Excavation of a Moated Site at Pivington

PLATE I *-» ij*r..*£ A. View of southern lialf of excavations from East. B. Threshold of south screen's door and spere-screen (Phase IT). | faff p. 20 PLATE II A. Hearth from north-east. B. South end of eastern outshut (Phase II) showing threshold. EXCAVATION OF A MOATED SITE AT PIVINGTON By S. E. RIGOLD (a) HISTORY PIVINGTON, or Pevington, is a manor and former parish in Pluckley.1 The parish was united with Pluckley in the sixteenth century. The manor, part of the Honour of Magminot (later, of Say), as for guard and service of Dover Castle,2 gave its name to a family of sub-tenants who were certainly there in 1253-54, probably as early as 1211-12.3 It passed to an heiress, Amabel Gobion,4 who in 1405-06 left it to her grandson John Spelsell. From him, through the Brents of Charing, it reached the Derings of Surrenden in 1612. After Spelsell's death it was probably always at farm except between 1644 and 1742, when it was occupied by a cadet branch of Dering. The present house on the capital tenement has a typical Surrenden facade : the carcase is mainly seventeenth century (1644 ff.?), but apparently contains sixteenth century work though nothing strictly medieval. There is no prima facie reason to doubt that it occupies the original site of the manor house and that the church was properly identified with that subsisting in Hasted's day among the farm buildings, though it cannot now be identified. 1 This account is based on Hasted (2nd ed.), VII, p. 473, and Philipott, Villare Cantianum (2nd ed.), p. 276. The later documentation, which is only marginally relevant, has not been verified. The author acknowledges assistance from communications to Mrs. de Seyssel from the late Dr. Gordon Ward and Mr. R. H. d'Elboux. 2 In Domesday Book it was held by Ralph de Curbespine of Odo, out of whose lands the Magminot barony was created on his fall. 3 Among the " Holders of fees in Kent, anno 38 Henry III ", in Arch. Cant., XII, 197 ff. No. 226, " Willelmus de Pyuintone tenet j . feod milit. in eadem de Willelmo de Say ". But in The Red Book of the Exchequer (Rolls Ser. No. 99), I I , p. 617, in 1211-12 there are two separate entries under "Custod.de Mamignot" ; " Farburne et Pointune " (the latter obviously Pivington ; the two are associated later, e.g. at the Knighting of the Black Prince the heirs of John de P. held " Farneburne " of Geoffrey de Say) and " Simon de Devintone " (sic—for Pevintone?). 1 Amabel held Pivington, with the advowson, and Parbourne. She and two brothers who predeceased her were children of John of Pivington (already tenant in 1305-06, recently dead in 1346), son of Ralph (occurs 1257), son of the William who held in 1253-4. Mr. d'Elboux has identified from wills a cadet branch (?) which survived in Charing until the late fifteenth century, including Thomas Pevington, who held the manor of Broughton in Charing, and his brother John, who left 4 acres called Brootonsmede, later held by a Brent. The Brents had the remarkable timber house in Charing oalled Pierce House. 27 EXCAVATION OP A MOATED SITE AT PIVINGTON Nevertheless, some 400 yards east of the present house lies a moated site, the subject of the present investigations. It was deserted in the later seventeenth century, considerably after 1644, and the occupation does not appear to go further back than the later thirteenth century. This would rather suggest that it may represent a subordinate (dower?) house of the manor, and the scale of the buildings found is more consistent with this hypothesis. (b) NATURE OF THE SITE Situated on even ground, shghtly lower than the present house (Nat. Grid. TQ/922465), the moat [Fig. 1] is roughly pentagonal, with its longer axis W.N.W.-E.S.E., but the north-east part of the enclosure lies very low and the higher, habitable part is oval with a narrow projection along the south arm of the moat, precisely on the axis (x-y) of the south wall of the earhest structure found on the site. This apparently comprised a single range which filled most of the elevated area. The coincidence of extent and orientation suggests that building and mound were contemporaneous. The anomalous shape of the PIVINGTON MOAT M PLUCKLEY, KENT PHASES l b IIOVERlb a PAVING 535 t l a IQRE-USEO lb, II O vmuflDin / VA Walls and paving \ > not necessarily \ \ \ LOW contemporary >H S, y V N /• •^ "N w J J V- GROUND ***$* »AL, «Ulr '<• in ~-« s ^ «<*« y «u -> 11 frr V - x oo i "-» •N % I U feel metres 3 O 5 IO 15 20 25 30 3 2 10 O r- i t i l l PIG. 1. 