The Roman Villa at Darenth

THE ROMAN VILLA AT DARENTH E.W. BLACK INTRODUCTION The villa at Darenth was largely uncovered in 1895. Some structural alterations were noted in the report by Payne in Archaeologia Cantiana 1 and these and the plan suggest the building had a long and complex history. In Archaeologia2 in 1905 George Fox proposed three periods, involving the operation of a fullonica on the site. In his article he gives some additional information not appearing in Payne's report. In 1969 an additional bath-house and an aisled structure were excavated. 3 Further excavations close to the west side of the villa were carried out in 1972 but have not been published in any detail. 4 In the following observations, the identification of blocks and rooms follows the system of Payne, not of Fox. For convenience, the detached bath-house excavated in 1969 will be referred to as building F and the sequence of buildings, including the aisled building, also excavated in 1969, as building G. Fig. 1 shows the general layout of the site and Figs. 2 and 3 show Blocks A, C and E in greater detail. THE COURTYARDS AND OUTBUILDINGS The most recently published comments on the plan of the Darenth villa have come from D.J. Smith.5 He has pointed out that room 26 ' George Payne, 'The Roman Villa at Darenth', Arch. Cant., xxii ( 1897), 49-84. 2 G .E. Fox, 'Notes on some probable Traces of Roman Fulling in Britain·, Archaeologia, Iix (1905), 218-32. ' B. Philp, Excavations in west Kent 1960-1970, .1973, 119-53. • Britannia, iv (1973), 322. 5 D.J. Smith, 'Regional Aspects of the winged Corridor Villa in Britain', in Malcolm Todd (ed.), Studies in the Romano-British Villa, Leicester 1978, 122-3. 159 A d 68 71 buildings excavated 1894-5 􀀂66 o,_ ____1.....0 0 ____2 ..00_ ____3 00'-' feet o._ __ --c-____ so. ________ 1;;:;oi,et Building F gravel tracK Building G Fig. l. Overall Plan of the Roman Villa at Darenth. THE ROMAN VILLA AT DARENTii in Block C of the villa, the gateway in the inner courtyard wall, and the possible shrine (66) to the south, were aligned so that the shrine would be visible from room 26. At a later stage, the gateway was blocked by the construction of room 43, originally a lined cistern holding water. The relationship between room 26 and the cistern is made obvious by the long structure (42) linking them. The relative chronology of the various elements is not certain, however, and requires some consideration. Plate I of Payne's report in Archaeologia Cantiana shows fragments of the apsidal wall of the cistern visible crossing over the thick south wall of the courtyard. The cistern does indeed seem to be later than the wall, but there does not seem to be any definite evidence that the courtyard wall was still standing above ground level when the cistern was constructed, although it would seem likely that it was. The shrine (66) is aligned with the courtyard wall, Block C, and the probably early building F, so that it is probably earlier than the laying out of the outer courtyard where building 65 on the west side seems to be aligned on the gravel track and building G. There is some slight indication that the south wall of the inner courtyard and its monumental gate were built later than the shrine. The east and west walls of the inner courtyard are clearly aligned on the west and east walls respectively of Blocks A and E. The gateway in the south wall does not lie directly opposite room 26, but its siting does provide a good view of the shrine outside the courtyard from that room. That symmetry, where possible, was important in the positioning of the gateway, can be seen from the fact that the asymmetrical placing of the gate seen from inside the courtyard was disguised from a visitor approaching from the south. A wall, presumably part of a building, ran south '.from the south courtyard wall east of the gate, some 16 ft. (4.80 m.) from where it joined the east wall of the courtyard. The fac,:ade of the courtyard was thus reduced to some ·70 ft. (21.35 m.) east of the gate, matching closely the 71 ft. (21.65 m.) between the gate and the western courtyard wall. The positioning of the shrine is not indeed quite central against such a fac,:ade, but it is not symmetrically placed in relation to any other element, and the positioning of the gate in the courtyard wall is only to be explained by the desire to link the shrine and room 26. There seems to have been an attempt to create an impression with the same elements from opposite directions, spoiled by the eccentric positioning of the shrine. From the south, the courtyard wall and the gate would appear symmetrical, but that the shrine was not quite centrally placed in relation to this fac,:ade. From the north, looking from room 26, the shrine was visible through the gate, but 161 E.W. BLACK the gate was clearly not placed centrally in the courtyard wall when viewed from this direction. We come next to the cistern and the long structure inside the inner courtyard. The cistern, as we have seen, blocks the monumental entrance to the courtyard. Since Fox's article in Archaeologia suggesting there was a fulling and dyeing establishment at Darenth, the long structure (42), described by Payne as a hall, has been regarded as having held water. The walls were 3 ft. (0.91 m.) thick and there was, in addition, an inner facing of tiles. Payne only mentions a floor consisting of some 4 in. (10 cm.) of yellow concrete which he states was on the same level as that of room 39 lying in front of Block C and joining Blocks A and E. The surviving height of the walls of room 42 is said to have been some 3 ft. (0.91 m.), but Payne states that they did not survive to ground level and therefore that no trace of doorways remained. He must have found the yellow concrete floor some 3 ft. (0.91 m.) above the bottom of the walls, and Fox was no doubt right in considering that room 42 was originally a structure sunk below ground level which was later filled in. Presumably he was also right in stating that the original floor must have been robbed. Fox suggested that room 42 was a tank connected with fulling; Philp has described it as an ornamental pool.6 It seems quite likely that it was filled with water: exactly the same tile lining was given to the walls of room 36, the 'swimming pool' in the bath-house (Block D).An objection to the theory that room 42 held water is the apparant lack of drains. There was a tile gutter, presumably for inflow, and an outflow, from which the lead pipe had been robbed, for'the cistern, but apparently the long structure had neither. It is possible, of course, that they were missed by the excavators or perhaps destroyed, or robbed for any lead, when room 42 was filled in. The walls of room 42 in Plate I of Payne's report, seem to stop where they meet the curved wall of the cistern, and also seem to broaden slightly to meet it. The cistern and room 42 are either contemporary or the cistern is earlier. Unfortunately, the walls of neither structure survived from the point described above to where they joined the courtyard wall, but their hypothetical continuation on Payne's plan must be correct. Structure 42 was eventually filled in, presumably after the removal of its original floor, and the yellow concrete floor was laid at the same level as that of room 39. Fox states that the concrete 6 Philp, op. cit., 119. 