Farningham Roman Villa II

^itteaLcrjp dfaniiana FARNINGHAM ROMAN VILLA H By LIEUT.-COLONEL G. W. MEATES, F.S.A. INTRODUCTION THIS Roman villa is situated beside the river Darent, 300 yards southwest of Farningham church, at TQ 545667. Located and excavated in 1948, it is the second major Roman building to have been found at Farningham. The following is a list of Roman sites in and around the village: (1) The villa under discussion. (2) Roman villa I, partially excavated in 1925.1 This is situated 275 yards east-south-east of (1). (3) Roman villa III, 400 yards south of Franks Hall, at TQ 554674 (to be published). (4) A system of Roman ditches, 100 yards east of (2) and excavated in 1947 (to be published). (5) A Roman building on Folly Hill, approximately at TQ 548677.2 There are thus at and around Farningham a Roman ditch system, three Roman villas, and a Roman building of unknown size, together forming a settlement where an ancient trackway running eastwards from the direction of Swanley crossed the river Darent. The discovery of villa II was, as in many such cases, fortuitous. The insertion of posts in the narrow field west of the river produced abundant Roman material, and tests were made near the east bank, the result of which suggested that a large building lay below the turf in that region. Permission to dig was readily given by the landowner, Sir Irving Albery, and the agreement of the tenant, Mr. William Alexander, having been obtained, work was commenced in June, 1948, directed by Mr. E. Greenfield, the late Mr. Edwin Birchenough, and the author. A grid-system was laid out as soon as the direction of the villa walls had been ascertained by preliminary tests, and this system was extended, finally containing over fifty 'boxes' (Plate I). Work continued 1 Trans. Dartford & Dist. Antiq. Soc, 1, December 1931, 66-73. s Arch. Cant., xviii (1889), 307 and map. 1 G. W. MEATES without intermission until December, 1948, during which the whole residential part of the villa had been uncovered. The house was a very large one, of tripartite design, 180 ft. long and 50 ft. wide, with a wing 35 ft. wide projecting eastwards at the south end for 30 ft. At the northern end was another eastward projection, apparently leading directly to a separate bath-house which was found 250 ft. east of the villa; but neither this northern projection nor the bath-house could be excavated, though part of the cold-plunge was examined. The long axis of the villa lay at 30 degrees east of True North, so that it lay with its frontage facing south-east, like so many villas in Britain. I t was found that the river had slightly changed its course since Roman times, having eroded away the extreme southwestern corner of the house. Otherwise, the villa plan remained virtually complete, though the northern extremity could not be excavated. OCCUPATION SUMMARY This falls into three main periods, the first sub-divided into two. These periods involve the original construction with subsequent additions and rebuilding. They are: Period I. c. A.D. 80. A long 'barn-house' of northern Germanic type, with a single wing containing a wide entrance, projecting eastwards at its southern end. Thick clay walls carrying painted plaster upon flint-and-mortar foundation-walls. A roof of thatch or similar material. Period IA. c. A.D. 100. Addition of a narrow front corridor, but with no entrance, though turning eastwards at both ends. The separate bath-house probably built at this time. Period II. c. A.D. 130-150. The front corridor was widened and a large central entrance constructed, the original entrance in the south wing being blocked off. A rear corridor was also added, divided into compartments, with a large hypocaust at the north end of the central range of rooms. Period III. c. A.D. 300. The northern part of the villa was disused. In the southern part, the clay walls were thrown down into the rooms, thus raising the height of the floors which received tessellated surfaces, except room 20, which retained its concrete floor, and the compartments along the back corridor. In addition, rooms 21 and 24 were embelhshed with mosaic floors, and a row of brick pilasters was placed at equal intervals along the exterior of the front corridor as far as the wide entrance, but not beyond it to the north. A small, high-level hypocaust was 2 FARNINGHAM ROMAN VILLA LT built against the south wall of the original large hypocaust, remaining separate from the truncated building. The villa continued in occupation until the second half of the fourth century. A few fragments of red colour-coated vessels and of painted wares were found on the destroyed surfaces of the mosaic floors and suggest the middle of the fourth century as approaching the terminal date; the complete absence of Valentinian and Theodosian coinage confirms this. The villa seems to have been abandoned; it was not destroyed by fire. Two small healths and a small oven, all sited north