An Unknown Medieval Site, Possibly a Manorial Chapel at Crabble Dover

An Unknown Medieval Site, Possibly A MANORIAL chapel, at Crabble, dover

keith parfitt

The Canterbury Archaeological Trust was engaged to undertake invest-igations on a new housing development (now called Mill Race) at the former Crabble Paper Mill, near Dover, in 2002. Significant results from the initial evaluation trenching led to further excavations on the line of the new estate road, followed by an intermittent watching-brief of the builder’s ground-works. Two previously unknown masonry structures and several pits and ditches of medieval date were recorded, together with evidence for earlier activity during the Neolithic-Bronze Age period (Parfitt 2002; Parfitt and Corke 2003). The present paper describes the medieval remains; the prehistoric discoveries have been the subject of a previous report (Parfitt 2006).

The nineteenth-century paper mill lies within the historic parish of River (see below) and occupies a roughly rectangular plot of ground at the foot of Old Park Hill, adjacent to the River Dour, about a mile and a quarter (2km) north-west of Dover town centre (Figs 1, 2 and 3). (NGR TR 2995 4311, centre.) The plot is bounded by the Dover-Canterbury railway line on the north-eastern (uphill) side, Crabble Road on the south-east side, the River Dour on the south-west and Kingston Close on the north-west. The elevation ranges from 30m od along the north-east side to 19m adjacent to the river (Fig. 3). The brick-built mill buildings, retained in the new development, are largely confined to the lower, southern half of the site, adjacent to the river.

The natural (surface) geology of the area is variable, consisting of river gravel, tufa, peat and head brickearth. The hill-slope deposits, principally in the form of head brickearth, are confined to the highest part of the site in the eastern corner of the plot and the bulk of the archaeological remains were discovered here. This is unlikely to be coincidence and must reflect the need to occupy well-drained ground above the flood plain but remain close enough to the river to make full use of its resources.

The Medieval Buildings (Structures 1 and 2)

The remains of two masonry structures (Structures 1 and 2; Fig. ⁴) were located in the eastern corner of the site, together with several pits (see below). These appear to relate to a previously unknown medieval occupation site, which included at least one stone building of some quality.

Structure 1

This was rapidly excavated and recorded ahead of its removal by the new estate access road (Trench 10). From its construction of mortared flint and Folkestone Greensand, and associated Caen stone architectural fragments, there can be no doubt that this building was medieval, although precise dating evidence is limited.

As surviving, the building consisted of the complete north wall (129) and portions of the east (141) and west (142) walls of a small, apparently rectangular structure (Fig. 4). The southern wall had been previously destroyed by deep terracing for the nineteenth-century mill but there was evidence for an original doorway in the west wall. Internal measurements of 9.36m (ese-wnw) by a minimum of 3.70m (nne-ssw) were recorded (30ft 9in. x 12ft 2in.). If the western doorway had been centrally positioned, as seems most likely, an original internal width of about 4.60m (15ft) may be postulated. Thus, almost half the building appears to have been previously removed.

On the north and west sides, the structure had been terraced into the hillside by up to 0.60m (Fig. 5, F. 122). On the east side, the wall would seem to have been free-standing, as the land naturally sloped away here (see below).

The foundations of the building consisted of large flint nodules and chalk lumps, loosely set in crushed chalk. These were generally between 0.45 and 0.60m deep (Fig. 5). Above the foundations, the lower courses of the main walls survived only on the west and north sides. These were well built and of similar construction, between 0.78 and 0.86m wide and survived to a maximum height of 0.75m (five courses). They were made from large flint nodules and water-rolled boulders of Folkestone Greensand (10-20%), set in a coarse cream-white mortar with small blue and brown flint pebbles. Traces of internal mortar rendering, up to 0.03m thick, survived on the wall faces at the north-west corner. Externally, the north-west corner of both the wall and its foundation had been neatly formed from unworked mortared Greensand slabs. Here, the stones were continued below the general level of the base of the foundation to give additional strength. The bottom stone consisted of a massive water-rolled boulder, measuring 0.98 x 0.76 x 0.30m. The north side of the western doorway had also been constructed from undressed Greensand boulders.