28 EXCAVATION OF A MOATED SITE AT PIVINGTON mound contrasts with the rectangular moats which are widely attested at least from the later thirteenth century, but is not referable to any other recognized pattern of earthwork. The purpose of the narrow extension is unknown, but no buildings were traced on it. (c) EXCAVATION By permission of the owners, Major and Mrs. Pym of Barnfield, the elevated part of the enclosure was excavated on a careful grid system by Mrs. Barbara de Seyssel with the assistance of Mr. Tom Semark, Mr. Leslie Wright and other local volunteers (PL I, A). The greater part of this work was done in 1955. The writer was not present at this period and is only responsible for making a large-scale plan, for recutting and extending the very precise sections in order to draw them and test his interpretation, and, with valuable suggestions from Mr. P. J. Tester, for analysis of the structures and finds. For the earher work he relies on Mrs. de Seyssel's notes and descriptions, on Mr. Leslie Wright's working plans and on photographs taken at the time. The latest occupation (Phase III) Removal of a thin topsoil revealed a stratum of building-debris with much roof-tile and occupation rubbish up to at least the third quarter of the seventeenth century. At the same horizon were the very fragmentary ragstone sleeper-walls of a timber-framed house. These are indicated by broken lines on the plan of the central area (Fig. 2). The only really substantial footings were : (i) running E.-W. along the north edge, in such a position as to suggest an outshut rather than a main wall, a line of large roughly squared ragstone blocks, and (ii) an internal chimney-breast on a different alignment from the wall, which suggests that it was secondary. A line of smaller stones, parallel with and 5 feet south of (i) may represent the north wall of the frame proper ; the south wall was not encountered at all. It may be taken that the final house lay roughly E.-W. on approximately the same axis as its much better preserved predecessor, and, probably, since the chimney was evidently an addition and there were traces of firing a little south of the site of the earlier hearth beneath, that even this house was built as a " hall-house " with a central hearth. Later it acquired glazed windows. It is unfortunate that the plan is too deficient to compare it precisely with any of the numerous late halls surviving. The low sleeper-walls of a timber-framed building laid, as in this case, on the ground surface, can be robbed without leaving any vestiges. The remains of this latest house lay on an evenly spread stratum (" blanket ") of made-up clay, averaging 1 foot in depth and completely concealing the earher remains, which it had preserved admirably [see 29 EXCAVATION OP A MOATED SITE AT PIVINGTON lilflii^ v - _?&&&£ c<£ro oQr>>0 C3 o n n i m i i , i n " ' , " i M i i M i ' i , i i i i i ) n i M t w i t ' i ' < " J J J Y M . ! 4 ' • ' • J . 7 ^ ' ' ' L - ' i - ? J . : ' , ' , ' ! ? T 7 T T T TTBSl33Tfml5B H la-'mt ib-ext. Io lb ] Clay of |___j Yellow ):-:-.-:-:| Ash/ I6&63 Sandy iMound Ur_=]Chalky Loam G 2 3 Fired Loam &S2S Lime b^CilClay . O 1 2 3 4 5 10« — i t t i i i i t -i T U t • Chalk nr LTD Clay (vTTl TTTTT] Soil/ Blanket t ^ V j Grovel \£*£ Debris Scale of detailed Sections'^ ^ metres o !<>! > H t—t O O K o !> M e co HH H i> H & H O 3 FIG. 3. EXCAVATION OF A MOATED SITE AT PIVINGTON continuity with the clay (Phase II) floor, from a lower paving of smaller cobbles which represented an aspect (but not the earhest) of Phase I. Important deposits of pottery and much oyster shell were sealed between the pavings, significantly different from the primary pottery Group " A ", but not from that among or beneath the lower cobbles, which lay either in dark yellow sand (particularly at the west) akin to the interior loam, or in the top of the primary clay. The lower cobbles, therefore, do not belong to the beginning of Phase I. These sealed wares (Group " B "), though evolved and practically free from shellfilling, are yet much closer to the thirteenth-century tradition than the late medieval wares associated with the laying-out of Phase I I I . Taken in conjunction with the archaic house-plan, they suggest a date for Phase I I at some time, probably quite early, in the fourteenth century. A section was taken across the western moat, as far as the water allowed, but owing to the quality of the clay it was almost featureless ; however, it gave no indication that any lower horizon of occupation existed. Trenches were also dug on the top and side of the eastward projection of the mound, revealing nothing except absence of stonework. (d) ANALYSIS OF THE HOUSE-PLANS The plan as revealed by stripping the clay " blanket " (Fig. 2) showed sleeper-walls of one or two courses, apparently comprising a squarish hall, with a cross-wing or " end " (not projecting beyond the walls of the hall), at the west (screens) end and a narrow outshut at the east. This plan, though simple enough, appeared from the sections to be composite and to contain work of two or three periods. Phase I The fact that the partition-wall (F) between the Hall (H) and the cross-wing (G) of the Phase I I house lay at an obhque angle to the rest led one to suspect that it survived from an earlier period. This was confirmed by the sections, but these also showed that the central part of the external west