of the Period I I I house, may suggest sporadic occupation subsequently. DETAILED DESCRIPTION Period I. It was due to the influence of Cnseus Julius Agricola, governor of Britain from A.D. 78 to 85, that the arts of civilization became of paramount importance to the inhabitants of the new province. Tacitus briefly recounts the policy of that great Roman governor, his introduction of a system of new and wise regulations, which included the building in durable materials of magnificent public buildings in towns, and commodious dwellings in the countryside.3 The disturbances attendant on the rebeUion of Boudicca, and its oppressive aftermath, were over, and people were at last ready to accept the benefits of civilization, and by increased efficiency to contribute positively to the requirements of the Roman government. The villa under discussion was founded at this time (Fig.l). Planned on a large scale in the style of a Germanic 'barn-house', it was 144 ft. long by 27 ft. wide, having an entrance 10 ft. wide at the eastern extremity of its south wing which projected for 43 ft. and was itself 23 ft. in width. Large flanking post-holes, each nearly 2 ft. in diameter, indicated heavy door-posts, upon which had hung the double doors. The long part of the house was divided into nine rooms with two transverse passages (1 to 11), the latter 5 ft. and 4 ft. 6 in. wide respectively, which probably contained wooden stairs giving access to the attic storey. The wing contained another transverse passage 4 ft. wide, one large room, and the vestibule (12 to 14). In five cases doorways connecting rooms with each other and two of the passages with the exterior were recognized. CONSTRUCTION AND WALL DECORATION All the wall foundations, including the internal wall partitions, were carefully constructed of flint and mortar, straight and standing upright for a height varying between 2 ft. 6 in. and 2 ft. throughout, the 8 Taoitus, Agricola, ch. xxi. 3 G. W. MEATES O B O 9 w w w ™ I D D SSSSSSSSSSSM Swsvw 4 FARNINGHAM ROMAN VILLA II surfaces flat and unbroken. An interesting feature of these wall foundations was the inclusion of 40 transverse slots, averaging 8 in. in width and extending down to the base of the wall foundations. As all these slots had later been carefully packed with selected flints mortared in, the total number was probably not obtained, but of the 26 incorporated in the exterior wall foundations, their average distance apart was 5 ft. All the slots were unblocked, and in some cases the mortar was discovered still in position which had clasped the wooden uprights that had originally been contained within them. A further interesting discovery was made. At the bases of 21 slots were found crushed down small handfulls of bird bones, placed therein before the heavy wooden uprights were inserted. They appear to have been of votive character, perhaps intended for the protection of the house. This phenomenon might suggest that the person for whom the building was constructed may well have been of Celtic derivation, with the superstitious characteristic of the Celt, a person slowly undergoing the process of romanization. The shape and thickness of the wooden uprights allowed them not only to support a heavy wooden roof-framework—the roof itself was probably of thatch or similar material—but their broad faces across the wall foundations provided vertical seatings for thick infillings of clay. It was noticed early in the excavations that an immense quantity of clay, mcluding within it much fragmentary painted wall-plaster, filled all the rooms and the front corridor over the southern half of the house to an average depth of 2 ft., upon the surface of which was laid a later floor, mainly of red-brick tesserae; it was found that the bottom of this mixed clay and wall-plaster rested upon the concrete floors of the original rooms. It was clear that this clay was the same clay that had once formed the walls, clasped by the wooden uprights and founded upon the wall foundations, and lavishly decorated with painted plaster. At the Period III reconstruction of the house the clay had been thrown down into the rooms to provide a treading surface at a higher level. When the high level tessellated floor of the passage (10) was removed —it lay within a foot of the modern surface and would have been destroyed by any subsequent ploughing—the clay above the lower concrete floor contained highly decorated plaster in slabs, all laminated at a steep downward angle, the result of slicing the clay walls and shovelling the spadefulls down into the space between the wall foundations. The clay having been removed from the rooms, their floors were revealed. In all cases these were of concrete, with large quarter-round mouldings at the bases of the walls, which for the most part retained their plaster dadoes to an average height of 2 ft. The method of construction of the internal wall partitions is well shown (Plate