Excavations at Each End, Ash near Sandwich, 1992

EXCAVATIONS AT EACH END, ASH, 1992* ALISON J HICKS With contributions by R. Anderson, M. Bates, B. Dickinson, P. Garrard, L. Harrison, E. Healey, N. Macpherson-Grant, D. F. Mackreth, A. Savage, V. Tatton-Brown THE EXCAVATION INTRODUCTION AND SUMMARY The site at Each End, Ash (TR 304585) was located by the Canterbury Archaeological Trust during trial trenching along the line of the Ash by-pass, now constructed and directing traffic along the A257 to the north of the village. No archaeological remains had previously been identified within the impact area. Upon the discovery of a single cremation burial and metalled surface within one of the trenches, a broad area of topsoil was stripped. The total area uncovered, in which archaeological features were visible within the impact zone, was 3,500 square metres (Fig.l). Excavation of the site was undertaken between April and July 1992. The occupation uncovered upon the site was principally of Roman date. However, the activity within the area since prehistoric times was suggested by the retrieval of a Mesolithic flint tranchet axe, a Neolithic leaf-shaped flint, a barbed and tanged arrowhead of Early Bronze Age date and sherds of prehistoric pottery. Iron Age activity was indicated by the presence of a ditch, gullies and a small quantity of pottery. The project was entirely funded by Kent County Council Department of Highways and Transportation. The Canterbury Archaeological Trust has contributed to the cost of publication. 91 ALISON J. HICKS Roman occupation was extensive across the entire excavation area. The most prominent feature was a metalled road, with associated drainage ditches. Thought to have been laid in the second half of the first century A.D., the road appears to have replaced an earlier trackway. A Roman settlement lay adjacent to the road. This was represented by a small wooden structure with associated yard surfaces, together with a hearth, three wells and several pits which perhaps indicated small scale industrial activity on the site. The predominant feature associated with Roman settlement, however, was a sequence of open gullies and ditches, cut in various directions across the site. Their complexity suggests that drainage of ground water was a persistent problem throughout the lifetime of the site. The major period of settlement appears to have been the mid to late second to third centuries A.D., although occasional fragments of early Roman and late first- and early second-century pottery perhaps point to earlier activity. Coin evidence suggests that the site may have contracted, or even been temporarily abandoned, in the later part of the third century. Three cremation burial groups were discovered, one to the north of the road line, the others to the south, yielding pottery which was largely mid to late second century in date. A number of the burials were rich in grave goods, some containing as many as ten ceramic vessels, whilst a lamp, copper alloy objects and remnants of footwear were also retrieved. A single amphora burial was discovered, as well as a box burial. In addition, an inhumation burial was excavated, set within a large rectangular grave, from which late second- to thirdcentury pottery was recovered. Evidence of later Roman activity was tenuous, although fragments of iate third- to fourth-century pottery were recovered. A deposit of dark silt covering the road and its ditches, yielded a large quantity of mid to late fourth-century coins. Thus by this period the road and its ditches were no longer maintained, although the line probably remained in use as a trackway. The close proximity of the major Roman site at Richborough (only 2.2 km. (1 Vi miles) to the north-east) means that any discussion on the nature of the settlement and the archaeological discoveries at Each End must include some consideration of contemporary events at Richborough. In Roman times Richborough was located on an island divided from mainland Britain by the Wantsum Channel (see below, The Topography). The site became a major port during the Roman period. As the bridgehead for the Claudian invasion, a ditched enclosure was built in the first half of the first century A.D.; this was rapidly replaced 92 nC.rrce mation burials G21 y Site boundary Road metalling G7 .• Location Plan Richborough Castle Remains n Sandwich SP Y trS^ . - • ^ G14 . / Well G9A " TT G19B Cremation burials O G22 O O \ \ \327\ > \G28 V » \ G32 \ \G26 V'A > \ P Drainage gullies G33 ,/' '».-*--' .^Structure G15 V s?