Roman Folkestone Reconsidered

ROMAN FOLKESTONE RECONSIDERED By S. E. RIGOLD, M.A., F.S.A., F.R.HIST.S. PAETS of this paper are frankly speculative and wiU have served their purpose if they can some day be tested and, if necessary, disproved. "Roman Folkestone is the title of a book by S. E. Winbolt,1 as discursive as his contributions to this journal2 were succinct. Both describe his excavation, in two months of 1924, of a large, but not excessively large nor luxurious, corridor villa, with an annexe containing a bathsuite, facing south-east over East Wear Bay (N.G.R. TR 242370: A on Fig. 1). Others are more competent to re-interpret his reports in modern terms: the only comment here is on the coin-list,3 which is misleading in premises and conclusion. It is adulterated with some third-century Alexandrian pieces, which came not from the viUa (such things never do) but, reputedly, from a near-by aUotment. No coins were found at the viUa between PhiHp and weU into the Constantinian period: the common base 'radiates', whether unreformed or reformed (in this country these are usually represented by Carausius and AHectus), were quite absent and that evidence, by itself, would suggest a long gap in the occupation. Immediately after the excavation the Borough built a capacious shed over the mosaics, at a third of the cost of Sir Charles Peers's recommendation to roof over the whole, and for a year or two the site was a 'money-spinner'. In fifteen years landshps had destroyed most of the baths. Then enemy and 'friendly' action together destroyed the sheds and the troops used the timber as fuel. FinaUy, after one halfhearted attempt to consohdate them, the crumbling and overgrown remains were back-filled and grassed over in 1957. A 'sensation' in its brief day, a 'burden on the rates' thereafter, the rise and faU of Winbolt's 'Roman Folkestone', easily within the compass of one memory,4 evokes Gibbonian meditations on the transience of some of our more pubHcized archaeological subjects. 1 S. E. Winbolt, Roman Folkestone, a record of excavation of Roman villas at East Wear Bay, with speculations and historical sketches on related subjects, London, 1925. 2 Arch. Cant., xxxvii (1926), 209-10 (interim); xxxviii (1926), 46-50. 3 Op. cit., in n. 1, 79-84: the stamped tiles, pp. 103-6, must also now be reconsidered, in the light of G. Brodribb's study, Sx. Arch. Coll., ovii (1969), 102-26. * The writer lived at Folkestone from 1921 to 1925. 31 7 S. E. RIGOLD THE WESTERN VILLA-COMPLEX (?) Winbolt knew that there was more to Roman Folkestone than his viUa in the eastern coomb and cited the report by Canon Jenkins5 on a group of buildings (N.G.R. TR 235367: B on Fig. 1), some 600 m. away, on the west slope of a ridge and south of the now defunct Folkestone Junction station. These had been revealed piecemeal in commercial works between 1869 and 1875 and a small section of the most interesting one was re-examined in 1952, by Mr. Frank Jenkins, F.S.A., whose comments are most gratefuUy acknowledged. They probably add up to a second viUa with its adjuncts but the main dwelhng was not identified. The northernmost, last found and best recorded was a hypocaust-chamber with a hexagonal room adjoining it, presumably part of a bath-block. Two others were fragmentary and beyond resolution, but the other two, of which one was seen again in 1952, lay a Httle higher up the ridge and certainly within Chapel Field, the site of St. Bartholomew's chapel, which Leland could stiU observe, with a 'Hkehhood of further building' around it.® These were not planned in detail but the observant Canon, comparing his recoUections with those of others, leaves Httle doubt about their general form: one was circular, the other rectangular, with a deeply sunken floor and a lighter addition at the west. Not to go beyond Kent, the analogies in size, shape and position with the pairs of sepulchral structures (not contemporaneous) at LuUingstone7 and Warbank, Keston,8 immediately suggest their original purpose, and the treatment of the rectangular building is strongly reminiscent of what happened to the rectangular tombs at LuUingstone and Stone-by-Faversham.9 The easiest solution is that a Roman tomb-house became the nucleus of St. Botolph's: a nave of sorts was added, burials continued in and around it, as both Jenkinses observed, and those 'in the walls' of the deep chamber may be accounted for by internal alterations.9* The Canon noticed a hard, white mortar, reminiscent of that in the surviving, early Saxon, building at Lyminge and a softer, red one, while the masonry had occasional bondingcourses of tile, as at Stone. Only the Botolph dedication is not conspicuously early, but that could have been changed. 