No walling survived in situ above the foundation on the east side. However, an extensive spread of fallen rubble following the slope of the ground, was recorded outside the building here and there can be little doubt that this represents the collapsed east wall. The rubble layer yielded several pieces of medieval roof tile, a pot-sherd dated c.1175-1225/50 and a number of fragments of Caen stone, including two large pieces from a claw-dressed window frame, probably of thirteenth-century date. It also partially sealed the remains of another medieval structure (Structure 2, see below; Fig. 6, 143).

A thick demolition deposit infilled Structure 1 (Fig. 5, 128) but this failed to produce any datable finds. It consisted of clean flint, greensand and mortar rubble, and contained a single unworked fragment of Caen stone. No roofing material was found in this layer. Below the rubble deposit, soil that had previously accumulated on the hillside formed the floor of the building, without any surviving evidence for a laid or trodden surface (Fig. 5, 127).

Structure 2

Following the excavation of Structure 1, a watching-brief was maintained as the builder’s prepared the new road-bed. This succeeded in locating the remains of a second medieval masonry structure (Structure 2), set at a lower level immediately to the east of the first (Figs 3 and 4). The structure was not fully exposed and its extent remains uncertain.

The new structure was represented by two separate walls (Fig. 4, Walls 176 and 177). The earliest and most substantial was Wall 176. This was 0.52-0.59m thick and consisted of roughly coursed medium and large flint nodules set in a coarse cream sandy mortar with moderate quantities of brown flint and chalk grits (Fig. 6). It was aligned on a similar axis to Structure 1, falling in line with its north wall (Fig. 4). At the west, the wall terminated in a simple square end. From here it was traced for a distance of 3.30m eastwards and survived to a height of 2.00m (Fig. 6). It was only faced on the southern (downhill) side and must represent a substantial retaining wall relating to a terrace cut into the hillside. Reflecting the eastward fall of the ground, the base of this terrace lay about 2m lower than the floor level of Structure 1. A thin mortar deposit recorded at the foot of Wall 176 (Fig. 6, 166) might represent the original construction layer.

Wall 177 was a later addition and joined Wall 176 at an angle of about 108 degrees, close to its western end (Fig. 4). The added wall was constructed of large flint nodules set in cream clay with chalk grit. It was about 0.65m wide and survived to a height of 1.08m.

The terraced area delimited by these two retaining walls had subsequently been in-filled with a series of soil, clay and rubble deposits (Fig. 6, 158, 159, 160, 175, 161, 162, 163, 164 and 165), probably representing a combination of natural collapse and decay of the walls, together perhaps with some deliberate in-filling. These fill deposits produced a small quantity of medieval peg-tile, together with a single pot-sherd broadly dated to c.1050-1225 from layer 160. Three large curved blocks of dressed Caen stone came from layer 159. These form part of a window, probably of thirteenth-century date and fairly certainly the same one as represented by the two blocks previously recovered from layer 143 (see above).

The final layer of mortar rubble infilling the terrace (158) was sealed by a thin clay floor (Fig. 6, 173), with an overlying occupation deposit (174). No datable finds were recovered but these deposits must be connected with later activity in the area. They were covered by the extensive rubble layer (143), which seemed to represent the collapsed east wall of Structure 1 (see above). A soil layer over this (Fig. 6, 130) had been cut by a circular pit, F. 115 (Fig. 4). The filling of this pit produced no finds but a late medieval or early post-medieval date seems likely on stratigraphic grounds (Table 1).

Too little was seen of Structure 2 to allow any detailed interpretation. It clearly contained more than one phase of work but overall the remains did not appear to form part of a building. More probably, they formed successive phases of retaining wall delimiting a larger terraced area, which lay to the east of Structure 1. Now mostly buried below Crabble Road, the exact purpose of this terrace must remain unknown but it may be reasonably surmised that it originally carried another medieval building, which had been set on a levelled platform cut into the hillside below, and to the east of, Structure 1 (Fig. 3).

Other Medieval Features and Deposits

Four pits occurred outside Structure 1 (Fig. 4, Fs 117, 119, 121 and 136). These were all dug in from the same horizon as the terrace for the building and so are likely to be broadly contemporary with it. The filling of three produced single sherds of pottery helping to confirm their medieval date. Based on this very limited evidence, F. 121, together with F. 115, noted above, may be the latest (Table 1).

