps ^^Wall-Plaster Debris \LZ^Prlmory Filling EXCAVATIONS AT ECCLES, 1964 Period IV. Thus, the north-west part of the earhest bath suite seems to have comprised an open courtyard (palaestra) in the well-known manner of most bath-houses; this may have been paved with tiles as some were found laid down on the subsoil in this area. Room 51, of unknown dimensions and purpose at this stage of the excavation, lies outside the palaestra of the earhest baths; its walls, built from the same materials as those used in these baths, were 1 foot 6 inches wide, with a doorway (2 feet 6 inches wide) at the south corner giving access to a room floored with a tessellated pavement of coarse tesserae, cut from red and buff tiles and set in opus signinum over a bedding of the same material and varying from 3-6 inches in thickness. This room could belong to the baths in Period III, but this cannot yet be shown beyond doubt. Pinally, little beyond conjectural reasons can be suggested for the demohtion of such a large and elaborate bath building and the construction of the second baths in the following period. Certainly, the first bath building was solid enough to have lasted much longer than the 55 years of its life, but there is always the possibility in a bath-house that fire may have compelled the various reconstructions and caused its final destruction; this is lent additional support by the fact that many of the mosaic fragments recovered in Room 46 showed signs of heat quite consistent with the fall of burning roof-timbers. (b) The Living Quarters Very httle has come to hght in the area north-west of Room 50 where the living accommodation, etc., of the first villa may lie. Two floors (Pig. 3, Section C-D, Layers 15 and 21), of opus signinum and hard white mortar (the former associated with a section of waU built of coursed bonding-tiles upon a ragstone foundation) and a drain channel cut through this floor which removed part of this wall, are the only signs of occupation in this area so far. Much more work, however, is needed in this area in order to establish the purpose of these floors and their relationship to the bath building. Period IV, c. A.D. 120-180: The Bath Building Excavation this season in the north-west part of the second bath building has resulted in the recovery of the north-west wall terminating Rooms 40 and 41. The complete length of Room 40 is now known to be no less than 46 feet 6 inches, that of Room 41 rather shorter at 34 feet; they both share the same north-west wall, which was built upon a length of wall belonging to the earlier bath building. One would expect Room 40 to be sub-divided into two smaller units by a partition wall, but none was found in spite of thorough examination for robber trenches; on the other 85 EXCAVATIONS AT ECCLES, 1964 hand, this part of the site had been completely disturbed several times in the course of the construction of the three bath suites. The northwest wall of these two rooms had been built of bonding-tiles, ragstone and tufa, all set in off-white mortar; it had fallen towards the southeast and must have been of considerable height as the material from it was scattered some 17 feet 6 inches beyond the line of the wall. A very small chamber, Room 45, measuring ,2 feet 6 inches by 4 feet,' was exposed at the north corner of Room 40; it belongs to one of the last phases of this period. The walls of this 'room' were built entirely of coursed bonding-tiles; the north-east wall was abutted against the north-west wall of Room 40 which was partly removed, so that both the floor and south-west wall of Room 45 were laid across the northwest wall of Room 40. The floor of Room 45 consisted of opus signinum, 3 inches thick, and was raised nearly 3 feet over the north-east cheek of the flue in the demolished Room 46 by the deposition of a large mass of loose ragstone, of which only the upper 8 inches were mortared together. The purpose of this room is not known, but the level of its floor (2 feet below the surface) suggests that the floor in Room 40 could scarcely have been much lower. Like its predecessor, the second bath building had an open courtyard to its north-west. Its courtyard wall, shown on the main site plan, (Pig. 1), was built almost exactly over the alignment of the earlier courtyard wall, which it used as a foundation; to the north-east, it was abutted against the south-west wall of the earlier Room 59 and, to the south-west, it extended beyond the limits of the excavation to enclose an area about two-thirds the size of the Period I I I palaestra. Period V, c. A.D. 180-290: (a) The Bath Building Excavation of the third bath building was completed with the recovery of the plan of the later furnace-room (Room 43) and its adjoining fuel store (Room 44) of Phases B and C; the bath building comprised in all seven heated rooms, two furnace-rooms and a large cold plunge-bath.18 Room 23 was confirmed as measuring 16 by 9 feet, and had three flues, each 1 foot wide, through its south-west wall allowing heat to reach Room 20. The construction of the suspended floor was described in the 1963 Report;16 the under-floor consisted of opus signinum, 2 inches thick (Pig. 3, Section E-P, Layer 43) laid over a foundation of rammed lumps of chalk (Fig. 3, Section E-P, Layer 44), which were deposited on the levelled debris of Period III. Room 20 was also completely cleared; it was a narrow room mea- 18 Arch. Cant., lxxviii (1963), 137-8. « Arch. Cant., lxxix (1964), 132-3. 