5 Arch. Cant., x (1876), 173-7. 6 Leland's Itinerary, ed. L. Toulmin Smith, London, 1909, iii, 64; the whole short passage on Folkestone is quoted by Hasted, quarto ed., viii, 171. 7 Full report in preparation: cf. Arch. Cant., Ixxvi (1961), li; Ixxxi (1966), 242. 8 Archceologia, xxxvi (1855), 120-28: notes on recent excavations in Current Archaeology, ii (1970), 73-5 and Kent Arch. Rev., 21 (1970), 21-3. 9 See below, pp. 38-9, and n. 34. "a Mr. F. Jenkins, F.S.A., reports that inhumations, lying east-west, had been inserted into the tops of the foundations. It is not suggested that the Roman building was visible in the final form of St. Botolph's. 32 ROMAN FOLKESTONE RECONSIDERED THE WIDER ASPECT OF ROMAN FOLKESTONE Winbolt had Httle to say on the several first- and second-century pottery-groups found within the borough, duly noted in the Victoria County History10 and Ordnance Survey records, and noted below in their geographical context. It matters Httle; the claim of this paper is that even the two villas are of secondary importance, that the real interest of Roman Folkestone Hes elsewhere, and that honest and admissible records of i t have been ignored. Winbolt's topographical speculations are sensible, as far as they go: he sees an advantageous position for signalling and his 'preferred' Hne of land-communication is essentially that supported by Mr. I. D. Margary.11 But the area must be seen as a whole and the relevant evidence is more than topographical. Fig. 1 is intended to convey the setting. I t is a sketch-map, not to be used for precise measurement, compounded from enlargements of early Ordnance maps, including that of 1819, details from Andrew, Dury and Herbert's map of 1769, and published excerpts, re-drawn from manorial plans of 1698 and 1782,12 and provided with sketchcontours, made entirely by eye, at intervals of about 9 m., the interval that best accommodates the features. The pubhshed 50-foot contours do not show the depression around Ingles demesne nor the summit of Sandgate hill. In the Gault clay bottom, between the bluff of Lower Greensand to the south and the chalk downs to the north is a miniature riversystem. The fast rivulet caUed 'Pent stream', though penning for mills has Httle altered its course, drains off the porous rocks, on the north by feeders which have cut guUies in the Gault, in one of which viUa B lay. Shght traces of first- and second-century settlement, and perhaps pottery-workings, have come from the clay bottom, while on the Greensand slope, at least three spots about the 36 m. contour,13 burials have been found (x on Fig. 1). The rivulet breaks southward to the sea through the Greensand cHffs, which were exposed to slow, steady erosion. This has been more or less arrested since the eighteenth century and can hardly have set the cHff back as much as 100 m. since Roman times. The Gault cliff, on the other hand, is stiU very much 'alive' and in solution at its base. The rate at which great shoes of landsHp have bitten into villa A in the last fifty years suggests that something over 500 m. has gone since 10 V.C.H., Kent, iii (1932), 114. 1 1 1 . D. Margary, Roman Roads in Britain, i, 2nd ed., London, 1967, 50. 12 Belonging to the Radnor estates; redrawn in Folkestone—Past and Present, Folkestone, 1954, 29 and following p. 24, and a portion of the 1782 plan much better reproduced in W. H. Elgar, A Record of a medieval House which stood until 1916 upon the Bayle, Folkestone, Folkestone, 1916, reprinted 1921, 46. 