A Roman enclosure system in the Little Stour valley at Ickham

A ROMAN ENCLOSURE SYSTEM IN THE LITTLE STOUR VALLEY AT ICKHAM AND WELL

christopher sparey-green

Observation of the cutting of a relief channel on the flood plain of the Little Stour between Littlebourne and Wickhambreaux has revealed part of a ditched enclosure system previously known from aerial survey and has also provided evidence dating part of it to the early Roman period. The watching brief was undertaken by the Canterbury Archaeological Trust between November 2001 and February 2002 on behalf of the Environment Agency. The channel was designed to alleviate flooding, the work comprising the cleaning of existing channels and the cutting of new sections of linking ditch over a distance of 1.5km (Fig. 1). The river valley here, between the hamlet of Seaton, near Ickham and the village of Littlebourne, is meadowland, now partly disused and overgrown. This report describes the features recorded in the excavation of the new channel and sets them in the context of the enclosure or drainage system previously seen from the air. Previous observations are here referenced to publications and/or Kent Sites and Monuments Record (SMR) numbers.

The course of the relief channel lay entirely within the floor of the river valley, starting on its eastern side with a link to the existing course of the Little Stour at Seaton Mill (TR 2255 5869) and then being conducted 2km up-stream to join with the river at Littlebourne (TR 2113 5752). The subsoil here is Valley Alluvium, the higher ground to either side recorded as drift deposits of Head Brickearth over a solid geology of Upper Chalk, with some intervening areas of Thanet Beds. River Terrace gravels also occurred along the eastern margin of the flood plain close to Ickham and Littlebourne. The channel revealed the Valley Alluvium here to be largely composed of coarse ill-sorted flint gravels and sand at least 2m deep. Part of a relict watercourse, an ‘oxbow’ of the Little Stour, was cut through at one point near Wickhambreaux and other sections of old river channel may also have been encountered near Littlebourne.

Archaeological background

The line of the relief channel lay close to several known archaeological sites, the most important dating to the Roman period although there are also some surviving late medieval or early post-medieval structures surviving within the present villages. The most important site is the Roman roadside settlement at Ickham, set in the valley beyond the northern end of the new channel at Seaton Mill (SMR TR 25 NW 46 and 136; Young 1981; Spain 1984).

The Ickham site was situated north-east of Seaton, downstream of the present works, and lay on the line of a Roman road first identified by J. Bradshaw and Dr J. D. Ogilvie in the early 1970s during observation of the expanding gravel quarries in this part of the Stour Valley. Bradshaw was directing the work of the Ashford Archaeological Group in the area of the Ickham pits, and Ogilvie working with a local group closer to the Wingham River. Their work is invaluable in recording features now destroyed and culminated in the identification of the site of at least two nationally important Roman watermills, the focus of formal excavation by Chris Young in 1974.¹

The Roman road at Ickham was identified from air photographs² taken over a number of years, the road-side ditches and the metalled surface registering as differential plant growth in the flood plain of the Stour between Wickhambreaux and the later quarry site (SMR TR 25 NW 58; TR 2250 5883 to TR 2327 5907). Its existence was confirmed during the quarrying, and metallings and road-side ditches were seen to coincide with the observed cropmarks. Slight earthworks also survive in the meadows north-west of Seaton Mill, air photographs clearly showing the road-side ditches (that on the north-western side interrupted), with the terminals returning to flank a track-way or the sides of two adjacent enclosures. To the south-west of Wickhambreaux the road line is lost in arable but the earthworks are again recognisable in Pine Wood, on the higher ground west of Littlebourne (Panton 1994, 10-12). Here the road makes a definite change of angle to continue towards Canterbury, for the most part along the line of the present road.

