Excavations at Eccles 1963

EXCAVATIONS AT ECCLES, 1963 SECOND INTBIOM BEPOBT By A. P. DETSICAS, M.A., F.S.A. INTEODUOTION Excavations were continued1 by the Lower Medway .ArchEeological Research Group during 1963 at the site of the Romano-British viUa at Rowe Place Farm, Eccles, in the parish of Aylesford (N.G.R. TQ/722605; O.S. 6in.SheetTQ/76S.W.). The work was undertaken once again by kind permission of the landowners, Messrs. Associated Portland Cement Manufacturers Limited, and with the most generous co-operation of the tenant farmers, Messrs. A. A. and A. E. SouthweU, to whom I am greatly indebted for their continuous help. The excavation was entirely financed by generous grants from the Carnegie United Kingdom Trust, the Kent Archseological Society, the Craven Fund and the Haverfield Trust of the University of Oxford, and by pubhc contributions. My thanks are due to the many volunteers who made this excavation possible, and in particular to the foUowing for their sustained support throughout a lengthy season: Misses A. M. Arnott, M. Bennett, L. S. Smith and M. Webster; and Messrs. D. Bartram, I. J. Bissett, P. Dives, D. Dorrington, L. A. Griffith, A. C. Harrison, B.A., T. Hetherington, W. A. Knowles, C. E. J. Martin, A. MUes, A. M. Rowland, B.A., E. R. Swain and P. Whiteoak. I am also greatly indebted to Dr. J. P. C. Kent, F.S.A., for kindly identifying the coins, and to Mr. R. G. Foord for taking charge of the monochrome photography and supplying the prints Ulustrating this report. THE EXCAVATION Work this season was carried on mainly to north and west of the 1962 area and feU into two distinct parts, (a) on the Bath BuUding, and (6) on the Living Quarters. In both areas the same three periods of construction were recognized as were noted in last year's interim report. 1 Arch. Cant., Ixxviii (1963), 125-41. I am indebted to Professor Sir Ian Richmond, P.S.A., and to Professor S. S. Erere, V-P.S.A., who kindly read the present report in draft form and contributed many helpful suggestions. 121 EXCAVATIONS AT ECCLES, 1963 Period I: c. A.D. 75-100. (a) The Bath Building Very Uttle was recovered during 1962 to suggest the purpose of the few structures known to antedate Period II. I t is quite clear now, however, that some parts of an early bath buUding could have been recovered under the opus signinum floors of Rooms 20 and 21 where considerations of space and time at the very end of the earlier season of excavation prevented methodical examination at depth. A part of this first bath buUding has now been exposed, but it is certain that much of it stUl awaits investigation to the west of the 1963 area. It was found partly under the successive structures of Periods I I and I I I but, where its rooms had avoided subsequent destruction, it lay close to the topsoU. The plan of several of its rooms has been completely recovered and presents the normal bath-house plan with one rather unusual feature for a vUla, a circular laconicum, Room 32, projecting beyond the main line of the buUding. Boom 32. This is a large laconicum, circular2 in plan, except where it meets the north-east waU of the bath buUding where aUowance was made for a doorway. Its waU was 1 ft. 6 in. thick, buUt of ragstone set in bright yeUow mortar, and the room had an internal diameter of 18 ft. InternaUy, the waU was rendered with bright yeUow mortar above the level of the suspended floor only, but no traces remained of any boxtiles; externaUy, the waU was similarly rendered and finished with a coating of whitewash (Plate I). The laconicum was heated by a hypocaust constructed on a floor of hard yellow mortar aggregate, which was laid on a very solid foundation of mortared ragstone resting directly upon the subsoU; the floor was 2 in. thick and the foundation 1 ft. The suspended floor, buUt of bridging-tUes none of which was recovered intact, was supported by pilae, consisting mainly of square tUes ( 9 | by 9£ by 2 in.) mortared together and rising, at a height of 2 ft. 4 in., to the level of the suspended floor; some larger tiles (of two sizes, 11 by 11 by 2£ in., and 16 by 11 by 2 \ in.) were also used as bases for the pilae. The ragstone wall of the laconicum was interrupted, at a height of 1 ft. 8 in., by three courses of corbeUed bonding-tiles, each course projecting a httle further inwards so as to afford additional support to the suspended floor at its weakest points. The hypocaust was originaUy served by a stokehole to its southwest. An original flue, 1 ft. 10 in. wide, pierced the waU which, at this point, was neatly finished with courses of bonding-tiles to give it a regular face; the floor of the laconicum was extended beyond the waU into this flue. Both the floor near the flue, the flue-waUs and the pilae nearest to it showed evidence of considerable burning. This original flue was subsequently blocked by ragstones mortared to its waUs, and the hypocaust was served by a new furnace to the north-west at a 2 'Ipsumque ad circinum fieri oportere videtur.' Vitruvius, De Architectura, V x 5. 