13 Julian-rd., 1918-20; Grammar School playing-field, 1934; Football-ground, 33 OS M O O 0 1 KM. FIG. 1. Sketch-Map of the Folkestone Basin; Sketch-Contours at approximately 9-metre Intervals. ROMAN FOLKESTONE RECONSIDERED Roman times. Both villas must be seen in secluded, relatively damp, and perhaps wooded, situations, while the Greensand shelf has always been exposed and well drained. Here, it is suggested, on the only sandstone headland east of Fairhght, was the focal point of Folkestone, Roman as well as Saxon. THE SAXON MINSTER AND ITS ROMAN ANTECEDENTS Kent had five, at one time perhaps seven, proportionately far more than any other kingdom, ancient nunneries of a pecuHar, Frankish type, each founded for the benefit of a lady of princely rank and her retinue, with an attached congregation of priests, which made them into 'double' houses.14 The tradition that Folkestone, founded for Eanswith, daughter of Eadbald, was the first of all, shghtly earHer than Lyminge, founded for the widowed Aethelburh, is confirmed by the precedence of the abbesses in witnessing charters.15 These houses and their male equivalents were bridgeheads of Christian and sub- Roman practice, and it is in keeping with this and with the precepts of Gregory that they and other bases of missionary activity were set up, where possible, in the physical protection of Roman structures —a civitas Dei, a turris fortissima was achieved by a kind of adoptive consecration. On the analogy of Reculver, Ythanchester, Burgh and Richborough, of Castor and Dover, of Whitby probably, of Walton16 more than possibly, it is reasonable to look for the same at Folkestone and there are material records to support it. Like others, Folkestone minster became untenable under Danish raids, and Christ Church enjoyed its revenues for a time. After 1095 an ahen ceU, of Lonlay, took its place.17 Always a poor house, something of its barely conventual-looking buildings lasted until the eighteenth century, south and west of the parish church,18 whither the community was moved in 1137 from the site, some distance to the east, held by all later tradition to have been that of St. Eanswith's minster. This (C on Fig. 1) was the castle-yard or Bayle, the triangle lying seaward of the street so-named. Leland in the 1540s19 noticed that one side was ditched, and a desire to refortify it in Stephen's day may have 14 See S. E. Rigold, 'The Double Minsters of Kent and their Analogies', J.B.A.A., 3rd ser., xxxi (1968), 27-37. 15 Dr. G. Ward noticed that the Kentish abbots attested in order of seniority of foundation; Arch. Cant., lix (1949), 19-25. 16 S. E. Rigold, 'The supposed See of Dunwich', J.B.A.A., 3rd ser., xxiv (1961), 55-9. 17 V.O.H., Kent, ii (1926), 236-7. 18 A barn, a dovecote and, in the 'mansion house', perhaps something of the undistinguished hall noted in the V.C.H. (see note 17); the site is discussed in the light of the manorial plans (see note 12) and plans of 1626 and 1628(?) by W. H. Elgar, The Ancient Buildings of Folkestone and District, 1st ser., 1921, 6-13. 10 See note 6 for the whole passage. 35 S. E. RIGOLD precipitated the move as much as erosion. In the later eighteenth century much of what remained of it was occupied, and much disturbed, by a fort,20 but part stiU remains unencumbered. But Leland also noticed within it a laden burial ground, torn open by the sea, as at Reculver, and impressive ruins, which he had no doubt were ecclesiastical and ascribed to the ancient minster ('a solemn old nunnery'), not to the Norman priory. These included much Roman bonding-tile-—'great and long Britons' bricks'. Lambarde,21 a Httle later and apparently independently, reported much the same. Erosion continued, but, as late as 1722, Stukeley22 saw 'two pieces of old wall' on the edge of the 'terrible' chff, 'seemingly of Roman work' and noted the common occurence of Roman coins. Whether what they saw was Roman, or early Saxon in Roman fashion, or early Norman re-using Roman material, or aU three in juxtaposition, Stukeley's opinion deserves respect. He could judge Roman waUing and here had slight reservations. Antiquaries can still argue whether or not there are Roman elements in St. Martin's and St. Pancras' at Canterbury. However, certain things can be said with confidence about site C: it produced Roman smaU-finds; it had a source of good bonding-tile, probably nearer than villas A and B, 1 -5 and 1 -0 km. distant; an early minster would prefer to make use of Roman structures; above aU, the position, commanding views from Shakespeare CHff to Fairhght Head, was one neither the Glassis Britannica nor the Comes Litoris could afford to neglect, at least for a Hghthouse or a signal-station, with adjuncts—there was hardly room for a sizeable fort. It merits at least a dot on the Ordnance map of Roman Britain. THE AQUEDUCT The origin of this, on present knowledge, is as debatable as the interpretation of the lost ruins in the Bayle. Both would admit of three possibiHties: Roman, probably early Roman; early Saxon; early Norman, But the concept is romanizing and the continuity of this Httle-known monument is, in any case, impressive. It is a contour-aqueduct, now mostly in a covered conduit, but open until the early nineteenth century, foUowing the south bank of the rivulet, about 100 m. from it, and dropping gently to the 30 m. Hne. Where its surroundings were visible in road-works beside Shellons Street in 1972, it was seen to run over unindurated sand. It would not work at aU unless heavily Hned with clay or masonry, and in Hasted's day much of it was aheady in a brick channel. The magnitude and accuracy 20 Shown on the 1782 plan; see note 12. 