The main Ickham site comprised firstly a length of road laid out early in the Roman period across the gravel and peat of the river valley, adjacent to the natural course of the river. Artificial channels, feeding a series of mills, adjoined the road on the east while, on the western side, enclosures containing occasional burials and other features were identified. The main settlement is not identified but may have lain to the north-east between the crossing of the Little Stour’s ancient course and the Wingham River, traces of buildings being reported as destroyed during the early phases of gravel extraction. At both crossings timber work from presumed bridge sites was observed during their destruction in the gravel extraction process. On the north side of the Wingham River the flanking ditches of a continuation of the road eastward are perhaps visible on air photographs at TR 2435 5918.³ A road at that point could then follow the high ground through Perry to Ash and link with the major road at Each End to provide a crossing to Richborough via a causeway or ferry (Hicks 1998). As Panton (1994) and Wilkinson (2002) have suggested, the more southerly route via Wingham and the present A257 is not plausible, the latter being an eighteenth-century turnpike.

Knowledge of sites other than Ickham in the Little Stour valley is limited although up-stream to the south-west casual finds are recorded within the villages of Ickham, Wickhambreaux and Littlebourne. The report of a Roman bronze vessel from near Ickham church may relate to Roman or Anglo-Saxon burials (SMR TR 25 NW 33). A late Roman coin hoard and Anglo-Saxon brooches, possibly grave-goods from burials, are reported from the parish but the exact find-spot is unknown (SMR TR 25 NW 28 and 134). The best-known site is the high-status building and find-spot of official lead seals, this lying near Britton Farm, on the higher ground east of the village (Detsicas 1983, 98; TR 25 NW 8 and 132-3).

Little is known from Wickhambreaux save a Roman cremation urn and Anglo-Saxon urns near the church, burials beside the Roman road line and peripheral to the main site (SMR TR 25 NW 11 and 35). A double ring-ditch in the same area might be a ploughed-out barrow (SMR TR 25 NW 139). A site at the northern end of Littlebourne, close to the church, has produced reports of early Roman pottery and wall plaster, which, if confirmed, might indicate a high-status building here also (SMR TR 25 NW 36). Late Roman coins are reported form the southern side of the village on the opposite side of the present Canterbury road (SMR TR 25 NW 6). Again, Anglo-Saxon graves are recorded from an unknown site within the parish (SMR TR 25 NW 26).

The main evidence for settlement in the area of the relief channel is from patterns in the grass cover of the flood plain shown on air photographs of the area east of Littlebourne (concentrated in the area TR 215 579 to 213 576) and also between Ickham and Wickhambreaux (centred at TR 218 584) (SMR TR 25 NW 59 and 60). Another area of ‘ladder-like’ enclosures was observed south east of the Littlebourne, beyond the area cut by the channel (TR 211 572; TR 25 NW 61). As differential growth marks in the meadow these may be of any period but, in the first area, a series of linear marks aligned at right angles or along the axis of the valley appeared to mark the line of double ditches of enclosures on the valley floor, these corresponding with some of the features in the relief channel which produced dateable finds. This confirmed that some at least were ancient features and not merely recent drainage ditches. In the second area it proved more difficult to identify the features seen on air photographs and no confirmatory evidence was obtained.

The channel did not cut through any major known post-Roman site but at the approximate mid point of its course it passed close to the medieval villages of Ickham and Wickhambreaux and it terminated east of Littlebourne. Reports of Anglo-Saxon urns and other finds from each parish suggest the presence of settlement from early in that period. Ickham is first mentioned in 724 as Ieccaham and thereafter occurs in charters of King Offa dated 785-786 and 791 (Wallenberg 1931, 32-3; Sawyer 1968, nos. 123, 125 and 1614), the latter granting the village to Christ Church, Canterbury. The name has been derived from Yoke-ham. The name of Wickhambreaux occurs in 948 as Wicham, later appearing as Wicheham in Domesday Book (Wallenberg 1931, 277). Gelling identifies this as a ‘wickham’ name of a type found elsewhere in settlements on or close to Roman road lines (Gelling 1978, 67-74). Although the Ickham Roman settlement lay within the boundaries of that parish, Wickhambreaux village is topographically closer to the site and lies on the projected line of the road to the south-west. This village is more likely to be the successor settlement to the known Roman occupation, the church set on the old road line where it gains the higher ground, and the mill on the adjacent river course. Littlebourne is first recorded in 696 as Liteburne, the name perhaps an early alternative for Little Stour (Wallenberg 1931, 21).