^. Well vG19A,'- VTV Cremation burial .G23 G32 • ^ Roman road Yard surface 10 20 30 Fig. 1. Main site plan, showing the location of all principal features EXCAVATIONS AT EACH END, ASH, 1992 by a wooden fort which remained in use until the late first century. The port flourished during the late first and early second centuries, with a civilian settlement encouraged by trade generated by the military supply base. The settlement declined during the later part of the second and the early third centuries, perhaps when Dover became the more dominant commercial centre. In the late third century Richborough was redeveloped as a site for the military and a new fort was constructed (Johnson 1970). The alignment of the road discovered at Each End runs towards the former Wantsum Channel at Little East Street Farm, from where it is postulated a ferry crossing could have connected with Richborough (see below, p. 102). Though the insubstantial nature of the road suggests it was unlikely to have been a major route it was undoubtedly the presence of the road and the proximity of Richborough which provided the impetus for occupation. THE TOPOGRAPHY The site lies just below the 10 m. land contour, upon Tertiary deposits of Thanet Beds, comprising sands, clays and silts, overlain by Head Brickearth. The underlying solid geology is predominantly Cretaceous Chalk. Between the Isle of Thanet and the Kent mainland exists the Richborough Syncline, a shallow basin formed by a warp in the underlying chalk. Following the general rise in sea level which occurred after the last glaciation, c. 15,000 B.P., the syncline was flooded by the sea and the Wantsum Channel was formed, creating a sea passage linking the River Thames and the English Channel. The Wantsum Channel is today represented by a broad belt of low-lying marshland, varying from c. 1.2-4.8 km. in width, dissected by man-made dykes and rivers flowing out to sea. This marshland formed over thousands of years, as the Great Stour and a number of smaller rivers flowed eastwards into the Channel. As they drained much of eastern Kent, they deposited silts, a process probably enhanced by the formation of the Stonar gravel bank to the east, further restricting the free flow of water through the Wantsum Channel. Silts would also have been produced by erosion of the cliffs at Deal and Reculver (Hawkes 1968). In the Roman period, despite the advanced state of the marsh formation, a sea passage appears to have still existed in the Wantsum Channel (Ogilivie 1968). This channel would have allowed ships to navigate from the south through to the mouth of the Great Stour 93 ALISON J. HICKS River, and so on to the Roman town at Canterbury. Further north, they would have met the River Thames. The forts of Reculver and Richborough were strategically sited at each end of the Wantsum Channel. The Roman port at Richborough lay upon an island (now known as the Isle of Thanet) to the east of the Wantsum Channel. The site excavated at Each End was situated on the fringes of the channel to the west. As has already been mentioned, the link between the two sites afforded by the newly discovered road, would have necessitated a ferry crossing somewhere in the vicinity of the existing Little East Street Farm (see also p. 102, below). Today, the land east of Each End drops away to form a broad, low-lying area of predominantly pasture, in places poorly drained. At Each End itself, lying upon a stretch of gently sloping ground at a height of c. 9-10 m. O.D., and upon the slightly higher, well drained ground to the north-west, west and south-west, the land is largely under arable cultivation, in part used for fruit growing. DESCRIPTION AND DISCUSSION OF THE ARCHAEOLOGY PREHISTORIC OCCUPATION Evidence of prehistoric occupation was slight. The excavations yielded a flint tranchet axe and obliquely truncated blade of Mesolithic date, a leaf-shaped flint of Neolithic date, a barbed and tanged arrowhead of Early Bronze Age date and two sherds of pottery tentatively identified as of Late Bronze Age/Early Iron Age transition type. These finds were all residual, but suggest activity within the locality. IRON AGE OCCUPATION (FIG.2) Iron Age activity was indicated by the presence of a large ditch and a number of smaller gullies, as well as by a small quantity of pottery, some residual. The large ditch (G.3)1 was located towards the 1 The stratigraphic context units (prefixed C.) were clustered into Sets (prefixed S.) and Groups (prefixed G.), indicating particular activities or groups of activities occurring upon the site during a particular time frame, using a hierarchical grouping system (Clark 1988,1992; Pearson and Williams 1992). Details of the Sets and Groups are provided in the site archive, available for study at the offices of the Canterbury Archaeological Trust, 92a Broad Street, Canterbury, CT1 2LU. 94 r" Site plan 1 J \ m. A1 :f^ A Section Will III IT A1 m 325 Ui G24 T Drainage channels Ditch G3 m X o 00 is > GO a; to 0 1 20m. Fig. 2. The Iron Age ditch (G.3) and drainage channels (G.24, G.25). ALISON J. HICKS southern side of the site. Up to 1.70 m. in depth and with a maximum excavated width of 4.80 m., this massive, linear cut ran east-west across the excavation area. The substantial nature of the cut (C.2033), with steep, undulating sides and a broad flat base, at first suggested that the feature was defensive in nature. However, its meandering plan and the nature of the infilling deposits did not support this interpretation. Instead, the cut appears to have been excavated for drainage. It would have contained water during much of its lifetime, so that as it lay open, it gradually infilled with erosion deposits, waterlain silts forming horizontally along its length. Smaller ditches and gullies probably funnelled water from the surrounding land into the ditch and then down towards the Wantsum Channel to the east. A sequence of these drainage channels (G.24) was identified during the excavation. Varying in depth from 0.15- 