162 TI-IE ROMAN VILLA AT DARENTH had been laid on a mass of clay. 7 Payne remarks that there was no indication that room 39 had ever been roofed. Its 'floor' was of flints grouted in mortar. It is possible, however, that room 42 was roofed in its final state, since the yellow concrete does not sound like the surface of an uncovered path and some remains of painted wall plaster were found in excavating the structure.s The cistern may have continued in use when room 42 was converted, perhaps to form a covered approach to it. Building G is clearly aligned on the gravel trackway, also found in 1969. So is building 65, though the two buildings are not quite parallel. Allowance must presumably be made for slight shifts in the line of the track or the buildings may only be approximately aligned on it. The odd lack of relationship between the alignments of Block C and Block E could also be explained if Block E were aligned on the trackway. Building F to the south of building 65, is not aligned on the trackway, but on Block C. The earliest structure on the site of building G produced no datable finds, but the excavator is inclined to place it in the middle of the second century in view of its place in the sequence of buildings and the lack of first-century occupation material. 9 If this is correct, the trackway must also have been in existence by then. Unfortunately, we cannot say whether or not it was earlier than this. 10 But if the fact that the bath-building (F) is not aligned on the trackway means that it is earlier than the trackway, then the date of construction of building F will provide a terminus post quem for the trackway. The bath-building is not securely dated, although a metalled surface adjoining it produced some second-century pottery, which Philp suggests may imply the building was in use in the second century. It also resembles a bathbuilding at Hayes which probably belonged to the first half of the second century. 11 Although not very strong, this evidence suggests that the track may be roughly contemporary with the first structure 7 Fox, op. cit,, 230. 8 Payne, op. cit., 65, and Fox, op. cit., 22-6, where it is stated, 'On the outer line of the flint rubble backing the brickwork of the tank (42) was built on either side a waIJ I ft. 5½ in. thick, leaving a gangway down each side of the tank with a width of 3 ft. 6 in.' This presumably supported a roof over the filled-in pool. • Philp, op. cit., 134. 10 The note in Britannia, iv (1973), 322, refers to a third-century road found west of the villa. No plan seems to have been published showing the relationship of this to the (undated) road found in 1969. However, even if they are parts of the same road, it must surely have been preceded by an unmetalled track on which the earliest building E was aligned in the second century. 11 Philp op. cit., 124. 163 E.W. BLACK on the site of building G, and therefore that building 65 and Block E should not be earlier than the middle of the second century. The west wall of the courtyards is a continuation of the east wall of Block E. It abuts the north wall of building 65, which projects slightly into the outer courtyard. This presumably implies that building 65 was originally free-standing. The date of building 71, apparently facing building 65 across the courtyard, is unknown. It is parallel to Block A and the east wall of the inner courtyard, although the inner courtyard wall does not seem to have been continued on this side to close off the outer courtyard as it was on the west. Payne's description of building 71 makes it clear that it was very similar to the first building on the site of building G. The walls of both buildings were only a few inches high and were smooth on top. They obviously acted as sleeper walJs for timber buildings. The minimum aTea of building 71 is some 2,427 sq. ft. (740 m2), although the total length of the building is not known. The earliest building beneath building G may have had a minimum area of 1,460 sq. ft. (445 m2), although it could have been 1,880 sq. ft. (573 m2).12 Again, the exact length of the building is not known, but it was narrower than building 71 and its area was probably rather smaller. The floor of building 71 was of rammed chalk, and that of the building excavated in 1969 of rammed chalk and loam. Both buildings, therefore, could have served similar purposes. Alterations were carried out to the earliest building G, probably in the second century, but it cannot be said that these definitely imply a change of use. Early in the third century, however, a completely new building was constructed on the site and supplied with an opus signinum floor. This must certainly signify a change of use. It is tempting to use the hlstory of building G to throw light on building 71. The first period structure was certainly similar to building 71, but does this imply that they were roughly contemporary, because the construction technique was the same, or that one succeeded the other because their function was perhaps the same? The walls of buildings 67, 68 and 70 were very poorly preserved. Their general alignment suggests that they did form part of the outer courtyard of the villa. It should be noted that the north wall of building 67 is in line with the south wall of the inner courtyard. 12 See Philp, op. cit., 124, where the length of the building seems to have been taken as 108 ft. without any explanation being given. In the text the possible figures mentioned are 73 ft. and 94 ft. 164 THE ROMAN VILLA AT DARENTI-1 Building 70, however, lay 4 ft. (1.22 m.) above building 71. If building 70 is Romano-British and not later, even if it belongs to the end of the period, the problem of the 4 ft. (1.22 m.) of soil between it and building 71 remains. Payne found a feature built of tiles in this soil which he connected with building 70, but he gives no detailed description of the soil which would allow us to say whether it was a single deliberate deposit to level up the area or represents cultivation across the west side of the outer courtyard. The gravel trackway to the east of building G had been sealed by a layer of black loam containing late Roman pottery and tile, and this is said to have been found over a wide area here by Philp. 