In Domesday each village is recorded similarly, Littlebourne notable for its lower population but, at 7 sulungs, almost twice the area of the others (Morgan 1983, 3,8; 5,124 and 7,5). All three had churches and meadowland but Ickham had the greatest number of ploughs and mills while Wickhambreaux had a larger number of specified animals, salthouses and fisheries. These differences might suggest Wickhambreaux had the most varied economy and Littlebourne was relatively the poorest of the three.

Since the work at Ickham in 1973-4 there has been no recent arch-aeological investigation in the area other than the work on Ickham Court Farm which produced evidence for early medieval occupation near the centre of Ickham close to the church (Sparey-Green et al. 2003). (See also pp. 266-7 of this volume.) The cutting of the relief channel was the first extensive ground disturbance since the time of the gravel extraction in the 1970s, and the most significant discoveries were made in the area of Littlebourne where it cut through a series of ditches, pits and possible river courses in the area of the vegetation marks observed from aerial survey.

Nature of works and method of archaeological observation

Work on the channel started at the north-eastern end, at Seaton Mill, and was at first directed along the eastern side of the flood plain and then along the centre of the valley between Wickhambreaux and Ickham. It then crossed over to the western side of the valley to link with the main channel of the Little Stour at Littlebourne. The watching brief allowed only for the observation of the features as exposed in cutting the channel, without further excavation. The cutting of the channel caused only minimal disturbance, removing only the uppermost valley gravel deposits into which features were cut, although the excavation at one point (TR 2127 5769) cut through a slight rise in the valley gravel, exposing more of the features in section. The channel was 6m wide and of varying depth to create a gradient to promote the correct water flow. Allowing for variations in the level of the valley floor the channel varied between 0.3-1.1m in depth, its profile rounded with sides gently sloping. No vertical sections of features were exposed and the plan of the features was also distorted by the curving floor of the excavation. In the following account figures for the depth of features will not normally be given in view of the lack of data for this parameter. For ease of reference the channel has been divided into six areas [A-F] shown on Fig. 1. Areas A-C, comprising the northern 1.4km of channel, were almost devoid of archaeological features but Areas D-F, the southern 550m nearest to Littlebourne, did reveal a variety of features deriving from river activity, drainage or ditching operations and more recent disturbances.

In Areas A-C the channel followed existing drains, the operation consisting largely of the cleaning out of an existing ditch to a wider profile. At one point a new section of channel was cut, one 100m stretch north of Ickham being cut between TR 2220 5840 and TR 2212 5843, across the site of an old rivercourse, described below. The other section of newly-cut channel in Areas D-F extended from TR 2142 5798 to TR 2113 5752.

The subsoil exposed along the length of this shallow excavation was almost entirely valley gravels but in the first 400m section in Area A the cleaning of the existing ditch on the south-eastern side of the flood plain revealed a brown clay silt overlying the natural gravel, this upper deposit possibly derived from erosion of the brickearth deposits on the hillside to the south-east. The following 100m section of newly excavated channel cut through a slight rise in the valley gravel deposits but also crossed a strip of lower ground containing dark silt at TR 2215 5842. Here a broad band of up to 0.5m of dark peaty humus was exposed over a subsoil of gravel and chalk. This feature could be traced as a curving strip of wet ground across the field to north and south and, from the pattern of drainage features on Ordnance Survey maps, it would appear that there was an ancient oxbow in the river here, looping south from the present course at Wickhambreaux towards Ickham.