EE), where the flint-and-mortar foundation is seen placed in a slot between two 5 G. W. MEATES rooms, the clay wall being then superimposed upon it, followed by the placing on the wall of the plaster, and finishing with the quarter-round moulding. In some cases the concrete of the floors was tinted to conform with the prevailing colour of the wall-plaster, as in Plate II, where the room on the right (4) has a white concrete floor with white plaster on the walls, while the room on the left (5) has a pinkish-red floor to agree with the red plaster which adorned that room. The clay baulks were everywhere very thick, about 18 in.; but they bore only their own weight, the stout wooden uprights founded in their slots supporting the roof. In only one room (1) was the floor not of concrete. Here it was of pebble-concrete, and this room may have been a small open yard; its south wall only contained slots, and here the fallen clay and plaster lay more thinly than in the rooms to its southward. The large room in the wing (13) contained a mass of white plaster on its concrete floor, rising from the north and south sides of the room, where it was only a few inches thick, to a low mound in the centre nearly 1 ft. thick. This was probably fallen ceiling-plaster, and is the only case in which it was observed. The west wall ofthis room contained the most lavishly decorated plaster of all (Plate IV). The design consisted of a greenish-yellow dado, with above it a trellis in dark blue against a cerise background. At each intersection of the trellis was a small white spot, and within each space had been painted a three-leafed flower in lemon colour. But the decoration was barbaric, though it conformed to classical motifs, for the whole had been splashed over its entire surface with red, yellow, white and blue paint flicked upon it from the brush. This is very suggestive of a semi-barbarous mind, as of an individual only partly civilized, not yet freed from his native artistic preoccupation and with only a vague understanding of classical ideas. A large section of this plaster was removed and is on permanent exhibition in Dartford Museum. The more usual decoration consisted of dadoes, yellow, grey-blue flecked with red, white and blue splashing (Plate III), and tomato-red. Above the dadoes were panels, white splashed with red; lavender splashed with red, yellow and olive-green; and pink, again with red, white and dark-blue splashing. One white panel was edged with red lines and contained small scrolls in the same colour, while floriated borders also occurred, in green, yellow, blue and black. Several panels could be re-assembled from the many fragments, and there was no evidence that any of them had contained pictorial or figure subjects. These thick clay walls, with the plaster keyed on to them by intersecting grooves mcised in the face of the clay, were new to Romano-British archseology when they were found in 1948; and it was interesting to find that at the same time Professor S. S. Frere was 6 FARNINGHAM ROMAN VILLA II discovering similar decorated clay walls in Canterbury. Since then they have been looked for at other sites with conspicuous results, as at Verulamium. It might be thought that walls of clay, apparently unsupported by wattlework as at Farningham, would not be impervious to weather. But the eaves of the thatched roof probably came well down over the tops of the walls, thus shielding them. In the present age, the cob-and-thatch buildings of the West Country stand up against any tempest. THE FOUNDATION DEPOSIT When the clay had been removed from room (4), a circular hole came to Ught in its north-east corner, cut through the concrete floor, and two large pieces of amphora were disclosed covering a vessel sunk below them in a vertical position (Plates VIA and B). This vessel was a finely potted 'bead-rim' cooking-pot of reddish fabric laced with minute slivers of flint and was complete and undamaged except for a small nick on the rim (Fig. 3, 5). I t can be securely dated to the reign of Vespasian, and this dating was confirmed by a dupondius of that emperor which lay inside the pot in contact with the base, a coin which is dated to A.D. 77-79. While inserted after the concrete floor had been laid, it conforms with fragments of samian vessels of Flavian date, and of other 'bead-rim' pots and vessels of the late first century, found elsewhere in the villa; and it may confidently be treated as a foundation deposit inserted shortly after the building had been completed. It is exhibited, with the coin and the amphora 'hd', in Dartford Museum. PERIOD IA About the turn of the century, and certainly near the start of Period II, perhaps c. A.D. 100, a verandah or corridor was built along the south-eastern aspect of the house, turning eastwards at both ends. I t was only 8 ft. wide, somewhat out of proportion with the length of the house; and it was quickly widened to 10 ft. But no entrance was