Earthwork survey and excavation at Boys Hall Moat, Sevington, Ashford

EARTHWORK SURVEY AND EXCAVATION AT BO YS HALL MOAT, SEVINGTON, ASHFORD* PAUL BOOTH and PAUL EVERSON INTRODUCTION Boys Hall Moat, more correctly historically known simply as 'The Moat', lies in Sevington parish near Ashford, close to its western boundary with Willesborough. It is a Scheduled Ancient Monument (Kent SAM 146) centred at N.G.R. TR 0 3004075. The site was surveyed and described by staff of the Royal Commission on the Historical Monuments of England in February 1990, fulfilling a request from Dr J.H. Williams, County Archaeological Officer.1 Subsequently in March-April 1993 the Oxford Archaeological Unit conducted a two-week excavation on a strip of land approximately 200 m. long (from north-west to south-east) and 4 m. wide, adjacent to the main Folkestone to London railway line but lying within the confines of the Scheduled Ancient Monument on its north-east side. The excavation, which was required as a condition of Scheduled Monument Consent before development, was commissioned by British Rail (Network South-East) in advance of construction work on the railway line. The excavation was directed for the Oxford Archaeological Unit by Miles Russell. The post-excavation work was carried out by Paul Booth, who is responsible for parts of the text not credited to other writers. The site archive and finds are currently held at the Oxford Archaeological Unit pending a decision on their ultimate place of deposition. • Published with the aid of a grant from the Oxford Archaeological Unit. 1 The fieldwork for a survey at 1:1000 scale was carried out by Paul Everson and Robert Wilson-North from the R.C.H.M.E. Keele office; a full report is deposited in the National Monuments Record under the reference TR 04 SW 2. The summary below is derived from the archive report by Paul Everson; Philip Sinton redrew the archive plan for publication. This account is published by courtesy of the Commissioners. 411 PAUL BOOTH AND PAUL EVERSON LOCATION AND ARCHAEOLOGICAL BACKGROUND (Fig. 1) The site lies some 3 km. south-east of the centre of Ashford on gently rising ground (between c. 40 and 45 m. above O.D.) on the north-east side of the river East Stour, in an area surrounded by pasture and setaside land at the time of the excavation. The underlying geology is principally Atherfield Clay, at the base of the Lower Greensand sequence, which outcrops locally in a narrow east-south-east to westnorth- west band parallel to the line of the river, between the underlying Wealden Clay cut by the river itself and the later Hythe Beds to the north-east. Until recently, knowledge of the archaeological background of the immediate area was confined to a number of finds of individual flint and stone implements. Sites now known include an extensive area of crop-marks, including possible ring ditches, a rectilinear enclosure and other linear elements of uncertain date, centred only 900 m. due east of Boys Hall. This complex is associated with field-walking finds of Neolithic and Bronze Age flint, Iron Age, Roman and medieval and post-medieval pottery.2 The certain presence of a number of Iron Age and Roman sites even closer to Boys Hall has also been demonstrated, in evaluations carried out by the Kent Archaeological Rescue Unit3 and the Canterbury Archaeological Trust.4 Two Late Bronze Age-Early Iron Age sites have been found at Waterbrook Farm, to the south of Boys Hall,5 and a further site assigned to the Late Iron Age lies immediately north-west of Boys Hall. 6 South-east of Boys Hall two Late Iron Age-Belgic sites were located in the area of the Ashford freight terminal.7 The more northerly of these sites may have extended north-westwards towards Boys Hall. s Belgic to Early Roman activity occurred to the south-west of Boys Hall and to the south at Waterbrook Farm.9 In addition, the Roman road which runs in a west-north-westerly direction from Lympne passes some 1.5 km. south of the site on the south side of the East Stour. Recent work in this vicinity has produced more Roman features as well as evidence for 2 Oxford Arch. Unit, P rovisional specialist report on historic and cultural impacts for Union Railways Limited (1994). 3 KAR., 104 (1991), 74-77; also 'An archaeological evaluation at the Boys Hall Industrial Area development', unpublished report by J. Willson. 4 Arch. Cant., cvi (1988), 2; Arch. Cant., ex (1992), 375-6. s Arch. Cant., ex (1992), 375-6. 6 Willson, (op. cit., n. 2), 3. Arch. Cant., cvi (1988), 2. Willson, op. cit., n. 2, 3. 9 Ibid., 3-4; Arch. Cant., ex (1992), 376. 412 EARTHWORK SURVEY AND EXCAVATION AT BOYS HALL MOAT LIA/Belgic Belgic/early Roman Fig. 1. The site in its local setting. 413 PAUL BOOTH AND PAUL EVERSON Mesolithic and Neolithic flint working.10 There are no early medieval finds from the vicinity. R.C.H.M.E. SURVEY OF 'THE MOAT' Paul Everson (Fig. 2) The moated site is one of a number of such sites in the area, but, according to the assessment of the Monuments Protection Programme in advance of the fieldwork, was reckoned one of the best examples in the county at large. The results of the survey, by confirming and documenting more clearly its late medieval status and adding an early post-medieval phase including the elaboration of associated garden earthworks, can only have enhanced its archaeological importance. The garden earthworks may be the first to be formally documented by archaeological survey and published in Kent of this monument class, which has been the subject of increasing study elsewhere in the country.11 They well illustrate the changing expectations of the setting of a gentry residence in the sixteenth century and the aspirations of families that formed the backbone of county society of that period. The distinct manor in Sevington of which 'The Moat' formed the residential centre is traced by Hasted to Sir John of Sevington in the time of Henry III.12 It passed by marriage in the thirteenth century to the Barry family. The evidence for their splendid series of fourteenthand fifteenth-century brasses formerly in St. Mary's Church at Sevington certainly supports the belief that they were resident at The Moat throughout the later Middle Ages. 13 Successive members of the family held important public office in the county in the fourteenth, fifteenth and sixteenth centuries, including knights of the shire, sheriff, commissioner of the peace and lieutenant of Dover Castle, and they intermarried with other leading county families.14 The male line failed on the death in 1588 of Richard Barry, M.P. for Winchelsea and Dover 10 Canterbury's Archaeology 1992-1993, 41-2. 