The Mount Roman villa, Maidstone

THE MOUNT ROMAN VILLA, MAIDSTONE D.B. KELLY The Mount villa (N.G.R. TQ 757563) is on the east bank of the River Medway in the north part of Maidstone, about 200 yards north-west of Maidstone East Station (Fig. 1). The northern part, the main subject of this report, lies within the old Cavalry Barracks, but the greater part of the building is south of the barracks wall, covered by a mound of earth up to 20 ft. in height on which a thicket of mature trees grows. The villa was built on the Weald clay, exposed by the River Medway cutting down as a result of successive falls in sea level. 1 Its south-west side looks across the river to what is now a semi-derelict industrial area, but would have been a pleasing prospect of the river and rising ground beyond, a factor in the choice of site as well as the convenience of the river. Three hundred yards to the east the Roman road from Rochester to the vicinity of Hastings passed the villa,2 so that access was good both by road and river. A mile to the south along this road was another Roman villa3 and beyond this the ragstone quarries at Boughton Monchelsea. Upstream, about three miles to the south-west, were two villas or farms at Barming and another at Teston.4 Downstream there was a building, perhaps a villa, at Allington5 and, three and a half miles distant, the great villa at Eccles6 on the right bank and another villa at Snodland on the left. 7 1 B.C. Worssam, Geology of the Country around Maidstone (Memoirs of the Geological Survey of Great Britain, H.M.S.O. 1963), 24, 115. 2 I.A. Margary, Roman Ways in the Weald (3rd. edn., London, 1965), 214-6. 3 Arch. Cant., x (1876), 163-172. 4 V.C.H. Kent, iii (1932), 104, 125 and refs. therein. 5 Ibid., 103. 6 A.P. Detsicas, The Cantiaci (Gloucester, 1983), 120-6; interim reports in Arch. Cant., lxxvii (1963) - xciii (1977), and Arch. Cant., cvii (1989), 83-8. 7 Arch. Cant., lxxxii (1967), 192-217, and forthcoming. 177 D.B. KELLY HISTORY OF THE SITE The Mount villa was discovered in 1843 when part of the river bank above the tow-path collapsed, revealing masonry. The adjacent garden to the east was excavated by C.T. Smythe in the same year, revealing the south wing of the villa, 8 but a newly planted orchard to the north prevented further work. This orchard was bounded on the north by the wall of the old Cavalry Barracks, built in 1797. Immediately inside the wall was the officers', later known as the commandant's, garden, separated from the large Barrack Field to its north by a hedge or row of trees. These features are shown on Daniel Alexander's map (for the new Maidstone Gaol) of 1810, Tootell's map of Maidstone (1848) and the 25 inch O.S. map of 1865. With the exception of a Tudor belt or dress hook nothing was found in the excavation, confined to the north of the barracks wall, that dated from the centuries after the Roman period to the seventeenth century. Immediately to the north of the barracks wall were three pits, dateable to the second half of the seventeenth and eighteenth centuries, used, to judge by the paucity of finds, for night-soil. Their position suggests that the boundary marked by the barracks wall was in existence by the seventeenth century. This boundary falls roughly along the north wall of what is here called the main block of the villa, although it diverges slightly to the north (Fig. 3), and it is possible that the visible remains of the villa were used to mark an early boundary here. The next boundary to the north, the row of trees, also falls roughly along the line of the buildings on the north side of the villa yard. Maidstone East Station was opened in 1874, with the arrival of the London, Chatham and South-East Railway, and the site of the villa became the property of the railway company, which built sidings over the eastern part of the garden and orchard, though not over the villa. In 1884, the line was extended east to Ashford and the spoil from the tunnel adjacent to the station and the cutting beyond was perhaps the source of the mound of earth which now covers the main part of the villa to the south of the barrack wall. In 1970, there was a proposal to build a telephone exchange near the site and the Maidstone Area Archaeological Group, directed by our member A. Miles, excavated by the tow-path adjacent to the part ofthe villa uncovered in 1843. Although the site had been disturbed by the laying of electric cables and a water main and damaged by a war-time bomb, the footings of the west range were found (Fig. 2), of 8 J.B.A.A., ii (1847), 86-8. i78 (/) ro w THE MOUNT ROMAN VILLA, MAIDSTONE (/) -"' (.) ro Cl) CJ) l C: 􀀊 C: ro z Cl) 179 o.i i:: 8 ·a :::E ..ci" 􀀇 E ;:l 0 :::E 􀀁 8 § ffi D.B. KELLY coarse wares, plain samian, a single sherd of figured samian and part of a colour-coated beaker. Only one identifiable coin was found, an AE of Gordian III. The plan gives the width of the walls: the outside south wall and the internal east-west wall next to it on the north were 2 ft. (0.60 m.) wide and all the others, including the outside wall to the east, 1 ft. 4 in. (0.40 m.). The footings of the west range uncovered in 1970 and 1978 were wider than this and the two standing walls observed in 1978 were 0.70 m. (2 ft. 4 in.) and 0.50 m. (1 ft. 8 in.) wide. The excavation by A. Miles in 1970 uncovered most of the west corridor, including the part of its east wall found in 1843. The site, next to the tow-path, had been disturbed by electricity cables, the water main and a war-time bomb, which had removed a section of the east wall. The east wall of the corridor was traced almost to the barracks wall, but disturbance had removed even the footings of the west wall for about 20 ft. south of the barracks wall. No trace was found of the buttresses shown on the 1843 plan and the south-west corner had completely disappeared. The east and north walls of room J, in the south-west corner, met the other walls as butt joints, but as these were footings they cannot necessarily be regarded as later additions. No stratified finds were found. The construction of a main sewer in 1978 necessitated the cutting back of a small part of the mound covering the main building in the area immediately south of the barracks wall next to the tow-path. The west range was here again reduced to footings, but a few stretches additional to those found in 1970 were uncovered and also a longer stretch of the narrow wall running east from it. To the north of this wall, the south wall of room L, another wall was uncovered, crossing the west range and continuing eastward. Where, as the south wall of room M, it joined the east wall of the west corridor, the footings, to about a foot in height, were butted against those of the north-south wall, but at the highest surviving level both walls were of one build. Presumably this was done for convenience in setting out the foundations and suggests that butt joints at the lowest level do not in themselves prove more than one building phase. This wall (Fig. 4, A-B) was 0.70 m. (2 ft. 4 in.) wide, of roughly coursed ragstone with yellow mortar, the bottom course of the footing set in clay. It was built at almost the same width as the footings. The part covered by the mound stood, measuring from the bottom of the footings, to a height of 1.80 m. (6 ft.). On the south, in room L, the layer above the natural clay was of a dirty brown clay, containing a few tiny fragments of unidentifiable pottery, and capped with ragstone cobbling. Above this was a second brown clay layer. These may be the make-up layers of successive floors, though the 186 THE MOUNT ROMAN VILLA, MAIDSTONE limited area of the cut precludes any certainty. In the north-east angle of room/corridor K, in a layer of yellow mortar above the natural clay, taken to be a construction layer, were found sherds from five pots, the only dating evidence for phase 1, or, indeed, any other phase, found in the main building. One of these pots, a grooved dog dish in a north Kent sandy reduced fabric (Fig. 13, 1) is dateable at the earliest to about A.D. 130/140 and lasted well into the third century.13 The others include storage jars of grogged ware and of Patch Grove ware with a double line of finger-tip decoration on the belly, a type lasting from the late first until the early third century. 14 In 1976, in the course of the excavation within the barracks, a large buttress was found at the south-east corner of the site, aligned north-south and continuing under the barracks wall (Fig. 3). Immediately to its west and adjacent to the phase 2 return wall, running west from the south end of the east wall of the yard, the edge of a wall running east-west projected from under the barracks wall and this is taken to be the north wall of the main building. East of the buttress was a substantial piece of masonry aligned east-west and mostly destroyed by a seventeenth-century pit, which had been cut through a thick mortary layer surrounding the masonry. Another seventeenth-century pit was found to the east of this, but further trenching along the line of the barracks wall revealed no more walling. The masonry is probably a second buttress and with the first buttress found marks the north-east corner of the main building, corresponding to the buttresses of the south-east corner. The presumed line of the north wall of the main building is shown on the plan (Fig. 2) by a broken line. This wall and the two east-west walls observed in 1978 correspond to the east-west walls of the south range - outer wall, wide inner wall, narrow inner wall- and suggest at least a degree of symmetry in the building, ranges of rooms surrounding an inner courtyard. No hypocaust was found in 1843, but two in the modern excavations (rooms 2 and 5). Unless there was a separate and so far undiscovered bath building or bath rooms in the unexcavated north or east range, the baths are likely to have occupied the north-west corner of the villa, now almost completely removed, and included the rooms (1-5), now inside the barracks wall, of phases 1 and 3. This quarter is nearest to the river and, given the natural slope of the clay towards the river, would necessitate less work in preparing the ground for hypocaust rooms and furnace. Moreover, its walls - 13 J. Monaghan, Upchurch and Thameside Roman Pottery, B.A.R. Brit. series 173 (1987), 150. (Hereafter Monaghan 1987). Type 5Fl. 14 R.J. Pollard, The Roman Pottery of Kent (Maidstone, 1988), fig. 13, 21. (Hereafter Pollard 1988). 