0.55 m., they yielded only struck flints and a single pottery sherd. Clearly well maintained, having undergone a number of re-cuttings, they were later replaced by a single, broad but shallow gully (G.25), also draining away towards the large ditch. The original date of the ditch is difficult to ascertain. Partial excavation of the feature provided no evidence of cleaning or re-cutting, suggesting that the deposits represented the entire original infilling. The single sherd of pottery of Iron Age date recovered from a deposit located just above the base may therefore indicate that the feature itself was cut during the Iron Age. Although no features specifically indicative of nucleated settlement were identified, residual pottery sherds, recovered both from the ditch and other features on the site, suggest that an Iron Age period settlement was present within the immediate vicinity of Each End, with its main period of activity occurring c. 650-450 B.C. The large ditch was probably related to this settlement activity. It may have been cut for land drainage within an area of crop cultivation. Sherds retrieved from the infilling deposits of the ditch suggest, both by their date and abrasion patterns, that the feature was open for a considerable period of time, perhaps extending into the early first century A.D. In addition, sherds of residual pottery recovered from other features and tentatively dated c. A.D. 25-75 may suggest multi-period pre-Roman occupation was present within the area of Each End. PRE-ROAD ACTIVITY (not illustrated) Features (G.4) were uncovered which lay directly below the line of the trackway. These consisted of two gullies, a post-hole and two 96 EXCAVATIONS AT EACH END, ASH, 1992 features of unknown function. In addition, a short length of gully (G.39) was excavated, aligned approximately north-south, lying below the line of the southern road ditches and heading beneath the line of the later track. What form the pre-trackway occupation took is not clear; most of the features provided little insight into the nature of settlement, although the two gullies suggest a process of land management, possibly related to the Iron Age settlement discussed above. No artefacts were recovered from the features and their date is therefore uncertain. Of importance, however, was the location of these features, which demonstrated that occupation occurred before the existence of the trackway. The trackway itself was therefore not the sole impetus for occupation within the area, although it may have been important during the Roman period. ROMAN OCCUPATION The trackway (not illustrated) Before the laying of the metalled road surface, a trackway existed which followed the same alignment. Remnants of this survived in the form of wheel ruts and puddle holes (G.5). In addition, shallow drainage gullies had been cut running parallel with the outer edges of the track. It was originally postulated that the trackway formed during the road construction process. This would have involved considerable activity; carts bringing gravel, tools and equipment, men laying the metalling and cutting ditches, and a substantial amount of earthmoving. Upon the clayey ground surface, this would have formed wheel ruts and a variety of irregular features. However, it seems unlikely that this process would have required the cutting of drainage gullies or whether it would have formed wheel ruts aligned solely in the direction of the trackway. It seems more probable that the trackway was an early route, later deemed inadequate to support the amount of traffic using it and consequently metalled. It may have connected small settlement sites or provided access to Little East Street Farm and a ferry crossing to Richborough. The date of the original trackway is not certain since the features yielded no finds. It was aligned towards Richborough, where Iron Age occupation has been uncovered so it may have existed prior to the Roman period. There was no direct evidence to suggest that it was associated with the large Iron Age ditch to the south. The 97 ALISON J. HICKS two features lay upon separate alignments, but this did not necessarily preclude an association. The trackway probably remained in use until the early Roman period and the laying of the metalled road. The road: metalling (Fig.l) The trackway was subsequently metalled to provide a more substantial surface. Initially, its construction involved the deliberate deposition of earth above the line of the earlier trackway (G.6). Within one area, a moderately compact clay loam had been used to level off a small hollow, whilst along the centre line of the road earth had been laid to form an agger, enhancing drainage from the surface of the road. The silty clay forming this central ridge probably came, at least in part, from earth upcast by the digging of the road ditches. The effect so created was of two 'lanes', to north-west and south-east, separated by a low central embankment. A further, mottled earth deposit was noted along the north side of the road (G.2) which seemed to have been formed by surface activity when the road was under construction, a process which would have involved much trampling across the ground by men and animals transporting materials and equipment. The road