11 cf. (Ed.) A.E. Brown, Garden Archaeology, C.B.A. Research Report no. 78 (1991), especially the papers by C.C. Taylor and P. Everson. 12 E. Hasted, The History and Topographical Survey of the County of Kent, Canterbury 1790, vol. III, 280. 13 H.L. Smith, 'Some account of brasses formerly in the Church of Sevington', Arch. Cant., iv (1861), 117-22, incl. pedigree p. 122. Also C.R. Councer, 'The medieval painted glass of Mersham', Arch. Cant., xlviii (1936), 81-90. 14 Hasted, lac. cit. ; W.B. Ellis, Early Kentish Armory, Arch. Cant., xv (1883), 15; J.C: Browne, Knights of the Shire for Kent from A.D. 1275 to A.D. 1831, Arch. Cant., xx1 (1895), 212; T. Philipott, Villare Cantianum or Kent Surveyed and Illustrated, 2nd ed. King's Lynn 1776, 317. 414 EARTHWORK SURVEY AND EXCAVATION AT BOYS HALL MOAT RCHM lNGLANO limit of 1993 o• -=::1....01 -= =::--==s,,o.... -= ==--==--.;;100 metres Fig. 2. Plan of the earthworks with the location of the 1993 trench (R.C.H.M.E. copyright). 415 PAUL BOOTH AND PAUL EVERSON and Lieutenant of Dover Castle during preparations against a Spanish invasion, and 'The Moat' passed by marriage to Vincent Boys of Bekesboume.15 Nevertheless, the estate remained in occupation, until it returned in the late 1620s by marriage to Thomas Boys of Willesborough, who belonged to a different branch of the Boys family. 16 Thomas 'pulled down this ancient seat and removed the materials of it to rebuild his house at Willesborough', either in 1616 (according to Hasted), or by 1632, 17 thereby creating the house known as Boys Hall that survives altered but with a seventeenth-century core at N.G.R. TR 0281 4120.18 The earthworks of The Moat evidently remained intact until the early 1840s and were labelled as 'site of the ancient Mote House' .19 The north-eastern periphery of the site was sliced through by the construction of the Ashford to Folkestone line of the South Eastern Railway in 1842-43. The site was first referred to in the archaeological literature in 1880.20 At the time of survey the earthworks were situated in a single field of old pasture. The moat itself, 'a' on plan, lies at the south-east end of the pasture field on almost level ground, with natural slopes rising away to the north and north-west. It is broadly rectangular with overall dimensions of 55-60 m. by 70 m. and is currently water-filled without any break or causeway. Early large-scale map depictions also show it water-filled but in two unequal sections that were separated by sharply defined breaks, perhaps in origin causeways. These occurred symmetrically in the north-west and south-east arms, not centrally but off-set towards the north-east end. The continuous sheet of open water that the moat now presents in contrast to its earlier depiction raises a suspicion that it has been dredged. The moat is cut slightly into the natural slope at its north-east end, and is tapered somewhat so that its upslope side is some 5 m. longer than the south-west side. The water sheet of the south-west arm is also markedly broader. The north-west and south-east arms appear to taper north-east from 13 m. to 9 m. in width, but this may rather originate in a single constriction in each arm at the point of the possible causeways. The moat encloses a rectangular island measuring some 40 m. x 15 P.W. Hasler, The History of Parliament. The House of Commons 1558-1603, H.M.S.O. 1981; (Ed.) R. Hovenden, The Visitation of Kent 1619-1621, Harleian Society 42 (I 898), 39. 16 Hasted, Kent, vol. Ill, 280. 17 E.P. Boys Richardson, 'A seventeenth-century Kentish proverb', Arch. Cant., xxx (1914), 81. 1s D.o.E. Listed Building List: Ashford District, TR 04SW 3/137. 19 P.R.O., IR 30/17/323. 20 W.M. Flinders P etrie, 'Notes on Kentish earthworks', Arch. Cant., xiii (1880), 8-16, esp. 10. 416 EARTHWORK SURVEY AND EXCAVATION AT BOYS HALL MOAT 28 m. lip to lip. No masonry or significant change of levels is visible on the island to give evidence of the house that stood there, but quantities of floor- and roof-tile and mortar were observed during the survey and a few sherds of late medieval pottery were collected. Hasted's description of the moat as 'filled with water . . . inclosing strictly the scite only of it ' might imply that the hall completely filled the island, its outer walls rising from the moat itself (as, for example, at Ightham Mote). The Barrys would have had a dwelling to match their pretensions, and this is likely to have been of fourteenthand fifteenth-century date, even if modernised or added to in the sixteenth century. The distinctive broadening of the water at the moat's south-west end may well be an adaptation of the medieval moat contemporary with the creation of a number of other water features to the south-west and west of it. These represent an elaboration of the site in the form of gardens and courts presumably in the early post-medieval period. The ornamental features are laid out on relatively level or slightly rising ground to the north-west