187 D.B. KELLY those of rooms 1-5 and the north wall of room L - are more substantial and survive to a greater height than any others. Like Thomas Charles, another local antiquary, the Rev. Beale Poste, was impressed by the 'preposterous size and thickness of the buttresses' and in his paper on Roman Maidstone in the first volume of our journal 15 allows himself some speculation on the matter. The reason for their building was almost certainly the unreliable nature of the Weald clay, especially when sloping towards the river. It cannot be shown at present whether or not these buttresses were planned from the beginning or were a later addition. If they are not part of a later re-building the predominance of the narrower gauge walls found in 1843, including the stretch of the external wall on the east side, may be due to the need to reduce the weight of the building and the buttresses may have been added after only some walls had been built at the greater width at the start of the construction. In two instances, at least, the footings of the narrow gauge walls - those between rooms E and H and the south wall of room L - were laid at the same time as those of the east wall of the west corridor. The 1843 plan shows that part of the villa in its final stage and without distinction of building phases. Although some of the narrow gauge walls were present from the start, the herring-bone masonry of some internal walls, not found elsewhere on the site, may indicate a partial rebuild or internal re-arrangement of rooms, as do the possible successive floor layers of room L. The 1970 excavation by A. Miles showed that the north and east walls of room J butted against the others, so this may be a later division, though these walls survived only as footings and are thus not necessarily later than phase 1. If rooms 1-4 in the north part of the building are taken as part of the original build, then the limited dating evidence from them for phase 1 is consistent with that from room K and points to a first building phase in the second half of the second century. The rarity of pottery dateable before the middle of the second century is remarked below. THE NORTHERN AREA Phase 1 The earliest building in the northern part of the villa is a range of four rooms (1-4) running north-south, roughly parallel with the river (Fig. 3). A modern water main had been laid in the natural clay 15 Arch. Cant., i (1858), 171-2. 188 THE MOUNT ROMAN VILLA, MAIDSTONE -----·-,,,, ,. ,l· .. I tu . .. 􀀏 r 0 ·, ,_ ____ ,, • N;: 􀀐= f-------=--=-􀀌􀀍􀀎cilr,i :::=.. .·) ·: :!:• _::11,􀀏il'_:_􀀐---􀀑 1------tl:d'C'f"""--i <;: t------of 􀀇 w------􀀋 -- • _______ I J 189 • 0 a, I "" i 0 "' '£ 􀀁 D.B. KELLY below the level of the villa foundations, running down the centre of the range and thus destroying the central parts of the internal east-west walls. The east wall of the range survived or was traceable for its entire length, but almost two-thirds of the west wall had been destroyed by a modern (? nineteenth century) brick terrace or foundation, including the whole west wall of room 1. These interruptions and the refurbishment of phase 3 had removed nearly all internal features and with them most of the dating evidence. Room 4 had no wall on its west side and was presumably a porch. The walls were of ragstone with a yellow mortar, very substantial and set into the natural clay, which slopes downwards to the west and north, at depths ranging from 0. 70 m. (2 ft. 4 in.) at the south end to 0.30 m. (1 ft.) at the north. The lowest course was unmortared and packed with clay. At the south end of the range the east wall survived to as much as 1.80 m. (6 ft.) from the lowest course (Fig. 5, K-L), but was reduced to a single course at its north end, north of the line of trees. The eastern north-south wall averaged 0. 75 m. (2 ft. 6 in.) in width, the western, in the short remaining length uncovered, 0.90 m. (3 ft.). The internal east-west walls were 1 m. (3 ft. 3 in.) wide. The internal lengths of rooms 2, 3 and 4 were, respectively, 7 m. (23 ft.), 3.10 m. (10 ft. 2 in.) and 3.80 m. (12 ft. 6 in.) and the internal width of room 3 and the north part of room 2 was 3.20 m. (10 ft. 6 in.). Room 2 was provided with a floor of white mortar, some 80 cm. (3 in.) thick, which lay beneath the heavy rubble make-up for the phase 3 hypocaust floor, and room 3 had a clay floor. Material from the phase 2 construction layers between the east wall of rooms 2 and 3 and the west wall of the phase 2 yard included window glass, T-shaped clamps, box-tiles and wall-plaster, not necessarily from rooms 1-4, but belonging to phase 1. The only certain evidence providing a date for phase 1 comes from a burot layer in room 3, immediately above the natural clay and below the make-up for the phase 3 floor, and this provides a terminus ante quern for the construction (T18/7 - Fig. 4, C-D). Among the sherds of half-a-dozen fabrics only three forms were recognisable: a samian Form 18/31 or Form 31 dish, a dog dish of Gillam type 328 (Fig. 13, 2) and a flange in a fine micaceous ware from what is probably a copy of a samian Form 38 bowl. Gillam 328 can be as early as A.D. 120-130 in the south, though lasting until the end of the century, 16 but the samian Form 38 copy belongs to the second half of 16 G. Marsh and P. Tyers, 'The Roman Pottery from Southwark,' in Southwark Excavations 1972-1974 (Joint Publication No. 1 of London and Middx. A.S. and Surrey A.S., 1978), 533 ff; their type IV.J.2. 190 THE MOUNT ROMAN VILLA, MAIDSTONE 􀀄-- ------------ t--=---=---=---=---=---=---=---=---=--= H . u ► z ... I S: -, I 191 I-" tS I j It> (:, ;-:·,:,l􀃩• 1 Not f e,c:eav.eted Buttress pl ,. .t> J :,,, 0 :I ., ·I Ot o, .... -, 111 ,·1·1-Lj--l·l 1•1'1 j.!.111111 rn, ,,,, ',',(1 1 • 1 •1'1 11 1 I•, t 1 1 1 I,•'• If•• ,•,•, • •'• ,•/) 1 1 • 1 t ! ,', 1' 1 j ,1, fr I••• If 1 1 • 0 C D ,:,:,;(:, !• 􀃪•􀃫•>: ,;,:,􀃬, :1:,:,􀃭,; ,; •􀃮• ,•􀃯••,:,; ,; ;,'(;,; ,1 ,;.,,:• •,:1,",(•!1;,1,: ,:,:,;1􀃰1􀃱,;!:l􀃲1:,:,;,:,;•!<􀃳rH r : !􀁮::::: ::: 􀁯 :::: ;::􀁰 ::: ::: !􀂀! 􀂁; :;: :; ::; 􀂂 :i;􀁷:: :: ::;;: 􀀘1 􀀙 '1 :􀆀 :;: >::: ! :::􀁸: 􀁹:: ;: :􀁺 􀁻 :; ::: :::: :􀀟􀀠; :: :􀀡 :'i::: 􀃕 :::: :::: :: :: I Wall p 3' I 11 1 1 II It1 11 :: 1 : I: I:,: I: I: I 􀀮I : t; l t l : /; I :1,, ,􀀐:,:, :, :, 􀂃I! I;,:,:, : •:I:, 􀀯I: 1 ; I :1: I; 1: I􀂄 I: 1/ I:•: 1 1 11 :l :f􀂅f􀂆I :, 􀂇 I􀂈 I: t :, : _ part de$troycd 1 ,1 111,1 , 11 1 1 111,,1,,,,, ,1,1 1,, • , 1 1 1 ,11,,,,,1,t,,,1,1 11,1 1 1 1 1 1 1 1, 1i 1111,,,1 11t 11 1111 1, 11 }i :,! ,:, 􀃴• 􀃵,􀃶,, ,: :􀃷 :􀃸:􀃹:;; f: ! ::: : :: :::;: !! 􀀢:;: ;; Pit 3 : : , : , : , t 􀁼􀁽 : ::: ::: : ; : ; ;!: :!: :: : 􀀣;:􀀤: :;, > :;: :􀀉:􀀊:􀀋!ii;􀀌; 1t! :: :; 􀆁f ! ,:.;,:I::: I: 1: I:,:•: I:; :1: I: I􀆂 :􀆃I:,: I; f: I: l􀆄l ;, 1I II I I 1tI , : !;:􀃺;: ,: ,:, : I:,:::: :1! :􀆅 ,: ,: ,: ,: : ; ! ; I: •:t :,:l :1 :I : I: I: I f Jl 1, , 1 1 1 , ,, ,,','/,', 1 , 1 1 1,,1 1 1 ,1 ,, ,'i',',', 1,1 ,1 , 1 1 1 t 1,1,,,,,11 1 1 1,1, 1 , , , ,', , ,1,11 11 1,1 ,1 , ,,,r,,, , 111 1 1 1 11,1t1,111 1 􀀑 t: 1: l :, ;, ; 1; I􀂉)􀂊 1 :, : 1: •:t \1 :I: I: l: I: I! l􀂋 I􀂌• ;1 !1: ,: •:,:I: I:';,;,:, ;1 􀀥1: •: 1! I: 1:' 11 l ( 􀂍• :, ; 1:• !1 ;, : 1: I 􀆆t :,: I: (:Ii!􀀕,; I: I; ; ; ::! ; :;: : ::: : :;: ; :j: ::􀃖!i :: ::􀃗 :::::!?*: 􀃘::: :: ::! i :I::􀃙:;:::,::;::::::􀃚:'.!􀃛::::::::::::􀃜::::::::: !:i! l 􀃝: 􀃞:; :􀃟; :!:;*,::, • _,1 11 1\l lll ll ll olt l lJ IJ l l l \ l l 111 l l 1 l !>l l1 o 1 • 1f1 l l t l,,•1 1 1 1 1 1 1111 111 1 1 1 1 1 i 1 1,f 1 1 1 1 1 J 1 1 1 1 l l ll ltlll1 lj l) ll ll ll) l ll 1 1 1 I fl1)\ l lll l1 11 Ji l> l l f! 11 tlf/111,tfl 111, I l•t•Jt•ltjlllll/1 II jl •.1: :1: :;􀁱:I::􀁲::,!􀁳::: :i :: :: :: 􀁴::: ::􀁵: ::: :: :: 􀁶: 􀁷 :: : : : :::􀁸 ;!;:􀁹􀁺;}:I􀁻:::;:;:;::􀁼,;::::::􀁽􀁾􀁿:􀂀:::::::;::::;:',:􀂁::: •} 􀀅: i: 1:::; :!: 􀃠􀃡􀃢;!:::􀃣􀃤:􀃥::ii!;;:;:􀃦:::;:;;!!;:;::::: ::: ::􀃧!:i: 􀀊::: ;:: :;::!1;: i!! i: ;: ;::: 􀀚:;::;:;::: ;:; ! :; ! I : ,:,: I 􀆇􀆈j ,:1􀁾 􀁿!1􀂀,t• ;•􀂁1;1 :1 :•!·!•!1 􀂂,:,:1;, :1:,: ·􀂃· :1:,: •􀂄1!: :, :,:1:,: 1:,􀂅❖: ,:,' 1: ,:, :1 :1: ,; • :•􀂆•: d . 􀀢 l􀀣 :!1::: l􀀤I :1: I :1 ;1 !i: I: \􀀥I 􀀦 I 􀀧I!,:, 􀃨i: • ! l :, ! il 1:t:1: :: t: •:I:;􀀐 l􀀨I :1:1 􀆉 1:1:1:1 '.:􀀅1; ,: •􀆊1:1 ;1: l :, :, ! • 1 ) I \ I I / 1 / 1/ l l ) l l ) l \ 1 I< I f11 1)1 )\1 ,.􀆋·􀆌lr l 1111 <' 􀀇 c:, § 􀀈 D.B. KELLY (Plate III A) ran to the north-west, passing under the north-east corner of room 10. The drain was wedge-shaped in section, its sides formed of rough slabs of ragstone, inwardly inclined towards the bottom and supporting capstones (Fig. 8). The bottom of the drain was unlined, consisting of the natural clay. Subsequently the tank was filled in and the gap in the north wall of the yard, through which the drain ran, blocked with ragstone, its white mortar contrasting with the yellow mortar of the rest of the north wall. The tank filling contained a worn coin of Hadrian and a small quantity of pottery sherds, including rims from three plain pie dishes, one of triangular and two of rounded profile, 18 none necessarily much after the first half of the third century. There is some evidence that the rooms on the north side of the yard were not in use after phase 2 (see below). If so, it is unlikely that the north wall of the yard would have been patched up after the drain and tank went out of use, if the yard and the rooms on its north side had been abandoned, so the tank was probably filled in during phase 2. Three groups of pottery are associated with the phase 2 construction: placed as foundation deposits by the east wall of the yard; between the east wall of the phase 1 range and the west wall of the yard; in the filling of the drain trench. Five pots were placed against the lowest, and, for most of its length, only surviving course of the east wall of the yard and one iI?, the course. (Fig. 3). They were in two groups. The northern had three pots, to the west and east of the wall and in the lowest course (Fig. 13, 4-6). All were shattered, though virtually complete when pieced together. The shattering was more complete than would be expected from crushing by the wall or later breakage and the pots appear to have been broken before deposit and jammed into the clay by the wall and into the lowest course. The other three pots (Fig. 13, 7-9), 2 m. to the south, were shattered, but incomplete and had been placed in a pit, 15 cm. deep, against the west face of the wall. With the exception of the globular beaker with the sharply everted rim (no. 4), not normally found after the early second century,19 all the pots fa!!within the late second to mid third century or a little later. The largest sealed group of pottery on the site was between: the east wall of the phase 1 range and the surviving northern part of the phase 2 west wall of the yard (Fig. 4, G-H).20 It was contained in two layers (3 and 4), resting immediately above the first back-fill 18 Cf Monaghan 1987, types 5C2 and 5Cl. 19 The form was found in a third-century deposit at the Marlowe Car Park, Canterbury. (Monaghan 1987, 72-class 211). 