metalling (G.7), aligned north-east to south-west, merely consisted of a single deposit of compacted flint pebbles set within a sparse sandy silt matrix, lying to a maximum depth of 0.14 m. but surprisingly 15.00 m. in width, extending up to 18.00 m. in isolated areas. The contour of the central agger was retained within the metalled surface, a deliberate feature of its construction, whilst at its edges the metalling was cambered, to enhance drainage from its surface into the flanking side ditches. Once laid, small depressions and hollows within the metalled surface rapidly filled with erosion silts (G.12), thus levelling out any irregularities. A number of long, linear depressions within the surface of the metalling were clearly identifiable as wheel ruts. There was evidence that some of the wear to the metalled surface was repaired by patching (G.13) but only within isolated, relatively small areas. Curiously, an area of patching identified along the south-east edge of the metalling consisting of small-medium flint pebbles (C.323) laid upon earth bedding (C.324), had the effect of levelling off the camber at the edge of the road and so would have restricted drainage into the road ditch. It is not known whether this was deliberate or merely a piece of poor workmanship. The patching along the north-west fringes of the road (C.369) petered out c. 3 m. before the edge of the earlier metalling, indicating that either the patching material itself 98 EXCAVATIONS AT EACH END, ASH, 1992 had later worn away or the full width of the road, as originally laid, was no longer required. The road: ditches (Fig. 1) Ditches were located to either side of the metalling, flanking the road. Cut primarily for drainage, they must also have served to prevent livestock from straying onto the road surface and possibly as markers, to denote the boundary of the property associated with the road. Immediately adjacent to the north edge of the road metalling lay a single ditch (G.8). It had been subjected to numerous re-cuttings during its lifetime (seven were identified during the excavation), no doubt necessary due to silt deposition from the road metalling and from the surrounding ground surface. The cuts, where they survived in any substantial form, varied in width from c. 1.70-2.50 m. and up to 1.50 m. in depth. In profile, most were bowl-shaped, with undulating sides due to successive episodes of erosion, slumping, cleaning and re-cutting. The base of one of the cuts (C.l 106, S.152) had been lined with a deposit of flint pebbles, probably in an attempt to enhance its drainage qualities. A single post-hole (C.1088) cut the base of one of the ditch cuts (C1063), whilst three stake-holes (C.1091, C.1097, C.1098) were identified cutting the base of another (C.1089). Since the ditch sequence was only excavated within a narrow slot, it is not known whether these formed part of a pattern of associated features or were merely isolated examples. Too many suggestions could be proposed for their function to make speculation worthwhile. Three lines of ditches were identified south of the road. The earliest (G.9) was located furthest to the south. Bowl-shaped in profile, with gently sloping and undulatory sides, the ditch cut (C.133/C.1029) was up to 2.90 m. wide and 1.20 m. deep. It displayed evidence of a single re-cutting (C.l 156), which reduced its depth to 0.75 m. A line of post-holes (C.l 133, C.l 134, C.l 135), each square in plan, cut the upper northern edge of the ditch, probably indicating the presence of a fence line. This may have been erected to prevent animals straying into the ditch when they were being herded along the road, no doubt when they were being taken to and from local markets. Subsequently, the ditch was deliberately backfilled by the dumping of rubbish deposits, containing quantities of charcoal, daub and pottery, within its upper levels, and a new ditch (G.10) was cut slightly to the north. This re-alignment was probably undertaken in order to insert a small structure (G. 15) and its attendant yard surfaces (G.16) (discussed below), which overlay the line of the former ditch 99 ALISON J. HICKS whilst respecting the line of the re-cut. Pottery of late second- to third-century date recovered from the primary ditch provides a terminus post quern for this re-location activity. The new ditch (G.10) ran roughly parallel with, but cut away the northern edge of, the former. With a roughly U-shaped base, the sides were gradually sloping and undulating. The ditch had a maximum width of 2.25 m. and a depth of 1.24 m. A single post-hole (C.l 18), 0.70 m. diameter, cut through the southern slope of the ditch. Since no other post-holes were identified cutting the ditch, it can be presumed that it did not represent a fence line. More likely, it was in some way associated with the settlement activity to the south. This secondary ditch was filled entirely with materials formed by silting and erosion. These lay in bands, indicating a number of erosion episodes and suggesting that the infilling occurred gradually. No evidence of re-cutting was found; the cut as excavated perhaps represented the final clearance of the feature. The third, and latest, ditch located south of the road (G.14) lay parallel with the others but was sited