of the moat, and rise up to the railway line at the north corner of the field. They comprise an elaborate arrangement of ponds and a number of square compartments, the details of whose internal arrangements are not everywhere clear, but which may have incorporated more water features. Precise definition and interpretation of these features are made more difficult by their differential preservation. Substantial ponds in a 40-45 m. strip along the west edge of the earthworks and slight details associated with them are well preserved. To the north-east of these the earthworks have been smoothed or obscured. The most prominent earthwork features are a deep angled pond, 'b - c', with its associated terraces, and the raised terrace and adjacent earthworks at 'd' that continue the alignment of 'c'. Together they extend for some 180 m. in length, divided into three equal sections, the unit for which is the length of the south-west arm of the moat. The straight section of pond, 'b', lies on the alignment of the south-west arm of the moat; the similar length, 'c', is continuous with it and, despite being on a different alignment, articulated with it by the presence of broad terraces on either side which run the length of the whole feature. This angular change effects a transition from the established orientation of the medieval moat to the orientation of the added gardens; 'd' continues the alignment of 'c' and is defined by the same broad terrace on the south-west side. It is separated from 'c' by an outlet channel, itself large enough also to have functioned as a pond. Within the terrace that defines the south-west side of 'd' and parallel to it lie two low banks, both exhibiting a drop south-eastwards halfway along their length. They may represent either planting banks or the 417 PAUL BOOTH AND PAUL EVERSON dividing banks between channels of an elaborate water system. A raised area at the south-west end of the broad terrace which defines the north-west edge of the site, is perhaps the site of a garden building or viewpoint that looks down the alignment of 'd' and 'c'. Closely related with 'd' is the block of earthworks to its immediate east, which also exhibit good form and regularity. This is clearest in the square platform, 'e', measuring some 7 m. by 6 m. and surrounded by a broad, shallow ditch and an outer bank. As a water feature, 'e' would have the appearance of a miniature moat and it links from its south comer with the channels of area 'd'. A curving outlet channel from the east comer of 'e' leads down over three slight earthen steps that might represent a form of cascade. The square compartment to the south-east has disturbed traces of what may have been a circular pond or basin for this to empty into. A series of water arrangements to the east of 'c' similarly take up the alignment of this middle section of the gardens. They comprise principally two less clearly marked ponds and the three parallel water features are separated by approximately equal broad terraces. Details of form in this area, however, are less clear-cut or reliable. Scarps to the east and north have the smoothed appearance already described. They are individually unintelligible but appear in places to represent platforming and also have a marked rectangularity. The north-west side of the site is defined by a disturbed and slumped but still imposing earthen terrace, standing up to 1.20 m. high internally. It exhibits marked changes in form. To the south-west, it is a broad but well-defined terrace, that initially steps down southwestwards but then holds a reasonable level for some 25 m. The final 20 m. again steps up slightly, and this, combined with the natural fall in the ground surface south and south-westwards renders the south-west end of the terrace a prominent feature. To the north-east, however, the terrace becomes broader and even more disturbed, with a marked increase in height at the point where a low terrace abuts it at right angles. The curve east of the terrace scarp before it is lost under the railway appears to be a genuine comer, probably marking the beginning of the north-east side of the earthworks, which otherwise lies beneath the railway. Access to the site appears to have been from the north-east. It is marked in the earthworks by a wide linear hollow, 'f' on plan, possibly a double-ditched road, whose scarps are quite clearly truncated by the railway. No convincing trace of it now exists in the disturbed land on the other side of the track. This direction of access lends weight to the interpretation of 'g' as a bridging-point or stub of a causeway to the island, which the presence of tumbled masonry in the moat bank here also supports. It might also suggest a function for some of the 418 EARTHWORK SURV EY AND EXCAVATION AT BOYS HALL MOAT unintelligible features and platforms to the north of 'g' as the sites of service buildings. The rectangular area to the north-west of 'f' might appear to lie outside the area of formal gardens as defined and to form a close, court or paddock flanking the entrance way. Yet, it contains faint traces of changes of level only a few centimetres high, whose rectangular layout and orientation are in close sympathy with the garden features to the south-west and with the north-west terrace. It may, therefore, be a further garden compartment. In summary, the site exhibits in its earthworks evidence for a medieval moated residence - 'The Moat' - which has been adapted with an elaborate layout of gardens that include raised terraced walks and water features. Particularly impressive in conception are the carefully contrived transition from the established orientation of the moat to that of the gardens and the way in which the gardens are organised in relation to the equal three-fold division of its south-west side. A