13 It may be that land previously used for other purposes was being cultivated in the late Roman period. The layer between buildings 70 and 71 may represent cultivation but, if so, its depth suggests it should extend over a longer period than that represented by the layer discovered by Philp, and it may be that building 71 in fact belonged to a fairly early period in the history of the villa. Building 71 is isolated and no courtyard wall links it to the eastern side of the inner courtyard as building 65 is linked on the west. Building 71 is aligned on Block A and therefore should not be earlier in date. Block A was relegated to little more than an outbuilding by the construction of the inner courtyard. The life of building 71 therefore probably matched that or was shorter than that of Block A. Reference to Table 1 and the discussion of Block A below will show that this means broadly the third century. This in fact ties in very neatly with the change in character of building G early in the third century, so that building 71 can be seen as an enlarged successor to the earliest version of building G. Building 71 was given up before the 'outer courtyard wall' on the west was built, or presumably it would have been linked to the inner courtyard by an extension of the eastern wall. Finally, there is the possibility that building 71 was succeeded by building 65 since the floor of this latter building was also of rammed chalk. But if building 71 was given up by the late third century, then for much of the period when the villa was in use there was probably no 'outer courtyard' at all. Instead, the working buildings of the estate were grouped to south-west of the main house, flanking the gravel track discovered in 1969, and the whole arrangement may have resembled the situation revealed by aerial photography at North 13 Philp, op. cit., 134. 165 TABLE 1: THE DEVELOPMENT OF THE VILLA AT DARENTH Period 1. Late First Century (?) 2. Mid-late second 3. Early third century Block A An independent suite of rooms. 4. Mid-third Room 9 heated (?) century Blocks C and D C an independent range of rooms (22-4, 26, 28-9) Block E A suite of rooms A large hall fronted by constructed at the east a porticus, but aligned end and baths (Block D) on the gravelled trackat the west end of C. way or an unmetalled predecessor of it. 5. Late third/ Southern half of A Swimming bath built. Baths added and hall subdivided. 6. early fourth probably disused. Hypo- Room 18 heated (?) century causts of remaining and room 16 built. heated rooms disused (?) Other Bath building (F) probably belongs to this period. F demolished. First structure on site of building G Building G, period III. Building 71 built. Building G, period IV. Building 65 now standing and building 71 disused Building 66 built now (?) Block B built and inner courtyard formed. A building attached to the south-east of courtyard. Gateway to inner courtyard blocked by construction of rooms 42 and 43. Entrance now through room 40 and corresponding structure in southwest angle of courtyard (?). Building adjacent to room 40 demolished. Building 64 created (?). 7. Second quarter of fourth century Remodelling of baths (?) Building G, period V. 8. Second Rooms 1-4 and 6 half of sealed off. fourth century Old baths converted for Baths disused (?) other use and new baths created out of swimming pool. Building G goes out of use. Room 42 filled in. Buildings 67, 68 and 70 possibly of this period. TI-IE ROMAN VILLA AT DARENTH Leigh in Oxfordshire. 14 It was Fox in his article in Archaeologia who first pointed out that the approach to the inner courtyard through the gate in the south courtyard wall was blocked by the building of the cistern (43). This point has recently been taken up by J.T. Smith, who has postulated two entrances, without indicating where he thinks they were. 15 One may have been through building 40. In this case, the building attached to the outside of the courtyard here would have been deliberately demolished at this time. This may explain why only two short lengths of its north and west walls were found attached to the external angle of the courtyard wall. In addition there is Fox's statement that after Payne's excavations 'another large square construction (was found) at the south west corner of the west courtyard', 16 exactly matching the position of building 40. Fox does not give the size of this, and its position only appears on the plan which accompanies his article as an asterisk. He thought it might have been 'another tank' connected with fulling (the use he conjectured for room 42), but the position of the room, or building, and its description as square, makes it much more likely that it matched building 40 both in size and function. If so, these may be the pair of gatehouses that replaced the gateway that was blocked by room 43. THE RESIDENTIAL QUARTERS. We now come to a detailed consideration of the main residential villa buildings. They will be dealt with as five 'blocks' (A-E), following the system used by Payne in his report. Block A Blocks A and C were clearly originally free-standing and linked at a later date by Block B (room 15). Only a short stretch is left of the original west fas;ade wall of Block A, at its northern end. Elsewhere, it had been replaced by the east wall of Block B and the east courtyard wall. The three different widths of wall are evident on Payne's plan. This implies that when the inner courtyard was created, Block A was no longer an important building. Both Block 1 • I.A. Richmond, 'The Plans of Roman Villas in Britain', in A.L.F. Rivet, The Roman Villa in BriJain, London 1969, 60-1 and Fig. 2.4. 15 J.T. Smith, 'Villas as Key to social Structure', in Malcolm Todd (ed.), Studies in the Romano-British Villa, Leicester 1978, 154-7. 16 Fox, op. cit., 226. 167 h • hypocaust c, concrete ffoor t , tile floor T, tess.,lated floor (t) indicates floor not original E oL-_-􀁛s---􀁜•o:_ _􀁝 •s:_ _ :ro􀁞--􀁟2􀁠􀁡---3􀁢0_ _􀁣 ¥m oL-----:L--􀁤--'-----'s,.o_ __ 􀀆􀀇􀀈􀀉- 􀀊􀀋=-􀀌-- feet This gap represents about 170 feet 1 15 64 Fig. 2. Blocks A and E of the Darenth Villa. 14 A (T)h 9 I 4h (t) THE ROMAN VILLA AT DARENTH A and Block E were rather excluded from the courtyard layout than incorporated in it. A gutter lay to the east of the courtyard wall along the southern half of the tiled corridor of Block A (room 13) and led into a cistern in room 14. This surely indicates a period when the southern half of the corridor was unroofed, and the purpose of the gutter was to receive water from a roof gutter where the roof presumably still remaining over the northern half of the corridor met the roof of Block B. One would like to know the date of the pottery found below the hole burnt through the floor of room 6, apparently by continuous fires. Those responsible were presumably not the 'wayfarers' suggested in the report, but the humbler occupants of the northern half of Block A in its final state. There are some suggestions of rebuilding, such as the odd kink formed where the south wall of room 14 meets that of rooms 5 and 10. Block A does not quite lie at 90° to Block C though it is not so obviously