122 EXCAVATIONS AT ECCLES, 1963 time when it would seem clear that some of the rooms were reconstructed (Plate II). The construction of this secondary furnace necessitated breaching the waU of the laconicum. The new flue had cheeks 1 ft. thick, buUt entirely of bonding-tiles set in a bright yeUow mortar, which appeared a httle darker than its original colour due to burning; both cheeks were very reddened by fire and in a very brittle condition. The floor of the flue was constructed of roofing-tiles, which were laid directly on the subsoU outside the laconicum and projected shghtly on to the room's floor; these roofing-tUes were covered with a thick deposit of soot and ash, in which were found pieces of charcoal and mineral coal, with much pottery and domestic refuse extending on to the floor of the laconicum beyond the mouth of the flue. The external waUs of the stokehole were 2 ft. thick, of ragstone set in bright yeUow mortar; the south-west waU of the furnace was buUt directly against the north-east wall of the bath buUding. The space between the south-west external waU and the south-west cheek of the flue contained much accumulated soot and ash as well as much domestic refuse. Circular baths such as Room 32 are normaUy a feature of miUtary establishments,3 whose laconica have independent furnaces and are entered directly from their frigidaria; these are usuaUy dated from the Flavian period to the middle of the second century A.D. Town baths, too were equipped with circular laconica or sudatoria,4, but there seem to be very few paraUels for such circular baths in viUa sites. Grenier,5 in discussing the Rouhling vUla,6 refers also to the vUlas at Friesdorf,7 WUtingen8 and Bubenheim;9 the Rouhling laconicum does not project beyond the line of the bath buUding, but is incorporated within it. The only close paraUel to Room 32 in a British vUla would seem to be at the Ashtead viUa,10 which had a circular laconicum offset at one end of the bath block, but of smaUer diameter (10 ft.) than Room 32. The area immediately to the south-west of Room 32 has not yet 3 For a general discussion, see M. C. Fair, Circular Bath Buildings in Connection with Cohort Forts, in J.R.S., xvii (1927), 220 ff., with fifteen British and German military parallels; also, C. M. Daniels, The Roman Bath House at Red House, Beaufort, near Corbridge, in AA*, xxxvii (1959), 85-176. 4 E.g. the Roman baths at Bath, Archceologia, lxxv (1926), 1-18, and at Wroxeter, D. Atkinson, Report on Excavations at Wroxeter, 1923-1927. 0 A. Grenier, Habitations gauloises et Villas latines dans la Cite" des Midiomatrices, Paris, 1906.1 owe this reference to Professor Sir Ian Richmond, P.S.A. 0 Op cit., 129. 1 Bonner Jahrbucher 81, 212. 8 Jahresbericht des Gesellscliaft fur nutzliche Forschungen, Trier, 1856, 61 ff. 0 Bonner Jahrbiicher 72, 126. 10 A. W. G. Lowther, Excavations at Ashtead, Surrey, Third Report (1929), in Surrey Arch. CoU., xxxviii (pt. II), 132-48.1 owe the original reference to Lt.-Col. G. W. Meates, F.S.A. 123 EXCAVATIONS AT ECCLES, 1963 been exammed, but it must clearly contain at least one room giving access to the laconicum through the doorway in its waU. Boom 30 is at present partly inferred and occupies an area, approximately 30 by 19 ft., containing Room 31. It is very hkely that Room 30 was the frigidarium of the bath buUding, giving direct access to the laconicum, and Room 31 the cold plunge-bath of the frigidarium. Part at least of Room 30 was floored with a mosaic, a few square inches of which were found in situ at the north corner of the room; this mosaic was laid on an opus signinum bedding, some 4 in. thick, buUt on a foundation of tightly packed yeUow mortar and gravel, which was found badly damaged by the later construction of the second period drain. No trace was found, within the excavated area, of a south-west waU continuing the aUgnment of the south-west waU of Room 31; but this is not necessary, if it can be accepted that the whole area was one room. Boom 31 is a smaU cold plunge-bath, 13 by 7 ft., with waUs of ragstone set in the usual bright yeUow mortar; these were found robbed down to their lowest foundation courses. The south-west waU of this room was only 1 ft. 6 in. wide; the other three waUs were of the standard 2-ft. thickness of the rest of the bath buUding. This room was floored with a mosaic, in the weU-known manner of other cold plunge-baths, laid over an opus signinum bedding, 1 ft. in thickness, itself laid on a mortared ragstone foundation resting directly upon the subsoU (Section K-L). The mosaic was almost completely destroyed—or perhaps its tesserae had been methodicaUy removed for re-use elsewhere—except in one or two places at the junctions of floor and waUs where a few square inches of tesserae stiU remained in situ. Nothing can, of course, be said about the pattern of this mosaic, except that it probably had a white border.11 Water from this plunge-bath was disposed of by a drain buUt at its west corner, which was later re-used in Period I I (Plate IV). I t was buUt with ragstone waUs, of which the northern was 1 ft. and the southern 1 ft. 6 in. thick; these waUs were set in bright yeUow mortar within a construction trench cut through the subsoU. The bottom of the drain consisted of tUes, but both bottom and sides were covered internaUy with a coating of opus signinum forming a very effective seal. The channel of the drain was found to be virtuaUy clean without any accumulation of sUt. The area between the south-east wall of Room 31 and the north-east waU of Room 28 was found completely unoccupied and fiUed with much debris resting upon the undisturbed subsoU (Seotion K-L). There was no trace of any kind of floor, nor any evidence for anything structural up to the south-east waU of the bath buUding whioh, though 11 Cf. the Fishbourne mosaics of same date, in Antiq. Journ., xliv (1964), 1-8. 