viewing point on the high south-west end of the north-west terrace gives a view both along the alignment of 'd', 'c' and 'b', and across towards the moated house with in the foreground at 'e' what may have appeared as a miniature version of that moat. On the evidence of comparable sites elsewhere, such elaboration would be expected to date from the later sixteenth or seventeenth century. Here, it clearly took place before the site's abandonment as a residence in the second or third decades of the seventeenth century. THE EXCAVATION (Figs. 3 and 4) The excavation was intended to examine and date features associated with the scheduled monument with the aim of clarifying the nature and sequence of the recorded earthworks, which were indistinct in this part of the site. Any sub-surface features were also to be examined. In the event, all surface features within the proposed development area were flattened before the excavation commenced. Clarification of the date and nature of the earthworks, therefore, proved impossible and effort was concentrated on the sub-surface features. The area was stripped down to relatively undisturbed archaeological horizons (the levelling already mentioned left a legacy of machine marks, particularly at the south-east end of the site) by a 360° excavator with a ditching bucket. All subsequent excavation was by hand. The excavation took place in generally poor weather conditions. A combination of this circumstance with difficulties of access and earlier machine damage meant that excavation of the north-west end of the site was not possible. The natural subsoil, variously numbered, usually consisted of 419 .i,.. N 0 see tren ch j􀀂- ---detail below ---1 10 1-·-·---·-·;;·r,·-·o:--;;-·-·-·-·ti-·􀀪----j·-􀀫\t·-·-·-·-·􀀬=􀀭􀀮c·􀀯f-·-􀀰4 ·-·-· -·-·-·-􀀨\_._._._______ --·-·- ---􀀩-􀀪--t1---·-·-·-·-/Lt . ..,r-\-·-·-·-·-·-·-·-·􀀫􀀬-- 10 0 trench detail _s 􀀺 e􀀻tion 1 ·-·-----·-·- 74 49 75 section 2 L _____ _ 10 scale 1:1000 0 47 65 18 36 50 m. 􀀁􀀂 section 4 --------·-·- j:_ e _ c t t ion 5 _ ___ :/.:' -·􀀃 .. -;; .! ·-·-·-·-'-· 10 m. i-·-; L_j ___6: :f!::􀀆 section 6 scale 1:250 Fig. 3. Trench plans. 􀀃 p to 0 0 􀀄 􀀃 p 0 z EARTHWORK SURVEY AND EXCAVATION AT BOYS HALL MOAT section 1 section 2 E s N 􀀋75 􀀌 75 ! section 3 SE NW 49'S,.....---=<......:.::c----, --------s - 67 -1,... ___ so_ _=-- -􀀿2 --1-. --------,.?'----"""---Ii-􀀍 􀀃 .. : 52 69 71 70 section 4 section 5 section 6 w E NW SE SE NW 455.Y" 64 S& 􀀎 0 1 m. scale 1:50 section 7 SE NW 26 24 􀁀 chalk 􀀏 􀁁 brick 0 3 m. scale 1:100 Fig. 4, Selected feature sections. 421 PAUL BOOTH AND PAUL EVERSON superficial deposits of yellow or yellowish-brown silty sand or silty clay over grey clay (one localised thin gravel deposit was also noted). Three possible tree holes (40, 56 and 77) were the only 'natural' features observed. The most significant archaeological features were concentrated in a group in the central part of the excavated trench. The largest feature, a north-south linear complex c. 4 m. wide, probably consisted of successive boundaries. The earliest of these was a wide, shallow possible cut (84), probably at least 3 m. across, the yellow-brown sand fills of which (52 and 71) survived to a maximum depth of 0.15 m. The existence of 84 is not certain, since it is possible that its fills represented subsoil disturbed by overlying features. It would, however, have served to establish the alignment which was then followed by a number of other features. On its west side the possible fill 71 was cut by a ditch (70/82), at least 1.20 m. wide and 0.45 m. deep, with an irregular profile. The fills of this feature and of a further north-south gully (69/80), some 0.65 m. wide and 0.30 m. deep, also with an irregular rounded profile, were barely distinguishable from the overlying deposit (50, etc.), of dark bluish-grey silty clay. This appeared to be the fill of a very wide, shallow cut (49), some 4 m. across but with a maximum surviving depth of only 0.22 m., which truncated all the other features. 50 and its analogous deposits produced a substantial group of Late Iron Age pottery, mainly in grog-tempered fabrics. Smaller quantities of comparable sherds came from the fills of the earlier features 84 and 69/80. The south terminal of a parallel feature (75) lay some 5.50 m. west of the north-south ditch complex. 75 was c. 1.15 m. wide and generally not more than c. 0.18 m. deep, but towards the terminal it deepened to 0.36 m. Its relationship to the larger boundary is unknown, but its fill of grey brown clay produced a single Late Iron Age sherd. East of the large ditch was a group of gullies typically c. 0.30-0.50 m. wide and with round bottomed V-shaped profiles up to c. 0.25 m. deep. One of these gullies, 45/47, ran parallel to the principal north-south boundary at a distance of some 8 m., with a second gully (47) roughly at right angles to it running east-west between the two. The fills of both these features produced Late Iron Age sherds. Slightly further to the east two gully terminals (64 and 65) projected from the north-east and southwest baulks of the trench, respectively, leaving an intervening gap of c. 2.20 m. Both gullies were quite substantial so there can be little doubt that the gap between them was intentional rather than being caused by differential preservation conditions. The alignments of these short lengths of gully suggested that they belonged to a curvilinear enclosure. Such an arrangement could have been contemporary with the north-south and east-west aligned features, but is perhaps more 422 EARTHWORK SURVEY AND EXCAVATION AT BOYS HALL MOAT likely to have been of a different phase. Neither gully produced any finds from its fill. North-west of the group of Late Iron Age features was a further linear feature (79), some 2.20 m. wide and 0.30 m. deep and aligned roughly