inserted in it, entrance being still presumably by the original doorway in the projecting wing. The fact that the narrow verandah or corridor turned at its north end and in the direction of the separate bath-house suggests that the latter was now constructed. THE BATH-HOUSB This was located 250 ft. away to the east on slightly rising ground, in direct line with the eastward turn of the verandah or corridor; but only part of the cold plunge could be excavated. This consisted of a horizontal concrete floor, ending in a shallow sill, whence the side of the plunge descended vertically for 3 ft., the wall being rendered with 7 G. W. MEATES pink mortar carefully smoothed. The bottom was level and was of buff-coloured concrete, upon which, laid in pink mortar, had once been a floor of mosaic. A double row of small white tesserae still remained at the base of the wall, and it was noted that the mosaic had been laid down before the vertical wall was constructed. No datable object was recovered from the filling, but the excavation was only 4 ft. square. Future investigation of this bath-house would undoubtedly add more to our knowledge of this villa. THE DOG BURIAL The date of this burial is not known, but it was made sometime before the filling of the rooms with clay at the beginning of Period III, and the note of it is inserted at this point for convenience. The burial was sited in the concrete floor of room (6) beside its east wall (Plate VII). When the clay was removed, a roughly circular white patch on the underlying concrete floor was observed, and it was further noticed that the edge of the patch had been cut round. On close examination, the slab of concrete was seen to be the 'hd' of a hole of similar dimension, which was found to contain the skeletal remains of a small dog. Nothing was found with the bones, but the whitening of the 'lid' was suggestive of a desire on the part of the people who then lived in the house to keep the memory and place of sepulture of their dog alive. PERIOD II. c. A.D. 130-150 The house was lengthened northwards for a further 36 ft., by the addition of two rooms (15) and (16), the former designed as a hypocaust with its furnace-room to the north. The pilae were arranged in nine rows, east to west, and consisted each of a massive construction of complete tegulae embedded in clay, resting on a concrete floor, those abutting on the walls being longer than those that were free-standing. The surrounding walls of the room were carefully plastered, though there was no quarter-round moulding at the base of the walls. I t also included a small rectangular compartment built against its west wall, 9 ft. long by 5 ft. wide. The purpose of this compartment is not known; it soon became disused, as the rows of pilae ran across its truncated walls. The upper floor of the hypocaust had completely disappeared, and the furnace-flue had later been blocked. There was Httle ash present or appearance of calcination, but the furnace-room could not be excavated. In addition to the northward extension of the house, a corridor 9 ft. wide was constructed along the western exterior, with four partitions dividing it into compartments, 27 ft., 22 ft., and 13 ft. in length. The original front verandah or corridor was widened to 10 ft., 8 FARNINGHAM ROMAN VILLA H perhaps during the construction of a new roof of tiles to replace the old one of thatch, the original timber roof-framework continuing in use, with pentice roofs along the east and west exteriors—quantities of broken roofing-tiles were found along the whole length of the back corridor. A wide and important main entrance to the house was now constructed centrally in the exterior wall of the new front verandah, an entrance 13 ft. wide, flanked by piers made of large rectangular brick slabs, each pier being 4 ft. in length. A pair of flint-and-mortar foundations to support the wide wooden portico projected at each end of the entrance, enclosing a treading surface 21 ft. across and cambered to north and south. The old entrance at the end of the south wing was blocked by a stout flint-and-mortar wall. PERIOD III. c. A.D. 300 There is no positive evidence that the occupation of the villa continued through the third century, but at the opening of the fourth century a radical reconstruction took place (Fig. 2). The northern part was demolished, a new cross-wall at the north being constructed from corridor to corridor through rooms (4) and (5); this new wall was edged on its north side by a row of chalk blocks, suggesting a dripcourse to an exterior wall. A denarius of Septimius Severus was found among the chalk blocks, but such denarii seem to have had a long life and could easily have still been in circulation even at the end of the third century. The wide entrance in the front verandah may now have been demolished; a row of seven brick pilasters, still remaining six bricks high, and founded on flint-and-mortar bases, was built up against