aligned on some other feature as Block E. The arrangement of the rooms is remarkable. Payne recognised that the hypocausts in rooms 2, 3 and 4, and 6, had been put out of use at some stage and that new floors of tiles had been laid in rooms 3 and 4. The original concrete floor in room 6 seems to have survived. No floor apparently survived in room 2, but Payne thought room 1 had been levelled up to the height of the second period floors in rooms 3 and 4, so the same thing may have happened in room 2. At some stage the doorways from room 7 into rooms 2 and 6 were blocked. Although Payne recognised the position of several doorways between rooms he clearly did not find them all and his remarks show that the surviving height of the walls varied considerably. It seems that they were much better preserved in the eastern part of the building. Plate B in the report seems to show that there was no other entrance provided into room 6 after the door from room 7 was blocked. The blocking must therefore have taken place after the floor of room 6 had been burnt through by the fires lighted on it which in turn presumably belong to the time after the hypocaust was disused. Plate C shows that there was no entrance from room 9 into rooms 3 or 4 and plate D that none existed from room 8 into room 3. Plate C also seems to show that there was no doorway between rooms 4 and 5. The blocking of the door from room 7 into room 2 must therefore have taken place after the hypocausts were filled in and tile floors laid in rooms 3 and 4, since it seems to have been the only entrance into these rooms. Room 7 was unheated and had a concrete floor. It seems to have been kept in use after the other rooms had been sealed off. Such as it is, the evidence from Block A suggests that concrete floors are earlier than tiled floors. 169 E.W. BLACK Room 9 had a red tessellated floor. The hypocaust below this is extremely interesting. Next to the west wall were two rows of tile pilae. The rest of the floor was supported on dwarf walls of chalk blocks. These lay on either side of a substantial, tile-built partition running across the middle of the room and extending. to the furnace, thus allowing the whole room or either half of it to be heated. It is not known whether this was reflected in a partitioning of the room above. The hypocaust is not an original feature of room 9 since the cheeks of the furnace flues in room 10 must have been built after the hypocaust in that room went out of use. A flue communicated from room 10 to room 5 so presumably room 5 also was no longer heated after room 9 was provided with its hypocaust. The furnace for rooms 10 and 5 will have been sited in room 14. The north wall of room 9 does not seem to have survived above floor level and so a door could have existed giving access from 8. Plate LXII, Fig. 1 in Fox's article seems to show that there was no break for a door in the west wall of room 9. Payne thought there was probably a door from room 9 into room 10, and another from room 10 into room 5, since the north, south and east walls of room 5 survived high enough to show they had had no doorways through them. The floor of room 5 was of plain red tesserae. That of room 10 was thick and consisted of 'layers of roofing tiles embedded in concrete'. Concrete seems t.o have formed the floor surface to judge by Plate D in Payne's report. The original core of Block A is the two large unheated rooms (7 and 9), separated by the passage 8. This is surrounded on three sides by a series of six heated rooms. All but room 10 seem to have had pillared hypocausts: that of room 10 was a series of channels opening at 90° from a central flue. The heated rooms resemble a continuous passage, but the subdivisions presumably indicate rooms with their own specific functions. However, the main purpose of the hypocausts in these rooms was to communicate heat to the large adjacent rooms (7 and 9). In this they have a close parallel in the north wing of the Fishbourne palace where in the mid-second century the passages (N6 and N8) flanking the important room N7 were subdivided: the northern part of each was given a pillared hypocaust and the southern part used as a furnace-room. 11 It seems probable that the degree of heat produced by a pillared hypocaust was, or was thought to be, too great for a room to be used for an " B. Cunliffe, Excavations at Fishbourne, i, Oxford 1971, 162-5. 170 TIIE ROMAN VILLA AT DARENTH extended period in comfort. 18 The alteration to room 9 at Darenth provided it with an early form of channelled hypocaust, while room N 1 at Fishbourne was being fitted with a commoner, and later, variety at the end of the third century. 19 The earlier practice, where a piJlared hypocaust was sited adjacent to the room to be warmed, is referred to by Pliny the Younger. Writing in the first years of the second century he mentions only one heated room heated by a hypocaust in his Laurentine villa, besides bathrooms: it was a passage which separated a library and a bedroom.20 An oven and a pit or soakaway in room 11 indicate a kitchen, so that room 7 is likely to have been a dining-room. This is confirmed by the fact that it was room 9 and not room 7 that was given its own hypocaust, and which will therefore have been a living-room or dayroom. In several villa houses an axially-placed room that can be identified as a dining-room is unheated whereas another, equally large room, often in one wing of the house and often, like the dining-room, floored with mosaic, is provided with a hypocaust.21 We can hazard a guess that bedrooms will have been the rooms furthest from the furnaces (room 3 and perhaps room 5). The roofing arrangements for Blor-1,: A are of some interest. The presence of hypocausts in the rooms on the north, south and east sides of the three central rooms, 7, 8 and 9, shows that on these sides at least we are not dealing with open verandahs, but with solid walled rooms. 22 Even if room 13, or rooms 12-14 were open, it seems likely that rooms 7-9 must have risen higher than the other rooms so that adequate lighting could be provided. Block B This consists of only one room, 15. It was clearly built to link the freestanding Blocks A and C when the inner courtyard was formed. Room 15 had walls 2 ft. 11 in. (0.89 m.) thick and there was a great 'buttress' outside the north wall placed opposite the central entrance 18 The smaller rooms with such hypocausts are comparable to the caldaria of bathsuites and the heat presumably was too great. Larger living-rooms with pillared hypocausts are found, e.g. in the villa at Folkestone (S.E. Winbolt, Roman Folkestone, London 1925, 53-4 and PI. X). 19 Cunliffe, op. cit., 171-2. 20 Epistulae, ii, 17.9. 21 Folkestone Block A is a good example. A series of fourth-century houses shows the same arrangement, e.g. Bramdean, Chilgrove 1, Downton. 22 The cill of a window was found in the north wall of room 6, 3 ft. 3 in. above the floor (Payne, op. cit., 54). 