north-north-east to south-south-west. It was filled with grey brown silty loam (78). The fill contained no finds apart from fragments of ceramic pipe. It is not clear whether the feature was dug to contain this pipe or if, as is perhaps more likely, the pipe was a later insertion. If the latter, the original date of the feature is unknown. South-east of the Late Iron Age features a large, irregular linear feature c. 7-9 m. across (36/19) ran in an east-north-east to west-south-west direction. This was partly sectioned at two points. It had a maximum depth of c. 0.95 m., with gently sloping sides and a flattish bottom. The sequence of fills within the feature was quite different on each side of the excavated area, but both sequences were consistent in producing small quantities of pottery and tile, probably of post-medieval date, from their lower fills. On the south-west side the primary fill (27), of grey clay, also contained irregular fragments of chalk rubble. Later fills (30 and 25) were similar in composition and character but with less chalk. These deposits, up to c. 0.30 m. in combined thickness, were overlaid by a substantial dump of bricks in a dark brown sandy loam (18), up to a maximum of c. 0.78 m. deep, the whole being overlaid by another greenish-grey clay layer (24). To the north-east the principal deposits (11 and 10) were yellowish-brown clay silt and sandy clay respectively. Fill 10 may have been in a localised, shallow recut (13), but this is unclear, its stratigraphic position and contour are similar to those of the brick filled deposit 18. 11, which was probably the primary fill in this part of 36/19, contained a small number of post-medieval tile fragments. At the south-east end of the site an ill-defined layer of yellowishbrown sandy loam some 0.10-0.20 m. thick (3 1/34) was the only deposit to produce artefacts not of post-medieval date. This layer may have been a medieval ploughsoil. The layer was cut by a gully (5, see below) and probably also by an apparently circular pit, perhaps originally c. 3 m. across (39) and up to c. 0.60 m. deep with a dished profile. Its single fill (12) was of brown silt loam and contained postmedieval pottery, tile, clay pipe and other material. This feature was cut by a south-east to north-west aligned gully (5) which extended just over 10 m. into the excavated area. After a gap of c. 8.60 m. this line was resumed by a further gully (16), c. 16.75 m. in length, which extended up to and over the upper fills of 36/19, which it cut. These features can be plausibly related to hedges along the railway line. A terminus post quem of at least the early nineteenth century is indicated by the contents of the features which they cut. More recent features were impressions derived from recent machine movements. These had cut 423 PAUL BOOTH AND PAUL EVERSON through the modern topsoil (2), which with other related layers (59 and 60), all of which contained post-medieval finds, overlaid the subsoil level exposed by the machine stripping of the site. THE FINDS The only finds recovered in any quantity were pottery, ceramic building material and animal bone. Other significant material consisted of a few flint flakes and two possible quern fragments. Small quantities of glass, three clay pipe fragments, five fragments of iron (four probably nails) and coal fragments, all probably or certainly of nineteenth-century or later date, were also found but are not reported on further. The overall quantity of animal bone was not sufficient to merit further study at this stage. Soil samples were recovered from some of the Late Iron Age features and sieved for carbonised plant remains, but were unproductive. The flint Philippa Bradley Eight pieces of struck flint were recovered from the excavations. All of the material is redeposited. The assemblage is summarised in Table 1. TABLE 1. ASSEMBLAGE COMPOSITION Context Flakes End and side Miscellaneous Total scraper retouched 2 - - 1 1 12 2 1 - 3 50 I - - 1 62 2 - 1 3 Total 5 1 2 8 The flint is of good quality. Two pieces are orange brown with cherty mottles and a fairly thick chalky cortex and the remaining material is mid-dark brown in colour. Hard hammers were used and none of the retouched pieces, scrapers and a retouched flake, are particularly diagnostic. Nevertheless, the material would not be out of place in a Neolithic or Bronze Age context. Neolithic flintwork has been found in the vicinity and field-walking by the O.A.U. recovered a scatter of Neolithic and Bronze Age flint from Willesborough approximately 800 m. to the north-east,21 21 P. Bradley, 'The struck flint', in Oxford Arch. Unit op. cit. in note 1, Cl5-C25. 424 EARTHWORK SURVEY AND EXCAVATION AT BOYS HALL MOAT The quern fragments Two small stone fragments, possibly from querns, were found in ditch contexts associated with 'Belgic' pottery. They were of a whitish (c. Munsell lOYR 8/1) sandstone, the quartz grains having a calcareous cement. This would be consistent with an origin in the Greensand, perhaps from a relatively local source. The Late Iron Age and Roman pottery (Figs. 5 and 6) Some 563 sherds of Iron Age and Roman pottery, weighing 4667 gm., were recovered. The material was generally in poor condition. The great majority was in grog-tempered fabrics of 'Belgic' type, which occurred in a relatively narrow range of forms. Fabrics The material was divided into fabrics and recorded for each context by sherd count, weight, rim count and EVEs (see further below). Ten fabrics were defined. Summary descriptions are given below. More complete descriptions are contained in the site archive. Fabric 1. Grog-tempered Grog is the major tempering agent, but burnt organic inclusions are also common. There are also sparse-rare inclusions of rounded quartz sand, sandstone, iron ore/clay pellets and (very occasionally) flint. Most of these appear to be incidental components of the basic clay. Fabric 2. Grog and sand-tempered As Fabric 1 but with moderate quantities of rounded quartz sand, sometimes augmented with rounded flint. The dividing line between this fabric and Fabric 1 was not always easily defined. Fabric 3. Sand-tempered Small rounded glassy quartz sand grains are the only significant inclusion type. Fabric 4. Grog and flint-tempered Similar to Fabric 2 but with moderate angular flint. Fabric 5. Flint-tempered Abundant coarse angular flint and moderate very fine sand. Fabric 6. Sand-tempered Abundant very fine quartz sand with sparse larger sand grains and ?sandstone lumps. Fabric 7. Fine sandy oxidised (Romanised) Fine orange-buff with sparse-moderate fine sand. A possible Canterbury product. 