the exterior verandah wall south of this entrance, but not continuing along the verandah wall north of it. There was no sign, however, of a bounding wall across the verandah at this point, though double-doors could easily have been hung here. The compartments in the back corridor remained in use, including compartment (17), which would now have projected northward from the new end of the house. A small, high-level hypocaust was constructed adjoining the large hypocaust, which may not have been still used as such; but this hypocaust system was now separated from the house itself, which was now designed as a luxurious riverside retreat. This small residence was almost completely rebuilt. The roof was taken down, the wooden uprights withdrawn from their slots, which were filled with selected flints carefully mortared to conform with the courses in the walls; and a further level of flint coursing was laid over the whole length of the walls, to provide a level surface for the horizontal sleepers into which should be mortised the new uprights to support the freshly designed roof. The clay walls, which had stood for a very long time, were sliced up and thrown into the rooms. This was then 9 G. W. MEATES EZZZZ22 S5S5S' zzSCTSz? Sssssss /77/;///// 10 FARNINGHAM ROMAN VILLA H levelled over, raising the floor-level throughout to the same height. Two transverse walls were then built across the original rooms (7) and (13), the former wall for some reason not exactly at right-angles to the east and west walls of the room. The new house now consisted of rooms (21) and (23) divided by the passage (22), the southern part of the front verandah, and room (25) now made equal in size with room (23), with a recessed vestibule or loggia between the two. All these received floors of red-brick tesserae, room (21) and the loggia being decorated each with a mosaic panel within the tessellated surround. Only the northernmost room (20) was not tessellated, having a floor of fine concrete. The recessed loggia was an interesting feature. Set back 5 ft. between its flanking rooms, and facing south with a pleasant aspect over the river, the whole of its southern fimit consisted of an entrance divided into three equal apertures, each 3 ft. 6 in. wide, flanked by brick columns 1 ft. square on small flint-and-mortar bases, the whole facade set upon a foundation of bricks in three courses, protruding along the inside like miniature steps. The two free-standing columns may possibly have been continued up in turned stone, for part of such a column, found built into a wall in the Farningham I villa not far away, was found to fit these brick bases accurately. This section of stone column can be seen in Dartford Museum. As has been said, this attractive loggia had a mosaic panel 9 ft. square in the centre of its tessellated floor; but of this panel only a double guiUoche in red, blue, yellow and white, divided by a blue band, all on a white ground, remained—the central motif was missing, as the tesserae had been removed, possibly for use elsewhere. Very Httle of the tessellated floor survived in room (25), and much of room (23) had been lost due to river encroachment, though a large part of its tessellated floor survived. The large room (21) originally possessed a mosaic floor, 16 ft. by 14 ft. 6 in., set towards the east side of the room, the whole with a tessellated surround. Again, this mosaic floor had been almost entirely destroyed, except for one smaU part approximately 1 ft. 5 in. square, showing part of a floral-like motif in blue, red and yeUow, but which could not be identified. Loose mosaic tesserae covered the floor, which was only shghtly below the modern field surface; but quantities had been ripped up for use elsewhere, as with the loggia mosaic. This smaU, re-designed house must have been a pleasant retreat in the fourth century, with a southern outlook along the valley beyond the nearby stream. Like the Lullingstone Roman villa, it provides yet another type of residence among those that existed at intervals along the valley of the Darent. Towards the end of its occupation, compartments (17), (18) and 11 G. W. MEATES (19) in the back corridor were used for the storage respectively of large dumps of chalk blocks, heavy flints, and gravel, the sort of material suitable for road-making or repairing; and a rough roadway of chalk and rubble was in fact found to extend from the original entrance in the front verandah over passage (2) into room (1). This roadway was of late Roman date, containing as it did much fragmentary pottery, and it may have been laid down for carts to carry away the demohshed walls of this part of the house. The northern part of the concrete floor in room (20) was found to have been destroyed, its edge conforming with the roadway, though over 40 ft. south of it. The viUa was finally demohshed, probably in the second half of the fourth century. ACKNOWLEDGEMENTS Thanks are due to Mr. D. B. Broodbank, A.R.I.B.A., for the plans, and to Mr. J. V. Ritson, of Dartford Museum, for making available aH the material from the excavation, which has been deposited there. The photographs are the work of the late Mr. M. B. Cookson, official photographer to the Institute of Archseology. The mechanical filhng-in of the excavations was very kindly undertaken by Mr. Gough-Cooper, of Wilmington. APPENDIX I THE SAMIAN WARE (Fig. 3.) The late Dr. Felix Oswald, D.Sc, F.S.A., very kindly examined the samian ware and was able to report on 27 examples from a total of 76 sherds, the majority of which were small, indeterminate fragments. His notes are incorporated hereunder: Form 18. 