21 W. Lambarde, Perambulation of Kent, 1826 ed. (reprinted London, 1970), 150-4. 22 W. Stukeley, Itinerarium curiosum, 2nd ed., London, 1776, I, 131. 36 ROMAN FOLKESTONE RECONSIDERED of the work was no mean feat at any period. The beginning of the course, in the broad head of the vaUey, has probably been disturbed by waterworks and brick-fields. Until the making of the by-pass, it could be traced a Httle north-west of Broadmead, but the true origin was surely a spring off the Greensand, not a field-sump. It wound along the contour, bending sharply into the Ingles depression, and then foUowed the length of what is now GuildhaU Street to a Cistern House on the site of the present Guildhall. This was its nearest point to the later Priory site, after which it swung in the other direction and ended in a pond on the edge of the Bayle. It was known throughout as St. Eanswith's Water-course, though the last stretch was sometimes caUed the Town Ditch.23 The hagiographical tradition is strong, and Lambarde aheady cites it out of the Nova Legenda Angliae,u how the saint 'hailed and drew water over the hills . . . to her oratory', for her personal use and not by a posthumous miracle—or, more wonderfuUy, made water flow uphiU. Certainly, the Town maintained tbis sacred service and had its cistern-house, and it is possible that we have no more than a common effort of lord, priory and burgesses in or after the Norman period, under the Saint's protection. It was not an ordinary monastic conduit, for the sole benefit of the later priory;25 its termination at the Bayle links it with the ancient site. Given that it was a weU-laid work that could be maintained or reactivated indefinitely, there is no need to dismiss the tradition associating it with St. Eanswith. If the tradition is sound, the question remains whether such a new work was possible in the seventh century, or whether it was a Roman work put back into use. Roman skiUs, even perhaps in finding levels, were not all lost in Merovingian Gaul. And what would the king of Kent not do for his daughter? Yet, on balance, the recovery of an existing work—perhaps the choice of site for that reason—seems preferable. Mracles depend more on prognosis than manipulation. This conjecture must stand by itself: it neither reinforces nor is reinforced by the aheady strong case for something substantial and Roman on the headland. Given the serious possibihty of a Roman origin for the aqueduct, comparable aqueducts, at Lincoln,26 or Dorchester, or, more pertinently, at northern mihtary sites, would point, as much else at Folkestone, to a relatively early Roman date. 23 As on the plan of 1782; see note 12. 24 See note 21. 25 I t does not seem to overflow in the direction of the claustral drain; in any case, a conduit for a small house would seem unlikely when that of Christ Church was such a wonder. Later, conduits became usual in urban religious houses, including friaries; the London Charterhouse possesses a medieval plan of its water-system. 20 F. H. Thompson, in Aroh. Journ., cxi (1954), 106-28, gives references for others, too. 37 S. E. RIGOLD The pond is now fiUed up, where cygnets played in the early 1920s, which, it is said, was boarded over to make a dancing-floor upon Gooseberry Fair.26a What may be the only Roman aqueduct in Britain that had never, with a Httle coaxing, ceased to flow Hes mortified with hygiene. THE NAME OE FOLKESTONE The primary form, Folcanstan, is weU attested; it occurs not only in the PrivUege of Wihtred, which is only known in later versions,27 but in two unimpeachable early ninth-century texts, the wiU of Abba the Reeve28 and a deed of Archbishop Wulfred.29 It is a stdn name: its second element is properly represented by the modern stone and, as such, quite rare. 