wall and part of the north wall of the wings were likewise from the earlier build, although they were aligned with the Phase I I walls. Of the Hall, the primary north wall had gone, unless a ghost was visible in the outline of the lower cobbles ; on the other hand much of the primary south wall (K) remained, at right angles to the partition-wall (F) on axis x'-y' (Fig. 1) : in this part it was the secondary wall that had almost disappeared. Of the east wall (J) the central part, which had a broad footing, was primary and there was a marked change of level here, indicating an original hall of approximately the same length as the later one. The hearth was in the same general area in 33 EXCAVATION OF A MOATED SITE AT PIVINGTON both phases, but a variation in the extent of firing was consistent with the shghtly different alignment of the hall. The first house covered more than just the hall and cross-wing. There were certainly loamfloored extensions, both to east and west, that to the east divided by a longitudinal wall (L), parallel with axis x-y with the chalk floor, already noticed, to the north ; the east end was perhaps open, and the floor sloped eastward, as though for drainage. The western extension was ill-defined, but showed traces of firing and most of the primary potsherds came from here, perhaps indicating a kitchen. There was a drainage-channel (M) running north, in or just beyond it. The clean condition of the loam flooring suggested that both extensions had been covered, and not churned up by cattle. Before the definitive re-flooring and re-orientation of the Hall represented by Phase II, the transverse walls, other than F, had apparently been reconstructed on another axis, which was to set the axis for the subsequent Hall. This work may be contemporary with the largely destroyed intermediate lime floor and the first set of external paving. All these intermediate works are here designed lb. The eastern extension was cut off and replaced by a narrow outshut, with its floor slightly below that of the Hall, and the east wall of the Hall and west wall of the wing, which are parallel to the outshut wall, must be of the same construction. Both walls retain broad but indefinite footing in the centre, from the earlier date, but in the east wall and, less certainly, in the west these are flanked by two large square stones (N), inset about a yard and joined to the lateral walls by footings of smaller stones. The spere walls (S) were reconstructed in Phase II but may have had predecessors projecting to the same depth as the large square stones in the end walls. If this was so, ignoring the outshut wall and wall F, which was now level with the floor, the frame would have comprised two equal bays of about 16 feet. These could be construed as having very narrow aisles, or, alternatively, as aisleless, but with the weight of the tie-beam carried on inset, not terminal, posts, giving the TT-shaped frame, noticed by Mr. J. T. Smith at Bagueley Hall, Cheshire.1 The narrow lateral (quasi-aisle) panes would have been filled solid. They might possibly have been spanned by the short base-cruck-like member such as still survives in the upper chamber of a few archaic Kentish houses2 but any sort of base-cruck construction 1 See J. T. Smith and C. P. Stell, " Baguley Hall : the survival of pre-Conquest building traditions in the fourteenth century " in Antiquaries Journal, XL (1960), pp. 131 ff., esp. p. 137 ; of these archaic motifs only the inset posts seem to occur at Pivington. 2 A base-cruck form, i.e. incurved members supporting a collar, not straight posts supporting a tie, is found in the solar at Burnt House, Benover, near Yalding, and in several houses in East Kent noted by Mr. E. W. Parkin, e.g. at Etchinghill (now demolished) and Frogholt, Newington by Hythe. These also have the inset post under the bearing. 34 EXCAVATION OF A MOATED SITE AT PIVINGTON at ground level is hitherto unrecorded from Kent.1 More probably they were spanned by an extended rafter and a short post, such as occurs at Fairfield Cottages, Eastry. This mteresting building, originally one house, provides perhaps the closest parallel to the Pivington hall : it has four inset posts, two in the end wall and two in the spere-screens, enclosing a squarish hall with quasi-aisles but no intermediate support. The first set of external cobbhng, ascribed to the same (lb) date, follows the outer wall-line fairly closely on the north, but on the south approaches the eastern (hearth) bay only, with a possible extension, lying directly on the loam, to the west of the house on the line of the earher south front, which was also the main axis of the mound. This would perhaps cover a kitchen, succeeding that suspected in the previous build. A provisional reconstruction is : in Phase la—a long range, covering most of the mound, and lying immediately north of the main axis ; a square hall and an irregular inner room (not yet, perhaps, a " service " wing) occupied only the second and third fifths from the east ; eastwards was an open(?) shed, with a medial partition, cutting off a chalkfloored room with traces of fire—this was replaced by a narrow outshut ; westward an extensive structure, or group of structures, perhaps containing an external kitchen on the north, and a byre on the south. One hesitates to see anything of the long-house sort, but it was clearly something much longer than the simple hall-and-bower plan that succeeded it. In Phase lb there was a two-bay structure with an outshut, apparently approaching that of Phase II. Phase II This plan is complete, except the easily reconstructed south wall. The work of this period is recognized by the relatively large and squared rag. It probably amounted to no more than a reconstruction of the Hall (H) itself (possibly following damage by fire in the northern half) and re-flooring and levelling of the whole. The inner room (G) and the outshut were there already in Phase lb, though the latter, which shows no signs of even having communication with the Hall, was now provided with a good step on the south (PI. I I B). The external paving was completely renewed, burying the previous work, but at the west end, it may be just a repair, as the stones, though large, he directly on 1 The nearest known base-orucks in a groundfloor hall are at Dunster's Mill, Ticehurst, within a few yards of the Kentish border (R. T. Mason, in Sussex Arch. Coll., XCVII (1960), p. 150), and at Chennells Brook, near Horsham, discovered by the author and to be described by Mr. Mason in Sussex Arch. Coll., c. (1962). 35 EXCAVATION OF A MOATED SITE AT PIVINGTON the loam. The layout of this shows that these three short compartments now comprised the whole plan ; the presumed western (kitchen?) extension was apparently now removed, as well as the eastern. The plan, Hall, bower or inner room, with a solar over it, and outshut at the other end, is paralleled in some relatively early existing small halls under a single hipped roof and approaching the " Wealden " type (e.g. Little Robhurst, Woodchurch), but is much simpler than the common, fully developed, double-ended " yeoman's house", the contrast being the more noticeable on a moated, and presumably relatively important site. However, the reconstructed speres and steps of Phase I I seem to imply a plan in which the Hall and crosswing are structurally distinct and probably were roofed at right angles to each other. It would seem more probable that the archaic house, thus roofed on a T-plan, and of comparable dimensions (though with a deeper outshut or store, communicating with the Hall), at Petham1 would provide a closer parallel than any house under a unitary hipped roof. The most remarkable feature in common to both houses is the very short passage bay—at Petham it is just too long to suspend a spere from the main tie, and the square hearth-bay. This allies both houses with archaic T-plan houses in other parts of the country, where the passage-bay is a mere transition between the hearth-bay and a structurally distinct wing.2 The Hall, 17 feet wide internally, and so comparable with only the smaller of the typical later hall-houses, was, even then, proportionately short. The passage-doors had neat stone steps and short spere-screens (PI. I B). Speres are now rare in Kent, though surviving in the " Highland Zone " throughout the Middle Ages : they occur at Fairfield Cottages, Eastry (see above), in the late thirteenth-century house excavated at Joyden's Wood,3 and the fragmentary fourteenthcentury hall now at Scadbury, Chiselhurst ;* dwarf speres survive occasionally in later halls, usually with an " undershot" screens 1 Formerly the " Thatched Cottage ", recently repaired, and brought to the author's notice by Mr. E. W. Parkin. Another of the rare Kentish examples of a structural cross-wing at the " low " end is Booting, not far from Pivington, a grander house with some archaic features in its framing. 2 E.g. the cruck hall attached to a contemporary cross-wing with refined detail, at Steventon, Berks., described by the author in Transactions of Newbury District Field Club, X (1958), pp. 4 ff. This plan is quite distinct from thatin which the solar lies in the cross-wing at the " high " end and the service end is in series with the hall—persistent towards Wessex (Chennells Brook, Horsham, The Old Deanery, Salisbury, both late thirteenth century), but probably the basic plan of certain stone halls in Kent (Southfieet Old Rectory, Battle Hall, Leeds, Old Soar). j 3 Excavated by Mr. P. J. Tester ; Arch. Cant. LXXII (1958), 18 ff. ; only one ! spere-screen but a similar projection towards the high end. 4 Brought from Manor Farm, St. Mary Cray, in 1936 : see Arch. Cant. LXIX (1955), 221. 