Another strip of low ground flanked by rises in the valley gravel deposit occurred in Area B, 250m to the west, where a more substantial cutting was made between Wickham Lane and Drill Lane, north-west of Ickham (TR 2185 5838). A depression, here filled with black peaty silts between two raised areas of natural gravel, may also have derived from the silting or shifting of an old water course although the line of any river course here was not easy to trace within the meadow land either side of the new cutting. Numerous linear features in this area of flood plain between Ickham and Wickhambreaux and visible on air photograph CUC BXM 38-43 were not identified in the course of the cutting of the channel and their date and purpose must remain uncertain. Some at least do not relate to the proposed old river course so are potentially of some age. To the south, in Area C, the new cutting followed an existing channel and no clean sections of the valley deposits were exposed.

The most significant observations were further south in Areas D-F, south-east of Littlebourne, containing significant archaeological deposits. Within this section, opposite the known sites of Romano-British finds within the village, the channel exposed numerous features crossing the trench and two possible sections of river channel (Figs 2 & 3). The description of these features is based on the visual examination of the base and sloping sides of the cutting without any hand excavation of their fill. With the nature of the operations it was not possible to investigate the full depth of features and the few finds were retrieved from the surface of the fills. The location of the more significant features in the Areas D-F is given by OS map reference.

Observations in Areas D-F

Only a summary of the main categories of feature will be given here. The most substantial features were extensive zones of dark silt and gravel which probably corresponded with sections of silted river bed. The first of these was encountered at TR 2130 5776 where an ill-defined area of dirty gravel (19) was cut through. Immediately to the south-west was a larger feature [24] which extended for a further 30m to the south. The eastern edge of [24] could be traced crossing the trench diagonally before returning at right angles, the fill consisting of dark grey-brown clay silt with lenses of clean gravel along the east edge and no obvious man-made inclusions. The outline of this deposit on the floor of the excavation and the nature of the fill suggested that this may have been the eastern side of a river course the greater part of which lay to the north-west, closer to the present river.

Further south-west, at TR 2124 5765, a band of dark soil (65 and 66) filled the north-eastern side of a feature [67] at the southern limit of this section of the channel. The south-west extent of this remains uncertain, this area being obscured by the concrete block surface of a ford laid across the channel and an area of disturbed ground, probably the site of modern sewer trenches linking with the now defunct sewage works on the eastern side of the valley. Further south-west, in Area F, an extensive homogeneous deposit of black greasy soil and pebbles (44) and (45) filled a large feature [46] which extended to TR 2121 5758 where the new channel crossed its western edge. The fill near this edge contained a significant amount of locally-made coarse ware of the late first century bc to the second century ad. This, again, is likely to have been an old course of the river, here originally heading almost due south.

A large feature centred at TR 2136 5788 consisted of a shallow hollow [83], 0.6m deep and filled with (82), a pale grey clay sand sealed by dark brown slightly peaty silt. It produced no dateable finds and might be some minor feature containing water, although not as substantial as the river channel.

In the southern end of the channel two discrete features, [48] and [50] were identified. The first appeared to be a parallel-sided trench 8m wide and lay close to the western edge of [46] while the latter was an oval pit of similar width and lay close to the existing river channel (beyond the limits of Fig. 3). Both contained pebbles and dark silt and contained no dateable finds. The most significant features were three linear features [33], [37] and [41] which crossed the channel diagonally at TR 2125 5768. The first of these [33], was traced for 14m crossing the trench on a north-east to south-west alignment, the feature approximately 0.5m wide and at least 0.8m deep, filled with (32) dark grey-brown silt. At its south-western end it joined [37], an apparent return which crossed the trench to the south-east. This feature was 0.7m wide, filled with similar soil and appeared to continue beyond the trench limits. Further south-west and 8m from this was a similar feature [41], comparable in fill and dimensions and approximately parallel to [37]. A very truncated linear feature [39] immediately south of [37] may represent a re-cut of the latter but on a slightly different alignment. Situated 40m to the north-east, an irregular U-shaped feature [30], was possibly a shallow hollow only 0.5m deep, the outline fortuitous and the result of its truncation. The dark grey silt (29) filling contained sherds of Upchurch ware and ‘Belgic’ ware besides a sherd of South Gaulish samian and traces of burnt clay and some small unidentified iron objects. Adjacent to this was an ill-defined area of soil (31), perhaps a disturbed feature, which produced another sherd of ‘Belgic’ pottery. Particular interest attaches to these features because they appear to correlate with more extensive linear features observed on air photographs, the latter forming an extensive enclosure system in the flood plain; this is further discussed below.