1 Nero-Vespasian, c. A.D. 65-70, stamp ROPPV. ROPPVS of La Graufesenque, 0 & P,4 XLV 12 (Fig. 3, S4). 1 stamp ( ) LI.M Domitian (Fig. 3, S5). Form 18/31. 1 Domitian. 0 & P, XLV 17. 1 Hadrian-Antonine, c. A.D. 130-140. Lezoux. 0 & P, XLVI, 14. Form 27. 1 Nero-Vespasian, c. A.D. 60-70.0 & P, XLIX 2-10 and 13. 3 Domitian, c. A.D. 80-90. 0 & P, XLIX 18 (Fig. 3, SI). Form 33. 1 Domitian, c. A.D. 80-90. O & P, LI 10. 4 late Antonine, c. A.D. 140-180. Stamp ELVILLI (Fig. 3, 82). ELVILLVS of Lezoux. * F. Oswald and T. D. Pryce, An Introduction to the Study of Terra Sigillata, London, 1920. Hereafter abbreviated to O & P. 12 FARNINGHAM ROMAN VILLA H " ^ SI S3 S + "SB*- S5 f T.-—••••-•-.-.T. ' , ssssaaaj^ijj BURNISH Fio. 3. Samian Ware (Sl-6) and Coarse Pottery (£). 6*. 13 G. W. MEATES Form 37. 2 Trajan, c. A.D. 110-120. 1 Hadrian-Antonine, c. A.D. 130-140. Probably IVLLINVS of Lezoux (Fig. 3, S3). 3 late Antonine, c. A.D. 160-190 (1 IVLLINVS of Lezoux). Form 31. 3 probably Antonine. Form 36. 1 Hadrian, c. A.D. 120-130. 1 Hadrian-Antonine, c. A.D. 130-140. O & P, LIII 13. Curie 21. 1 Antonine, c. A.D. 140-150, 0 & P, LXXIII1. Walters 79. 1 Hadrian-Antonine, c. A.D. 130-140. Walters 81. 1 Antonine, c. A.D. 140-150. O & P, LXI 7. The samian ware faUs into two groups, Nero-Flavian (A.D. 60-90), and Hadrian-Antonine (A.D. 130-190), with a very small intermediate group Trajan-Hadrian (A.D. 110-130). The first group may be associated with the Period I building, the second with the Period I I enlargement of the viUa; the intermediate group may be placed as survivals in the second group. The foUowing analysis-table shows the datable vessels represented: Nero- "I Vespasian (• A.D. 60-70 J Domitian \ A.D. 80-90/ Trajan A.D. 110-120 Hadrian A.D. 120-130 Hadrian- ~| Antonine 1 A.D. 130- f 140 J Antonine "1 A.D. 140- V 160 J Form 18 1 1 Form Form Form Form Form Form Curie 18/31 27 33 37 31 36 21 1 1 3 1 2 1 1 1 1 4 3 3 1 Walters 79 1 Walters 81 1 The samian ware poses an interesting problem. I t consists exclusively of products of South and Central Gaul, East GauHsh material being entirely absent, the latest dating being c. A.D. 150. Where an occupation continues through the latter half of the second century, such East GauHsh material is usuaUy present; its absence here suggests a tem- 14 FARNINGHAM ROMAN VILLA H porary cessation of occupation, though the coarse pottery, bead-rim and straight-sided pie-dishes, does not confirm this. And there are only two rims of Rhenish ware and one fragment of colour-coated 'hunt-cup', which can be dated not earlier than c. A.D. 170. This absence of East GauHsh samian, and the paucity of Rhenish and early colour-coated material, is negative, and a temporary cessation of occupation may probably not be proved by such evidence in the face of the apparent continuance of Antonine-type coarse pottery. Nevertheless, the value of an analysis of even a smaU quantity of samian ware, as in this case, does produce a useful guide-line to the occupation or otherwise of a building. APPENDIX II THE COARSE POTTERY (Figs. 3-5) The coarse pottery confirms the samian ware as regards the Period I and Period I I occupations, except that the latter would seem to be extended by the inclusion of a considerable quantity of Antonine material. Period I I I is amply confirmed by the fourth-century pottery. PERIOD I The pottery of this period can be divided into that before c. A.D. 80 and from then to c. A.D. 120. The categories are Hsted below: Store-jar This is the earfiest vessel and is represented by a rim fragment (Fig. 3, 1). Heavy and coarse, of fight-brown fabric impregnated with sHvers of flint and showing traces of burnishing, this very large vessel was built up, the sharply everted rim probably finished on the wheel and 11 in. in diameter. There has been a tendency to describe these huge, coarse store-jars as Belgic products, and this is understandable as fragments of them sometimes appear stratified with fine GaUo- Belgic pottery. But this attribution, as also for a date well within the earlier Roman period in Britain, is unlikely to be accurate. They are almost certainly of native tradition, stemming from the late Iron Age; and the fact that very few fragments, even large rim fragments, are found in Roman contexts is very suggestive of a date before the Claudian conquest. Four examples with similar sharply everted rims were found at the Lullingstone Roman villa, where they are placed in the late Iron Age occupation of the site before the viUa was built c. A.D. 80-90. 