'Stone', as in Tilmanstone, may often be a disguised tun. Stan can cover many things: a mass of rock, whether a natural outcrop or a megaHth; a workable source of stone (Folkestone beds, though not very good, were quarried for such expendable things as baUs for engines of war) ;so a paved, and therefore Roman, road, or a milestone beside it, as in the many Stone Streets and Stane Streets, Watton-at-Stone (Herts.) or Staines (Middlesex, until it was partitioned). The contention here, hinted at by EkwaU31 and more firmly by Cameron,32 is that, in Kent, if only there,33 it almost invariably impHes a stone, and therefore Roman, building (compare the current Flemish usage of Steen for a domus lapidea, keep or castle) and that this can often stiU be identified, in some cases as the exorcised and consecrated nucleus of a church and, by extension, of a parish or chapelry. It will not take long to examine every Kentish stdn in turn. Stone-by-Faversham. This, after conclusive excavation34 by Lord Fletcher and Colonel Meates, has become the locus classicus for the incorporation of a Roman building, almost certainly a tomb-house, into a Christian church. This seems to have been effected in or before the 20a The fair, held on 28th, 29th June, was granted in 1383. Mr. C. P. Davies, whom I thank for his comments, thinks that it ceased to be held on the Bayle c. 1857, and knows of no firm authority for the dancing-floor tradition. 2' Birch, Cartularium Saxonicum, no. 91; P. H. Sawyer, Anglo-Saxon Charters, no. 22. 28 Birch, no. 412; Sawyer, no. 1482. 29 Birch, no. 378; Sawyer, no. 1454. 30 Arch. Cant., x (1876), civ, citing Rymer, Foedera, iii, pt. 2, 316. 31 E. EkwaU, Oxford Dictionary of English Place-names, 3rd ed. (1947), 416; 'sometimes . . . a stone monument'. 32 K. Cameron, English Place-names, London, 1961, 116; 'occasionally', Roman buildings. 33 But Cameron cites Stansteads in Essex and Herts, and Mr. Anthony Fleming has argued the same independently in a thesis about sites in north Bucks., e.g. Stanton Low. 34 Antiq. Journ., xlix (1969), 273-94. 38 ROMAN FOLKESTONE RECONSIDERED eighth century. It may weU have been this building that gave its name to the chapelry, rather than Wathng Street or what remained of the Roman ribbon-township beside it. Stone-by-Dartford. The parish contains a scatter of Roman remains, of settlement as weU as burial, including a building in Stone Castle Quarry,85 but its nucleus—church and manor-house—lay 1-7 km. from Wathng Street. The famous quire of the church is faced in fresh material, but the nave, nearly as fine and shghtly earHer, contains much re-used tufa and Roman tile, plentiful enough to suggest a substantial Roman building on or near the site. Indeed, the slope on which the church stands would suit a mausoleum. Stone-in-Oxney. The church was totaUy rebuilt in the later fifteenth century, but it contains a Roman altar with a buU, presumably Mithraic,36 which has been moved around, but never far from the church, since the early eighteenth century, and, if folk-memory is to be respected, was dug up in the chancel.37 If so, this is hardly a case for one monohth making a stdn name. Conceivably, it might have been imported as a block for the rebuilding and rejected, but if in situ it would not have stood alone. The easiest solution is that the church occupies the site of a mithraeum. Though no other finds are reported, the strategic site above the Limen estuary caUs for a subsidiary military occupation at the right period. Stanes in Thanet, chapel of St. John of.38 This is apparently an early reference to what is now Margate parish church; if so, c. 1200 Margate was known as Stanes. Nothing Roman is reported from the spot, but the church and surroundings of the inescapable haven are so altered by building that this is worth Httle as negative evidence. Stane, where in 991 Olaf Tryggvason attacked before going on to Sandwich, Ipswich and, finally, to the battle of Maldon.39 This is usually taken to mean Folkestone itself; Margate seems a possible alternative. Cuxton (Gucolanstan). The church contains conspicuous Roman materials and finds, including structures, are persistently reported from the extended churchyard and just below it, which has inhibited proper excavation.40 The evidence seems to indicate a house, but the steep site of the church itself is more suitable for a mausoleum. 86 Arch. Cant., Ixxxi (1966), 136-90. 30 Arch. Cant., xlvii (1935), 1-12. 37 Ibid. 3. 