124 £ C £ S 6 3 ..BURIAL BURIAL glwOOD^ i |?ENFjPE%. § BURIAL OPUS SIGNINUM FLOOR ,-f I0!!5?!!1! -mm ^ ^ ^ OPUS SIGNINUM STOKEHOLE I I^Mi ///S/S/f y//s///s, ///////S///S.S ',**,,SS,/,(Sff w»m m^n°°m m Aa" 3:i" aL Ja L Si Ua. Ac^:;^- > - » - : c : D • " • n a : ; : : % ; . . X - STOKEHOLE DOORWAY WfczzA © ^ M ^ ^ S \ \ \ ^ ? v \^ ^J^^M»^^^W; n_a • i: RUBBISH PIT B HHHfflW ELEEPER t = T BEAM STOKEHOLE LEAD PIPE k-\3L•J UnLc>h-LcJi Ua aUdLfsiX ILALJUAj , UAL.. 1V$J nn n n n rWi n in ri rj in r; f PHASE A ETfH PERIOD I PHASES 3 n a LA J D D :>z: z:::z g g | INFERRED OR FOUNDATIONS nn CH: oac I PHASE C E H PHASE A DRAIN [ 2 3 INFERRED OR FOUNDATIONS BURNT %%\ INFERRED f^\ ROBBED PHASE B PERIOD II PHASE C P%g] INFERRED OR FOUNDATIONS PHASED PTg INFERRED PHASE E g g ] INFERRED PHASE A PERIOD 111 E ^ INFERRED PHASE B S B ^ METRES Fio. 1, EXCAVATIONS AT ECCLES, 1963 robbed down to its lowest foundation courses, could be traced to Rooms 28 and 29. No explanation can at present be offered for this vacant space, unless it can be considered as an entrance or corridor leading into the larger Room 30, with a floor which need not have been other than mortar and was completely destroyed in the subsequent robbing. This possibUity is reinforced by the fact that the subsoil in this space was encountered at the same depth as the lowest level of the robbing of the waUs of Rooms 28 and 31; this clearly suggests that the floor lay at a level only very shghtly higher than the foundation courses of these waUs. Boom 28 was found to have been completely demohshed. It measured 12 ft. 6 in. by 9 ft., with waUs of ragstone and bright yeUow mortar; these were 2 ft. thick, except for the south-east waU dividing it from Room 29 which was only 1 ft. 6 in. wide. This room had contained a hypocaust, of which only three tUes (9 by 9 by 2 in. thick) of a pila were found in situ set upon an opus signinum floor with yeUow mortar; the floor was 4 in. thick and laid on an extremely solid foundation of mortared ragstone to a thickness of 10 in. lying on the subsoU in identical manner to the foundation of the floor in Room 32 (Section K-L; Plate V). This room was so thoroughly destroyed in subsequent periods that it was impossible to deduce the height of the suspended floor over the hypocaust. The function of this room is clearly that of a tepidarium, in view of its distance from any stokehole and of the presence of Rooms 38 and 39 to the south-west. A short length of drain was found turning a right angle immediately outside the north-west waU of Room 28, sealed beneath a second-period waU (Fig. 1). This drain had sides and bottom of bonding tUes, set in bright yeUow mortar, but not rendered with opus signinum. Boom 29, a very smaU room measuring 2 ft. 6 in. by 9 ft, is undoubtedly a warm plunge-bath buUt over a hypocaust of which only the base tUe (9 by 9 by 2 in.) of a pila was preserved in situ at its north corner where it was mortared to the opus signinum floor of the room. The waUs of this room, except for the shghter waU (1 ft. 6 in.) dividing it from Room 28, were the bath buUding's standard 2 ft. in thickness, of ragstone and bright yeUow mortar; they were almost entirely robbed to their lowest courses and then overlaid by the successive structures of Periods I I and III. Boom 39, is largely unexcavated and only its extreme south-eastern portion was exposed below the later waUs and floors. It contained a hypocaust over an opus signinum floor, which was 2-3 in. thick; only the mortar traces of its pilae remained on this floor. The known waUs of this room were 1 ft. 6 in. thick, but of the same ragstone and bright yeUow mortar construction. This room is very likely to be the main caldarium of the bath suite, particularly in view of its apsidal end, a 125 EXCAVATIONS AT ECCLES, 1963 feature fairly usual in such rooms; its praefurnium no doubt hes further to the south-west in the area beyond the present excavation.12 Boom 38,—smaU apsidal room (10 ft. at its widest and 4 ft. in maximum depth), opens off Room 39. Its waU was 2 ft. thick and built entirely of bonding-tiles set with bright yeUow mortar, which was also used to render both the inner and outer faces of the waU (Plate III); some ragstone was also found in situ on the uppermost course of tUes which clearly suggests that, at a higher level, the construction was of ragstone and mortar. This room had been floored with yeUow mortar compacted to the consistency of opus signinum, without any traces of pilae, and it is very probable that it was not hypocausted; it may have contained a hot-water basin (labrum) but, owing to the thorough demolition of this room, it is not possible to say what was the height of an upper floor. It would seem that, in the origmal plan of the bath buUding, Rooms 38 and 39 formed one single unit, and were later separated by the construction of a waU, 1 ft. 6 in. wide and of ragstone