Excavations at Reculver, Kent, 1951

EXCAVATIONS AT RECULVER, KENT, 1951 By F. H. THOMPSON INTRODUCTION THE surviving sections of the walls of the Roman fort at Reculver, Kent, now constitute an ancient monument in the guardianship of the Ministry of Works and the excavations which form the subject of this report were the outcome of a programme of consolidation which has recently been put in hand. As the result of continuous ploughing over many years, the ground level outside the walls is now much lower than in antiquity so that the outer face is exposed to below the foundation of beach pebbles. This has the happy effect of revealing the structural features but unfortunately also means that there is some peril of settlement, or even collapse, of the walls through pressure from within, where the ground surface is level with the present top of the walls. The danger is being met by underpimxing at the foot with concrete which is ultimately to be concealed by a turf-covered bank. Along the western side, behind the King Ethelbert Inn, the wall is not only fully exposed on its outer face but also split into sections, two of which appear to have an outward tilt. It was decided to clear a short stretch of the inner face at this point to determine whether the tilt affected the whole thickness of the wall or only its outer " skin," and so to discover the best method of preservation. As archseological supervision would be necessary, the opportunity was taken of extending the excavation into the interior of the fort and of obtaining a measured section of the defences at this point and, if possible, a reasonably close date for the construction of the walls. Exposure of the inner face revealed that not only had the wall as a whole tilted outwards through an angle of approximately 30 degrees but had also slid outwards bodily for a distance of approximately 2 ft. This had caused some mutilation of the stratification immediately against the inner face, and the section as a whole, though structurally informative, was singularly unrewarding in terms of datable finds. In an effort to obtain clearer results, a second section was cut behind the south wall where it was virtually intact; it is this section which is figured and described, with corroborative detail from the first section. The writer is indebted to Mr. G. C. Dunning, F.S.A., for help and advice both in the course of the excavation and during the preparation of this report. 52 PLATE I (a) RECULVER : Mortar mixing floor inside south wall. Inches (b) RECULVER : Belgic combed ware from pre-fort occupation. [face p. 52 EXCAVATIONS AT RECULVER, KENT, 1951 PREVIOUS EXCAVATIONS Our knowledge of the defences of Reculver rests largely on an examination of the inner face of the east wall carried out by Mr. George Dowker in 18781 and on a section cut by Major Gordon Home in 1927 against the inner face of the south wall.2 Both showed that the inner face of the walls had two offsets and the second that there was a contemporary earth bank, but neither shed any light on the date of their construction. The fort certainly formed part of the Saxon Shore defences and despite certain features in its design which have been held to point to its construction at an earlier date (viz., its almost square shape, the internal bank, and the absence of bonding tiles and bastions), it is generally thought that it was built towards the end of the third century specifically to thwart Saxon raiding.3 However, the coin and pottery evidence points to an earlier occupation of the site in the first century, though possibly only of brief duration ;4 the present excavations have thrown some light on the character of this earliest settlement. THE EXCAVATIONS The first section, measuring 24 ft. long by 5 ft. wide, was cut at right angles to the west wall immediately behind the King Ethelbert Inn, its southern edge being approximately 10 ft. north of the top edge of the scarp which here forms the continuation of the south wall. It was excavated by the writer and three workmen during the first fortnight of April, 1951. The second section, measuring 29 ft. long from the inner face of the wall by 6 ft. wide, was cut at right angles to the south wall, its western edge being 190 ft. east of the boundary wall of the coastguard station (i.e., 50 ft. west of the section cut by Major Home in 1927). It was excavated during the last week of April and the first week of May, 1951, and is the section which is now described (Fig. I).5 The description reverses the order of excavation and follows the historical sequence. On the natural Thanet Sand (1), which here slopes gently upwards from south to north, rested a layer, approximately 2 ft. thick, of dark brown sand (2) flecked with charcoal and containing scattered potsherds and animal bones. The pottery from this layer falls broadly into two groups : firstly, fragments of Belgic bead-rim 1 Arch. Cant., XII (1878), pp. 1-13. 2 Arch. Journ., LXXXVT (1929), p. 260. 3 V.C.H. Kent, III (1932), p. 23, and R. G. Collingwood, Archceology of Roman Britain (1930), p. 64. Mr. R. E. Jessup's paper in Antiquity for June, 1936, gives a useful summary of the known facts. 1 V.O.H. Kent, III, pp. 23, 24. 5 This seotion has already been figured and described in J.R.S., XLII (p. 102, Fig. 13 and PI. XIII) and I am indebted to the Editor, Miss M. V. Taylor, for the loan of the block. 