36 EXCAVATION OF A MOATED SITE AT PIVINGTON passage ;x there is evidence that these were once common, but were in the nature of fittings—at Pivington they may well have been structural. The position of the hearth in the centre of the less-than-square 14-foot space within the speres suggests that this was a single bay with no medial truss. In Phase II at least, the speres did not project sufficiently to provide inset quasi-aisle-posts exactly corresponding to those in the end wall. The hearth (PL II A), set askew, was roughly 3 feet 6 inches square and of more or less squared stones : other known instances of central hearths are usually circular and rougher.2 No trace could be found of the earher hearthstones. The later external paving is informative. A broad path veers left from the back (north) passage-door, but nothing approaches the front door ; instead, it leads to the south front of the end, which was floored with chalk in its southern part, as though replacing the northern part of the former east extension. Though the passage threshold is unmistakable, there was evidently another door in the inner room. To the west was a broad paved base-court 22 feet wide and apparently an independent outhouse or byre, rather than a kitchen, beyond it. The stones of the base-court are of the larger sort; as they lie directly on the loam, they would appear to go no further back from Phase II. The area at the extreme west is unpaved and burned ; a further outbuilding, destroyed by fire, evidently stood here until Phase III. Phase III Little can be said with certainty here. The house had an outshut along the back, and perhaps finally, a stair-compartment or other extension beyond it. It appears to have ended a yard or so east of its predecessor, and its western extent may be deduced, if, as usual, the chimney was inserted in the screens passage, and if, as seems probable, certain large stones (T), based directly on the westward paving, are in fact part of the footings of the final house. (Note the common use of large stones, preferably sarsen, as corner blocks : there was hardly any sarsen at Pivington.) This gives an external length of some 48 feet, and, allowing for room to enter south of the chimney, a width, excluding 1 I.e. when the chamber over the service end jetties into the hall by the width of the screens passage. A structural dwarf-spere in this position is seen, e.g. at Sundridge (Arch. Cant., XXXVII (1925), 167—see especially the section on p. 173) ; another has recently come to light at East Mailing. One of a pair of independent dwarf-speres survives at the Old Vicarage, Headcorn, similar in construction to the dwarf screen that occasionally survives by the parlour door (e.g. the house from Benenden, now at Northiam ; N. Lloyd, A History of the English House, Figs. 100-101, p. 197). 2 The hearth is square at Joyden's Wood (Arch. Cant., LXXII (1958), 18 ff.) and the Old Deanery, Salisbury, both late thirteenth century ; circular in late examples, e.g. Sundridge ; Durlock, Minster-in-Thanet, excavated by Mr. Parkin. 37 o EXCAVATION OF A MOATED SITE AT PIVINGTON the outshut, of just over 20 feet. These are normal and consistent dimensions for a large double ended hall-house of the famihar later type. It is possible that the southern cobble pathway was retained and now led to the passage of the third house. The northern was certainly partially covered over, and the southern certainly preceded the final build, but it is possible that these pathways survived to determine the obhque setting of the Phase I I I screens passage, which the position of the chimney-breast would suggest. I t was clear from the sealed deposits of each phase, that each had a peg-tiled roof with un-crested ridge-tiles and was probably hipped, as is usual in most comparable Kentish sites from the thirteenth century onwards.1 (e) FINDS (I) POTTERY. Four groups can be distinguished : A. No sealed deposits—small sherds scattered over charred areas over the chalk floor beneath the eastern outshut and in the unpaved part, perhaps the site of the kitchen, at the exterme west. To be associated with Phase la. B. Represented by deposits sealed beneath or among the lower cobble paving (B (i)), associated with Phase lb, or (B (ii)) sealed between the upper and lower paving, usually in the intervening clay, and thus associated with the Phase II reconstruction. These two sub-groups are not clearly distinguishable in fabric, but any difference may be obscured by the fact that B (i) consists largely of cooking pots, B (ii) mainly of jugs. C. Best represented by large fragments sealed by the clay " blanket " under the total reconstruction of Phase III. B (ii) and C thus have a precise structural context and absolute dating would be desirable. D. Associated with the final destruction of Phase III. Late seventeenth-century forms predominate, but there is some earlier, e.g. Elizabethan, material. Group A (Fig. 4) Cooking pots : sandy wares with slight shell-filling. The full shell-filling, which lasts in Kent until at least the mid-thirteenth century, is absent except for one piece. The characteristic flat, horizontal rim, with the