in orange-coloured mortar, buUt across the opening of Room 38. The south-east waU of the bath buUding continues to the south-west beyond the part of Room 39 so far excavated and so does the opus signinum floor with its traces of mortar for pilae, suggesting that Room 39 was a fairly large hot room. In this area some sections of waUs have been exposed which, structuraUy, antedate the buUding of Room 39, as does a drainage system which is probably connected with the early drain found under structures of Period III in 1962 and traced only as far as the east corner of Room 19.13 However, the area in which these features were first encountered at the very end of the excavation was so restricted that their examination had to be postponed. (b) The Living Quarters The general area of the villa's Uving quarters in this first period is thought to Ue north-west and west of the bath buUding, but the evidence recovered so far is so shght that Uttle of a positive nature can be reported. A few trial trenches yielded evidence of two successive opus signinum floors in the area beyond Rooms 32 and 35; the pottery sealed under these floors is generaUy of the same types and fabrics as found in comparable layers in the bath buUding area. A section of waU was associated with these opus signinum floors, but it was found completely robbed out; no other structural evidence was recovered in this area. 12 Confirmed in 1964. 13 Arch. Cant., lxxviii (1963), 128. 126 EXCAVATIONS AT ECCLES, 1963 Period II: c. A.D. 100-150 Work was aimed at completing the plan of this period's bath buUding, and initiating the examination of the part of the viUa's living quarters known to be contained within the area under investigation. (a) The Bath Building Boom 26, belonging to the latest phase in the building's structural history, was almost completely unknown untU it was first exposed this season as only its south-east waU had been partiaUy cleared in 1962. It is a large hypocausted room (20 by 15 ft.) served by its own separate furnace, Room 27. Of its ragstone waUs (2 ft. in thickness) in the characteristic off-white mortar of this period, only the south-east waU was preserved in good condition, and only short sections of the north-east waU remained on either side of the flue from Room 27; the rest were completely robbed out, although their construction trenches could be identified. Some courses of bonding-tiles were used in the construction of the south-east waU as weU as some lumps of tufa, particularly below the level of the suspended floor, and the inner face of the waU was rendered with a coating of off-white mortar, which did not survive in situ above the suspended floor (Plate IX). The hypocaust was buUt on a floor of hard yeUow mortar, 2 in. thick, laid on a foundation of ragstone above the subsoU; the pilae buUt on this floor were composed of a larger base (12 by 12 by 2 in.) and several smaUer tUes, but none of the latter were found in situ intact. The floor was suspended over these pilae at a height of 2 ft. and rested upon shght ledges in the south-east and north-east waUs of the room. The suspended floor was completely destroyed, and only very few lumps of opus signinum were recovered in the mass of debris filling the room; it was also noticeable that very few of its pila tUes remained, clearly pointing to robbing and re-use elsewhere, very probably in the Period III hypocausts. The lowest level over the hypocaust floor in this room was fiUed with a deposit of domestic rubbish, which extended beyond the robbed north-west waU of this room (Section K-L: Layer 21) and contained much pottery of characteristicaUy late-A&ntonine date, wlule a dupondius of Antoninus Pius was found lying in the mortar debris at the extreme south-west end of this room. Boom 27 (6 by 8 ft.) is the furnace-room for Room 26, with rather thin ragstone walls, 1 ft. 6 in. wide, set in yeUow mortar. The door, 2 ft. wide, was at the eastern corner, and was floored with bonding-tiles. The stokehole itself was floored with bonding-tiles over the subsoU; these were covered with a thick layer of soot and black ash. The flue, leading into Room 26, was also floored with bonding-tiles and had cheeks 127 EXCAVATIONS AT ECCLES, 1963 simUarly finished; whether it was arched or not was found impossible to estabhsh due to demolition (Plate IX). Boom 8. The north-east waU of this room was exposed immediately to the south-east of the furnace-room, Room 40. This room is now known to be rather larger (8 by 9 ft.) than thought in 1962 as the ahgnment of its robbed south-west waU was found to be sUghtly further to the south-west than previously inferred. Boom 10. Re-consideration of the evidence in the hght of further results to the north-west of the 1962 area has led to the estabhshment of the correct dimensions of this room (18 ft. 6 in. by 7 ft.). It was observed in the earlier report14 that the south-east waU of this room, found then completely removed to foundation courses, was much thicker (4 ft.) than the rest of the bath buUding's waUs (2 ft.); it is now certain that this is due to the fact that this waU is partly overbuUt by the south-eastern waU of the Period I I I bath buUding, which formed the limit of this later structure and, in effect, incorporated part of