15 G. W. MEATES Bead-rim cooking pots These should be divided into two types, those of so-caUed 'native' fabric, somewhat thick but usuaUy with soft surfaces, and probably wheel-made, and those of hard fabric, weU-made and finished on a fast wheel, and these may be described as of 'Roman' fabric and generaUy dating later than the first type. There are 35 examples of this first type, three of which are iUustrated (Fig. 3, 6-8) to show the considerable variations in rim profile. Of the later type, the complete pot from the foundation deposit is representative (Fig. 3, 5), a vessel of hard, reddish fabric impregnated with flint grit. This is dated in the reign of Vespasian (A.D. 69-79), and there are 40 more examples ofthis type. Cordoned bowls Like the bead-rim vessels of the earher type, these may be placed in the native British tradition, wheel-made, and usually with soft surfaces. They are similar to those excavated by the late A. W. G. Lowther, F.S.A., at Purberry Shot, Surrey, which he dates to the end of the Iron Age and continuing for a time after the Claudian invasion.5 Of the 10 examples, three are iUustrated: Fig. 3, 2. Light-brown with grey burnishing on exterior, soft surfaces. Spin-marks at intervals and slight cordons. Diam. 8 in. Fig. 3, 3. Light-brown, soft fabric with traces of lightly incised trelfis on exterior, with smooth zone between trelHs and cordon. Hand-made, the rim finished on the wheel. Diam. 5 in. Fig. 3, 4. Dark-grey fabric with exterior burnishing in black. Smooth exterior, rough interior, but smoothed on inside of the rim. No cordons. Diam. 7 in. Patch Grove ware Only three examples of these large vessels, with a zone of finger-tip impressions high on the body, are present. Patch Grove ware occurs in abundance on other Romano-British sites in the Darent valley and on the North Downs, and its paucity here is unusual; its virtual absence is aU the more strange as it continues weU into the second century and ought to be quite common. Further pottery of Period I is iUustrated below: Fig. 4,9. Light-grey with concentrio incised circles and vertical combing. GaUo-Belgio type moulding and rim. c. A.D. 50-100. Fig. 4, 10. Cordoned bowl, fine, hard, grey fabric, c. A.D. 80-120. 6 A. W. G. Lowther, 'Purberry Shot, EweU, Surrey', Surrey Arch. Coll., 1 (1946-47), 26-7, F16, 17 and 20. 16 PLATES I AND II :x'"j; ;-• Plate I. Grid .System, looking South. Plate II. A Clay WaU, showing Method of Construction. [face p . lti PLATES III AND IV Plate III. Clav Wall with Plaster, and Periods I and III Levels. . Plate IV. Example of Plaster Decoration. PLATES V AND VIA AND VIB Plate V. Period I Wall, showing Slots for Uprights. /» Plate VIA. Foundation Deposit. Plate VIB. Foundation Deposit uncovered. PLATES VII AND VIII :•-. r*» <*« 3k Plate VII. Dog Burial. Plate VIII. Period III Loggia, looking South-east. FARNINGHAM ROMAN VILLA II in» //f//////////t//s///i tt i Hum r "i^\x IS lb 17 18 % • » 21 ii • 23 ^ 2+ FIG. 4. Coarse Pottery (£). 17 G. W. MEATES Fig. 4, 11. Fragment of 'rustic' ware, grey fabric with shp on exterior worked with the fingers. (GiUam 97)6 c. A.D. 80-130. Fig. 4, 12. Globular vessel with sharply everted rim. Light pinkishbrown, thin fabric, rather soft, with zones of Hghtly incised rouletting on exterior. Late-first to early-second century. Fig. 4, 13. Rough-cast beaker, brown fabric, sanded on exterior, blackish interior. (GiUam 72) c. A.D. 80-130. Fig. 4, 14. Disc-mouth flagon. Light buff, with groove round top of body. (Gfflam 13) c. A.D. 80-130. Fig. 4, 20. Mortarium. Coarse, granular texture, greyish-cream with pinkish tint. Grey, white, black and brown grit. SHght concentric scoring on interior and on top of flange. Stamp MARIN repeated right and left of spout, and reading from interior. MARINVS is fisted as from the Verulamium region and dated c. A.D. 75-105 (Richborough V).7 Fig. 4, 21. Mortarium. Cream fabric, white, red and black grit. (GiUam 242 type) c. A.D. 90-130. PERIOD II The usual Antonine pottery assemblage was present, but is not illustrated, except for five pie-dish rims which are suggestive of rim development (Fig. 4, 15-19). They are detailed as foUows: 15. (Gillam 308) c. A.D. 130-180. 16. (GiUam 307) c. A.D. 125-160. 17. (GiUam 310) c. A.D. 170-210. 18. (GiUam 225) c. A.D. 190-240. 