171 h , hypocaust c s concrete floor cm :cement floor t.tile floor T, tessellated floor 32 C 0 1p ¥> 5 10 1􀀑 2p 2􀀅 C 27 29 28 26 24 T T C 25 T 30 39 42 Fig. 3. Blocks B, C and D of the Darenth Villa. 120 feet 3p 3; metres N I 13 tr1 􀀁 THE ROMAN VILLA AT DARENTH in the south wall. The room had fine painted wall-plaster and a concrete floor and was evidently monumental in character. The 'buttress' may have been a podium of some sort. Traces of fires were found on the floor of the room but these presumably belonged to the final stages of the villa's occupation. Unfortunately, we do not know what function room 15 had. Its size and lack of communication with other rooms suggest that it was communal or public in character. Block C The original nucleus of Block C seems to have been rooms 22-4, 26, and 28-9, where the partition walls were of timber, set on a foundation of flint at floor level, and plastered. 23 The width of rooms 30, 31 and 34 on the west matches that of rooms 20 and 21 on the east, and these two groups of rooms, both projecting to the south, were clearly designed to balance one another. It is possible that large wing-rooms, the size of room 20, formed part of the original plan and that the main range of rooms extended further both to the east and the west, but if so most of this has been removed by later alterations. It is more likely that the original building resembled the first-century houses at Eccles and Farningham II, and that it lacked a winged fa<;ade. 24 Although at the time the inner courtyard was formed room 26 was evidently the dining-room, in the original house this may have been situated in room 29. Its position at the end of a simple range of rooms and separated from the others by a passage, 28, is reminiscent of similar rooms in other villas that are thought to have functioned as triclinia. 25 Room 22 had apparently been reduced in size by the construction of room 18. As it appears on the plan it is simply a passage which could serve no purpose at the end of the range from room 22 to room 29. 26 It seems most likely that room 17, heated by a pillared hypocaust, was similar to rooms 2-6 and 10 in Block A. It has a very close analogy in room 6 in the villa at Ashtead in Surrey and, u Payne, op. cit., 59. 24 Eccles: A.P. Detsicas, 'Excavations at Eccles, 1970', Arch. Cant., lxxxvi (1971), 32-3; Farningham: G.W. Meates, 'Farningham Roman Villa II', Arch. Cant., lxxxviii (1973), 3. 25 I.A. Richmond, Roman Britain, Harmondsworth 1963, I 13. 2• There are, however, houses where passage-like rooms closing a range of rooms were possibly stairwells. Bramdean in Hampshire is one example and the first house at Eccles may be another. 173 E.W. BLACK like it, was presumably designed to warm all three of the adjacent rooms. 27 It was also presumably itself a room with a set function, requiring a minimum area, otherwise it would not have projected beyond the east end of the block. As with room 9 in Block A, the provision of a channelled hypocaust in room 18, drawing its heat from room 17, will have been a later development. This provision indicates that room 18 was a living-room or day-room, and room 19, opening from it, could well have been a bedroom. If room 20 was then a dining-room, the provision of rooms at the east end of Block C matches that in Block A, though on a smaller scale. Room 22 presumably gave access to room 18. The original provision of pillared hypocausts to heat adjacent rooms in Block A and the east end of Block C, and the later channelled hypocausts below a living-room in both blocks suggest that these buildings are contemporary, as well as being functionally identical. The east wall of the furnace-room for room 17 (room 16) must be Jater, however, since it is clearly contemporary with or later than the construction of room 15, and this is later than, or more probably contemporary with, the disuse of the southern part of Block A. The whole east end of Block C was given a second storey, to judge by the buttresses added to the north wall of room 18 and the south wall of room 20, but what rooms it may have contained cannot be known, nor is it clear where the stairwell was sited. Rooms 26 and 29 had red tessellated floors and rooms 23 and 24 concrete floors. These rooms as well as the passages, 22 and 28, all produced painted wall-plaster, indicating their domestic character. Room 30 looks very much as if it was formed out of an out-turned length of corridor similar to room 21. On its south side was a stock of unused red tesserae, but it is not clear if they were intended for reflooring this room. In the suite at the east end, room 20 has a red tessellated floor, and room 18 one of concrete and it produced painted wall-plaster. Room 17 had had a floor surface of red cement and the floor of room 19 was of tiles. Block D Most of Block D is, in fact, the western end of Block C, but it was certainly distinct in terms of function and comprised a bath suite. If room 31 was a corridor, it is difficult to see what function room 30 27 For a plan of Ashtead, see Surrey Arch. Coll., xxxviii (1930), opp. 148. I hope to publish a full re-assessment of this villa. 174 THE ROMAN VILLA AT DARENTII might have had in relation to the baths, unless it was the apodyterium. There seemed to Payne to have been a door between rooms 31 and 29 which might suggest that when Block D was built there was no corridor 25 in existence. Room 32 served as a furnaceroom for the hypocausts in rooms 33 and 35. Its east side contained a roughly built 'hypocaust', with low walls of tiles. It is not possible to be certain of the plan of this, but it seems to have been like an upturned L and to have had comparatively wide flues. One is tempted to think of a later corn-drying oven inserted where the woodstore of the baths hypocaust had been. Room 34 seems to have been the cold room in the original baths with a cold plunge-bath attached (floored with red tesserae and with opus signinum walls - it was 1 ft. 5 in. (0.43 m.) deep). The plunge-bath was filled in and the wall dividing rooms 34 and 36 built across it. 28 The more modest plunge-bath was evidently done away with in favour of room 36, 39 ft. 6 in. (12 m.) long by 9 ft. 11 in. (3 m.) wide, and 4 ft. (1.22 m.) deep, a swimming pool. The caldarium must have been the part of room 33 nearest the furnace and the tepularium the part adj􀂇cent to room 34. The hypocaust below room 33 is in three sections. Nearest the furnace the upper floor was supported on pilae, and on the east side of the room these rested on a chalk platform that rose upwards towards the east wall of the room. Then came two rows of large tiles taken to represent an extra support for the floor of the room above. The next section contained a slanting chalk platform on both the east and the west, but no pilae were found set on them. A flint-andtile wall divided this section from the third where the floor was supported on tile pilae. This flint-and-tile wall probably reflects the division of the room above into caldarium and tepidarium. Fox says that there were slight remains of a platform on the west side of the section nearest the furnace, 29 and if this was so, then the two parts of the caldarium were provided with these platforms but the tepidarium was not. In his report Payne quoted the opinion of a manager of a cement factory that these slanting platforms might have served to influence the draught through the hypocaust.30 George Fox, however, noting that the pilae of the hypocaust had been built over the platform in the first section, took this to indicate that the hypocaust was not an original feature of room 33. He saw the platforms surviving from an 2 • Payne, op. cit., 61. 