38 Arch. Cant., xx (1893), 66. 30 Anglo-Saxon Chronicle, ed. Thorpe, London, 1861, sub anno, 993 (Parker- Winohester), text only. 40 Aroh. Cant., xxv (1902), Ixvii, and information from Mr. P. J. Tester, F.S.A. 39 S. E. RIGOLD Keston (Cysse stanes gemcero; Cystaninga mearc). The church is on the boundary (mearc), but near Keston Court and the nucleus of the Cyssestan settlement is the large viUa with its well-known mausolea,41 which surely gave it its name. Lullingstone. Assuming that the Textus Roffensis4-2 is deHberate in distinguishing between LuUingestuna (for St. Botolph's) and LoUingestann (for St. John's) the now-destroyed church and now-deserted settlement just above the villa has a stdn name. It is probably to be identified with the capella de Stanes attached to Eynsford;43 at least, there is no other known candidate. The church stood on the site of the 'temple-mausoleum', making use of its materials and probably part of its fabric, at least as a foundation. Maidstone. This had a concentration of Roman settlements, first surveyed by Canon Scott Robertson,44 but no hint of a township. A building near the East station was re-examined recently, but the grander southern one, in Stoneborough, approached by Stone Street was probably the eponymous building. Nothing Roman reported near the church. Teston (Terstan) had an extensive building some 500 m. south-west of the church. The Bredenstone, Dover. This, the second Pharos, on the Western Heights, as late as HoUar's time45 was a recognizable structure, not the shapeless lump of rubble it had become by the late eighteenth century. In recent centuries Lords Warden have sometimes been invested beside it and tradition points to it as a place of justice for the Borough in earHer times. These facts are adduced to suggest that it has long been considered a suitable place of assembly and seems as good a candidate as any for the Fracta Turris where, according to WiUiam of Poitiers, the Conqueror encamped after receiving the submission of the men of Kent.46 The directions given are imprecise and need not imply that the Conqueror deigned to go half-way to Canterbury to receive them. Breden and fracta are open to guesses, but stdn (cf. Steen) seems a possible Old EngHsh equivalent of turris. What other word might be suggested? 41 See note 8. 42 Arch. Cant., xliv (1932), 47—Dr. Ward's comments on this text. 43 The Domesday Monachorum of Christ Church, Canterbury, ed. D. C. Douglas, London, 1944, 108. 44 Arch. Cant., xv (1883), 68-80. 46 I t appears in the foreground of his view from the Western Heights, c. 1640 and more conspicuously in a view of c. 1540 (Hatfield Coll.). 40 Qesta Willelmi ducis . . . , in Scriptores Rerum Qestarum Will. Conquestoris, (ed.) J. A. Giles, (1845), 140. 40 ROMAN FOLKESTONE RECONSIDERED Stansgate Creek passes through the drowned potteries in the direction of Lower Halstow, which had intensive Roman occupation with at least one sohd building. Stanford, on Stone Street, is self-explanatory. Stansted, despite its 'Coldharbour', has produced nothing and is not very promising, unless for a flimsy building, as at North Ash. At Chiddingstone nothing Roman is known or expected. At Stonar the stones are certainly natural. But these negative 'controls' do not invaHdate the remarkable coincidence between stdn names and some of the most prominent Roman sites in Kent. On the strength of building B alone Folkestone would quaHfy as a typical Kentish stdn, with possible re-use of a mausoleum in a Christian chapel. This, however, must probably yield in antiquity and prominence to whatever St. Eanswith found on the chff and adapted to her purposes. What of Folcan% Despite Wallenberg's47 hankering after an unevidenced 'weak' form, modern toponymists are generally agreed that the traditional interpretation—lapis populi—is impossible. Folc (folk) would produce folces, not folcan, and the late Saxons, writing Folces stane, etc., had already fallen into the folk-etymological trap. A person with the rare name Folca was therefore conjured up.48 He cannot, technicaUy, be dismissed, but a survival of the Romano-British placename is not out of the question and as Hkely to have been remembered as Regulbium in Raculf. A name such as Volcania or Volcanium could easily have produced Folcan and would be not unsuitable for a beaconpoint. 47 J. K. Wallenberg, Kentish Place-names, Lund, 1931, 23. 48 E.g. by Ekwall, op. cit., in note 31, 174-5. 41