further north so that it cut through the edge of the road metalling. Roughly U-shaped in profile, with undulating sides, the maximum width of the cut was 1.85 m. and it was 0.75 m. deep. This ditch was seen to terminate in blunt ends to both the south-west and north-east. That this ditch cut the road metalling may imply decreased use of the road or possibly a need for greater space between the ditch and the settlement to the south, already postulated for the re-location of the primary ditch. That the ditch terminated within the excavation area may suggest that access was required from the road into the settlement, suggesting that occupation was beginning to expand into new areas to which direct access had not previously been necessary. It should be noted, however, that it could not be conclusively determined whether the secondary ditch (G.9) to the south had gone out of use by the time this more northerly ditch was cut, although it cannot be envisaged why a double-ditch system would be required unless the ditches served separate functions (for example, one a drain, the other a boundary); there was no evidence, from the nature of the ditches, for such an assumption. The final ditch cuts to both the north and south of the road contained erosion deposits in their lower levels, suggesting gradual neglect. The upper levels, however, contained quantities of domestic refuse, and pottery of late third- to fourth-century date, indicating that some form of settlement remained within the vicinity even once the ditches, and therefore probably the road, had ceased to be maintained. 100 EXCAVATIONS AT EACH END, ASH, 1992 The road: Discussion The date at which the road and its associated ditches were constructed is far from certain. Few finds were retrieved either from the metalled surface or from the ditches, and those from the latter merely provide a terminus post quern for the infilling of the features whilst giving no indication of their date of cutting. Similarly the finds recovered from the metalled surface, consisting of a copper alloy horse harness toggle and cheek piece of first-century date (SFN2 1262; C.l 15), a single sherd of second- to third-century pottery and a coin of mid to late fourth-century date, cannot assist with the primary dating of the road for material could easily have become impressed into its fabric by trampling upon the surface during its lifetime. It seems feasible to suggest, however, that the road was laid during the second half of the first century A.D., shortly after the Roman conquest. This period saw the expansion of the site at Richborough, which began life as the bridgehead for the Claudian invasion. The ditched enclosure at Richborough was almost immediately replaced by a wooden fort, which remained in use until the late first century A.D., when the military base was superseded by a civilian settlement, enticed by the thriving port (Cunliffe 1968). The growth of the port would have been accompanied by the growth of trading links, along roads and waterways, with surrounding settlements, including the towns of Canterbury and beyond to London. Smaller sites, like that at Each End, may have benefited from this increased trade, initially perhaps by the replacement trackways with more substantial metalled roads. The metalled surface appeared to remain in use until the late third to early fourth century, although to what extent it continued to be maintained during the later part of this period is uncertain. A deposit of dark silt, allowed to form across the line of the metalling and across the infilled road ditches, yielded a large number of coins of mid to late fourth-century date. Thus it appears that by the late Roman period the road and ditches had been neglected yet the line remained in use as a trackway, the coins attesting to continued activity. The road uncovered at Each End is in some ways rather curious since it appears to provide conflicting evidence. The depth of the road metalling, a maximum of 0.14 m., would have been insufficient to support great volumes of traffic, yet the metalling width of c. 15 m. is comparable with, if not considerably wider than, a number of other Roman roads in Britain (cf widths of c. 7 m. at Fleet Farm (Ogilvie 1968); just 2 SFN = Small Find Number. 101 ALISON J. HICKS under 7 m. at the west gate of Richborough (Cunliffe 1968); c. 8.5m. for the widest road excavated at Springhead (Penn 1965); whilst the portion of Watling Street uncovered at Springhead was only c 5.2 m. in width (Penn 1958)). In addition, the cutting and re-cutting of the road ditches to both north and south of the metalling suggests a considerable amount of effort was expended both in the construction of the road and in its maintenance. The width of the metalling has led to suggestions that the feature was not a road, but a courtyard. However, the ditches, flanking the feature, and the presence of the central agger, do not support this theory. Neither do the features uncovered beyond the ditches. If a courtyard existed, it would be expected to be surrounded by buildings of a substantial nature, especially as the 'courtyard' would have been at least 35 m. long from south-west to north-east, and perhaps considerably longer, since it was not bounded at either end of the excavation. The