A thin, localised soil deposit (Fig. 4, 98) was discovered lying on the surface of the natural brickearth at a point about 6m to the north of the Structure 1. This contained large quantities of oyster and dog whelk shell, together with some fifteen sherds of medieval pottery datable to the period c.1150/75-1250 and a sandstone hone. The deposit must represent a dump of domestic refuse and may well be derived from the adjacent building. Also related may be a deposit of dark grey-brown clay loam found overlying the natural tufa in Trench 9, some 12m to the west of the Structure 1. This contained animal bone and seven fresh, partially conjoining, fragments of sandy medieval pottery datable to the period c.1050-1150/75.

Probably broadly contemporary with the medieval remains discovered at the eastern corner of the site are two successive ditches located in Trench 4, on the river flood plain about 100m to the west (Fig. 3, Fs 86 and 88). They were aligned roughly n-s and ran diagonally across the width of the excavated trench. Ditch F. 88 was continuous but F. 86 terminated about half-way across; it appeared to represent a direct replacement of F. 88 and was wider and shallower than the preceding ditch. A few medieval pot-sherds, ranging in date from the late eleventh to the early thirteenth centuries, were recovered from the associated fillings. It seems quite likely that these ditches were connected with water-meadows lying adjacent to the river, below the main occupation area. An undated pit close-by (Fig. 3, F. 84; Table 1) may also be broadly contemporary.

Dating and Discussion

Although there is sufficient evidence to establish their medieval date, it is not possible to be entirely certain as to the purpose of either Structure 1 or 2 from the surviving remains. However, they clearly cannot relate to any early water-powered mill because they are too far from the river and are set too high above river level. Moreover, the substantial masonry construction, apparently with some Caen stone window dressings, suggests that Structure 1 was a building of relatively high status.

Structure 1 does not appear to have been an isolated building and it would seem that at least one other building lay further to the east, set on a substantial terrace which was retained by the walls forming Structure 2 (Fig. 4). Presumably, this postulated second medieval building lies under what is now the line of Crabble Road (Fig. 3). Other related structures to the south could have been destroyed by the large-scale terracing for the paper mill and, indeed, local antiquary E.J.G. Amos recorded early post-medieval building remains here in the early twentieth century (Amos 1939), which might have been connected with the later phases of such a complex.

In the light of the above, the origin and date of Crabble Road itself requires some consideration. Branching off the Roman Watling Street leading from Dover to Canterbury, Welby (1997, 28) shows this route as being in existence in the late eighteenth century and it seems likely that it formed part of a much older trackway which led to the site of St Radegund’s Abbey on the western hills at Bradsole (see below). In the area of the paper mill, the road was widened and the adjacent railway bridge rebuilt in the 1938 (Welby 1997, 56). It was at this stage that a large brick and concrete retaining wall delimiting the present site was erected, cutting through Structure 2 (Fig. 4).

Situated adjacent to what would appear to have been a medieval route-way, perhaps the most likely interpretation for the medieval structures located at Crabble Paper Mill, with their associated pits, is as part of a manor house complex. Set a little back from the road, the substantial stone construction, shape and orientation of Structure 1 suggests that this building could represent the manorial chapel. Alternatively, it might have formed a wayside chapel, placed near the junction of two important medieval roads, which would have been regularly used by pilgrims and travellers (see discussion below). From the limited ceramic and architectural evidence, a thirteenth-century date for the erection of Structure 1 may be tentatively suggested from the Caen stone window dressings recovered. These most probably derive from its east window.

The medieval building – a documentary investigation

by S.M. Sweetinburgh

The site of the high-quality medieval building (Structure 1) is in the parish of River and presumably also in the manor of the same name, lying within the Hundred of Bewsborough. Unfortunately, the surviving documentary sources for medieval River are extremely limited. They are primarily confined to the records of St Mary’s Hospital (Maison Dieu) and St Martin’s Priory at Dover, St Radegund’s Abbey at Bradsole, and those of the Crown. There are a few other materials, such as wills for the late Middle Ages and the field names from the 1844 tithe assessment, but they do not provide any useful information.