the earher Room 10 into the later Room 42 (see below). Likewise, work to the north-west of the 1962 area, below Period III levels, has exposed the ahgnment of the north-east waU of Room 10, which is now known to end at a north-west waU the position of which has been inferred from indications under Period I I I structures at the point shown on the plan. In the original plan of the Period I I buUding, Room 10 was served by the furnace, Room 40, through two flues; of these, only the south-eastern one was retained in the foUowing period whUst the other was blocked with bonding tUes behind the south-east waU of Room 24. Boom 40, whose complete dimensions are not yet known, contains the furnace serving the hypocausts in Rooms 10-12. Its north-east wall was found partly demohshed to make room for the later stokehole and was pierced, west of Room 26, by a drain buUt directly on the hypocaust floor of the earher Room 28 (Plate V); this drain was used for the disposal of water from the area of Rooms 2 and 4 through an outlet in the north-west waU of the bath buUding. The channel of this drain was thoroughly removed in the later reconstruction of the furnaceroom, but its construction trench could be traced immediately outside the north-east waU of Room 40. The flue of the stokehole had cheeks buUt of tUes set in mortar, which were found badly worn through burning; it was 1 ft. 6 in. wide and had a clay floor baked hard and red over which was much black soot and ash. No evidence was found for a testudo emplacement, but this may have been removed in the subsequent re-buUding of the furnace. Boom 41, as yet incomplete, is represented by two lengths of waUs built directly over Period I structures and sealed under the opus signinum floor of Room 24; they were both buUt of the standard " Loe. cit., 134. 128 PLATE I «• IflPP 552 Photo: A. P. Detttivan Period I: Room 32. General View to West. [face p. 128 PLATE II •^ J . ^^^^jfi\f*?f^™[^ , <7 . . «•>•- ..~y' -A. -• ,,•_ - Period I: Room 32. Flues. BL Photo: R. (I. Foord PLATE III j » y^jWj f -• •S T _.*Jfc-. tip -*;,. Photo: R. G. Foord Period I: Room 38, under Period III Room 24. PLATE IV - ••• a * ,"' 'tl i% *•• #5.*, *r- -A. PAoto: R. G. Foord Junction of Drains of Periods I and II. PLATE V Photo: R. G. Foord Period I: Room 28, and Period I I Wall and Drain. PLATE VI ^ ^ W ^ l .4 .•J Off ' '* -"•• *> M - *• -«.' Period I I I : Room 25. Flue. ,5. ••- .. PAoto: R. G. Foord PLATE VII A» 1 ^ / ^ " - jfy A" • ^ • i a - . J I ' ^ M II j A • J l " M ? ^ ^ ^^ 4 f M : fl. G. Foord Period I I I : Room 23. PLATE VIII Photo: R. G. Foord Period III: Room 24. Detail of Box-Tiles. EXCAVATIONS AT ECCLES, 1963 Period II ragstone and off-white mortar to the usual 2-ft. thickness. Nothmg much can be suggested about the likely function of this room, though the definite absence of an opus signinum floor, which if present could have been used for the later Room 24, suggests that Room 41 was not heated. (6) The Living Quarters Several lengths of waUs, structuraUy earUer than Phase B, have been exposed in the area between the villa itself and the bath buUding belonging to this period, north of Rooms 26 and 27. The north-west waU of Room 33, for instance, is earUer than the rest of this room's waUs; of ragstone and off-white mortar, 2 ft. wide, it was buUt to the north-west directly against the ruins of Room 32, its north-west face remaining unrendered and resting against the debris of mortar and tUes from the demohshed Room 32. Likewise, some lengths of foundation courses of loose ragstone were found under Room 33 and northeast of Room 27 as well as the robber trenches of waUs north-east of Room 37. However, as aU these waUs do not, at present, appear to belong to any definite structure, they wUl not be considered in detail until later work has elucidated their structural relationship. Rooms 33-37 are a range of four rooms at one end of the vUla reached by means of a long corridor. Boom 37, the corridor, was 12 ft. wide, but its length is not yet established. Its waUs were 2 ft. thick and buUt of ragstone and offwhite mortar; the north-east waU was almost completely robbed, but its position was definitely estabhshed thanks to the robber trenches. The corridor was floored with a tesseUated pavement of plain red and buff tesserae, cut from tUes, which was exposed immediately below topsoU; the tesseUation was laid on an opus signinum bed, 2 in. thick, which showed in places signs of subsidence to the north-east (Section M-N), and over an earlier floor of white mortar. Boom 33, the largest of this range (22 by 10 ft.), had waUs of ragstone and off-white mortar, 2 ft. thick; its north-west waU, as mentioned above, is earUer than the rest. The floor of this room was badly destroyed, but appears to have been of hard yeUow mortar upon which were found some fragments of mosaic and loose tesserae, though it is thought more hkely that these were used in the make-up of the floor rather than in situ. This mortar floor was cut through by the trench of a wooden pipe some iron coUars of which were recovered; it ran from the north corner of the room towards the main drain on the south side of this room. (Section M-N: Layers 47-48.) The whole room was fiUed by a large deposit of domestic rubbish (Section M-N: Layer 55), clearly denoting the end of this phase. Boom 34 (11 by 6 ft.), with the usual ragstone and off-white mortar m 129 EXCAVATIONS AT ECCLES, 1963 walls, was floored with a thin layer of white mortar, 2 in. thick, compacted upon the Romano-British topsoil. A channel, for a wooden pipe with some iron coUars stUl in situ, had been cut through this topsoil from the north corner of the room; it led through the waUs at the south corner of the room towards the main drain immediately outside the south-east waU of the room. A burial was inserted in this room close to the east corner; probably a male, it was fully extended on its back, except for the head which, at right angles to the body, was resting against the south-east waU of the room. No grave goods were recovered, and it is by no means certain that it is a Romano-British burial belonging to a later stage in the viUa's history. .Another burial was found inserted in the area immediately outside the south-east waU of Room 34 beyond the drain (Section K-L: Layer 4), but very httle can be said of this as only the lower part of the skeleton was exposed within the excavated area; it would appear, however, that this burial, too, was interred in a position simUar to that of the burial inside Room 34. Boom 35 (20 by 9 ft.) is largely inferred, and its north-east wall was completely robbed.15 I t was floored with the same compacted layer of white mortar laid directly upon the Romano-British topsoil. The area immediately outside the robbed north-east wall contained a very solid aggregate of ragstone and yellow mortar to a thickness of some 2 ft. over the subsoil, which showed signs of heavy oxidation and probably denotes an industrial working surface of Period III. Boom 36 (12 by 11 ft.) contained much painted wall-plaster debris and rubble upon the same white mortar floor. The presence of these floors of white mortar16 laid down over the Romano-British topsoil in Rooms 33-36 and underneath the tessellated pavement in Room 37 clearly suggests that Room 37 was re-floored with its tessellated pavement in a later phase, but no floors were found in Rooms 33-36, which could be contemporary with the tessellation. A probable explanation may be that the corridor required re-flooring because of its heavy use whereas the mortar floors in Rooms 33-36 may have remained in good repair and quite suitable for their use in rooms, which could be considered as servants' quarters at the extreme northwest end of the villa; alternatively, it may be thought that even Rooms 33-36 were re-floored at the same time as Room 37, but with a less durable material than the opus signinum bed of the tessellation, which would not have survived under the slight depth of topsoil in this area. A drain was exposed outside the south-east walls of Rooms 33, 34 and 36, which has been traced below the pavement of Room 37 and 18 The south-east wall of this room was confirmed in 1964. 18 These floors of white mortar could be the under-base for timber floors. I owe this suggestion to Professor Sir Ian Biohmond, P.S.A. 130 E S 9 6 ^ STOKEHOLE PERIOD I • PHASE A mPHASE B S3 PHASE C y / / ; ^ ^ ^ / / » | B A T H •///;;/; B U n G EEZZZZZZZZZZZZ %ZZZZZZZZZ2ZZZZ£ '//;;;/// PERIOD II M«MOSAIC H-HYPOCAUST t ^ PHASE B E 3 PHASE C EH PHASED EZ1 PHASE E STOKEHOLEFEtTte& s *&=. JfiMETRES ^ < N \jLmiiN^2T^^^g - IMMK I M I V V PERIOD III ^ PHASE A ES3 PHASE B *£° Eio. 2. EXCAVATIONS AT ECCLES, 1963 beyond it to the north-east. The sides of this drain were built of bondingtiles and roofing-tiles were used to floor it; presumably, similar tiles would have been used to form a cover, but none were found in situ. Its sides were made waterproof with a thick coating of opus signinum (Plate X). About halfway along its course outside the south-east wall of Room 34, the drain channel deviates shghtly to the south-east and becomes slightly wider; it continues under the south corner of Room 33, turns a little further to the west to avoid the thick floor of Room 31, and then joins the earher drain from this room at a point a little beyond its west corner (Plate IV). Though this drain is known to continue beyond Room 37, its beginning is yet to be discovered; it is quite certain that it was already in existence before the building of the tessellated floor over it in Room 37, and that it served as an outlet for the wooden pipes leading to it through Rooms 33 and 34. Period III: c. A.D. 150-290. (a) The Bath Building The massive bath buUding of this period has now been almost completely exposed, with the exception of the later furnace still to be examined north-west of Rooms 20 and 23. Boom 25 is the original stokehole of this period and, in effect, a reconstruction of part of the earlier stokehole, Room 40. It measured 11 by 9 ft. 9 in.; its north-west and north-east walls (2 ft. in width) were new constructions, as was the south-west wall which is the full 3-ft. thickness of the bath building just beyond the flue of the furnace. The north-east waU was abutted against the north-west wall and necessitated the partial removal of the north-east wall of the earlier Room 40. The door of Room 25 lay near the north corner and was found containing a large accumulation of soot and ash; further ash had been raked outside