425 PAUL BOOTH AND PAUL EVERSON Fabric 8. Fine buff-white (Romanised) Fine buff-white with moderate fine sand. Fabric 9. Fine buff (Romanised) Fine buff with sparse-moderate burnt organic fragments/voids. Fabric 10. 'Chaff tempered' ware22 TABLE 2: QUANTIFICATION OF IRON AGE AND ROMAN POTTERY FABRICS Fabric No. sh. %sh. Wt. (gm) %wt. No. rims %rims EVEs %EVEs 1 506 89.9 4406 94.4 52 96.3 3.75 94.5 2 25 4.4 185 4.0 2 3.7 0.22 5.5 3 5 0.9 13 0.3 4 2 0.4 11 0.2 5 1 0.2 2 - 6 1 0.2 8 0.2 7 3 0.5 10 0.2 8 1 0.2 1 - 9 1 0.2 3 0.1 10 18 3.2 28 0.6 Total 563 4667 54 3.97 The assemblage was dominated by the grog-tempered Fabric 1. Subdivision of this fabric might have been possible but was not considered to be worthwhile at present. The distinction between Fabric 1 and Fabric 2 ( on the basis that the latter contained more sand) appears to be valid but of little significance since the repertoire of decorative techniques and vessel forms appears to have been the same in both. Sherds in these fabrics exhibited a wide range of colour. While most were dark brownish-grey to black, light and medium grey and orange-brown sherds also occurred. Most, if not all, vessels in these fabrics appeared to be hand-made. The condition of the sherds made assessment of surface treatment difficult. A relatively small number of sherds was recorded as having been burnished, particularly on rims and shoulders, but this proportion was probably originally much higher. The only significant decorative techniques were the use of grooves and cordons and of irregular 22 KAR, 61 (1980), 2-3. 426 EARTHWORK SURVEY AND EXCAVATION AT B OYS HALL MOAT combing or 'furrowing'. Some 98 sherds in Fabrics 1 and 2 (18.5 per cent of all sherds in these fabrics) had furrowed decoration. The remaining fabrics were of minimal importance. The single heavily flinttempered sherd (Fabric 5) might have been residual in a late Iron Age context. The sherds in Romanised fabrics were very small and the only apparently significantly stratified fragments, in 50/62, the upper fill of the late Iron Age north-south ditch complex, could have been intrusive. The occurrence of fragments of 'chaff tempered ware' is noteworthy. This fabric, probably to be associated with the manufacture and distribution of salt,23 is often found in fragmentary condition and the Boys Hall sherds are no exception. Forms Vessel types were quantified by rim count and EVEs (rim percentages). Many of the rim sherds were small and it cannot be certain that each one represented a different vessel, particularly since most of the vessels, being hand-made, were, therefore, potentially irregular in rim form. The figure for EVEs, though small, may provide a more reliable indication of the breakdown of vessel types, though it seems that there are no significant differences in the data produced by the two methods. The range of vessel types, confined to Fabrics 1 and 2, was relatively limited. Probable jar types amounted to 96.7 per cent of the total EVEs from the site. Just over half of the jars were of a simple barrel-shaped to globular bead-rim form continuum, with no clearly-defined cut-off point between the two. Medium mouthed jars with curving everted rims were almost exactly half as common as the barrel-bead rim types, and most of the remaining jars were of unspecified (but not bead-rim) types, usually with the rim as the only surviving part. There was a single rim of a larger jar, probably a storage vessel. Other types were very poorly represented. There was only a single rim certainly of carinated bowl or cup form, though other carinated sherds did occur (without rims), and a possible bowl or dish rim fragment was also seen. A strainer was represented by a perforated base. The upper part of this vessel could have been of a standard jar form. There was little significant evidence for use or re-use of the vessels. Carbonised deposits were noted on the interior of a few sherds, but their variable condition meant that it was not possible to record such evidence systematically. Glossy black ?carbonised deposits were noted on the exterior of a number of vessels, usually just below the neck. The consistency of application of this feature suggests that it may possibly have been decorative in function. One sherd was partly covered with a dense black substance of bituminous appearance. This extended along one edge of the sherd and may have been intended as an adhesive for repairing a broken vessel, as has occasionally been noted elsewhere.24 23 Ibid., 2-4. 24 Cf. P.M. Booth, 'The Iron Age, Roman and Anglo-Saxon pottery,' in N.J. Palmer, Excavations at Tiddington, near Stratford-upon-Avon, Warwickshire, (forthcoming). 