River is not recorded in the Domesday Book, but Paul Cullen agrees with Wallenburg that it may be referred to in Domesday Monachorum as Burnan (OE burna ‘a stream’). By 1199 the name had changed to Rip’ia or Riveria from ME rivere ‘a river’ (Vulgar Latin riparia), a continuing reference to the river Dour which ran through the parish (Fig. 2). The parish church, dedicated to St Peter, similarly may be identified in Domesday Monachorum, under Ewell (two churches) and it is possible that one of the mills listed under Ewell in Domesday was in the parish of River (Welby 1977, 11). If that was the case, then the mill was held by Hugh de Montfort from the Bishop of Bayeux, presumably becoming a royal possession following the bishop’s disgrace (Morgan 1983, 5:185, 192). The parish of River comprised two small river-side settlements, Kearsney and Crabble, the latter further downstream (Fig. 2). About 2 miles to the south-west, at Bradsole in the parish of Poulton, was the Premonstratensian house, St Radegund’s Abbey, founded in 1192/3 by Hugh, its first abbot (Page 1926, 172).

The early history of the manor of River is difficult to ascertain, but Hasted believed that during the reign of William I it was held by Hugh de Montfort (under Ewell in Domesday), remaining with the family until the exile of his grandson, Robert, when it passed to Henry I. Thereafter it was held by Robert, son of Bernard de Ver, constable of England and later Henry de Essex, constable of England, before returning again to the Crown when it was confiscated following Henry’s cowardice in battle (Hasted 1800, 438). Remaining in royal hands during John’s reign and the early years of Henry III, it was stated in Testa de Nevill (1216) that the manor was held in three parts: by the canons of St Radegunds’s Abbey; by Dover Castle; and by Soloman de Dovere – which later became known as Archer’s Court, the land lying in the north of the parish, and as such does not need to be considered further here (ibid.).

The canons had received their part of the manor in 1204 from King John, comprising 100 acres, and four years later he gave them the local church, so allowing them to relocate their house from the head of the valley to a more attractive site by the river (Page 1926, 172). However, the community decided not to move and in 1215 John again granted the parish church to them for their maintenance and that of pilgrims.¹ This suggests that their land was close to the church, in the centre of the parish, rather than in the southern part at Crabble (but if they had been contemplating building at Crabble, they might have started there, which could account for the excavated medieval structure).² Yet they did hold a mill at Crabble, apparently as a result of the sale and gift of Alan Corbell, though, interestingly, Alan’s grant is not listed in the abbey’s register. The Crown also had interests in the mill; it received 20s. from the rent, and in 1227 Henry III granted this in frankalmoin to the canons. In addition, he granted the site of the mill to them, similarly in frankalmoin, the canons to celebrate daily in their church (presumably at the abbey) for the king and his ancestors, gifts which were recorded in the abbey register (Calendar of Charter Rolls 1226-1257, 21).³

The part of the manor held of Dover Castle was under the control of Hubert de Burgh, as constable of the castle, during the early thirteenth century, and in 1228, at Hubert’s request, Henry III granted it to the master and brothers of St Mary’s hospital, Dover.⁴ This charitable institution for the sustenance of paupers and pilgrims had been founded by Hubert, but was soon to come under royal patronage. Considering the hospital’s proximity to its manorial holding, it seems highly likely that some of the land was direct-farmed, the rest leased to local peasants. Thus St Mary’s revenue from River would have derived from a number of sources: its own farming activities, rents from its tenants, dues from both the court leet and court baron held at the manor. It is difficult to gauge the extent of the manorial buildings required at River, but presumably the courts were held there, not at the local parish church which was under St Radegund’s, and there may have been at least one barn and some livestock accommodation, apparently all enclosed by a wall.⁵ The court may have been accommodated in a hall (possibly part of a manor house) or in a chapel, if one had been built but there is no evidence as to where such structures may have stood. Because of the short distance between the manor and the hospital (about 1¼ miles), it seems improbable that the master would have felt it necessary to provide a chapel there for whichever of the brothers had been sent to oversee the manor, but it is feasible that a small wayside chapel had been built during the late 1220s or 1230s for pilgrims travelling between the Channel port of Dover and Becket’s shrine at Canterbury. Such a chapel would not be dissimilar to another (the surviving St Edmund’s Chapel) held by the hospital from the mid-thirteenth century, which was sited immediately to the south-west of the hospital and is known to have held a relic of St Richard of Chichester (Tanner 1968, 8). Nevertheless, as a chapel held by a religious house, it would not have been recorded in the papal taxation list of 1291, and equally it would not appear in episcopal visitation records.⁶ Furthermore, very few of the medieval records for St Radegund’s Abbey and St Mary’s Hospital survive (the lack of accounts and court records is particularly unhelpful), and even the mid sixteenth-century returns produced during the Reformation period do not give any detail beyond the name of the manorial holdings (Valor Ecclesiasticus, vol. 1, 57).⁷ Furthermore, none of the parishioners, nor those from neighbouring parishes who made their wills in the fifteenth and early sixteenth centuries, gave anything to a chapel in River; instead all bequests were centred on the parish church. However, the general absence of pious bequests to other than the testator’s parish church (or occasionally those in surrounding parishes) is a feature of the wills of Dover and its hinterland. Thus the apparent ‘neglect’ of local wayside chapels was common in the area.