2° For an analysis of this group, see Pollard 1988, 236-7. 200 THE MOUNT ROMAN VILLA, MAIDSTONE layer of the construction trench for the phase 2 wall. This layer varied in thickness and composition, from a thin layer of greyish clay at the northern end to a layer of loam and rubble, 25 cm. thick, at the south. Likewise layers 3 and 4 were of variable thickness throughout, L4 being composed of brown soil with ragstone rubble and quantities of mortar, L3 of a dark black soil with domestic refuse. Both L3 and L4 are regarded as contemporary with the building of the phase 2 wall. L4 is in places level with the lowest two courses and the striking variability of the depths of both layers suggests that they were thrown in haphazardly. The contemporaneity of the two layers is confirmed by the pottery, since when some of the pots were being reconstructed the sherds of the same pot sometimes came from both. The pottery (Figs. 13-16, 10-41) is late second to mid third century in date, though it contains some forms which continued in use until the end of the third century. Three of these - plain pie dishes, plain and grooved dog dishes in north Kent wares - are well attested in late second century contexts; production of a fourth, the flanged bowl in 'BB2'21 or north Kent reduced sandy ware is thought not to begin before about A.D. 230,22 though in a number of instances where it appears to have occurred before this23 the earlier date has been called in question on the grounds that it does not appear at northern military sites, where the importation of 'BB2' ceased at about the middle of the third century. The form appears in other fabrics in the early third century24 and a starting date for its production in Kent even before A.D. 230 is likely. A single sherd from a Hadham ware flagon was found in L3. The type is that discussed by C. Green,25 who suggests a date within the first half of the third century. The absence of Alice Holt and later third-century Oxford wares supports a deposition date within the first half of the third century and the high proportion of pie dishes to flanged bowls contrasts with later third- 21 See Monaghan's comments on the use of the term in Monaghan 1987, 171-2. 22 R. Pollard discusses the date in Britannia, xiv (1983), 133-4. 23 Monaghan 1987, 136; Snodland villa, Arch. Cant. lxxxii (1967), 206; Canterbury kiln, St. Stephen's Road, Antiq. Journ., xxxvi (1956), 48. 24 In 'BBl' though perhaps only in early form with grooved flange (Gillam 226-7). The early Alice Holt form, 5B1, is given a starting date of A.D. 200, the fully developed forms, 5B2 and 5B3, of A.D. 220. (M.A.B. Lyne and R.S. Jefferies, The Alice Holt/Farnham Roman Pottery Industry (C.B.A. Res. Rpt. no. 30, 1979), 46. 25 Bird, Chapman and Clark (Eds.), Collectanea Londiniensia (Special Paper No. 2, Lond. and Middx. A.S., 1978), 170- 2. The sherd, not illustrated, comes from the neck under the handle, of which the stump only remains, and the top of the shoulder. As on t?e Minories jugs the shoulder has a highly burnished bright orange-red slip and a smgle groove at the top, but the area under the handle is not burnished. 201 D.B. KELLY century deposits where flanged bowls greatly exceeded pie dishes in number. (below, p. 214-26) A denarius of Caracalla, issued in A.D. 199-200, was found at the top of L3. Although it was damaged at the edges both sides are sharp, suggesting a fairly short period of circulation. It could have been lost before the deposition of L3 and dumped with the pottery and other refuse, but its position at the top of the layer is in favour of its loss at the time of the phase 2 construction. The pottery found in the drain trench packing (Fig. 16, 42-48) was again mainly of late second- to mid third-century date and included at least seven plain pie dishes, though only a single sherd of a flanged bowl. A few sherds were found in the construction trench of the east wall of the yard at its northern end, the only identifiable ones coming from a jar with a thickened, everted rim (Fig. 16, 49) and a pie dish. Sherds from the phase 2 occupation were found under the dirty clay make-up of the phase 3 cobbled yard, both in the gully immediately west of the shed and in the burnt layer within the shed (Figs. 16-17, 50-56). They included an Oxford mortarium (M14) and nothing that is necessarily later than the mid third century and would support a date for the phase 2 construction within the first half of the third century and perhaps around A.D. 225-230. Phase 3 The third building phase saw the refurbishment of the phase 1 range with the addition of a small room with a hypocaust on the east side (rooll?- 5) and a corridor to the east of this (6). The east wall of room 2 was partly removed and replaced by a flue. The phase 2 west wall of the yard was completely removed, except for a stretch at its northern end, to allow the construction of the new hypocausted room and corridor (Fig. 5, I-J). On the south side of the yard a building divided into two rooms (11 and 12) was built against the north wall of the main building, its east wall overlying the footing of the west return of the yard wall. The shed was removed, the yard resurfaced and the east yard wall demolished. It is likely that the phase 2 rooms on the north side of the yard fell into disuse, though they were not demolished as completely as the east wall of the yard. Only the west wall of the range, that is, the continuation of the west wall of the yard, rose above the clay levelling of the phase 3 yard, which in places overran the dividing wall of rooms 7 and 8. Elsewhere modern interference and the line of trees obscured the stratification. No late third-century pottery was found in this area, even unstratified. 202 THE MOUNT ROMAN VILLA, MAIDSTONE The hypocaust (room 5) (Fig. 7) measured 3.50 x 2.10 m. (11 ft. 6 in. X 7 ft.) internally and 5.10 x 2.90 m. (16 ft. 9 in. X 9 ft. 6 in.) externally. Its ragstone and yellow mortar walls were 0.80 m. (2 ft. 7 in.) wide and had survived to heights of 0.45-0.85 m. (1 ft. 6 in. - 2 ft. 9 in.) from the hypocaust floor level or 0.60-1 m. (2 ft.- 3 ft. 3 in.) from the lowest course. The hypocaust floor was of opus signinum. The walls were faced for the lowest 50 cm. (1 ft. 8 in.) with a coarse mortar, 4-5 cm. (2 in.) thick, containing a high proportion of crushed tile and small pebbles and above this with an off-white or pinkish plaster with a smooth surface. The ends and faces of the walls on each side of the flue, that is, the remaining parts of the phase 1 wall forming the west side of the hypocaust, were rendered in a similar fashion. In places a second, thin facing of plaster had been applied. In the centre of the east wall were the remains of a vent, lined with white mortar, at 50 cm. (1 ft. 8 in.) above the hypocaust floor and giving onto the corridor (70 cm. above the lowest course) (Plate III B).The surviving wall drops in height to the south, but the bottom of the vent, 25 cm. (10 in.) across, and part of one of the sloping sides survived. Its height is not known. Squarish or rectangular marks in yellow mortar on the hypocaust floor marked the position of some of the pilae, six of them surviving next to the east wall and two to the east of the pillars of the flue. There were probably five rows of eight pilae each, running from north to south, but none remained in situ. The pilae were probably box-tiles. In the heavy mortary fill (L3) were numerous fragments of box-tiles, but over half of these had mortar adhering to the keyed surfaces and must have come from the jacketing of the room above the hypocaust. In the southern part of the east wall, at a height of 72 cm. (2 ft. 4 in.) from the floor, were the stump of an iron nail and three nail holes. The upper line of the tile and pebble mortar marks the level of the floor above the hypocaust, the plaster and nail holes, for securing box tile jacketing, the walls of the room above. No sign of this floor remained, but the demolition of the floor had removed the mortar fillets; a few lengths of fillet, perhaps from this or the adjacent room 2, were found in the mortary fill (L3). An iron hammer, found in the south-west corner on the hypocaust floor may have been discarded during the demolition. The vent would have opened into the room over the hypocaust, immediately above floor level. Room 2 also had a hypocaust, its floor of opus signinum at the same level as that of room 5, but only surviving in the part adjacent to that room. The rectangular marks of the bases of three flue pillars were clearly visible (Plate II B), their eastern ends in line with the inner face of the west wall of room 5, the western projecting for 25 cm. 203 D.B. KELLY into room 2. The opus signinum floor was laid on a foundation of large ragstone rubble, 50 cm. thick, which rested on the mortar floor of phase 1. No trace of the floor