427 PAUL BOOTH AND PAUL EVERSON Discussion The pottery forms a small but quite consistent group of material of late Iron Age to early Roman date. The great majority of all the sherds (95.7 per cent) came from the group of ditches and gullies in the central area of the site and may be presumed to provide a terminus post quern for the filling of these features. The upper fill of the main north-south ditch contained a small amount of post-medieval tile, but this was probably intrusive. The three small Roman sherds in the same feature may have been introduced with the tile, but this need not have been the case. Most, if not all, of the material may be assumed to have a fairly local origin. The grog-tempered tradition was firmly rooted in east Kent in the first centuries B.C. and A.D.25 and may be presumed to have been maintained at a number of local production centres within the region. Flint-tempering has also been seen as characteristic of the region,26 but its almost total absence at Boys Hall may indicate the location of this site at the periphery of the flint-tempering zone.27 The date of the inception of grog-tempering in the region is still unclear, though it presumably extended well back into the first century B.C. Its demise falls within the later part of the first century A.D.,28 in part corresponding with the expansion of sandy ware production at Canterbury.29 If the small Fabric 7 sherds from the north-south ditch fill 50/62 are in situ they may represent the very beginning of this transition. If intrusive, the otherwise unadulterated nature of the grogtempered assemblage might suggest that it had terminated before this transition began, perhaps, therefore, not much after the middle of the first century A.D. and almost certainly by about A.D. 75. The current estimation of the date range of chaff-tempered ware would be consistent with this.3o There is no clear evidence for chronological distinction between the feature fills in which the pottery concentrates. The assemblage is homogeneous throughout and its condition does allow consideration of questions such as abrasion. Indeed the average sherd weight, 8.3 gm., demonstrates that the material was well fragmented. The relatively abraded character of some of the sherds was probably, however, in part a function of the soil conditions on the site rather than indicating extensive redeposition. The overall quantity of material suggests that it derives from more or less immediately adjacent activity, though this clearly did not include the deposition of primary rubbish. Illustrated vessels (Figs. 5 and 6) Because of the lack of major ty pological variation and of any clear chronological distinction between individual feature groups the illustrated 25 R.J. Pollard, The Roman pottery of Kent (1988), 42. 26 I. Thompson, Grog-tempered 'Belgic' pottery of South-eastern England, BAR 108 (1982), 12-14. 27 Cf. ibid., 7. 2s Pollard, op. cit. n. 10, 67. 29 Ibid., 66. 30 Arch. Cant., ex (1992), 298. 428 EARTHWORK SURVEY AND EXCAVATION AT BOYS HALL MOAT pieces are presented as a single group, arranged in a typological sequence based mainly on rim forms. In each entry the context number follows that of the fabric. All the vessels are handmade except for no. 20. 1. Fabric 1. 62. Barrel-shaped jar with incipient bead rim. Groove on shoulder and furrowed decoration. 2. Fabric 1. 62. As last, but more globular in form. Burnished on shoulder. 3. Fabric 1. 52. Bead-rim jar with crude thickened rim. Groove on shoulder and furrowed decoration. 4. Fabric 1. 50. As last, without groove on shoulder. 5. Fabric 1. 62. Small bead-rim jar with simple thickened rim. The shoulder is burnished and has a (?deliberate) oval indentation. Furrowed decoration. 6. Fabric 1. 50. Bead-rim jar with groove on shoulder and furrowed decoration. Traces of both exterior and interior sooting. 7. Fabric 2. 62. Bead-rim jar. Possible groove on shoulder. 8. Fabric 1. 62. Bead-rim jar with groove and burnishing on shoulder. 9. Fabric 1. 62. Small bead-rim jar. 10. Fabric 1. 53. Globular jar with well-defined bead-rim and groove on the shoulder. Burnished on the shoulder, possibly overall. 11. Fabric 1. 50. Bead-rim jar with burnished shoulder and furrowed decoration. 12. Fabric 1. 62. Bead-rim jar with large and well-defined rim. Groove on shoulder. 13. Fabric 1. 50. Jar with short upstanding elongated bead-rim. Burnish and groove on shoulder, furrowed decoration below. 14. Fabric 1. 52. Jar with short curving everted rim. Burnished on shoulder. There is a shiny black deposit in a narrow band under the rim. 15. Fabric 1. 62. Jar with short curving thickened everted rim. A very slight groove defines the base of the neck. Above this line is a fragmentary band of a black deposit as on no. 14. In addition, lumps of a similar material occur along one edge of the sherd, both on the surface and partly on the break. These may represent an accidental post-breakage occurrence, but might suggest an attempt at repair. 429 PAUL BOOTH AND PAUL EVERSON Ii I I 1 I I / ---,- 7 --1 --------\ I I I I 11 1\ \1 \ \ 0 10 I I I I 4L􀀃􀀄 I \ / I \ 8 l======l,\ \ \ I / II -----,-I- - --::-::-_-::-:::-_-_) "' Fig. 5. Late Iron Age pottery nos. 1-12. 430 EARTHWORK SURVEY AND EXCAVATION AT BOYS HALL MOAT ,t 13 I-=-􀀃 \ I ' ( 14 16 ' \\ II I I 15 II I ) t 􀀂 17 II I I} ( 19 II 18 ' 0 100 mm ) I 􀀂: H 20 \ I I I ... I / ,) ( I/ 21 <, 23 ' / Fig. 6. Late Iron Age pottery nos. 13-23. 431 PAUL BOOTH AND PAUL EVERSON 16. Fabric 1. 50. Jar with short curving slightly everted rim. Grooves on shoulder. 17. Fabric 1. 62. Jar with thickened curving everted rim. Grooves on shoulder. Possibly burnished. 18. Fabric 2. 62. Jar with curving everted rim. 19. Fabric 1. 62. Jar with curving everted rim, a cordon at the base of the neck and grooves at the girth. The neck has a fragmentary shiny black deposit as on nos. 14 and 15. 20. Fabric 2. 62. Jar similar to last, but larger. 21. Fabric 1. 62. Large jar with curving everted rim. Very slight traces of possible furrowed or combed decoration. 22. Fabric 1. 62. Small carinated bowl or cup with slightly everted rim. Burnished above carination. 