bevel entirely on the underside (Fig. 4, ii) already 1 Compare Arch. Cant., LXXII (1958), Fig. 3, p. 28 (Joyden's Wood) ; tiles comparable in size and texture are Icnown from thirteenth-fourteenth century contexts at Eynsford, Temple Manor, Strood, etc. 38 EXCAVATION OF A MOATED SITE AT PIVINGTON ^ \ : K K ^ ^ i ' i i * \ kv 39 EXCAVATION OF A MOATED SITE AT PIVINGTON predominates (compare groups of c. 1300 from Eynsford Castle1 and Joyden's Wood2), but some thicker rims (Fig. 4, i) have a shght upper bevel and should carry the group further back into the thirteenth century. Whereas at the West Kent sites just quoted the ware is predominantly grey, at Pivington a pinkish, oxydized surface predominates. This seems to be a persistent characteristic of East and Mid Kent sites, as well as of the wares from the excavated kilns at Potter's Corner, Ashford3 and Tyler Hill, near Canterbury.4 But, in spite of the proximity of the Ashford kilns and the general resemblance of texture (with only a slight admixture of shell) and profile, the Ashford material is of a distinctly brighter pink-buff than most of the Pivington material of this group or the next. Pots from Ashford seem to find their way to New Romney5 but Pivington, being in a clay district, could evidently rely on another source, even nearer. Potter's Forstal, near Egerton Forstal, is a suggestive name. Fine wares : a few wall-sherds of grey (unglazed, as predominantly at Eynsford), or buff (green-glazed) jugs of late thirteenth-century type—too fragmentary for close comparison. The group as a whole would seem to cover at least the last quarter of the thirteenth century and perhaps a little earlier. Group B (Fig. 4) Cooking pots and bowls : mostly from Phase B (i), among the lower cobbles ; a sandy, oxydized, generally pink-buff fabric, with a greyer core, similar to Group A, but shell-filling now practically eliminated. This ware must derive from the same source as the Group A cooking pots and follow closely on them. Including jugs of the same fabric, it comprises about 75 per cent, of the total ; the grey sandy or gritty wares, common in the west of Kent, perhaps 15 per cent., mainly from jugs. The same flat, horizontal, or more often slightly down-turned, rims predominate (Fig. 4, iii, iv, v),6 but some, including one still containing shell-filling, are slightly concave-rimmed, as seems to be the general tendency as the fourteenth century progressed. In one grey, sandy pot the rim is slightly convex (Fig. 4, vi). A pink-buff pot 1 Yet unpublished : the difference in rim-forms noted is exemplified by two groups, one c. 1300, the other mid-thirteenth century and still characterized by much heavily shell-filled ware. 2 e.g. Arch. Cant., LXXII (1958), Fig. 5, 32. 3 Arch. Cant., LXV (1952), 183 ff. 4 Arch. Cant., LV (1942), p. 57 ff. : the " comparative pottery ", mainly from Maidstone is grey. 5 The characteristic " Potter's Corner " colour is shown by a proportion of the sherds from the " leper Hospital " site, found in recent, unpublished excavations. 6 The profiles of group B are particularly close to the Ashford ones (compare Arch. Cant., LXV (1952), Fig. 4, p. 185, perhaps dated a trifle early) ; only the colour and lack of shell distinguishes them. 40 EXCAVATION OF A MOATED SITE AT PIVINGTON certainly from B (ii) (Fig. 4, vii) has a convex everted neck with an inner bead in the early thirteenth-century tradition ; this conservativelooking pot, though in a paste like the others, may follow a fashion from Wessex—similar pots occur frequently, e.g. at Farnham Castle and Bishops Waltham (both late in a thirteenth-century context), but it is not easily paralleled from Kent (it is absent from the Ashford kiln)1. None of these wares are far removed from thirteenth-century types, but the same general forms seem to persist for some time after. Bases are still " sagging ". There was also part of a rectangular dish, 6 cm. deep, with orange internal glaze. Fine wares (i.e. jugs, the fabric is not always superior to the cooking pots) : four fabrics, but the same general shape occurs in all—rim narrow and flat or nearly so ; bulbous body ; handle flat and, as in the quoted groups of c. 1300, pricked, not slashed. A round handle and narrower baluster-like shape is also found in fabrics i and ii, and an old-fashioned carinate rim in ii. Fabric i, pink-buff and sandy, like the cooking pots (local) (Fig. 4, xi). Fabric ii harder but slightly gritty, vermilion to grey, but with a thin mauve-red slip and sometimes striped decoration in thicker, white slip;2 orange (on decorated jugs especially) or yellow-green glaze (as on Fig. 4, viii). Fabric hi, the typical West Kent and Surrey grey sandy ware, found throughout the thirteenth century,3 now sometimes with a splashy, olive-green glaze (Fig. 4, ix) ; note the wider distribution of jug wares then cooking pots. Fabric iv, pink body, sometimes grey exterior, and overall thick white slip (Fig. 4, x). This seems definitely a fourteenth-century innovation ; something similar is represented in the foundation levels of the Westminster Jewel Tower (1365). One smallish, pear-shaped jug has thumbpressed handles (Fig. 5). i and iii are conservative, like the coarse lph,n FIG. 5. 