29 Fox, op. cit., 222. 30 Payne, op. cit., 61. 175 E.W. BLACK earlier use of the rooms in Block D connected with fulling,31 but similar ramps or platforms have now been found in hypocausts at Lullingstone and Little Chart. 32 Whether their purpose was to protect the walls from heat or control the draught or 'to block a functionally useless part of the hypocaust which would otherwise serve only for the accumulation of relatively cold air', 33 it is clear that these ramps are a feature of hypocaust construction, and that there was no pre-baths period in Block D of the Darenth villa. In any case, Fox's suggestion that room 36 was a vast rinsing tank connected with fulling before the block was converted to a bath34 ignores the fact that room 36 replaced a cold plunge-bath. It is interesting that the tessellated floor of the original cold room was retained when the cold plunge bath was filled in and room 36 constructed, but that the part of the floor above the filled-in bath was laid with tiles. This must confirm to some extent the impression from Block A that tiles were used as flooring rather than concrete or red tessellated paving at a later stage in the villa's life. Room 35 apparently received hot air from room 33 at the north end through their common north-south wall, while there may have been another passage at the south end, communicating with the tepidarium. 35 The exact extent of the hypocaust in room 35 is not clear, however. It is possible that it occupied only the northern part of the room. Some sort of wall existed in line with the wall dividing the caldarium and tepidarium in room 33; this is marked in outline on Payne's plan and is clearly visible on plate G of his report. Another partition perhaps existed where the west wall of room 33 terminated. All the pilae and vertical box flue-tiles which can be specifically located from Payne's description were found in the area of room 35 north of the termination of the west wall of room 33. Clearly, the ramp that originally existed on the west side of the first section of room 33 was largely destroyed when the hypocaust was extended westwards to form part of room 35. That the same thing did not happen on the west side of the second section of the hypocaust under room 33 suggests strongly that only the northern third of room 35 was heated. The large number of vertical box flue-tiles " Fox, op. cit., 223. •12 Lullingstone: G .W. Meates et al., 'The Lullingstone Roman Villa', Arch. Cant., !xiii ( 1950), 9, and lxv ( 1952), 36, 38; Little Chart: J. Eames, 'A Roman Bath-house at Little Chart, Kent', Arch. Cant., lxxi (1957), 130-46 . . u Eames, op. cit., 136. l• Fox, op. cit., 225. 3 s Payne, op. cit., 62. 176 THE ROMAN VILLA AT DARENTH found in situ there (three in the north wall and nine in the west wall) and the drain in the west wall above the level of the box flue-tiles suggest that this extension may have been a bot plunge-bath, something that is otherwise missing from the bath-suite. The tiles found in line with the end of the west wall of room 33, crossing room 35, may not have been part of a floor as Payne thought, but the base of the south wall of this hot plunge-bath. The addition of room 35 to the suite presumably belongs with the creation of room 36. At a later date, the hypocausts in rooms 32, 33 and 35 were filled in and a floor of coarse white cement laid over the top. At the same time, presumably, a less elaborate bath-suite was constructed out of room 36, the swimming pool. A wall was built across room 36 dividing it into two and rooms 37 and 38 were added to the north. Since room 37 was a cold plunge-bath, presumably the eastern half of room 36 now became a cold room. Room 38 was a hot bath. Its hypocaust had a flue on either side of a central wall, running the length of the room, reminiscent of the construction of the hypocaust below room 9 in Block A. Payne thought that room 38 was stoked from the western half of room 36 and that the two roughly-made walls projecting east from the east wall of room 36 formed an entrance to the furnace room. 36 It would be more acceptable if, as Fox postulated, 37 the western end of room 36 was also heated by a hypocaust at this stage and the two rough walls formed the stokehole for this. A quantity of burnt debris from hypocaust firing was found in room 49 in Block E and this could perhaps have come from a stokehole in room 46. Block E In its final state, Block E was clearly used as a store area and workplace, red tile tesserae, clay and pounded tile being found in various rooms. The stokehole of the hypocaust serving the caldarium in the western half of the converted room 36 impinged on Block E, and the building was presumably of no importance even earlier when the swimming pool was constructed abutting it. It was suggested above that Block E was aligned on the trackway found east of building G. The west wall of room 64 which joins building 65 and Block E is contemporary with or later than the addition of a set of baths to Block E since it changes course halfway along its length and passes west of the west wall of the block to join the south wall of room 61 36 Ibid., 64. 37 Fox, op. cit., 229. 177 E.W. BLACK of these baths. The south wall of room 50 was then continued west to meet it and enclose a new, wedge-shaped room 63. North of Block E the new alignment of the west wall of room 64 is maintained from the north-west comer of room 55, though it seems to have moved slightly west. In the bath-suite, rooms 56-8 seem to have been a set of cold baths, originally floored with tesserae and later with tiles. A large part of room 54 was occupied by a foundation of rammed chalk with which was associated a fine drain which led through rooms 51 and 52 and which is shown on Plate J of Payne's report. The chalk foundation and the drain must have been parts of a cold douche in the frigidarium of the baths. 38 A lobby separated the frigidarium from room 53 to the north. In the north wall of room 53 there was 'an opening 1 ft. 5 ins. wide, faced with tiles laid in courses on each side'. 