of the room above survived, but it was presumably at the same level as that of room 5. In two places on the inner face of the east wall were stretches of an off-white wall plaster, which survived although the wall itself had been robbed. It is not known how rooms 2 and 5 were heated. Whether the furnace, if built, was to the west of room 2, to its south by room 1 or in the north-west corner of the main building, modern construction or excavation has removed all traces. Neither the walls nor the floors of these hypocausts showed any trace of the burning or blackening which would have been expected even if they were at the furthest remove from the furnace, and it is possible that they were in use for a very short time only or even not completed.26 Demolition had removed the pilae, the flue pillars and the floors above completely and the opus signinum floors, including the marks of pilae and flue pillars, were covered with a thin layer of black soil (L4). In the heavy, mortary fill (L3), which lay above the layer of black soil, were, in addition to box-tile and fillet fragments, numerous fragments of wall plaster, mostly dark red, but some off-white or pink, like that still in situ. A few of these have small tile fragments on the surface, which give, intentionally or not, a marbled effect. This was by far the largest amount of wall plaster found on the site. The same layer also contained a T-clamp and a lump of mortar (with tile fragments) showing a semi-circular hollow, as though it had enclosed a pipe. The mortared box-tiles, fillets and wall plaster suggest that the rooms were completed, since it is unlikely that walls would be plastered before they were roofed. Their quantity, however, is only a tiny part of the jacketing and plastering of room 5 alone. The possible explanation of the thin layer of black earth is that the rooms were demolished, left open for a period and later levelled with the unused rubble, some perhaps coming from them. The areas of re-plastering were perhaps done at the time of construction to correct any unevenness. To the north room 3 was provided with an opus signinum floor, at the same level as the hypocaust floor of room 2, but was not heated, so that it was presumably reached by a step from room 2. Room 4 had a mortar floor at the same level. Whereas in phase 1 its west side was open it was now given a west wall of ragstone and white mortar, 0.75 26 For another unused or uncompleted bath-house see the review by A.P. Detsicas of the excavation report on the bath-house at Poverest Road, Orpington, in Arch. Cant., cii (1985), 278-80. 204 THE MOUNT ROMAN VILLA, MAIDSTONE m. wide, whiclr butted against the north wall of room 3, its footings at a higher level. The west wall of the range, for a short distance to the north and south of the dividing wall of rooms 2 and 3, and the walls dividing rooms 2-3 and 1-2 were also rebuilt at the narrower width of 0. 75 m., two courses of this phase 3 wall surviving. The corridor (6) was to the east of rooms 1 and 5, one entrance to the yard at the north, the other apparently in the main building. Its outer (eastern) walls had been completely removed except for the footings of ragstone, packed with clay (Fig. 5, I-J), which survived to heights of from 20-40 cm. The corridor would have been 2 m. (6 ft. 6 in.) wide. The floor make-up was of ragstone rubble and hassock and, in one place, chalk blocks and rested on the natural clay. The building on the south side of the yard containing rooms 11 and 12 (Fig. 3) was built against the north wall of the main building, although this could only be shown for its east wall. It measured 8.50 x 2.70 m. (28 ft. 2 in. x 8 ft. 10 in.), externally, and was divided internally by a north-south wall with a doorway at its north end. The walls of ragstone and white mortar were 0.55 m. (1 ft. 10 in.) wide. The larger room, to the west, measured 4. 70 x 2.20 m. (15 ft. 5 in. x 7 ft. 3 in.), internally. The entire part excavated was filled by an eighteenth-century pit, presumably for night-soil judging by the paucity of finds, dug to a depth of at least half a metre into the natural clay. This pit extended to the north and east walls of the room, but not to the west wall; it had left them intact except for the inner side of the partition wall. The smaller east room (12) measured 2.20 x 2.20 m. (7 ft. 3 in. square), internally. Occupying the whole of the east side were the remains of an oven (Fig. 9, B), built against the east wall and extending west for 0.90 m. The two bases or piers were built of ragstone, with two large pieces of an orange/brown sandstone above, faced with re-used and broken tiles. They were 0.45 m. (1 ft. 6 in.) high with a gap of 0.60 m. (2 ft.) between them and built on the natural clay. At the top were layers of re-used tiles (tegulae) and mortar, overlapped to reduce the width of the gap between the piers. The top part of the oven was missing, but the gap was perhaps spanned by an iron grid. There was no trace of burning or blackening on the sides of the piers on each side of the gap, though a patch of clay against the south pier was heavily burnt. The floor in the rest of the room was of pebble and mortar. The east wall of the building ran over the remaining lowest course of the return wall, running west, of the phase 2 yard east wall and overlapped the edge of the footings of the north (phase 1) wall of the main building (Fig. 6). Overlying the floor of room 12 was a thin layer of black earth and above this a layer of tile and mortar, about 20 cm. (8 in.) thick. 205 D.B. KELLY The yard, or yard area, was resurfaced. A layer of dirty clay, in places with an admixture of white mortar, presumably from the demolition of the yard walls, ranging in depth from 20-30 cm., was laid over the phase 2 yard surface, sealing beneath it the phase 2 pottery mentioned above. It ran over the remaining lowest course of 'the phase 2 east yard wall. In several areas cobbling survived, mostly sparse, but substantial towards the north side. Two groups of pottery are attributable to the construction period of the hypocausted room 5 and the corridor. The larger comes from the back-fill of the foundation trench dug for the north wall of room 5, on its north side under the corridor floor (Fig. 17, 57-65). There are sherds from five flanged bowls, two plain and two grooved dog dishes, a flange-necked flask, seven everted rim jars with squared or undercut rims, two cooking-pots, nearly all in local sandy reduced fabrics, Patch Grove and grogged jars and three Much Hadham sherds. The group is third-century and would not be inconsistent with a date in the second half of the century. At the west end of this same foundation trench one of two sherds found (Q21/8) was from an Alice Holt flanged bowl with white slip on rim and flange. The smaller group comes from beneath and within the ragstone floor make-up of the corridor, to the east and south of room 5. It includes sherds of two flanged bowls, a sherd of an Alice Holt jar, which should be later than A.D. 270, and two adjoining sherds of a New Forest grey ware bowl (Fig. 17, 66). One of a small number of trial trenches, dug before the excavation to see if the remains of the villa existed within the barracks, was sited in the corridor at the south-east corner of room 5. In it were found, in the floor make-up, third-century sherds from three flanged bowls, two plain dog dishes and four jars, one in Alice Holt fabric. The proportion of flanged bowls to plain pie dishes in these groups is high (10:1) compared with that from the large group of pottery from the phase 2 construction. Immediately east of the east wall of room 12 the surviving single course of the return wall of the east yard wall and the edge of the footings of the period 1 wall were covered by a layer of dark brown loam (Fig. 6, L6), taken to be levelling after the building of the rooms and thus giving a terminus ante quern for their construction. This layer was sealed by a black, burnt layer. It contained sherds of flanged bowls, grooved dog dishes, jars with undercut rims, including the 'swan's neck' form, and a Patch Grove jar, a group comparable with the pottery found in the foundation trench of the north wall of room 5 • The construction of the phase 3 buildings thus appears to have been in the second half of the third century and probably in the last quarter of the century, given the presence of Alice Holt and New Forest wares. 206 l:5 -...J N IA N I Wall p3 0 1 Metre 'B Pit? / NYC s I 􀀁 10a Ji)t�{􀀅qf 􀀂 r·······1 . . . . · · · · · - Stone base Rag stone Mor tar Tile Sandstone Tufa Yellow clay Fig. 9. A: section of stone base 3 and pit; B: section showing front of oven. 􀀃 trl a:: 0 C 􀀄 :i::i 0 􀀅 z < - t""' t""' ► a:: ► 8 en ...:i 0 z trl D.B. KELLY THE END OF THE OCCUPATION The apparent lack of use of the west range indicates demolition shortly after its construction. The dark earth layer covering the hypocaust floors of rooms 2 and 5, on average 2 cm. thick, suggests that they were left open for some years. They were then filled with soil containing a great deal of mortar, with tile, plaster, opus signinum and fragments of ragstone up to the level of the top of the surviving walls (Fig. 7, L2/3). Where there had been no modern interference this thick mortary layer occurred at the same level above the corridor and its robbed east wall. At the south end of the corridor an undisturbed layer with large quantities of broken tile overlaid the disturbed corridor floor, its top level with that of the mortary layer (Fig. 5, I-J). The appearance was of a deliberate dismantling and subsequent levelling, rather than the casual depredations of later stone robbers; usable tiles and stone had been removed. Only two sherds of pottery were found in the mortary fill of room 5, both a few inches above the layer of dark earth (Fig. 17, 67) and none in the undisturbed part of room 2. On top of the east wall of room 2, beneath the mortary fill of the robber trench, was a coin of Tetricus I. No pottery was found in the undisturbed mortary layer over the corridor wall. Covering the pebble and mortar floor of room 12 (Fig. 5, Q-R) was a very thin layer of earth, as in rooms 2 and 5, suggesting again that it had been left open for a time. Above this was a layer with tile and mortar, containing dog dishes, a pie dish, a rouletted beaker and an Oxford mortarium (M17) (Fig. 17, 68-71). There were two coins in this layer, one of Claudius II and a barbarous radiate (Tetricus I/ Spes Augg.) The evidence points to a demolition of the phase 3 buildings before the end of the third century. The latest coin found on the site was of Constantius I, dated to A.D. 295, found unstratified in the yard. Occupation of the site, however, is likely to have continued into the fourth century. The proportion of late wares found unstratified in the top soil or in areas of modern disturbance is noticeably higher than that in the undisturbed phase 3 layers. They include Alice Holt, Oxford and grogged wares and a small group, perhaps local, whose source is as yet unidentified (Figs. 17-18, 72-82). 