23. Fabric 1. 62. Angle of sharply carinated bowl or cup with a cordon defined by deep grooves above the carination. The later pottery Lucy Bown Thirty-four medieval and post-medieval sherds, weighing c. 200 gm., were recovered. Twenty-two sherds were assigned to six medieval fabric types and the remaining 12 to seven post-medieval wares . Six sherds belong to a distinctive quartz and flint tempered fabric which could be late Saxon or early medieval, dating from c. A.D. 1000-1200. All these sherds are heavily abraded and fragmented and the only diagnostic example is the inturned rim of a bowl, probably paralleled at Canterbury by forms dated A.D. 1050-1150.31 Seven sherds are in a coarse quartz-tempered ware used for cooking vessels. This fabric has been recognised as a common component of the Union Rail Link field-walking collection32 from the vicinity of Ashford and is thought to date from the eleventh to the mid thirteenth century. Of the remaining nine sherds four are less than 10 mm. in size and, therefore, too small to identify closely and the remaining five belong to four different fabric types thought to be of twelfth- to fourteenth-century date. These are a single sherd in a fine pinkish/cream fabric, three sherds in two hard fired reddish yellow quartztempered fabrics (one moderately tempered with larger quartz and the other abundantly tempered with fine quartz); and a single jug rim in a coarse quartz- 31 N. Macpherson-Grant, 'Local and imported wares at Canterbury: Late Saxon, SaxoNorman and Medieval, a provisional Guide. Canterbury Arch. Trust (1981), figs. 20 and 25. 32 Cf. note I. 432 EARTHWORK SURVEY AND EXCAVATION AT BOYS HALL MOAT tempered sherd of light red colour. These well thrown, light firing sherds are all likely to be from jugs and are, therefore, placed stylistically in the twelfth to fourteenth century. The earliest post-medieval wares present are two sherds of late medieval/early post-medieval red earthenwares of sixteenth-century date and four fragments of imported Cologne/Frechen Bellarmine dating from the mid sixteenth to seventeenth century. Later post-medieval wares include eighteenthto twentieth-century Creamware, English Porcelain, English Stoneware, Transfer printed wares and modem white earthenware. The contexts containing medieval sherds also had either post-medieval pottery or tile present. All the medieval material is likely to have been residual. The building material Just over 3.5 kg. of ceramic building material, probably all of post-medieval date, was recovered from the site. This total excludes the post-medieval brick deposit (18) in the ditch/hollow 36 (material from this context was not retained). The assemblage contained a little brick in coarse sandy fabrics, but consisted primarily of fragments of thin peg-tiles (c. 10-14 mm. thick) in finetextured fabrics with at most sparse-moderate rounded quartz sand grains. This material occurred principally in topsoil and other contexts where it was associated with post-medieval pottery. A few fragments were, however, found in the tops of the fills of two late Iron Age features (fills 46 and 50) but they are assumed to have been intrusive in these contexts. DISCUSSION The excavated late Iron Age and early Roman features do not constitute a coherent settlement entity, though the quantity of finds, particularly of pottery, and the presence of items such as quern fragments indicate that domestic settlement lay closely adjacent to the excavated area. In view of the discrete character of the features in the central part of the excavated area, it is not possible to determine their relationship (if any) to the apparently contemporary 'site 2' identified in the Kent Archaeological Rescue Unit work immediately to the south-east. This site in turn may be seen as a continuation of that located in the northwest corner of the Ashford Freight Terminal by the Canterbury Archaeological Trust. The occurrence of so much activity of 'Belgic' to early Roman date within the immediate area (see summary above) raises the question of whether this should be seen as representing a typical rural settlement density for this region and period, or whether the concentration of 'sites' was unusual. It is perhaps possible that the location of broadly contemporary features across a relatively wide area indicates the presence of a number of dispersed component elements of a larger settlement complex. 433 PAUL BOOTH AND PAUL EVERSON Damage to the site prior to the commencement of fieldwork meant that it was not possible to achieve the principal objective of the excavation, to examine the earthworks in detail. However, two major features encountered within the trench may have related to these earthworks. Feature 79, towards the north-west end of the excavated area, links closely with the north side of the probable entrance road (f) into the moat complex, though its fill was unfortunately undated. A ceramic pipe within the feature was presumably a later insertion in the way that field drains tend to be laid along the furrows of ridge and furrow fields. Towards the south-east end of the trench a further substantial feature (36 and 19) up to 9 m. wide clearly related to an irregular hollow which aligned with the south-eastern arm of the moat. The interpretation of this feature in the context of the earthwork layout is unclear. The excavated evidence suggests that it was substantially, if not entirely, of post-medieval date. The upper part of the sequence of fills, consisting largely of brick rubble, may have been deposited immediately prior to the construction of the railway embankment. The function of the feature is uncertain, unless it formed some kind of feeder channel for the moat. Even if this was the case it need not have formed part of the original configuration of the moat. 434