In contrast, mills are sometimes recorded, a corn mill called Rakestravis Mill and another called Ash Mill, were the subject of a leasing agreement between St Mary’s Hospital and its lessees in 1472 (Welby 1977, 113-14). Crabble mill has already been mentioned and it is possible that ‘Cripelemelle’ (subject of an agreement between Dover Priory and John de Dover, recorded in the section of the priory’s register covering its activities with St Mary’s Hospital) was in the same area.⁸ In the mid and late sixteenth-century ministers’ accounts for the manor of River (initially it passed to the Crown at the Dissolution, then became the property of the Archbishop of Canterbury before returning to the Crown, changing hands again in the mid-seventeenth century) included a mill that was called ‘Chappellmylle’, ‘Capellmyll’ or ‘Crabhallmyll’, which may have been scribal mistakes but is suggestive.⁹ A possibly sixteenth-century list of the possessions of St Mary’s Hospital does not shed any more light on the subject, because it says the manor of River included (part of the demesne, perhaps) two watermills, both farmed out, a tenement, a barn?, a fulling mill, and a piece of land called ‘Bedelland’ to the north of this mill. The tenants at River and their holdings are then listed, and then the house’s holdings elsewhere, including a small chapel called the ‘Ladys of Pitties’ chapel, probably at Archcliffe, west of Dover, and churchyard.¹⁰ Unfortunately none of the later documents consulted provided any further clues.

To conclude, the remains of a high-quality stone building found during the excavation presumably belonged to either St Radegund’s Abbey or St Mary’s Hospital, forming part of a manorial complex associated with the religious house’s part of the manor of River. Whether it was a small chapel or was associated with the manor house where courts could be held is open to question. If it was a chapel linked to St Mary’s it might be seen as similar to the small chapel of St Edmund near the hospital; and if to St Radegund’s, part of that house’s response to the royal grant of River parish church for the sustenance of pilgrims. However, it may never be possible to say exactly what the building was used for because of the loss of the medieval archive and the building itself.

St Edmund’s chapel, situated on the western side of the main road out of Dover and close to St Mary’s Hospital, comprises a flint- and Greensand-built, single-celled structure without a chancel arch. Access is through a centrally placed doorway in the west end wall. Consecrated in 1253, the chapel’s internal dimensions are 8.15m (e-w) by 4.26m (n-s) (26 ft x 14 ft), only fractionally smaller than those suggested for Structure 1 at Crabble (see above). Overall, this chapel seems to provide a very close local parallel for the present site both architecturally and from the documentary perspective.

acknowledgements

Thanks are due to a number of people who assisted with the project in a variety of ways. First and foremost, the contribution of the developers, Crabble Mill Developments Ltd, who allowed access and funded both the fieldwork and post-excavation analysis must be gratefully recorded. On site, Richard Howells and his construction teams gave every assistance, despite a busy work schedule which allowed little time for any delays. Simon Mason of Kent County Council’s Heritage Conservation Group monitored progress of the work throughout and provided much assistance and encouragement. Paul Bennett, Director of Canterbury Archaeological Trust, also gave much help and encouragement throughout. Barry Corke produced the site drawings for publication and Sheila Sweetinburgh has kindly looked into the available documentary history of the site.

bibliography

Amos, E.G.J., 1939, ‘Report from Dover for 1939’, Archaeologia Cantiana, li, 212.