208 THE MOUNT ROMAN VILLA, MAIDSTONE THE COINS27 1. Hadrian, sestertius, R.I.C. 759. Phase 2, fill of tank. 2. Caracalla, denarius, R.I.C. 47b. Phase 2 construction. (X22/3). 3. Claudius II, antoninianus, R.I.C. 48. Phase 3 demolition layer, room 12. (D32/30/2). 4. Tetricus I, antoninianus, Elmer 789. Phase 3 demolition layer, room 2. (Q21/7). 5. Barbarous radiate, Tetricus 1/Spes Augg. Phase 3 demolition layer, room 12. (D32/30/2). 6. Barbarous radiate minim, Tetricus II/ Pax Aug. Surface of phase 3 yard. (V33/3). 7. Barbarous radiate, illegible. Surface of phase 3 yard. (K33/35/2). 8. Barbarous radiate minim, illegible. Surface of phase 3 yard. (K33/35/2). 9. Barbarous radiate, deified Claudius 11/Consecratio (eagle). Top soil, north end of corridor. (S24/1). 10. Barbarous radiate, Tetricus I/Pax Aug. Topsoil, south end of corridor. (F24/1). 11. Barbarous radiate, Tetricus 11/Spes Aug. Top soil over northeast angle of corridor wall. (T27/1). 12. Barbarous radiate minim, illegible. Topsoil over east wall of room 3 and west wall of yard. (W21/l). 13. Barbarous radiate, Tetricus I/ illegible. Topsoil, west end of room 11. (E26/1). 14. Constantius I, Caesar, antoninianus, Trier mint, R.I.C. 657 ('􀁎). Topsoil, yard. (N33/l). PTR THE SMALL FINDS (Figs. 10--12) 1. Tweezers, copper alloy; length 68 mm. unstratified. (U/24/1). 2. Bracelet, copper alloy, plain, D-section, pointed terminals; approximate diameter 55 mm. Fill of foundation trench of north wall of room 5, phase 3 construction. Second half of third century. (N24/S24/3). 3. Pelta-shaped mount or stud, copper alloy, a single rivet projecting from the back; 33 x 23 mm. Unstratified, top soil above north end of corridor. (N24/S24/1). 4. S-shaped suspension hook, copper alloy, the terminals 27 I am much indebted to Dr John A. Davies for identifying the 'barbarous radiates'. 209 D.B. KELLY decorated with snake heads; decoration the same on both sides; length 39 mm. Although unstratified and reminiscent of modern snake belt-fasteners on boys' and army belts of about the time of World War I, the crude decoration on both sides and its rich green patina suggest that the piece is Roman. (B/1/29/1). 5. Hammer or hammer-pick, iron; length 205 mm. On opus signinum floor of hypocaust of room 5. Late third century. (J21/3). 6. Candle-holder with spike for driving into wall, iron; height of socket 55 mm.; length of spike 75 mm. Two similar iron candle-holders were found at the Roman settlement at King Harry Lane, Verulamium, in the fillings of cellars 28 and 34, where the latest coins were Severan. (I.M. Stead and V. Rigby, Verulamium: The King Harry Lane Site, H.B.M.C. Report No. 12 (1989), 32 and fig. 23, nos. 210, 211). From phase 2 construction layer (Q21/3) and thus phase 1, second half of second or early third century. 7. Part of an iron latch-lifter; the flat handle with the looped end is 100 mm. long. From phase 2 construction layer (Q21/3); second half of second or early third century. 8. Loop-headed spike, iron, the loop closed by welding; length 110 mm. (W.H. Manning, Catalogue of Romano-British Iron Tools, Fittings and Weapons in The British Museum (1985), 129-30). From make-up of phase 3 yard (K33/E ext./2). Second half of third century. 9. Top part of bone or antler needle with flat, spatulate head; length 43 mm. Burnt layer under upper stone surface of cobbled area east of the yard. Third century. (P28/3). 10. Bone or antler pin with spherical head; length 69 mm. N. Crummy's type 3. (Britannia, x (1979), 157-163). From make-up of phase 3 yard (K33/E ext./2). Second half of third century. 11. Spindle whorl, bone; diam. 33 mm. Burnt layer above phase 2 yard surface and below phase 3 yard floor make-up. Second half of third century. (J30/K33/3). 12. Spindle whorl made from base of a black burnished pot; diam. 32 mm. Fill of foundation trench of north wall of room 5, phase 3 construction. Second half of third century. (N24/S24/3). 13. Ten pieces of red deer antler showing stages in the making of pins. The antler was sawn into lengths and some sides of the off-cuts pared with a knife to give a smooth, flat surface. Grooves were then cut longitudinally with a saw at intervals of about 10 mm. to a depth of 7-8 mm., leaving rough-outs, still attached, with a width of 5-8 mm. These were then cut or 210 : I 111111 THE MOUNT ROMAN VILLA, MAIDSTONE 1 3 2 11I 1 //1111• I - ! ,, ---􀀇􀀈 -- - 5 􀀃-􀀄i 􀀆􀀇9illillliJ W􀀈 Fig. 10. Roman small finds: 1-4 actual size, 5 (½). 211 D.B. KELLY 12 -=-􀀅- Ro=nsm,U fimk 6-8 10 11 . (½), 9-12 14 , actual size. 212 THE MOUNT ROMAN VILLA, MAIDSTONE J, J, h Fig. 12. Roman small finds: 13 (½): Antler working debris. 13 levered from the core. The three rough-outs have not been worked and still retain the porous part of the antler. a. Tip of tine; length 80 mm.; unworked. b. Junction of beam and tine; 1.83 mm.; unworked. c. Junction of beam and two tines; 1.96 mm.; the longest side pared. d. Piece of beam; 82 x 39/33 mm.; pared on one wide and one narrow side. e. Piece of beam; 72 x 47/42 mm.; pared on both wide sides. f. Piece of beam; 84 X 37/36 mm.; pared on both wide sides and with two saw cuts, 20 mm. apart, on one of these (Fig. 12, 13f). g. Junction of beam and tine; 93 x 37 mm.; pared over almost its entire surface, with four saw cuts on one side and one on the other. Two rough-outs (13h, below) have been removed and the third partially separated (Fig. 12, 13g). h. Two rough-outs; 50 x 6 mm. and 48 x 6/5 mm.; taken from g. (Fig. 12, 13h). 213 D.B. KELLY i. Rough-out; 66 x 8 mm. (Fig. 12, 13i). From make-up of corridor floor, phase 3 construction. Second half of third century. Other antler fragments were found unstratified and in the phase 2 construction trench between the phase 1 east wall and the west wall of the phase 2 yard (Q21/3), so that the craft was also practised in phase 1. The same layer contained a small ring of antler made from a cross-section of a tine, the porous centre removed (diam. 23 mm., 4 mm. thick). A similar domestic industry making antler pins was found at Chalk, where a tine, three rough-outs and many pins were found in the cellar. (Britannia, iii (1972), 137-40). 14. Melon bead, turquoise frit; 14 x 17 mm. Unstratified, topsoil above room 12. (D32/1). THE POTTERY (Figs. 13-18) By far the commonest pottery found was, as might be expected, that made in north Kent, in the marshes and along the Thames, a reduced sandy ware frequently burnished and occasionally slipped and burnished. In his quantification table of the coarse pottery found between the east wall of the phase 1 rooms 2-4 and the west wall of the phase 2 yard, dating the phase 2 construction, Pollard28 gives 52.5 as the percentage of reduced sandy wares and 32.2 for 'BB2'. Monaghan29 has pointed out that the term 'BB2' is almost meaningless in describing Kent pottery and Pollard,3° though retaining the term, is clearly unhappy with its use. No distinction is made here between 'BB2' and North Kent reduced sandy wares. NKRS: North Kent reduced sandy ware. Monaghan : J. Monaghan, Upchurch and Thameside Roman Pottery, BAR British Series 173 (1987). Pollard: R.J. Pollard, The Roman Pottery of Kent, (Maidstone, 1988). Oxford type numbers are as given in C.J. Young, Oxfordshire Roman Pottery, BAR 43 (1977). Alice Holt type numbers are as given in M.A.B. Lyne and R.S. Jefferies, The Alice Holt/Farnham Roman Pottery Industry, C.B.A. Research Report No. 30 (1979). Phase 1 construction (main building). Second half of second century. 1. Grooved dog dish. NKRS; all over black burnished surfaces. (EM/1/1). 28 Pollard 1988, 237. 29 Monaghan 1987, 171-2. 30 Pollard 1988, 88, and in G.W. Meates, The Roman Villa at Lullingstone, II. (Maidstone, 1987), 198-9. 214 THE MOUNT ROMAN VILLA, MAIDSTONE Phase 1 destruction. Late second - early third century. 2. Dog dish. NKRS; all over grey burnished surfaces with decoration of single wavy line. (T18/7). 3. Plain pie dish. NKRS; all over grey burnished surfaces. (M15/4). Foundation pots by east wall of the yard, phase 2 construction. First half of third century. 4. Globular beaker or jar with sharply everted rim. Hard grey fabric with sparse mica, exterior partially oxidised. Usually late first-early second century, but see note 19. (R35). 5. Jar with recurved rim. NKRS; exterior partly oxidised; decorated with lightly tooled vertical lines; outside of rim and top of shoulder covered with pale grey/white slip, which has dribbled down on one side of the body. (R35). 6. Jar with recurved rim. NKRS; exterior partially oxidised; rim, shoulder and body above base burnished; body decorated with lightly tooled lattice pattern. (R35). 7. Folded beaker with recurved rim. NKRS; grey surfaces; rim, shoulder and body above base lightly burnished. (P35). 8. Large bag beaker. Reduced micaceous ware; grey interior, brown-grey exterior; all over rouletted decoration except for burnished band below neck. (P35). 