39 This was presumably a stokehole and room 53 was a laconicum. 40 The tepidarium, caldarium, and hot bath, if one existed, will all have occupied the southern part of room 54, where the bases of five pilae were found, and room 61, with a furnace somewhere at its south end. If sited beyond the south wall of room 61, the construction of room 63 must have put it out of commission, but it may well have been further to the north. The west wall of rooms 52 and 53 and the three cold baths are on a n alignment slightly different from the other walls which follow the alignments of Block E. This may mean that the cold baths belong to a period of reconstruction. There is another indication of this in the wall constructed within room 53 west of the stokehole and the drain which flows west from here and through room 55. The western part of room 53 may have become a latrine, perhaps replacing one that had existed earlier in room 51, flushed by the water from the cold douche. The two short lengths of wall within room 61 may belong to this secondary period and have formed a new tepidarium and caldarium. If these suggestions are correct, it is evident that in the earlier period the baths were on a grander scale and offered greater facilities. The only element missing is a cold bath. It is possible that the douche was a substitute for such a bath but this seems unlikely. One may have existed where the set of three baths was later built to the west, or it may have been in the south-east corner of room 54. '" Compare the similar feature in the cold room of the Angmering bath-house (Sussex Arch. Coll., lxxix (1938), plan opp. p. 6). " Payne, op. cir., 66. •° Fox states that 'the bases of the wall flues' remained visible (Fox, op. cit., 220). 178 I r -=- . 􀀁 , .. 300 42:0 􀀆· ,.. • - I - ' ·􀀅 ISO - I - 600 .... 2Po #11'1.tfU • I 􀀂 - - ' ■ - = F .. - - S 6 7 • 8 Fig. 4. The Development of the Darenth Villa: The Numbers refer to the Periods in Table I THE ROMAN VILLA AT DARENTII The southern wall of the room here is 3 ft. (0.91 m.) or more in width and there is a short parallel stub of wall leaving the east wall which could mark the other side of a bath. Apart from those in the plunge-baths no other floors of the rooms in the bath-suite survived. The only remaining traces of domestic comfort in the block were a possible tiled floor in room 59 and painted wall-plaster from room 63. There were no living-rooms with hypocausts. The block undoubtedly began as a rectangular building comprising what appears on the plan as rooms 45-50, 54, 59-60 and 62. The irregular walls around room 46 are certainly connected with the use of this area as a furnace room for the late bath-suite formed out of the swimming pool to the east, and it is likely that some of the other walls are not original. The more regular north and south walls of room 46 were presumably built when the swimming-pool was first constructed. There is a foundation shown on both sides of the wall between rooms 49 and 60. If it is drawn to scale it measures some 5 ft. sq. (1.52 m2). It is described as the foundation of a pier and is said to have had tufa quoins and to have been otherwise faced with tile. It lies not quite centrally between the north and south walls of the building, some 31 ft. (9.45 m.) from the former and 36 ft. (I I m.) from the latter, and was presumably structural.41 The wall into which it was incorporated may not have been part of the original building, which may have consisted simply of a vast hall with a porticus on the east side. The date of the internal sub-division of the hall is not known, though the fact that the eastern limit of room 54 coincides with a major north-south division wall suggests it may be contemporary with the construction of the bath-suite. In the final scheme, the large central room ( 49) must have retained some of the functions of the original workhall. 42 The Development of the Villa Plan The various developments suggested to have taken place here are illustrated in Fig. 4 and are set down in Table 1, grouped into periods. This does not imply the exact contemporaneity of the items under each period but the groupings do satisfy a scheme of logical and connected development for the buildings of the villa using the •• A row of posts occupied a similar position towards the rear of the hall in two German hall-type villas discussed by S. Applebaum (in H.P.R. Finberg (ed.), The agrarian History of England and Wales, i, 1972, 127-8). 42 J.T. Smith, op. cit., 157. 179 E.W. BLACK evidence available and set out above. Unfortunately, none of these structural periods is dated, except by association with a dated period in the life of building G. The 53 coins recovered from the 1894-95 excavations ranged from Domitian t o Gratian. Continuous occupation from late-first/early-second century to late-fourth century is probably indicated. There is also an early Saxon presence on the site which is not discussed here.43 If any reliance can be placed on analogy with other villas, the addition of the bath-suite to Block C might be dated to the second century. 44 The earliest nucleus of Block C would not be out of place in the later first century. Slightly more faith may be put in an analogy between the swimming pools here and at Eccles. At Eccles, the pool (Room 17) seems to belong to the late third/early fourth century.45 Emulation by the owner of one or other of these not too distant houses might seem reasonable. If this is correct it will be seen from the table ·that the building of Block A probably belongs to the later seconrl or third century. Periods III and IV of building G also belong tn ·.he lhird century, so that this was evidently a time of great prosperity for the estate. The relationship between Blocks A, C and E, and its social implications, are discussed in detail in the next section. Despite the apparent attempt to harmonize the various buildings into an integrated scheme around the inner courtyard in the late third or early fourth century, there is little sign of the wealth one would expect in a fourth-century villa: mosaic floors are not present. 46 The red tessellated floors do not, of course, imply a lack of wealth in a pre-fourth century villa, since mosaics in the domestic rooms of villas are comparatively rare before the fourth century. But at Darenth in the fourth century tessellated floors are replaced by tiled floors. • 3 B. and E. Philp, 'Archaeological Excavations in the Darent Valley', (1973), 7. More early Saxon structures have been found at the Keston villa in west Kent, and the question of 'continuity' between the latest Romano-British and earliest Saxon occupations of these sites will require careful consideration when the evidence is published. 44 Detached bath-houses are certainly not an exclusive first-century phenomenon, but almost all first-century bath-houses that are known were detached. Secondcentury bath-suites are known attached to houses, e.g. Newport, l.o.W., and Lullingstone, Kent. •s A.P. Detsicas, 'Excavations at Eccles, 1973', Arch. Cant., lxxxix (1973), 127-8. 