53 EXCAVATIONS AT RECULVER, KENT, 1951 ollae with combed outer surface (Plate 1(b)) and of hand-made vessels with backing of crushed flint (Fig. 2 (2, 4 and 5)), and secondly, fragments of Roman pottery of first century date in direct association (Fig. 2, (3 and 6)). No structural features were noted in this layer with the exception of two parallel shallow trenches cut into natural at the north end of the section. Similar combed ware appeared in Claudian contexts at Richborough1 and at Bigberry Camp, the late Iron Age hill-fort near Canterbury, where its association with similar heavily 1 Richborough, II, p. 97, where there is a disoussion of the origin and incidence of this type of ware. 54 EXCAVATIONS AT RECULVER, KENT, 1951 gritted pottery was held to mark Belgic dominance over a native population with lingering Iron Age A traditions.1 The conclusion which may be drawn from the appearance of these wares at Reculver is that they represent occupation of the site on no impressive scale by a native population, probably in the decades immediately before and \&-l 10 ^ II FIG. 2. Reculver : pottery from section through south wall, 1951 (£). after the Roman conquest. The same layer was present in the first section at a distance of 380 ft. to the west which gives a slight indication of the extent of the occupation. The top of this layer represented the ground surface when the fort wall was built. The structural sequence appeared clearly in the section and, for as much of the wall as still survives, was in two stages. 1 Arch. Cant., XLVHI (1936), p. 167. 55 EXCAVATIONS AT RECULVER, KENT, 1951 Firstly, a foundation trench was cut down through the old ground surface just into natural, and on the bottom was laid a layer of beach pebbles some 15 in. thick at this point, to serve as a foundation. On this were laid the first three courses of flints to carry the wall up to a height of just over a foot; here the inner face was stepped back to give an offset of nearly a foot, after which the wall was carried up for a further six courses to a height of 4 ft. 3 in. Light brown sand (3) was then heaped against the face to this height and the second stage reached. This was marked by the appearance on the old ground surface of a thick mortar-mixing floor (4) (Plate I (a)), droppings from which led to the face of the wall on the surface of layer 3. The wall was now carried up to a height of 5 ft. 6 in. at which point it was again stepped in a further foot to give a second offset, marked by a sandstone course at its outer angle, and finaUy to its present height of 7 ft. 8 in., where a sandstone course again appears. How much higher the wall stood in its original state it is impossible to say but it is likely that it was at least 20 ft. in all.1 A similar line of mortar droppings on a low bank of sand was observed by Major Home in 19272 and can be interpreted as an economical means of wall-building without the use of scaffolding.3 It did not appear in the first section but there was there, some 24 ft. east of the inner face, a small mound of beach pebbles on the old ground surface, representing unused building material. Next came the construction behind the wall of an earth bank but whether as an intermediate stage in the building of the wall or after its completion it is impossible to say. It was composed of alternate layers of yellow and brown sand (5-9) which were capped by a thick layer of light brown clay (10). On the tail of layer 9 rested a thin layer of tightly packed mussel shells, possibly cookhouse refuse. Finds from these layers were so rare (despite the hopeful sieving of the mortar floor) as to prompt the conclusion that the earth which composed them had undergone no previous occupation. The same conditions held in the first section. Behind the bank appeared a thick unstratified filling of brown clayey sand (11) which again was relatively sterile, the single datable find being a fragment of terra sigillata of Antonine date (Fig. 2 (8)). Overlying all was a layer of black topsoil (12) containing a few probably late Roman sherds (Fig. 2 (9, 10 and 11)). The persistence of the military tradition at Reculver finds expression in the slit trench (13) which was probably cut in the 1914-1918 war. To summarize, the excavations disclosed original occupation of 1 V.C.H. Kent, III, p. 20. 2 Arch. Journ., LXXXVI, p. 263, Fig. 9. 3 The same technique appears at Colchester (from information supplied by Miss K. M. Richardson, F.S.A.). 56 EXCAVATIONS AT RECULVER, KENT, 1951 the site by a people with lingering Iron Age A traditions but under Belgic influence, an occupation which seems to have ceased not very long after the conquest, at least at the two points examined, and to have been followed by an intermission of unknown length before the construction of the fort wall. Originally this probably stood to a height of over 20 ft. and was about 10 ft. thick at the base, narrowing to about 8 ft. above the two inner offsets. Behind was a clay-covered earth bank which originally may have reached half-way up the wall. In general, it does not seem necessary to modify the conclusion that the fort belongs to the Saxon Shore period. Its closest parallel is the Shore fort at Brancaster, Norfolk, which had similar early-looking defences ; but excavation pointed to their construction in the mid-third century.1 ' THE DITCH SYSTEM (The following section is based on the observations of Mr. H. E. Gough of the Heme Bay Library and Museum who has most kindly made his notes and drawings available to me for inclusion in this report.) One would have expected, by analogy with other Saxon Shore forts, that the walls at Reculver would have been surrounded by at least one ditch. The only direct statement that such existed is by Gough, who, in his expansion of Camden's Britannia, speaks of the church at Reculver being " environed by the vallum . . . surrounded by a ditch " but it is not known on what he based his remarks. In any event, the lowering of the surrounding ground level aheady referred to has successfully obliterated any surface indications of its existence, at least to observers on the ground, and it was not possible to continue the sections described above in order to trace it. In 1948, Mr. Gough noticed on an air-photograph taken in 1920 (by Aerofilms Ltd.