9. Poppy-head beaker. Reduced micaceous ware; exterior burnished and decorated with lozenge-shaped groups of barbotine dots. (P35). Phase 2 construction. A large quantity of pottery ranging in date from the late second to mid third century, found between the east wall of the phase 1 range (rooms 2-4) and the west wall of the phase 2 yard (Q21-X22/3-4). Pollard, 236-7, gives quantification tables by vessel rim equivalents proportion for the fabrics and his quantification table for forms is given in appendix B. 10-29: North Kent reduced sandy wares. 10-11. Dog dishes. Black/grey burnished surfaces. 12. Grooved dog dish. Black/grey burnished surfaces. 13. Plain pie dish. Brown/grey burnished surfaces. 14. Plain pie dish. Black burnished surfaces. 15. Plain pie dish. Dark grey surfaces, burnished on exterior and internally on base. 215 \ J D.B. KELLY / \. __________ 􀀌JV VVVJ 7 Fig. 13. Roman pottery, 1-15 (¼). 216 4 ;1m!!r!!!!1 􀀈·•JjJ;; - . 􀀈 i;;􀀉:t' THE MOUNT ROMAN VILLA, MAIDSTONE 19 22 I I i ' 26 27 Fig. 14. Roman pottery, 16-27 (¼). 217 23 7 331 \ mf----------r--' ? 351-􀀇\ } 38f 􀀅 􀀆 D.B. KELLY 36 _/; Fig. 15. Roman pottery, 28-38 (¼). 218 30 ,. , ,, ' 1 1, I ,',I \ f/ f' ,'1 I,. THE MOUNT ROMAN VILLA, MAIDSTONE , 45r ; \ 46 R ""􀀉 = 50 􀀊􀀋􀀌== 􀀂􀀃? \ \ 471 ) 1 ) 49F - Z 48 51 Fig. 16. Roman pottery, 39-51, 53 (¼). 219 53 68 -"'·· \ 6􀀂 D.B. KELLY 54 \'--_55L---t _7 .. .. 7 . ·-1 .. - - -1 Fig. 17. Roman pottery, 52, 54-73 (¼). 220 ) THE MOUNT ROMAN VILLA, MAIDSTONE 741 () 􀀇􀀈-7----.---􀀌 5-=-f------1 􀀍--,---'/----- 7 6 \1-------,--􀀇; -+ - nlf--------,--,;􀀆 b -( 78 1==R1=􀀇􀀈􀀉􀀊􀀋􀀌\ -\ l ) / , I ) 81 􀀎-==-----=----=------_------,-,----1( ,:, :· ' 'o • 􀀅o .. . . 82 · · Fig. 18. Roman pottery, 74-82 (¼). 16. Flanged bowl. Highly burnished, giving brown and grey banded surfaces. 17. Flanged bowl. Black/grey burnished surfaces. 18. Flanged bowl. Grey burnished surfaces. 19. Flanged bowl. Brown surfaces, the rim and interior highly burnished. 20. Jar with recurved rim. Rim and shoulder covered with pale grey/off-white slip; grey/brown surface with lightly tooled lattice decoration. 21. Jar with recurved rim. Grey surface, the rim and shoulder highly burnished to appear dark brown. 22. S-profile bowl. Grey surface, the rim, neck, shoulder and lower half of body highly burnished to appear brown; single line of wavy tooled decoration on reserved band at bottom of neck. 23. S-profile bowl. Grey surface, the rim and neck burnished and bands of burnishing on shoulder and above base. Owner's (?) mark IA. 24. Necked jar with rolled, undercut rim. Dark grey surfaces. 25. Necked jar with rolled rim. Brown/grey surfaces. 26. Globular rolled-rim, necked jar with seating for lid. Exterior highly burnished giving a banded brown surface. 27. Globular rolled-rim, necked jar with seating for lid. Patchy grey/buff surface. 221 D.B. KELLY 28. Large flask. Grey surface with rim, neck and bands on shoulder and body highly burnished to appear brown; two bands of wavy line decoration on reserved background. 29. Flask. Grey surface, burnished down to lowest cordon; band of rouletted decoration. 30-34: reduced micaceous ware. (?) Upchurch. 30. Necked jar with bifid rim. Mottled grey surface. 31. Poppy-head beaker. Black/grey surfaces, the exterior burnished between bands of rouletted decoration. 32. Dimpled, funnel-necked beaker. Grey/brown surfaces, the exterior slipped and burnished. 33. Funnel-necked beaker. Black/grey surfaces, the exterior burnished. 34. Funnel-necked beaker with (?) seating for lid. Grey surfaces. 35. Necked jar with undercut rolled rim. Hand-made; coarse, sandy fabric with grog and flint; mottled grey/buff, rough, pimply surfaces. 36-40: grogged ware, hand-made. (See Pollard, 124 and Appendix A). 36. Large jar with everted rim, a cordon at its base. Combed decoration on body and two tooled, irregular, horizontal lines; exterior partially oxidised. 37. Jar with everted rim. Exterior patchy black/grey/buff. 38. Jar with recurved rim. Exterior roughly burnished and partially oxidised. 39-40. Dog dishes. Exteriors partially oxidised; rough burnishing on upper part inside and out. 41. Shallow bowl or dish. Oxidised orange/red fabric and surfaces, burnished on the inside in streaks. The fabric and appearance match the description of 'Streak-burnished' ware. (Marion Green, Romano-British Streak-burnished ware, Kent Arch. Review, 66 (1981), 128--130; Monaghan, 173).31 Pottery of the early second to mid third century from the phase 2 drain trench filling. First half of third century. (E/1/29). A large quantity of pottery, nearly all late second- to mid third-century in date. The earlier second-century pottery was re-deposited from a disturbed pit. Apart from the vessels illustrated it included half-a-dozen sherds of shelly ware and single sherds of a sandy bead-rimmed jar, a decorated pie dish and a rough-cast beaker, in all perhaps 1 per cent of the group. 31 I am grateful to Marion Green for examining this sherd. 222 THE MOUNT ROMAN VILLA, MAIDSTONE 42. Mortarium. Orange/red, fine sandy fabric with grey core; orange/red surfaces, the flange partially covered with an offwhite/ cream slip. Canterbury or east Kent origin; Hartley's fabric 2 (?C). (The Archaeology of Canterbury I: Excavations at Canterbury Castle (1982), 152). 43. Cup, a copy of Form 30. Grey core and orange/red surfaces; sparse mica. (?) Upchurch. 44. Nene Valley colour-coated 'Hunt beaker'. White fabric, brown exterior, dark brown interior. 45. Flanged-rim bowl, with undulating flange and bifid rim. Fine, hard sandy grey fabric with sparse mica; oxidised orange exterior and rim; (?) Upchurch. Monaghan type 5B2, early second century. 46. Grooved dog dish. NKRS; grey burnished surfaces; decorated with a single wavy line. 47. Plain pie dish. NKRS; grey/brown burnished surfaces. 48. Jar. Hand-made, grogged; grey/buff exterior, grey interior; rim and shoulder burnished; strongly tooled lattice decoration. Foundation trench of north-east corner of east wall of phase 2 yard. First half of third century. 49. Everted rim jar. NKRS; grey/brown fabric with sandy surface; traces of slip remaining on rim. (B/1/29/3). Phase 2 occupation. Pottery found in gully in front of shed (nos. 50-54) (N33/M30/4) and in burnt layer. Second-third quarter of third century. 50. Mortarium. Oxford M14; buff fabric and surfaces, the colour perhaps due to burning. 51. Bead-rimmed jar with seating for lid. NKRS; grey/brown mottled surface. Cf. Oakleigh Farm no. 33. (P.D. Catherall, 'A Romano-British Pottery Manufacturing Site at Oakleigh Farm, Higham, Kent', Britannia, xiv (1983), 127). 52. Poppy-head beaker. Sandy ware with sparse mica; surfaces red/brown through burning, grey core; exterior burnished; triangular decoration of barbotine dots remaining on one sherd. 53. Lower part of jar. NKRS; grey/brown surfaces; lattice decoration, band of burnishing above base. 54. Part of jar. NKRS; sparse mica; reddish brown surfaces; light burnishing on shoulder and band of lightly-tooled vertical lines above. 55. Grooved dog dish. NKRS; burnished surfaces, the component sherds grey/brown to bright red through burning. 223 D.B. KELLY 56. Neck of flanged-neck flagon. Oxidised ware with white/cream slip. Phase 3 construction. Pottery from the back-fill of foundation trench of north wall of room 5 (nos.57-65) (S24/N24/3) and floor make-up of corridor (66). Second half of third century. 57. Flanged bowl. Hand-made, coarse grey sandy fabric with a little grog; black/dark grey burnished surfaces; decoration of tooled intersecting chevrons on exterior, curlicues and vertical lines on interior. 58-60: Flanged bowls. NKRS; grey fabric, black/dark grey burnished surfaces. 61. Necked jar, squared rolled rim. NKRS; grey surfaces. 62. Necked jar, undercut rolled rim. NKRS; black exterior, grey interior. 63. Necked jar, squared everted rim. NKRS; dark grey surfaces. 64. Neck of flanged-neck flagon or flask. NKRS; exterior black burnished, interior grey. 65. Girth beaker. Fine micaceous ware; rim, neck and shoulder burnished, interior dark grey. Perhaps a survival from the first or early second century. 66. A thick-walled vessel with a flat, heavy reeded rim, decorated with nicks between the reeding and thumbnail impressions on its lower part. New Forest grey ware, nearest to Fulford's type 10,32 which is a bowl with a reeded rim. A type 10 rim was found in Staple Gardens, Winchester, in a pit filled in the late third century.33 (Adjoining sherds from F24/3 and L24/5). Phase 3 demolition. Pottery from the mortary fill of room 5 (no. 67) (L24/N24/3) and the tile and mortar layer, room 12 (nos. 68-71) (D32/30/2). Late third century. 67. Mortarium. Coarse, sandy buff fabric with grey core and surfaces; rim burnished. 68. Mortarium. Oxford white-ware, type M17; off-white/cream fabric and surfaces. 69. Dog dish. NKRS; black burnished surfaces. 32 M.G. Fulford, New Forest Roman Pottery (B.A.R. 17, 1975), 94 and fig. 31. The sherds were identified by M.A.B. Lyne. 33 B.W. Cunliffe, Winchester Excavations 1949-1960, I (Winchester, 1964), 177-9, fig. 61, 14. 224 THE MOUNT ROMAN VILLA, MAIDSTONE 70. Grooved dog dish. NKRS; black burnished surfaces. 71. Plain pie dish. NKRS; black burnished surfaces. Unstratified pottery, late third - (?) early fourth century. 34 72. Mortarium. Oxford white-ware, type M22; cream fabric and surfaces. 73. Flanged bowl. Alice Holt type 5B6; black slip on rim, flange and interior. 74. Jar. Tilford ware; coarse grey sandy fabric and surfaces. (see Trans. London and Middx. A.S., xxviii (1977), 35; Orton's 'Portchester D', fabric iii). 75-78 are of a fine, smooth pale grey fabric with very little sand and sparse shell. Perhaps a local ware; it does not occur at Canterbury. 75-77. Flanged bowls. Pale grey surfaces. 78. Jar, thickened upright rim. Dark grey burnished exterior, grey interior. 79. Flanged bowl. Hand-made, grogged ware with some mica; dark grey surfaces. 80. Globular (?funnel-necked) beaker. Thick, dark grey ware with sparse mica; girth groove and rouletted decoration. 81. Large jar, cornice rim. Coarse, sandy grey fabric with rough, grey surfaces. The fabric is reminiscent of Tilford/Portchester D ware, but coarser. Probably a local product; it does not occur at Canterbury. 