1 This is a more advanced form than, e.g. Arch. Cant., LIX (1946), Fig. 8, p. 22. 2 Not enough jugs from Ashford to settle the question. 3 Such jugs are numerous at Eynsford, but, like the illustrated example, are unglazed, 41 EXCAVATION OF A MOATED SITE AT PIVINGTON wares, ii and iv progressive and not always distinguishable from the next group. The group as a whole is tentatively ascribed to the first half or three quarters of the fourteenth century, though much of the coarse ware looks little after 1300. Group C (Fig. 6) Cooking or pickling pots and bowls : two distinct fabrics : (i) conservative, soft and sandy but fairly smooth ; rather pallid pink. The rim is always gently concave and everted, the extrados of the rim curving more steeply to meet the interior in a single angle (Fig. 6, vii, viii and x—bowl), (h) Progressive—very hard and ringing, but gritty ; brick-red, or lighter, often with a grey or mauvish slip. The rim of the cooking pots is very sharply incurved to form a ledge for a lid (Fig. 6, i, ii, iii and xi—bowl). Fragments of lids in similar ware are found (Fig. 6, iv, v). A comparable, but earlier(?) piece from Temple Manor, Strood, has the ledge formed separately. In all vessels the extrados and interior run fairly parallel towards a squared-off rim. A splashy treacle-brown internal or external glaze occurs sometimes. A comparable form, in even harder grey ware with a red lining, is found in the Dissolution d6bris at St. Augustines, Canterbury. Fine wares : (i) Jugs and " cisterns " (" gotches " or spigot-pitchers)1 in ware similar to cooking pots of type ii, with the same grey or mauve shp and sometimes curvilinear designs in white slip (Fig. 6, ix) ; these were also found in the same context at Temple Manor.2 They usually have a light red lining, and indeed, the white slip patterns can occur without the grey shp. Some jugs have the splashy treacly brown or olive glaze over the grey shp. The rim and pricked handle are much as in Group B, but handles regularly have thumb-impressions at the top. (ii) Thin white " Tudor" ware with a deep green glaze, all apparently from small jugs, (iii) Stoneware. The one reconstructable Rhenish jug has quite a thick grey glaze, but a vertical neck and medieval-looking carinate rim (Fig. 6, vi)—not the common early sixteenthcentury form with plain, shghtly inturned rim.3 It was found right against the Phase I I north wall. In general the indigenous wares remain essentially medieval: the 1 Compare Sussex Arch. Coll., LXXVI (1935), 223 (Bodiam), Antiquaries Journal, XXXV (1955), 64-66, and, especially, Colchester and Essex Museum Annual Report 1929, p. 19 and pi. x, 2 (cistern associated with a cooking-pot, called a storage-jar, with ledge for lid, precisely as at Pivington, fabric ii). 2 A late fifteenth-early sixteenth-century midden, unpublished. 8 E.g. tho vessel of the Maidstone coin-hoard, Arch. Cant., LXV (1952), 190, Fig. 6. 42 EXCAVATION OF A MOATED SITE AT PIVINGTON X ^ \J\ V Si V ft 43 EXCAVATION OF A MOATED SITE AT PIVINGTON precise chronology remains uncertain, but there seems to be nothing to prevent as late a date as an open-hailed house will allow, even into the second quarter of the sixteenth century. Distinctly fifteenth-century wares are hard to identify, but some may be included in Group C. Group D The later sixteenth and seventeenth-century deposits were not sealed. From the sample preserved it would appear that wares were plentiful down to the third quarter of the seventeenth century, but that later material was negligible. The final element was a selection comparable to that from the filling of the moat at the Jewel Tower, Westminster (c. 1660), and included pipe-bowls and stoneware " Bellarmines " of evolved seventeenth-century form. The Delft included the usual forms coarsely painted in blue or blue and orange, some probably Lambeth, and a fine, imported plate, diameter 25 cm., with a design in blue of distantly Chinese inspiration.1 (Fig. 7). [run FIG. 7. (II) BUILDING MATEBIALS Roof-tiles throughout were of the almost unchanging two-pegged Kentish type, and ridge-tiles, as usual, without crests. A proportion of each had an orange glaze.2 Floor-tile : one fragment from stone-debris west of hall, i.e. probably from the demolition of Phase II. Thin (1 -4 cm.) with grey core, bronze-green glaze and white shp print, not inlay, showing cusping and part of a human figure, or a stag (Fig. 8, v). The fabric suggests an import ;3 such things are known in Kent from richer domestic sites in the fifteenth century (e.g. Tonford), but are unexpected here. 1 Compare C. H. de Jonge, Oud-Nederlandsche Majolica en Delftsch Aardwerh, Fig. 176. 2 Compare Joyden's Wood, Arch. Cant., LXXII (1958), 28, Fig. 3. 