46 Fox, op. cit., 219, refers to the finding of a mosaic pavement, but this is probably the same discovery as that mentioned by Payne (op. cit., 49) who talks of a tessellated pavement. 180 THE ROMAN VILLA AT DARENTH Later in the fourth century the pool ( 42) in the courtyard was filled in and the house may have again been approached through the central gate in the south wall. Perhaps buildings 67, 68 and 70 were laid out at this time. The swimming pool had been converted to provide a new suite of baths and the baths attached to Block E may have gone out of use. Evidence of the lighting of repeated fires was found on the floors of room 6 and room 15, and this implies a decline in the level of civilisation of the occupants of these rooms of the villa, although the sealing of room 6 and therefore the continued occupation of room 7 must belong to a later date. The latest coins recorded are two of Gratian and the latest fourthcentury issues to reach Britain are absent from the list. This presumably implies that the villa had ceased to participate in the money economy of the province before the end of the fourth century and that perhaps it was deserted. 47 The Functional Elements and Social Organisation of the Villa In period 1 Block C clearly formed a house, and there was a detached bath-building (F), a considerable distance away. Apart from the possibility that room 29 was a dining-room, no other rooms can be assigned particular functions. Probably in the mid-second century, to judge by the analogies of Fishbourne and Ashtead, Block A was built. The functional similarity between 17 in the east end of Block C and rooms 2-6 and 10 in Block A and the succession of first unheated and later heated living-room in each block (rooms 18 and 9 respectively) make it clear that the two blocks were occupied simultaneously. Both Block A and the east end of Block C were suites of domestic rooms, comprising a living-room, probable dining-room (certain in Block A because of the kitchen), and other rooms of uncertain use, but probably including bedrooms. The baths at the west end of Block C balance the east end and, if the Lullingstone baths are an analogy, could be the same date. They are rather more conveniently placed for the use of the occupants of Block C than those of Block A. Block E could belong to the same time, though no earlier to judge by its alignment. In its original state it was probably an undivided work-hall fronted by a porticus. What 47 The significance of the absence of late coinage is uncertain. It was reaching some villas, and so presumably a money economy was still operating in the countryside. Dr. G. Webster has suggested that in the case of some villas a sort of self-sufficiency obtained, without the need for money ('The future of Villa Studies', in A.L.F. Rivet (ed.), The Roman Villa in Britain, London 1969, 221-37). 18 l E.W. BLACK are we to make of these period 2 buildings in terms of the villa's social organisation? Philp has apparently referred to Block A as a guest-house. 48 However, the scale of the rooms in Block A is grander than what must then have been the owner's suite in Block C. J.T. Smith has suggested that Blocks A, C, and E were each an independent house, and that in effect there were three 'joint-owners' of the villa.49 While this has a certain attraction if you look at a small scale plan of the villa (such as Fig. 1), the distribution of functional elements makes it most unlikely. Looked at functionally, the three blocks in period 2 comprise: A, domestic suite; C, domestic suite; the older house (use in this period uncertain, perhaps accommodation for servants/ workers); bath-suite (Block D); E, large work-hall. We have seen, however, that Block A is on a grander scale than the domestic suite in Block C. I suggest that Block A was the house occupied by the owner of the villa when he was present, but that he was usually living elsewhere. The east end of Block C would then have been occupied by a vilicus or manager. The servants could have occupied the older portion of Block C, and the siting of the baths at the west end of this block would be because the owner only required to use them occasionally, on his periodic visits, but they were in regular use by the vilicus and workers on the estate. Block E was a workhall and nothing more. The creation of the inner courtyard sees a drastic change in the buildings of the viUa. The southern part of Block A was disused; the older portion of C was again the focus of the villa, with room 26 becoming a dining-room. Presumably, a new owner was now permanently resident in the villa and had taken over the vilicus' quarters. The new owner reserved the baths in Block D for his own use and a new suite for his workers was built onto Block E. At the same time Block E was divided up to provide them with living accommodation now that they had been displaced from the older part of Block C. A swimming-pool was constructed against Block E on the east, emphasising its subordinate character. This is also brought out by the absence of heated living rooms from the block, for even if Block E, which resembles a 'Hall type villa' in this period, represented a different social unit from the type which had earlier occupied Block A and was now occupying Block C, this would not affect the provision of heated rooms. This was a matter of 48 Philp, op. cit., 119. 49 J.T. Smith, op. cit., 154-7, 174. 182 THE ROMAN VILLA AT DARENTII comort, not of social organisation. There is one more functional element within the courtyard which requires some comment. This is Block B (room 15). It is not a workhall, but something public and formal. Perhaps the best parallel is provided by room 10 at Whittington. so This was floored with a fine geometric mosaic and was linked to the rest of the house, or more particularly to the kitchen, by two mosaic-floored corridors. It must have been used for feasts, but not just by the occupants of the house, for corridor 11 had doors in both walls by which people could enter the hall without going through the house. Room 15 at Darenth has similar arrangements for access, and may have had a similar function. The guests of the owner may have been his workers and dependents. ACKNOWLEDGEMENTS I should like to express my thanks to Mrs. Jennifer Steel, for preparing the plans which accompany this article, and to Prof. A.L.F. Rivet, for his help and encouragement in the field of villa studies. My debt to the writings of Mr. J.T. Smith will be obvious to readers who share his conviction that villa plans can be made to yield information about their occupants' social relationships as well as their material circumstances. so H.E. O'Neil, 'Whittington Roman Villa, Glos.', Trans. Bristol and Gloucestershire Arch. Soc., boci (1953), plan opp. 24. 183