82. Large jar, like no. 81 in form. Coarse, sandy grey fabric with fairly large amount of flint; grey surfaces. Identified as Rettendon (Essex) ware by M. Lyne. SAMIAN POTTERY There was surprisingly little samian pottery found, about three dozen sherds representing probably no more than a dozen or so vessels. Sherds of Forms 37 and 18/31 were found in the phase 2 construction layer (Q21/4) and of 18/31 and 33 in the filling of the phase 2 drain trench. Otherwise the samian came from demolition layers, was embedded in the opus signinum floor of room 2 or was unstratified. Two stamps were found, both of East Gaulish potters: 1. OF CALVI on a Form 18/31R. 34 M.A.B. Lyne kindly commented on the Alice Holtfrilford sherds and others which were comparable but not from that group of kilns. 225 D.B. KELLY 2. VICT////RINV (retrograde), stamped obliquely on the body of a Form 37 bowl, below ovolo band. The Victorinus sherd, the Form 37 sherd from Q21/4, which showed only a small part of the bottom of a medallion and a Form 37 rim sherd were the only sherds of decorated samian found. The plain forms were 18/31, 31, 18/31R, 31R, 33 and 45. OTHER IMPORTEDPOTIERY Two sherds of amphorae were found, both of Peacock and William's class 25 (Dressel 20) (D.P.S. Peacock and D.F. Williams, Amphorae and the Roman Economy (London, 1986), 136--40); one came from the fill of the drain trench (B/1/29), the other was unstratified. Fragments of four or five Central Gaulish colour-coated beakers (A.D. 150-250) were found, in a fine, hard orange-red fabric with metallic black surfaces. In the fill of the drain trench were sherds from a bag beaker, as in K. Greene, 'Imported Fine Wares in Britain to A.D. 250', in (Eds.) P. Arthur and G. Marsh, Early Fine Wares in Roman Britain, B.A.R., British Series 57 (1978), 15-25, Fig. 2.3, no. 4., and an indented beaker like Fig. 2.3, no. 5 (ibid.). From trial trench B ( = Q or T/21/3-4) came fragments of similar bag and indented beakers, the fabric of the bag beaker appearing as a red and grey sandwich. Finally a fragment of a decorated bag beaker, as Fig. 2.3, no. 9 (ibid.) was found unstratified in trench T27/1. BUILDING MATERIALS Window glass Phase 1. Five fragments including a rounded corner piece and two edge pieces; colourless, 3 mm. thick. (Q21/4, phase 2 construction). A single edge piece was found embedded in the phase 3 opus signinum floor of room 2. Wall plaster Phase 1. Seven fragments were found in the phase 2 construction layers (Q21/3 and 4, T21/3). Two were white, three dark red and one white and red, presumably from a white panel with a red border. The remaining fragment was orange, dark red and white, the red a thin line 5 mm. wide separating the orange and white. 226 THE MOUNT ROMAN VILLA, MAIDSTONE Phase 1 or 2. From the floor make-up of the phase 3 corridor (L24/5) came a single corner fragment, one side dark red, the other having a dark red band as a border to white. From the foundation trench of the north wall of room 5 (phase 3) (N24, N ext./3) came another single piece, in dark red with an orange line at the edge. Phase 3. The largest quantity of plaster came from the phase 3 demolition layers, particularly from room 5. Most fragments were dark red, but some were white including a few with a black line, 5 mm. wide, dividing areas of white. There were also pieces of both the red and white plaster with small pieces of tile or brick left level with the surface of the plaster, giving a marbled effect. Amongst the unstratified plaster was a corner fragment, one side red, the other white (127/1). Tile and brick Most of the tile and brick was found in the phase 3 demolition layers, though fragments occurred in the phase 2 and 3 construction layers. The greatest concentration of roofing-tiles, mainly tegulae, was in the undisturbed demolition layer at the south end of the corridor (F24/2) and there were large amounts at the north end of the corridor (Q21/2- S24/N24/2) and over room 12 (D32/2), another undisturbed layer. No complete tegula was found and no fragment larger than about one sixth of its original size. Likewise only a few fragments of imbrices survived. Taken as a whole the fragments of roofing-tile represented only a tiny proportion of the tiles that would have been used. 35 In the construction of the oven in room 12 larger fragments of tegulae were used, their flanges knocked off. The largest piece showed the original length of 450 mm. The largest concentration of box-tile fragments was found in the demolition fill of the hypocausted room 5, but these again represented only a tiny proportion of those needed. With the exception of a half box-tile the fragments were parts of sides only and of four measurable pieces two were 160 mm. wide and two 170 mm. The half tile measures 180 x 130 mm., one side survives to a height of 260 mm. and the incomplete side vents were 180 and 190 mm. from the end. It is decorated with shallow arcs made by a four toothed comb. Combed patterns on other fragments included arcs, diagonal lines 35 For estimates of the number of tegulae required on buildings, see G. Brodribb, Roman Brick and Tile (Gloucester, 1987), 11-12. 227 D.B. KELLY forming a lozenge pattern, transverse arcs bisected by central vertical lines, vertical wavy lines and vertical wavy lines each side of straight lines. The combs used had from four to seven teeth. Two at least of the four bricks used for a hearth in the phase 2 shed were lydion bricks.36 One, though shattered, was complete37 and measured 460 x 330 x 40 mm.; it is among the largest examples of the type (N33/3). In the fill of the construction trench of the north wall of the phase 3 room 5 were five bessales. 38 These bricks were commonly used in the construction of pilae, but none showed traces of mortar and they may have been used for flooring. The measurements are 200 x 190 x 30 mm., 200 x 200 x 40 mm., 190 x 190 x 40 mm. (2) and 195 x 190 x 40 mm. There was also the corner of a larger brick which had been roughly squared to the size of a bessalis, 180 x 180 x 47 mm. Stone Apart from the ragstone of which the villa was built and the use of chalk rubble as make-up for part of the phase 3 corridor floor only two other kinds of stone were found. These were identified by Mrs. S. van Rose of the Natural History Museum. 1. Two large blocks, each about 60 x 45 x 20 cm., of orange/brown sandstone from the Hastings Beds. Used in the construction of the late third-century oven (phase 3) in room 12. 2. Two fragments of Millstone Grit of a pinkish colour, which is found in Derbyshire. The larger fragment has two flat surfaces and is 7 cm. thick. Re-used in drain at front of shed (period 2) and from a millstone (N33). Fragments of Millstone Grit have been recorded from Roman sites in Kent at the Darenth villa,39 West Wickham and Hayes. 40 A millstone fragment of a similar pinkish Millstone Grit has been found at a Roman site at Shuart, near St. Nicholas at Wade.41 THE SEVENTEENTH- AND EIGHTEENTH-CENTURY PITS Excluding modern pits and trenches and pit 1, Victorian, all the pits found dated from the second half of the seventeenth century to the 36 Ibid., 37-40. 37 I am grateful to Richard Krejsa for reconstructing this. 38 See note 35, 34-6. 39 B.J. Philp, Excavations in West Kent 1960-1970, (Dover 1973) 143-4. 40 Ibid., 63, 90. ' ' 41 Shown to me by the excavator, David Perkins. 228 THE MOUNT ROMAN VILLA, MAIDSTONE I 4 2 ---------- - - --- 􀀄 6 Fig. 19. Post-medieval pottery, etc.: 1-4, pit 2; 5, pit 3; 6-7, pit 6; 8, belt or dress hook. (1, 2, 7 (¼); 3-6 (½); 8 actual size). 229 D.B. KELLY mid eighteenth century. Five of them, numbers 3, 4, 6, 7 and 9, contained so little pottery and other finds, despite their large size, that they must be presumed to be night-soil pits. The clay pipes are classified by the types used in Atkinson and Oswald, London Clay Tobacco Pipes, J.B.A.A. (3rd. series) xxxii (1969), 171-227. The suggested names of the makers are taken from the Kent list in A. Oswald, Clay Pipes for the Archaeologist, B.A.R. 14 (1975), 174-6. Pit 2. Bowl of Border ware (Fig. 19, 1); white fabric, internal speckled apple-green glaze. Incomplete. Chamber pot (Fig. 19, 2); hard, sandy red ware, internal glaze of chocolate or very dark brown. Incomplete. Sherds of London tin-glaze from five or six different vessels. Clay pipes: London type 18 (2), c. 1660-80 (Fig. 19, 3); London type 20, c. 1680-1710 (Fig. 19, 4). Date of deposit end of seventeenth century. Pit 3. Clay pipes: London type 25, initials I.C. (Fig. 19, 5). (? James Cutbush, West Malling 1756/ Strood 1758-61); three bowl bases, initials E.M. (2) and E.-(Elizabeth Middleton, Maidstone, 1724-32). A few sherds of eighteenth-century pottery. Deposit of mid eighteenth century. Pit 4. London tin-glaze charger (Fig. 20), decorated with fruit, presumably pomegranates, and leaves; c. 1650-60; incomplete. Deposited during second half of seventeenth century. No other finds. Pit 5. Only partly excavated. Sherds of a large jar of coarse, sandy red ware with internal orange/brown glaze with black flecks; heavy rim (diameter 9 in.) of semi-circular section; a local ware such as High Halden. Base of a grey stoneware tankard. Other sherds include Border white ware and a Staffordshire slipware cup or posset pot. Clay pipe bowl of London type 18 and pipe bowl base with initials I.H. First half of eighteenth century. Pit 6. Only partly excavated. The only finds (Fig. 19, 6 and 7) were the top of a glass wine bottle and a clay pipe bowl of London type 25 with initials RB (Robert Bewley, Maidstone, 1732, apprenticed to Elizabeth Middleton - see pit 3). First half of eighteenth century. Pit 7. Only partly excavated. A few sherds and pipe stem fragments. Eighteenth century. Pit 8. Small pit; sherds and pipe fragments, eighteenth century. 230 THE MOUNT ROMAN VILLA, MAIDSTONE UNSTRATIFIED POST-MEDIEVAL FINDS 1. Belt or dress hook (Fig. 19, 8); copper alloy, length 30 mm. First half of sixteenth century. For similar examples see Arch. Cant., cvii (1989), 407-8. 