the door (Section K-L; Layer 35). The flue of the stokehole had cheeks built of large river boulders set in clay and some bonding-tiles; its floor was of clay baked a deep brick-red colour and laid on a foundation of broken tiles resting upon the ash of the earlier flue of Period II. The flue itself was slightly longer than the earlier one of Period II and may have had a boiler placed over it; there was enough space for it and, indeed, the whole area had all the appearance of having had water-tanks placed over the flue, but of these no direct evidence was found (Plate VI). This furnace is undoubtedly the original source of heat for the hypocausted rooms to its south-west as shown by the single flue pierced through the south-west waU of Room 42. I t is possible that no provision for another stokehole had been made in the original plan of the bath building, and that Room 25 proved insufficient for the needs of such a large heated area, thus necessitating the construction of another furnace; but, at present, this is a matter for surmise. 131 EXCAVATIONS AT ECCLES, 1963 jRoom 42 is Room 10 of the earlier Period IT reconstructed rather smaUer in size (16 ft. 6 in. by 10 ft.). Its walls were badly preserved; they are partly new constructions and partly rebuildings of existing Period I I alignments. The room was extended to the north-east in aUgnment with the outside waU of the bath building, and this meant the removal of the north-east wall of Room 10 and the extension of the floored area over it. A new north-west wall was built shghtly inside the inferred position of the north-west wall of Room 10, blocking the north-western of the two earher flues. The south-eastern wall was breached by a flue to allow for the heating of the new Rooms 19-23 to the south-east. I t would appear that the hypocaust pilae of the earlier Room 10 were mainly retained during this reconstruction as they are of a different size from those used in the exclusively Period I I I rooms. Boom 24 (9 by 7 ft. 6 in.) was a small heated room. Its south-east waU did not contain any flues and very Uttle heat would have reached this room from Room 25 until the building of the presumed later furnace to its north-west. Its external north-east and north-west walls were the standard Period IEE width of 3 ft., of ragstone set in yeUow mortar, and the north corner of this room cuts through the north-east waU of Room 41. The hypocaust was constructed on an opus signinum floor, 2 in. thick, which was laid on a foundation of lumps of chaUc deposited on the levelled debris filling the Period I I rooms; it also sealed underneath it the robbed remains of the Period I I walls. The suspended floor, of which many large fragments were recovered in the debris filling this room, was of opus signinum, some 4 in. thick and supported by pilae. These consisted of one base tUe (10| by 10| by 2 in.) and several smaller ones (71 by 7J by 1 in.) set in yellow mortar, varying in thickness from £ to | in.; the suspended floor was supported by these pilae at a height of 2 ft. The walls were lined with box-tiles (about 6 in. in width), which were affixed to the waU of the room and held against one another by a layer of opus signinum, to a thickness of If in. between each box-tile (Plate VIII); the box-tiles were supported on a slight ledge projecting from the walls of the room, and only about one-third of their depth was set on this ledge so that heat would circulate through them. Such boxtiles, or their opus signinum imprints, were found on aU waUs of this room, except for the robbed south-western partition wall. Room 24 was divided from Room 23 by this partition wall, which was 2 ft. thick and constructed of ragstone set in yellow mortar; it was rendered with a thick coating of yellow mortar and contained three fines,17 each 1 ft. wide for the heating supply from Room 23. Boom 23 was a larger room, which was only partly exposed; its dimensions are probably 16 by 9 ft. as the wall dividing it from Room 20 17 Not shown on main plan, but see Fig. 2. 132 PLATE IX ^•^**'t ;™ Photo: R. G. Foord Period II: Room 26. Hypocaust. [face p. 132 PLATE X :S*4 jt « \ A £ Jff^B^r^w^, Photo: ft. <;. Fount Period II: Drain under Room 37, EXCAVATIONS AT ECCLES, 1963 is hkely to be on the same ahgnment as the south-east walls of Rooms 21 and 22.18 Its walls and hypocaust were of identical construction with those of Room 24; only the size (7£ by 1\ by \\ in.) of its smaller pilatiles was slightly different. The north-west waU of this room had a flue in it conveying heat from the unexcavated stokehole outside; this flue was lined with clay, which was found baked to a hard brick-red consistency (Plate VII). Of this later stokehole, only a small length of its north-east wall was exposed; built of ragstone and yellow mortar to the standard 3-ft. thickness, it was constructed on the opus signinum floor of Room 39 and is presumed to be built directly upon part of the south-west wall of Room 41. Rooms 23 and 24 were both filled with a large deposit of opus signinum and tile debris as well as much decayed mortar from the waU faces, which lay directly upon the soot on the hypocaust floors; this deposit provided evidence for interments as in Rooms 20 and 21 in 1962.19 Again, and as observed during