86 EXCAVATIONS AT ECCLES, 1964 suring 6 by 24 feet and was supplied with heat through the flues in the walls dividing it from Rooms 21 and 23. Its hypocaust was of identical construction to those in the other heated rooms of the suite. Room 43 (Plate IVB), the later and larger of the two furnaces of the baths,20 measured 13 by 15 feet 6 inches and was abutted against the north-west wall of Room 23. The north-east and south-west walls of this furnace-room were of the standard construction for the baths, 3 feet wide and of ragstone set in yellow mortar, and were both abutted against the south-east wall of the Period IIIB-E Room 64, which was retained in use as the north-west wall of the furnace-room. . This furnace-room was originally entered through a door (4 feet in width) in its north-east wall, but no evidence was found for any fuel stores in the area beyond this wall. The floor of the furnace-room was a compacted layer of yellow mortar and the fire, serving the hypocausts to the south-east, was laid on a platform built of two courses of claybonded bridging-tiles and constructed against the north-west wall of Room 23. This firing platform measured 8 by 6 feet 6 inches and faced the 3-foot-wide flue into the hypocausts. Room 44, measuring 14 feet 6 inches by 7 feet, was abutted, in Phase C, against the walls of Rooms 43 and 20; its construction was the standard ragstone and yellow mortar, but its walls were only 2 feet thick. Access to this room was through an entrance 4 feet wide, in its south-west wall. Its construction involved the blocking of the original entrance into Room 43 (Plate IVB, behind the vertical ranging-rod) and the breaching of its south-west wall for a doorway, 5 feet in width, into Room 44. Room 44 was not floored, but contained much soot and ash deposited directly on the subsoil and probably the result of treading in and out of the furnace-room; this room was used very probably as a fuel store. (b) The Living Quarters The length of wall,21 north-east of Rooms 36 and 37, thought to belong to the living accommodation of the villa in this period, was traced further in an exploratory trench north-east of the part exposed in 1963, but it has not been possible to determine so far either its full length or whether it is a courtyard wall rather than one of the building proper. The absence, however, of any floor north-west of this'wall suggests that the living quarters of Period V must lie to the north-east. A further burial was found in this trench, with its head lying close to the wall and its body at right angles to it. This is the fourth instance of interments in this general area, very likely following the destruction 20 For the smaller furnace, ibid., 131. 21 ibid., 133. 87 EXCAVATIONS AT ECCLES, 1964 of the Period V buildings in the fourth century; it is interesting to note that similar burials were found at Pishbourne.22 Periods IV-V, c. A.D. 120-290: The Ditch The north-east part of the site at some distance from the nearest evidence of occupation was bounded by a ditch (Pig. 3, Sections K-L and I-J). It had a maximum width of 10 feet and a depth of some 4 feet at its deepest; it was irregular in profile, tending to have less steep sides to north-west (Pig. 3, Section I-J) than to south-east (Pig. 3, Section K-L), and it was found filled with a very large deposit of domestic refuse and building dehris. There is no doubt that this ditch was cut during Period IV as it • had been dug through much building dehris, including painted wallplaster of the kind found in Rooms 32, 59 and 67, which was deposited in this area following the destruction of the Period III structures; likewise, it is quite certain that it remained in use throughout Period IV, gradually filling in with domestic rubbish to be finally abandoned in Period V; for the pottery it contained was entirely of the second century and showed a total lack of forms and fabrics exclusively first- or third-century A.D. in date. DATING This third year's excavation has provided evidence for a partial revision of the tentative dating suggested in the two preceding reports. Period I, the Ditch: Work on this part of the site is not yet sufficiently advanced to estabhsh a date for the cutting of this ditch, but it is known that it began to get filled with domestic rubbish about the middle of the first century A.D.; indeed, the pottery types contained in its filling show strong Belgic influence and can hardly have lasted in use much beyond Claudian times. In consequence, the dating of the close of this period to c. A.D. 55 is based on secure grounds. Period II, c. A.D. 55-65: The Granary: The only definite structural evidence in this period is the small granary built partly over the filledin ditch of Period I and, probably, Rooms 53 and 54 against which part of the first bath building is abutted. The closing date of this period is based on the pottery stratified in the destruction layers of the granary, which is largely of Neronian date, with some earher fabrics and forms surviving from Period I. Period III, c. A.D. 65-120: The First Bath Building: This period is extended at either end; it was already abundantly clear early in the season that the former dating, though • based on sufficiently strong evidence, was not wide enough to allow a useful lifetime for the bath 28 Antiquity, xxxix (1965), 183. 88 EXCAVATIONS AT ECCLES, 1964 building, particularly in view of its various reconstructions. The closing date is arrived at by adopting the later of two possible dates for the stratified pottery and is further supported by two fairly worn coins of Nerva, an as and a dupondius, found in the destruction layers of Room 46. Period IV, c. A.D. 120-180: The Second Bath Building is also dated rather later in the second century by again adopting, at either end, the later of two possible dates for the stratified pottery. Period V, c. A.D. 180-290: The Third Bath Building is shortened as the result of the later dating for the closing of Period IV, but no evidence has been forthcoming to suggest any reconsideration of the date for the closing of this period. SUMMARY After three seasons' work, many questions posed from the very beginning of the excavation remain still unanswered and the hypotheses, suggested in the earher reports as possible interpretations of some of the problems, are still valid. One of the gains of this season's