Caley, J. and Hunter, J. (eds), 1810, Valor Ecclesiasticus, vol. 1 (London).

Hasted, E., 1800, The History and Topographical Survey of the County of Kent, vol. IX, reprinted Canterbury, 1972).

Lyon, Rev, and Owen, T., 1787, History and Antiquities of St Radigund’s or Bradsole Abbey, near Dover, no. 42, Bibliotheca Topographica Britannica (London).

Morgan, P. (ed.), 1983, Domesday Book: Kent (Chichester).

Page, W. (ed.), 1926, Kent, Victoria County History, vol. ii (London).

Parfitt, K., 2002, Crabble Paper Mill, River, Dover: an Assessment of the Archaeological and Geoarchaeological Discoveries (unpublished CAT report August 2002).

Parfitt, K., 2006, ‘A Prehistoric ‘Burnt Mound’ Site at Crabble Paper Mill, near Dover’, Archaeologia Cantiana, cxxvi, 219-237.

Parfitt, K. and Corke, B., 2003, ‘Crabble Paper Mill, Dover’, Canterbury’s Archaeology 2001-2002 (CAT 26th Annual Report).

Tanner, T., 1968, Saint Edmund’s Chapel, Dover and its Restoration (Dover).

Welby, D., 1977, The Kentish Village of River (Dover).

Welby, D., 1997, Reflections of River the Kentish Village (Dover).

endnotes

¹ Bodleian: Rawlinson MS B336, 168.

² According to a letter from Rev. Lyon to J. Nichols dated 4/6/1785, St Radegund’s had a mill near the manor house of River and other lands purchased by the king from Alan Corbel [this may be his translation of the grant in the abbey register which says the mill was under the court of River, that is under its jurisdiction]; Rev. Lyon and Owen, T., History and Antiquities of St Radigund’s or Bradsole Abbey, near Dover, no. 42 Bibliotheca Topographica Britannica (London, 1787), 458.

³ Bodleian: Rawlinson MS B336, 139.

⁴ British Library: Add. MS 6166, f. 215v.

⁵ Lambeth Palace Library (LPL): MS 241, f. 42.

⁶ Taxatio Ecclesiastica Nicholai IV database (University of Manchester). In the Dover Priory register there is a copy of the document detailing the privileges of Pope Nicholas of the possessions, churches and tithes of St Mary’s Hospital. Dated 1278, it does mention the manor of River but no chapels; LPL: MS 241, f. 44.

⁷ St Radegund’s register does contain an undated (early) list of rents from River, among the tenants were Laurence son of Hewe, mason, and the heirs of Simon the tiler; Bodleian: Rawlinson MS B336, 248.

⁸ LPL: MS 241, f. 42v.

⁹ LPL: ED 2054, 2055, 1378.

¹⁰ TNA: SC 12/20/22.

Fig. 1 Map showing location of Crabble in relation to relief, medieval religious sites and Watling Street.

Fig. 2 Map showing location of Crabble in relation to River parish boundary, medieval sites and trackways (symbols as for Fig. 1).

Fig. 3 Plan showing location of excavated trenches and main medieval features.

Fig. 4 General plan of medieval structures and features in Trench 10.

Fig. 5 Section across Structure 1, adjacent features and deposits (Section 23).

Fig. 6 Section across Structure 2 and adjacent deposits (Section 26).

Table 1. Details of excavated pits

F. No.	Shape	Length (m)	Width (m)	Depth (m)	Sides	Base	Notes
84	Oval	1.22	1.10	0.32	steep	flat	Animal bone and daub only
115	Circ.	dia. = 1.00		0.45	steep	round	No finds
117	Oval	1.30	1.10	0.50	steep	flat	No finds
119	Circ.	dia. = 1.60		1.00	steep	flat	1 pot-sherd, c.1050-1225
121	Oval	1.71	1.53	1.05	Slop-ing	dished	1 pot-sherd, c.1375-1525
136	Circ.	dia. = 1.80		0.80	steep	flat	1 pot-sherd, c.1075/1100-1250