This area also revealed discrete oval patches of soil [69], [75] and [81], which were probably the fill of pits, approximately 1.0-1.5m in diameter. Other irregular features [77], [35] only partially exposed along the eastern side of the channel may represent further pits, the bifurcated outline of the first suggesting it had been re-cut. These may have formed part of a dispersed cluster of small pits none of which produced any dateable finds.

At TR 2125 5767 two irregular features, [54] and [56], extending beyond the limits of the trench may have been two more extensive pits or hollows. These lay immediately south of the boundary ditch [41], the former up to 18m in diameter and containing a gravely lower fill (55), the upper fill consisting of dark silt and pebbles (42). An area of similar dark soil (53) on the eastern side of the trench was the truncated remnant of the fill of [54]. Four irregular features [58], [60], [62] and [64] immediately to the south, all contained similar dark pebbly silt and may have been smaller pits.

The remaining features, all undated, consisted of groups of narrow gullies or ditches which occurred at intervals in groups of two or three in Areas D and E between TR 2134 5784 and TR 2126 5770. This category of feature comprised six groups, [3] and [5]; [7] and [11]; [14], [16] and [18]; [21] and [23]; [26] and [28]; [71], [73] and [79]). They were mostly between 0.5 and 1m wide and at least 0.3m deep filled with dark-grey-brown silt and gravel. The majority crossed the channel on a north-west to south-east alignment but [7 ], [11] and [14] differed in that they followed a more northerly alignment; [7] was also 5m wide. The majority, again with the exception of [7], were in groups at intervals of between 20 and 30m. Only [26] produced a sherd of Roman pottery, this in an area close to [30] and [31] which produced most of the pottery from Area E. The character of these features suggests that they formed part of a regular system and are more likely to be recent drainage ditches, although feature [7] may have belonged with the earlier boundary system. One ditch [9] in Area D was notable in following a north-east to south-west alignment, its irregular outline suggesting that it had coincided with a small pit on the north-west side. The dark silt fill (8) provided no clue as to its date, but the alignment was closer to that of the main boundary features 30m to the south-east.

Discussion

Although only the most superficial observation was possible of the features exposed in cutting the new relief channel some information was recovered both relevant to the study of sites in the vicinity and of significance to the wider study of river valley archaeology in eastern Kent. The character and extent of the deposits observed allows two broad categories to be distinguished. Firstly there are the more extensive areas of dark soil where the nature of the more organic rich silt and their location in relation to visible features and slight earthworks on the floor of the valley suggests that they were old silted river channels. The first of these was in Area A where, as described above, an ancient oxbow could be traced in the valley floor. Although there is no direct dating evidence for this feature the presence of a mill 250m to the north, in the village of Wickhambreaux and fed by a stretch of canalised river channel cutting across the top of the bow, would suggest this pre-dated that mill and therefore was of early date. Another section of possible river channel encountered in Area B was not easy to trace but it could have continued north-east to return and link with the oxbow just described.