3 For the fabric compare London Mus. Medieval Catalogue, Fig. 81, 65, 66, and Antiquaries Journal, XVII, p. 442 : a duplicate of one of these was found at Tonford near Canterbury. 44 EXCAVATION OF A MOATED SITE AT PIVINGTON (III) NUMISMATICA 1. Latton coin-weight for Angel, 1-5 cm. square, bevelled, weight 5-122 gm. (=79-05 grains—the Angel weighed 80 grains at all times between 1464 and 1604). Obverse type only, St. Michael and dragon in circular beading, closest in style to that on late Henry VII and early Henry VIII coins (Fig. 8, vii). Apparently sealed in demohtion of Phase II, and confirming this to the early sixteenth century.1 I i i Ic m i ~i 1 m IV ^ I ii ,I in X / V////////A A I \ cm 1 1 / (no? vi and vii) VII v FIG. 8 2. Jetton, diameter 2-3 cm. ; early Nuremberg for the " Franco- Flemish " market. Obv. stylized ship. Rev. Lozenge with French royal arms. Legends garbled (cf. F. P. Barnard, The Casting Counter and the Counting Board, pi. XXIX, 8, 9). A common type of c. 1500- 1520,2 found among " rock paving ", i.e. before the end of Phase II, thus confirming the same dating. 3, 4. Nuremberg jettons of late sixteenth-century types, from topsoil. Obv. Reichsapfel in trilobe. Rev. three crowns and three lys. 3 (cf. Barnard, op. cit. pi. XXXIII, 84), diameter 2-2 cm., of Hans Krauwinckel (active 1586-1612), reads " Gotes Segen macht reich " ; 4, diameter 2-3 cm. of Hans Schultes, reads " Glick kumpt von Got ist war ". 1 Square weights, including this, are generally of Low-country make : compare Mateu y Llopis, Catalogo de los Ponderales monetarios del Museo Arq. Nacional (Madrid), PI. XVIII, 158, wrongly ascribed to Henry VI, and PI. XV, 132, etc— sixteenth century Antwerp weights for coins of Ferdinand and Isabella. Round weights for Angels are known, e.g. from Somerby, Lines. 2 This is the commonest type for the series ; it has been found at Hampton Court (not before c. 1515), and the reverse type at Wharram Peroy, Yorks. (not after c. 1516). 45 EXCAVATION OF A MOATED SITE AT PIVINGTON (IV) OTHER OBJECTS OF METAL OR BONE 1. Copper mounting (Ifrom a belt), silvered on one face, 3-9 by 1 -4 cm. with two rivets and central perforation and incised with schematic leaf-pattern (Fig. 8, vi). From upper layer around hearth (Phase II—looks late fifteenth century). 2. Bronze scabbard (or possibly belt)1 chape, 4-7 cm. long, bent from one piece of sheet-metal, with cusping on top edge (Fig. 8, iii). From loam floor, so presumably Phase I. 3. Thin bronze sheeting showing an angle-fold. Phase II, between the two external cobble paving layers. 4. Bronze ring, diameter 2-7 cm. Phase III? 5. Iron plate for fixing handle to wooden bucket (Fig. 8, ii). In mortar of Phase n i wall. 6. Iron scissors, total length about 14 cm. Phase III. 7. Half of unfullered ox-shoe (possibly a horse-shoe, but looks too broad). Phase II, immediately above lower cobble paving. 8. Flange of iron knife, riveted to bone handle-plates ornamented with brass pin-heads arranged in fours (Fig. 8, iv). Probably Phase II2. 9. Bodkin of fish-bone, length 8-5 cm. (Fig. 8, i). Phase II, north corner of hearth area. ( / ) CONCLUSIONS (a) Nothing, either from stratigraphy or single finds, suggests any habitation of the site before the construction of the moat. (b) The earliest remains (Phase la), which lay immediately on the body of the mound were assignable to the mid- to late thirteenth century. (c) The plan of the Phase la structures is uncertain but these appeared to be a range following the axis of the mound, with a central hall, the position of which survived subsequent modifications, and outbuildings at either end. (d) The buildings were modified, and the eastern extension reduced in Phase 16 (?c. 1300), and again in Phase I I (Nearly or mid-fourteenth century). (e) The Phase I I plan was largely intact : the house a simple hall with short cross-wing and eastern outshut. Remarkable features, which may in part derive from the earlier phases, include the inset 1 Compare London Museum Medieval Catalogue, Fig. 86, 8 and 88, 2. 2 For the form of mounting compare London. Mus. Medieval Catalogue, PI. XI, 10, Pitt Rivers, King John's House, Tollard, Wilts., PL XVIII, no. 8 (flange only) and Records of Bucks., XVI (1957-8), p. 161, an exact parallel save for the ornamentation, in a fourteenth-fifteenth century context, from the Mount, Princes Risborough. 46 EXCAVATION OF A MOATED SITE AT PIVINGTON (possibly aisle) posts in the terminal walls and the spere screens. The evolution of the western outbuildings (byre and kitchen) is less clear. ( / ) Apparently the Phase II building lasted throughout the fifteenth century. There was no trace either of desertion, or of modification. %• (g) Early in the sixteenth century the Phase II building was demolished, the level raised by a deposit of clay and a new hall-house built which lasted, with modifications until deserted after the middle of the seventeenth century. 47