—Ref. No. : 531 ; Title : Heme Bay) a suggestive pair of parallel dark markings on the south and east sides of the fort which prompted him to make further investigations. He was helped by the action of a storm which had partially stripped the protective facing of clay and faggots from the cliff outside the east wall, revealing the original soil for some distance. In the section thus exposed appeared a wide V-shaped excavation with a filling of dark earth containing fragments of red tile and animal bones, in distinct contrast to the surrounding sand. The inner lip was approximately 20 ft. out from and 2\ ft. below the bottom of the wall and the bottom apparently about 36 ft. out and 10 ft. below, though it was not possible to determine the latter precisely. The outer face of the ditch was observed but not the outer lip ; probably the maximum width in its present wasted state 1 Antiq. Journ., XVI (1936), p. 452. 57 EXCAVATIONS AT RECULVER, KENT, 1951 is between 30 and 35 ft. Confirmation of its presence on the south side of the fort was obtained in the summer of 1950 when a pipe-trench and drainage pit were dug some 50 yards west of the east wall; the evidence for this was only verbal, however, and it was not possible to take any detailed measurements. Apparently the pit had cut through a layer of oyster shells serving as a line of demarcation between two kinds of soil, and the lowest point in this layer was again about 35 ft. out from the foot of the wall. Probably the pit had been dug into the centre of the ditch. It is thus possible to say that there is strong evidence for the existence of at least one ditch and there may in fact have been two, as at Richborough,1 if the air-photograph evidence is accepted. If so, it is likely that the ditch observed was the inner as the horizontal distance between the inner lip and the wall is insufficient to allow the interpolation of the second ditch. What its original dimensions were it is impossible to say because of the lowering of the ground-level, but if it is conjectured that the ground surface was originally 3 ft. above the base of the wall, this would give a berm of 10 ft. and a ditch 55 ft. wide and 13 ft. deep. These dimensions presuppose a level berm but from the disproportion between the conjectured berm and ditch widths one is tempted to suppose that it actually sloped downwards from the wall. THE FINDS Pottery (Fig. 2 (1-11)) 1. Incomplete rim sherd of wheel-made pot in hard grey ware with polished black surface. The internal thickening of the rim suggests a native cooking-pot of the first half of the first century, e.g., Camulodunum, Form 256. 2. Rim of hand-made pot in black ware with brown outer surface and backing of crushed white flint. Vertical rim, flattened on top, with small external bead and internal folding. Cf. Bigberry,2, Fig. VI, 107, where this type of rim is held to denote a lingering Iron Age A tradition. 3. Upper part of wheel-made pot in light red ware ; plain everted rim on nearly vertical neck. The ware is Roman so that it can probably be assigned to a mid-first century date. 4. Fragment of a handle in reddish-brown ware with dark brown outer surface and backing of crushed white flint. 5. Base of hand-made pot in grey ware with dark brown outer surface and backing of crushed white flint. 1 Richborough, IV, p. 66. 2 Arch. Cant., XLVIII, p. 151. 68 EXCAVATIONS AT RECULVER, KENT, 1951 6. Beaded rim, slightly undercut, in hard grey ware with slightly polished surface. The ware is thoroughly Roman indicating a mid-first century or later date. Cf. Camulodunum, Form 266. (1-6 came from layer 2 and were associated with the sherds of combed ware illustrated in Plate I.) 7. Everted rim in hard grey ware. (Layer 8.) 8. Samian sherd—Curie Type 21. Antonine. (Layer 11.) 9. Moulded undercut rim in hard grey ware with polished surface. 10. Rim of flanged dish in hard grey ware with dark brown surface. A common third and fourth century type. 11. Rim of lid in hard grey sandy ware with dark brown outer and light brown inner surface. (9-11 came from the topsoil-layer 12.) Animal Bones A few animal bones were found in the make-up of the earth bank and these have been identified by Miss J. E. King of the British Museum (Natural History), who kindly exammed them, as the remains of horse, sheep and ox. Mortar Samples Four samples of mortar were taken during the excavation of the section through the south wall of the fort and submitted to Dr. Norman Davey of the Building Research Station who kindly examined and reported on them in detail. Sample A was taken from a coat of mortar covering the lowest three courses of the inner face of the wall, B from the mortar mixing floor (layer 4), C from between the lower courses of the wall (i.e., beneath layer 4), and D from between the upper courses (above layer 4). The purpose in taking these samples was to determine whether there was any significant change in the mix associated with the introduction of the mixing floor. In the event this was found not to be the case. Samples B, C and D were found to be of similar composition, viz., lime with an aggregate of white sea sand and shingle. The proportions varied to some extent and samples B and D contained a trace of tile but a graph of the particle distribution of the aggregate showed much the same outline for all three samples. They were also similar in composition to samples taken from the east wall by Dr. Davey in 1948. The aggregate in sample A, however, was a buff-coloured pit sand and its particle distribution dissimilar from that of the other samples. The purpose of the mortar coating from which it was taken is hard to estimate but it may have served as a protective rendering. 59