In Areas E and F two extensive zones of dark silts could represent a further section of old river course. Firstly, deposits (19) and (24) at TR 2130 5776 may have been deposited in an old bend of the river channel, east of the present canalised course. Two hundred metres south the extensive dark silts (44) and (45) in Area F between TR 2124 5765 and TR 2121 5758 were almost certainly part of a watercourse heading south, 70m east of and almost parallel with the existing river. The original course here cannot be reconstructed exactly but in these areas the new channel may have been cut across two bends in a sinuous river course which occupied the ground between the new channel and the existing canalised river. The more northerly bend was represented by (19) and (24) in Area E and the southern curve by [67] and [46] in Areas E and F. The presence of Roman pottery in the latter river channel suggests that it was of some antiquity and a feature of the early landscape.

Similar riverine features were identified at Ickham forming part of a somewhat broader stream, flanked there by man-made channels serving the water-mills and associated with timber structures and quern-stones (Young 1981; Spain 1984; CAT forthcoming). No finds of that nature were recovered here so an interpretation as natural watercourses can only be proposed.

The densest cluster of features in Area E lay between the two postulated sections of river, occupying the inside curve of the double bend in its course. In this area some correlation is also possible between the recorded features and the linear boundary system recorded from aerial survey. The available evidence from air photography shows a number of linear boundaries, the most significant from an apparent trackway flanked by double ditches and describing two right angle bends 185m apart, the northern one returning to the north-west for a distance of at least 70m, the other to the south-east for 100m. Other linear features may be boundaries defining enclosures either side of the axial trackway. These subdivisions appear to suggest the existence of field ends approximately 45 and 60m long. This system coincided with the line of the new channel in Areas D and E, but another more irregular pair of boundaries can be traced heading north-east, away from the line of the new channel. The general axis of the system coincides with that of the valley and does not reflect that of the Roman road which here lies some 600m to the north-west.

The majority of the discrete features, probably small pits or hollows occurred in the southern part of Area E in the region of the trackway. The more localised but extensive features with curving outlines, (19) and [56], both of which adjoined the larger channel-like feature (24) and [46], could result from animals creating hollows or from the erosion of river banks in gaining access to the water. The smaller features could indicate settlement features, the pottery of the late first century bc to second century ad suggesting limited activity in the area of more than one phase, as suggested by the coincidence of these pits with the apparent line of the axial trackway.

The numerous narrow gullies or ditches crossing the cutting remain undated and could also relate to the early boundary system. Their number and spacing, however, could suggest they belonged to more recent enclosure and drainage patterns perhaps connected with water meadow management.

The small quantity of dateable finds was mostly of the first and second century ad, with some sherds perhaps as early as the first century bc and others of types continuing to Late Roman times. The majority came from feature [30] and deposit (31), coinciding approximately with the line of the axial trackway. The other group was derived from (45) on the western side of the putative river channel [46]. This can only give the most general indication of activity in the early Roman period, contemporary with the earlier phases of road and riverside settlement at Ickham, but would suggest that the main enclosure system recognised from air photography was of early Roman date. The lack of later material is in marked contrast with the riverside site at Ickham where the majority of activity was of the late fourth and early fifth centuries and may suggest that at the present site occupation had ceased or shifted from the flood plain. Changes in ground-water levels may have affected it, causing its abandonment in the Roman period and a shift to the slightly higher ground occupied by Littlebourne. This may parallel the situation lower down the Little Stour valley, where, as suggested above, the Ickham site may have been succeeded by Wickhambreaux. Occupation at Ickham continued longer possibly because of the need to exploit the mills and river course and because the road crossing the flood plain would have aided access even if the river levels had risen. Only with the eventual flooding of the road and the mill sites would that area have been abandoned and replaced by settlement at Wickhambreaux and Littlebourne, higher up the valley, the flooding preserving the mill structures for which the former site was so notable. At the present site the trackway alignment and the few boundaries either side indicate extensive rectilinear enclosures of approximately 185m long by 30, 45 or 60m wide. This contrasts with the system 600m further up the valley and south of the village where a row of smaller rectangular enclosures, approximately 50 by 70m, can be recognised adjacent to two more extensive linear compounds similar in size to the present site (TR 25 NW 61).