work has been to establish what is often only conjectured on Romano-British sites in Kent, a pre-Roman occupation; though this is now quite certain at Eccles, what is enclosed by the ditch of Period I can only be guessed. On the other hand, it does not seem unreasonable to postulate a direct development from a native farm to a fully Romanized style of living. One of the questions still to be resolved is how it became possible for this change over to Roman concepts of building and standards of living to take place so soon -after the conquest, and for the first baths to be already destroyed and replaced by the second bath building at a time when, at other sites, simple villas were only beginning to become estabhshed. Por at most villa-sites, baths were a luxury added only in the second century or later. The suggestion made earlier23 of a philo- Roman landowner adapting himself to the Roman example within a decade of the invasion is, in the absence of any conclusive evidence for a military occupation of the site, still the most likely interpretation; this landowner could well be one of the ancestors of the Bellicius Ianuaris who, thoughtfully, inscribed his name for posterity on a silver spoon and a bowl of samian ware, which were found this season.2* In general, the picture that emerges so far is of a site with a large bath building, built by an architect with a military background, replaced by another with marked civilian characteristics though with 23 Arch.Cant.,lxxix (1964), 135. 24 Kindlyreadby Mr.R.P. Wright,F.S.A.;cf.«7.ftS,lv (1965), 224-6, no. 18 and pi. XVI, 3, and 226, no. 33. 89 EXCAVATIONS AT ECCLES, 1964 some reduction in size; and a third bath building on an even more grandiose scale than its immediate predecessor. What changes in fortune or ownership these alterations may reflect, it is impossible to say; but there is every likehhood that some of these problems may be resolved in the excavation of the villa itself in future years. APPENDIX THE FRIGIDARIUM MOSAICS (Pig. 4 and Frontispiece) by David S. Neal On examination of the fragments, two types of workmanship can be distinguished (1) Fragments associated with a panel containing two gladiators, and (2) a possible aquarium scene. The tesserae of the possible aquarium scene are of finer workmanship and different in colour from those used in the gladiatorial pavement; again, the mortar backing on the former is only one inch thick, compared with a thickness of two to three inches on the latter. So far during reconstruction, none of the one-inch fragments have joined to the two-inch ones. Designs associated with the Gladiator Mosaic (a) Fragments of perspective box-pattern with internal lozenge measurement of 7 inches. (6) Remains of four-leaf rosettes originally set into boxes with an internal measurement of 15 inches (Pig. 4, 2). " (c) Single leaf design with stalk set into boxes with internal measurement of 9 inches (Fig. 4, 4). (d) Two types of guiUoche. Nine-inch wide double-braid guilloche and 6-inch wide single-braid guilloche. (e) Fragments of two gladiators with armoured cod-piece and arm bands (Fig. 4, 1 and 3). On examination and reconstruction of the fragments, a pair of legs can be seen standing on an irregular grey line suggesting a ground, line or shadow. At the top of the leg is a point, dropping down and looking very much like the tail of a stag, but which is almost certainly an armoured cod-piece with the hem-line of a skirt running behind it. Another fragment shows an arm and wrist bound with leather thongs. A red sword-pommel can be seen, and part of the right foot of the second gladiator. All these features are to be seen on two of the gladiator panels on 90 am CDDDDDDDDDmCODDDDDDd • D U O 4 FIG. 4. Reconstructed fragments of the Frigidarium Mosaics EXCAVATIONS AT ECCLES, 1964 the large Gladiatorial Mosaic from Reims.25 No fragments resembhng a face have yet been seen, but the face could well be obscured by a helmet as in the gladiators from Reims. Orange, yellow and blue/grey tesserae predominate in the gladiatorial panel, but these tesserae are not to be found in the fragments associated with the aquarium scene. The workmanship is not so fine, and the floor surface gives the impression of having been ground down and smoothed by the mosaicist. The grinding could not have been caused by general wear, suggesting that the mosaic was in use for only a short time. General Scheme of the Gladiatorial Mosaic The pair of gladiators would have probably featured in the square central panel, bordered by a single band of guilloche. Many fragments of perspective box pattern survive, and this design may have surrounded the main panel. Set into the perspective box pattern would have been a number of boxes containing rosette designs with tendrils radiating from between the leaves; smaller boxes contained a leaf and stalk motif. A wide band of guilloche and a white border would have surrounded the design; some of this white border was found in situ. (Frontispiece). Possible Aquarium Scene Very few fragments survive, and from these I can only interpret the reconstruction as a dolphin. The tesserae used in this work are much finer than those used in the gladiatorial scene, all measuring approximately I inch square. None of the orange, yellow and bide/grey stones appear, and the black outline tesserae are of a different shade. A purple/ red stone features in some fragments, but this colour is not to be found in the gladiatorial panel. The surface of these stones is also different, and they appear to have either had much wear or, more probably, to have been subjected to water, which has dissolved the mortar from between the tesserae; it is, therefore, possible that this mosaic came from the cold plungebath (Room 31). 25 EC. Stern, Recueil Qdniral des Mosaiques de la Qaule, I . Province de Belgique, 1. Partie Oueso, QaVia, X" Supplement, Paris, 1957, pis. XIII, 16, and XTI, 9. - 91