2. Part of clay pipe bowl decorated with flutings and dots, initials IG on spur; first half of nineteenth century. (James Green appears in 7􀀂 ,_ ,..., [I]]] yellow m orange 􀀑 green • blue Fig. 20. London tin-glaze charger (¼), pit 4. 231 D.B. KELLY Maidstone Poll Books from 1761-1812, perhaps father and son). For similar decoration see an early nineteenth century pipe by a Chatham maker (Arch. Cant., xcv (1979), 235, no. 23). 3. Incomplete clay pipe bowl with design of eagle with outstretched wings on each side. Pipes with this design are commonly connected with inns named 'The Spread Eagle' and there was a tavern of this name in Brewer Street, fairly near The Mount (Melville, Kent Directory, 1858). For similar designs see another Chatham pipe (reference above, no. 24) and, for London, Atkinson and Oswald, London Clay Tobacco Pipes, J.B.A.A., (3rd. series) xxxii (1969), 189 and 201. 4. Part of stem of clay tobacco pipe impressed with, on one side ... POLITICAL & HOME RULE; on the other RD. No. 4. 8149. The registration number dates to 1866. APPENDIX A The Mount villa, Maidstone: the evidence of the Roman pottery for trade and the economy Richard Pollard The main period of occupation of The Mount villa, from the late second to the end of the third century, corresponds to the peak of production of sandy wares, including 'BB2', along the Thames estuary in north-west Kent and the Mucking-Tilbury area of Essex (Pollard 1988). The closing date for this main period of the villa's occupation is based on the suggested date for the filling in room 12 (D32/30/2). This period also witnessed the collapse of the Gaulish samian industries, giving the local north Kent mar&hes fine ware industry a potentially strong marketing position (Monaghan 1987). The location of Maidstone, with a navigable waterway availi)ble along the trade routes from the Hoo Peninsula and Upchurch Marshes, focal points of the two north Kent industries, renders the domination of The Mount villa's pottery by local suppliers unsurprising. This is demonstrated by the quantified group from Q21-X22/3-4 (phase 2 construction) and supported by a more subjective analysis of other assemblages. That Maidstone lies within the heartland of these potteries in the third century is demonstrated by the presence of forms with localised distributions such as the folded everted-rim jar/beaker and the decorated necked jar bowl (Pollard 1988, Fig. 46 and Fig. 50, 192 and 194) and by the high proportion of 'BB2' (ibid., Fig. 48) and other local sandy wares in The Mount quantified group. The town lies at the eastern edge of the second-century distribution of Patch Grove ware and within that of the first - second century shell-tempered storage jars of north-west Kent, south Essex and Surrey (ibid., Fig. 31, Fig. 13, 17-21 and Fig. 12, 16). The relationship of The Mount to the numerous other villa sites in the Medway Valley is difficult to gauge - the publication of comprehensive reports on the Church Field, Snodland, and Eccles sites would be invaluable - but the high density of such sites implies a strong magnet for trade. The weakness of assertions drawn from simple geographical models of site density is demonstrated by the evidence of both fine and 232 THE MOUNT ROMAN VILLA, MAIDSTONE coarse wares, however. The quantities of both exotic and local fine wares are quite low at The Mount when compared with other sites in the area of a similar date range. A later second-early/mid third century group from Rochester (Pollard 1982, group 3; Pollard 1988, Rochester (4)) contained over 13 per cent fine wares, including some 5.5 per cent possible local products, quantifying by EVES of rims. A mid second- to mid third-century group from Springhead (Pollard 1988, Springhead (3)) produced some 17 per cent fine wares including over 7 per cent local products. The Chalk cellar (Johnston 1972) gives two sets of useful figures: layer 8 (third century) included over 12 per cent, and over 5 per cent, respectively, of all and local fine wares and layer 7 (late third century) gives figures of over 15 per cent and over 12 per cent (Pollard 1988, Chalk (1) and (2); all figures from appendix 5). The late second- mid third-century group from The Mount (Q21-X22/3-4) in contrast included a mere 3 per cent fine wares, the great majority (2.6 per cent of the total assemblage) being potentially local products. The pottery from the Cobham Park villa (Tester 1961) was not quantified, but clearly contained a wider range of fine wares than The Mount, including at least two Trier 'Rhenish' beakers, a ware absent from the deposits from The Mount examined by me. The comparison is particularly telling as the excavated structure at Cobham Park is small, comprising only eight rooms, including a furnace room for the heating system. A number of topsoil deposits from The Mount were examined, some of which contained late second-early third-century fine wares. These deposits did not contain large amounts of fine wares, however, and a more directly comparable group from S24/N24/3 (phase 3 construction) contained only fine reduced micaceous wares of probable local origin. This group may belong to a period which saw a hiatus in the supply of exotic fine wares to Kent, between the era of Lower Rhineland and Colchester colour-coated ware importation ending in the earlier third century and that of large scale trade in Oxfordshire and Lower Nene Valley wares from the end of the third century or a little earlier. This hiatus was only partially made good by supplies of Lower Nene Valley beakers and Trier 'Rhenish' ware, the latter being of exceptionally high quality and considerable rarity. Petrological analysis of colour-coated beakers has suggested that the Lower Rhineland exported a significant proportion of the later second - mid third-century barbotine decorated vessels found in south-east England, in competition with the Nene Valley (Anderson et al., 1982). Whether or not the S24/N24/3 group occupies this hiatus, it is clear that The Mount was using comparatively few fine wares during its lifespan. This anachronistic situation - a well appointed property with a humble range of pottery - is reflected in the presence of a small but significant quantity of grog tempered hand-made ware in Q21-X22/3-4 (phase 2 construction). The forms include recurved, everted rim jars and straight-sided dog dishes and are undecorated. They cannot be considered residuals from the first century A.D., when grog tempered wares were common, as the dog dish was not produced in that period and no other first century type wares were recorded from any deposit. The absence of flanged bowls suggests that a fourth century intrusive nature can also be ruled out. Occasional sherds were found in other contexts, including everted rim jars or bowls, but dog dishes seem to be confined to this group. Elsewhere in north-west Kent grog tempered wares do not appear in any quantity in the later Roman period until the fourth century; local manufacture or importation from east Kent are both possibilities (Pollard 1988, 143-150). The existence of a localised, short term and small scale product so close to the major north Kent industries suggests some temporary inability of the latter to satisfy demand, but whether this shortcoming was one of productivity or pricing is unclear. Going (1992) has suggested that the early-mid third century represented a time of recession in the pottery industry of Britain as a whole, with 'BB2' production declining from the end of the second century. His model provides one plausible context for the Maidstone grog tempered ware. 233 D.B. KELLY The absence of stratified fourth-century deposits precludes discussion of the trading status of The Mount villa during the later stages of its occupation. The most common imports in west Kent are present - Alice Holt grey ware and Oxfordshire wares including white ware mortaria and red and white slipped fine wares. The failure to record rarer wares, for Kent, such as Lower Nene Valley late colour-coat and Argonne ware may reflect the small quantities of fourth-century pottery examined rather than its non-appearance at the site (Pollard 1988, 138-143). Late Roman grog tempered ware, including the flanged bowl, is present, and Malcolm Lyne has identified a sherd of Essex Rettendon ware. REFERENCES A.C. Anderson, M.G. Fulford, H. Hatcher and A.M. Pollard, 1982, 'Chemical Analysis of Hunt cups and allied wares from Britain', Britannia, xiii (1982), 229-38. C.J. Going, 1992, 'Economic 'Long Waves' in the Roman Period? A Reconnaissance of the Romano-British Ceramic Evidence', Oxford Journ. Archaeol., xi (i) (1992), 93-116. D.E. Johnston, 1972, 'A Roman Building at Chalk', Britannia, iii (1972), 112-48. J. Monaghan, 1987, Upchurch and Thameside Roman Pottery, BAR British Series 173 (1987). R.J. Pollard, 1982, Roman Coarse Pottery, in A.C. Harrison, 'Rochester 1974-75', Arch. Cant., xcvii (1981), 108-118. R.J. Pollard, 1988, The Roman Pottery of Kent, Kent Arch. Soc. Monograph Series v (Maidstone, 1988). P.J. Tester, 1961, 'The Roman Villa in Cobham Park, near Rochester', Arch. Cant., lxxiv (1961), 88-109. APPENDIXB Quantification table by EVES of forms in the phase 2 construction deposit in Q21-X22/3-4 and B11121-22/3 R. J. Pollard Fabric Form Presence Total % N. Kent reduced roll-rim necked jar 955 21.6 sandy ware. roll-rim jar/bowl 681 15.4 flask 215 4.9 everted lid-seat jar 122 2.8 everted/recurved rim jar 111 2.5 bead-rim beaker 25 0.6 'B82' everted-rim jar 404 9.1 plain pie dish 190 4.3 dog dish 686 15.5 flanged bowl 147 3.3 Grogged wares everted/recurved rim jar + necked jar 323 7.3 dog dish 121 2.7 Patch Grove everted-rim jar 45 1.0 234 MOUNT ROMAN VILLA, MAIDSTONE Sand + grog + flint roll-rim necked jar 56 1.3 Fine reduced roll-rim necked jar 12 0.3 micaceous poppy-head beaker 35 0.8 roll-rim beaker 10 0.2 bead-rim necked beaker 5 0.1 Fine oxidised bead-rim beaker 30 0.7 white col.coat. everted-rim vessel 12 0.3 *Fine oxidised platter 13 0.3 C.G. Rhenish bead-rim beaker 16 0.4 • 'Streak-burnished ware'; shape and slip like Pompeian red ware. 235