the first season of excavation,20 a deliberate effort appears to have been made to seal this deposit of rubble with a layer of large lumps of opus signinum from the destroyed suspended floors. (6) The Living Quarters Very little is known of the hving accommodation of the viUa in this period of occupation. The only evidence found is that of a length of waU exposed immediately north-east of Rooms 36 and 37. This wall was constructed of the standard ragstone and yeUow mortar of the period to the usual thickness of 3 ft., and lies immediately north-east of its second period predecessor, the edge of which it slightly overlaps; it was abutted on to the north-east wall of the Period II corridor, but no further evidence is at present available about the purpose of this wall. .Another burial was found at the eastern corner of this trench, south-east of the Period LT drain; a set of bronze toilet instruments, found in direct contact with the skeleton, suggests that this burial at least may be of Romano-British date. DATING Period I: c. A.D. 75-100. The tentative dating for this period suggested in the 1962 report is stiU retained for the time being, though evidence is graduaUy accumulating that the initial date of c. A.D. 75 may be a little late. Some of the material stratified in layers immediately above the undisturbed subsoil may weU have become deposited rather 18 Confirmed in 1964; with three flues, w Arch. Cant., bcxviii (1963), 140. 20 Ibid. 133 EXCAVATIONS AT ECCLES, 1963 earher than c. A.D. 75. A fairly large deposit of coarse pottery found in the ash of the secondary stokehole of Room 32, which was sealed by later debris and is clearly the result of breakage by those tending the hypocaust, is consistently late first-century in date. The dating suggested for the close of this period depends on pottery and, in particular, on figured samian ware, of types not usually thought to have been current much later than c. A.D. 100-120.21 However, to extend the first period of occupation of the site much beyond about A.D. 100 would seriously curtail the already rather short span aUotted on secure grounds to Period II, and if Period I is to be made longer, it must start earlier than the suggested date of c. A.D. 75; an earher starting date would suit the evidence at present available. Period II: c. A.D. 100-150. This tentative dating, very slightly modified in one respect from that suggested in the earlier report, has been given additional support by pottery and coins found stratified in the destruction layers of Periods I and II. The closing date for this period is based on the rubbish deposits (Section K-L: Layer 21; Section M-N: Layer 55) found in Rooms 33 and 26, which have produced coarse pottery and samian ware characteristically Antonine to late-Antonine; it is further supported by figured samian stratified below the hypocaust floor of Room 23 and by a worn dupondius of .Antoninus Pius of A.D. 145 deposited on the floor of Room 26 under the debris filling it, which provides a definite terminus post quern for the destruction. PeriodHI: c. A.D. 150-290. The initial dating of this period is based on the pottery stratified under the floor of Room 23 in the levelled debris under the chakk lumps which supported the opus signinum; it is clearly of mid-second-century date and supports the pottery and coin mentioned above. The closing date, suggested in the 1962 report, is confirmed by the pottery and stratified coins found in the destruction layers of Room 25. Though the pottery was rather scanty, enough was recovered of late third-century fabrics and forms, and in full accordance with the pottery types recovered from the 1962 Rubbish Pit C.22 Furthermore, the soot and ashes raked out of the furnace (Section K-L: Layer 35) yielded a coin of Victorinus, and the lowest destruction layer in Room 25 four coins of Carausius. SOTQIIARY AND DISCUSSION Mter two seasons' work on this site, it is clear that the complete extent of this large villa is yet to be determined, as is also its place in the history of the area. Considering the bath buildings alone, the striking feature is the 21 Further supported by two dupondii of Nerva of A.D. 98 found in 1964 stratified in the destruction layers of this period. 22 Arch. Cant., lxxviii (1963), 140. 134 / %i 1 y • ^ '>• '^' i>' •>• '>' '>• '>• '>• •>' •>• •.> ">' '>• •>• •> ">' •> Room 37 Mn!T%iWiM t Period lib 3 2. |ii|il|iiilli|H|ii|i II II ui 11 iTmli HI MI in 1111111111111111111111111111111n111111111<—"~~"—«—II—1 — & - Drain side v Bottom of drain ^ Burial m^m^m^mm'h^m U! I ' i i© i ' ' '-I I Ml lo.MHiiinn rrrt-; rrTTfTTT ^^^S^Br^mm^^ Period lla M Coin of Hadrian m •wmf-mrtfmxrrrrrm 11 I I I ^ I i i i ' i M 1111 M 11 u n i i 11 i ' IT 11,-v ' ' ! i i i i i 11 111 i 11 L L 1 U J . • . ^ h ± - y < i K 1 1 i i M 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ® i 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 in Period lib piTnwn MTnTrnTrnrnTrrTWn^titTrr^^ fTmnnutni r r r r r r ^ r r r r r r rr Period lla Room 33 N m i i ! ! ! ! " ! i " ! f l i n i i ii Room 31 Period Ib Period Ib nrnTmnTi rff i 1 ri h "j *t "l 1 TTTTTTfrrT f I f t f l fi" f*TT 1111U1111 ff 'TTTTfTTTTrTTrrT-rrrTTi i' m i u i i i i i i i i i ii II Mill Illllllllllllllll I (a Ml! INIIII Hj " ^ ~ i