The limited extent of the recorded enclosure system allows little detailed discussion of the comparanda, such sites being rare in Kentish river valleys. It is only in the study of, for instance, the river gravels of the Thames Valley that extensive systems in a similar environment have been plotted. The main enclosures would be larger than the complex small field systems at, for instance, Claydon Pike, Gloucs., or Hayton, East Yorks. (Miles 1989, 124-5; Taylor 2001, 53) but the southern system is superficially similar to the ‘ladder enclosures’ noted in the northern lowlands (Dark and Dark 1997, 95-6). The axial droveway on the present site is reminiscent of the system at Brancaster (Norfolk) where parallel roadways east of the fort are set a similar distance apart in an enclosure system probably dating to the early third century (Hinchliffe with Sparey-Green 1985, 176-7). There, however, the intervening space is filled with plots as small as 25 by 40m, each possibly containing structures facing onto the roads. The present site is unlikely to have been as intensively occupied, perhaps serving as meadowland bisected by droveways, the individual units being on a larger scale than in the case of arable land or settlements. The tracing of this archaeological landscape over a larger area combined with environmental study of the associated deposits would be necessary before making further comment on the past land use.

acknowledgements

The writer is grateful to the Environment Agency for funding the watching brief and publication. Thanks are also due to Colin Demeza for arranging the watching brief and Kit McLean who supervised operations on site. Also to Paul Bennett who managed the project, Peter Atkinson who drew the site plans and Mark Davey who identified the pottery. The report was edited by Jane Elder.

endnotes

¹ Young 1981. This site will be the subject of a major monograph (CAT forthcoming).

² Air photograph HSL Kent 67 16, 17 July 1967, Run 25, photo 6637.

³ Ibid.

bibliography

CAT, forthcoming, The Roman Watermills at Ickham, Kent.

Dark, K. and Dark, P., 1997, The Landscape of Roman Britain (Stroud).

Detsicas, A., 1983, The Cantiaci, Gloucester.

Gelling, M., 1978, Signposts to the past: Place-names and the History of England (London).

Hicks, A.J., 1998, ‘Excavations at Each End, Ash, 1992’, Archaeologia Cantiana, 118 (1998), 91-172.

Hinchliffe, J. with Sparey-Green C. et al., 1985, Excavations at Brancaster 1974 and 1977, East Anglian Archaeology Report 23.

Miles, D., 1989, ‘The Romano-British Countryside’, in Todd, M. (ed.), Research on Roman Britain 1960-89, Britannia Monograph Series 11 (London).

Morgan, P. (ed), 1983, Domesday Book, 1, Kent (Chichester).

Panton, F.H., 1994, ‘The Canterbury-Richborough Roman Road - a Review’, Archaeologia Cantiana, 114, 1-16.

Sawyer, P.H. 1968, Anglo-Saxon Charters (Royal Historical Society, London).

Spain, R.J. 1984, ‘The Second-Century Romano-British Watermill at Ickham, Kent’, History of Technology, 9, 143-180.

Sparey-Green, C.J, Linklater, A. and Sweetinburgh, S., 2003, ‘An Archaeological Excavation at the Former Site of Ickham Court Farm, Ickham, Kent’, unpub. CAT Client Report.

Taylor, J. 2001, ‘Rural society in Roman Britain’, in James, S and Millett, M. (eds) Britons and Romans: advancing an archaeological agenda, CBA Research Report 125 (York).

Wallenberg, J.P., 1931, Kentish Place-Names (Uppsala).

Wilkinson, P., 2002, Practical Archaeology 6, Summer 2002, 10-11.

Young, C.J., 1981, ‘The Late Roman Water-Mill at Ickham, Kent and the Saxon Shore’, in Detsicas, A. (ed.) Collectanea Historica: Essays in Memory of Stuart Rigold, 32-9.

Fig. 1 Site Location plan.

Fig. 2 Little Stour relief channel showing Areas D and E.

Fig. 3 Little Stour relief channel showing Areas E and F.