19. (Gillam 222) c. A.D. 170-210. Also, one mortarium (Fig. 4, 22). Cream fabric, shghtly pink, black and white grit (Gillam type 251). c. A.D. 130-160. PERIOD i n No identifiable third-century pottery is present; the two mortaria (Fig. 4, 23 and 24) which begin towards the end of the century also continue for some decades •vyithin it. The same apphes to the red colourcoated ware and the painted ware, which continues at least towards the middle of the fourth century though beginning before the third century closed. Indeed, some sherds of red colour-coated and painted wares, the interesting fbHated bowl (Fig. 5,32), and the mortarium rim • J. P. Gillam, 'Types of Roman Coarse Pottory Vossols in Roman Britain', Arehceol. AeKana, 4th series, xxxv (10C7), 180-251, ' (Ed.) B. Cunliffe, Fifth Report on the Excavations of the Roman Fort at Richborough in Kent, Oxford, 1968, 183 under Fahrio B, and List at p. 178. 18 FARNINGHAM ROMAN VILLA H (Fig. 4, 23), were found directly upon the broken-up mosaic panels in room (21) and the loggia (24), and on the treading-level immediately south of the loggia. This material would, therefore, seem to have been already out of use by the second half of the fourth century. Some Castor ware of fourth-century date is present, including part of a rouletted box-Hd, which was found in the make-up of the flint and rubble roadway. It is similar to GiUam 341, where it is dated c. A.D. 180-320; and at the Roman viUa at Park Street, St. Albans, such Castor boxes and Hds are confidently assigned to the first half of the fourth century.8 A Httle riUed ware, flanged bowls, and rough, cordoned vessels, like those of the first century, are present as usual in this fourth-century context. The latter are often exceedingly difficult to distinguish from their first-century prototypes, and secure stratification is the only sure guide to differentiate them. The following pottery is iUustrated: Fig. 4, 23. Mortarium. Red fabric, black and white grit. Diam. 6 in. (Gillam type 279) c. A.D. 270-350. Fig. 4, 24. Mortarium. Cream fabric, brown and grey grit, groove on top of rim. Diam. 9 in., typical fourth century form. (Gillam type 275). c. A.D. 250-330. Fig. 5, 25. Bowl, creamy-buff, shghtly soft fabric. Band of reddishorange paint on exterior and thin band of deep reddishbrown paint round top of rim. (Richborough II, PI. XXXI, 162, late third to early fourth century.) Fig. 6, 26. Creamy-buff, shghtly soft fabric. Interior of base, decorated with concentric zones in brownish-chocolate paint. Fig. 5, 27. Side of bowl, very Hght pinkish-brown fabric, shghtly soft. Two zones of rouletting towards the base, with curvilinear decoration in fight-brown paint. Fig. 5, 28. Light creamy-buff, slightly soft fabric. Interior of base near the side, with two concentric bands of orange-brown paint, and a band of same paint round exterior of the side next the base. Fig. 5, 29. Creamy-buff, orangy fabric, shghtly soft. Interior of base, roughly decorated in brownish chocolate paint. Fig. 5, 30. Rim and side of bowl, red colour-coat, smoke-blackened on part of exterior and on interior. Zone of rouletting below rim and remains of ourvifinear decoration in shp, probably white, on exterior. Fig. 5, 31. Side of cup, with impressed rosettes and billets, red colourcoat remaining on base and on interior. 8 H. O'Neil, The Roman Villa at Park Street near St. Albans, Herts., 90 and Fig. 19, 42. 19 G. W. MEATES " ^ ZS •\::~si r *\ ? 27 .IP FIG. 6. Coarse Pottery (£). Fig. 5, 32. Bowl in grey fabric, with shallow indentations and vertical fohated stalks incised between the indentations. Similar decoration, though the fohated staUts are inverted and in white shp, is iUustrated in New Forest pottery.9 Fig. 5, 33. Store-jar, AHce Holt Forest. Grey, hard fabric, with darkgrey shp in zones and combing decoration on exterior between the zones of shp. Second half of fourth century. APPENDIX m THE COINS By BERTRAM W. PEAROE, M.A., F.S.A. Vespasian Dupondius Obv. IMP/CAESAR VESPASIAN A.D. 77-79 COS VIII. Head r. radiate. Rev. CERES AVGVST Ceres 1. holding corn ears and sceptre. 0 H. Sumner, Excavations in New Forest Roman Pottery Sites, London, 1927, 116 and Fig. xxxv, 1. 20 FARNINGHAM ROMAN VILLA II Septimius Severus Licinius I. Theodora ConstantinopoHs Constans Radiate Denarius A.D. 205 AE2 A.D. 307-24 AE2 A.D. 307-24 AE3 A.D. 330-7 i AE3 A.D. 330-7 AE3 A.D. 337-50 AE3 Obv. SEVERVS PIVS AVG Head r. laureate. Rev. PM TRP XHI COS H I P P Rome standing 1. holding Victory and Spear reversed. Obv. IMP LICINIVS PF AVG Bust r. cuirassed and laureated. Rev. SOLI LNVICTO COMITI Sungod I. raising hand and holding globe. SIP PLON Obv. IMP LICINIVS PF AVG. Bust r. cuirassed and laureated. Rev. GENIO POP ROM. Genius 1. with patera and cornucopia. SF PLN Cohen 49. Obv. FL MAX THEODORAE AVG. Bust r. draped. Rev. PIETAS ROMANA. Empress nursing child. Cohen 4. Obv. CONSTANTINOPOLIS. Bust 1. helmed, cuirassed, with sceptre. Rev. Victory 1. with foot on prow. Obv. CONSTANS PF AVG. Bust r. draped with diadem. Rev. VICTORIAE DD AVGG QNN. Two Victories facing. Indeterminate. Third century Total Roman coins: 8. Post Roman Venetian German AR fourteenthfifteenth century Jetton Obv. Lion of St. Mark. (SANCTV)S MARCVS VEN. Rev. Kneeling Doge 1. with flag. PR 1. DVX r AC P ROL . . . HANNS KRAVWINCKEL IN NVR. Crown and Fleur-de-lys. 21