Roman villa at Lullingstone, a substantial Roman site, is only 38 m. in length (Meates 1979). No such buildings were identified at Each End; the only building uncovered was the small timber structure to the south, clearly not grand enough to have warranted such a major courtyard. The cremation burials, bounding to north and south, also suggest that the metalling represented a road, graves often being sited flanking such routes during the Roman period. The road was aligned in a north-east to south-west direction, apparently heading towards the Roman fort at Richborough (Fig. 3). It seems probable that the Wantsum Channel was crossed by ferry; Margary (1948) proposed a crossing site at Little East Street Farm which is on the protected alignment of the Each End road. A ferry terminus may have been a simple affair. Ferries plying waterways in parts of Asia today would leave no trace archaeologically; they simply comprise a man with his boat pulled up onto the muddy foreshore, ready to push off when traffic arrives to be taken across (author's observation). It is quite probable that the Little East Street Farm-Richborough ferry crossing was little more elaborate. However, Margary postulates that the principal road leading to Richborough was located further north-west and there seems no reason to dispute this. The evidence of an earthen agger just north of the village of Ash and traces of flint metalling near Cooper's Street suggest that this principal road was heading towards a site near Fleet Farm, located at the closest point to Richborough from the mainland. This was confirmed by excavation in 1957 (Ogilvie 1968), the road revealed being c. 7 m. wide, up to c. 0.6 m. deep, and comprising of a double thickness of flint pebbles each up to 0.15 m. in diameter. Remnants of the road were also found to the east leading towards the 102 EXCAVATIONS AT EACH END, ASH, 1992 RICHBOROUGH Fleet Fm ^ / ASH / CANTERBURY EACH END^ Wingham Woodnesborough o 1 10km Projected alignment of Roman road uncovered at Each End Roman roads as proposed by I D Margary (1948) Fig. 3. The proposed alignment of the road uncovered at Each End, and its relationship to other principal roads within the area west gate of the Roman fort. Thus it would appear that this was the main road connecting Richborough with the mainland, its method of construction demonstrating that it would have been capable of supporting a large volume of traffic. The road metalling uncovered at Each End was of a far less substantial nature, despite the impressive construction of the road ditches, and could never have supported any quantity of traffic. It was probably a minor roadway serving a local function, perhaps connecting with the junction of the main Dover and Canterbury roads located to the south-west. At its south-west end, the road probably joined the eastward extension of the Canterbury-Ash road, again postulated by Margary (1948). From this Each End junction, a road or trackway may have existed which led directly south of Woodnesborough, and from there straight towards the port of Dover, rather than detouring to the west via Ash before turning to Woodnesborough. Either way, the small Each End road would have provided an additional link between Richborough island and the mainland, and from there on to the towns of Canterbury and Dover. The settlement activity: Features and deposits (Fig. 4) The most convincing evidence of nucleated settlement activity at Each End was the discovery of a sequence of post-holes thought to be the remnants of a small wooden structure (G.15) located south of the road metalling. Seven post-holes (C.1072, C.1073, C.1074, C.1075, 103 o Structure G19A G18D G18G Hearth > X o Fig. 4. The settlement features. EXCAVATIONS AT EACH END, ASH, 1992 C.1076, C.1077, C.1087) were positioned in two parallel rows to form a rectangular structure 4.60 m. long and 2.30 m. wide, its width thus exactly half its length. Substantial deposits of tile recovered from nearby abandonment or demolition material (G.71) suggested it had a tiled roof. Associated with the wooden structure were three patches of compacted cobbling (G.16). Immediately adjacent to the structure, a smooth cobbled threshold was formed by flint pebbles and occasional chalk and tile fragments (C.385). North-east of this lay an area of more uneven cobbling (C.386), comprised of both flint pebbles and angular flints, combined with tile and slag fragments, probably a rough yard surface. A narrow strip of this yard, adjacent to the structure and set at a slightly lower level, may indicate the location of a small trampled pathway running alongside the building. The third patch of cobbling (C.387) was located on the south-west side of the structure. Consisting of medium flint pebbles, large flint nodules and occasional sandstone, tile, pottery and bone fragments, this yard surface was smooth and compact in places but undulatory, irregular and patchy elsewhere. As discussed above, the wooden structure and its associated yards were inserted into an area created by the re-location, slightly northwards, of the southern road ditch. Whilst overlying the line of the earliest road ditch, the structure and yards respected the line of the re-located ditch, implying contemporaneity. Pottery recovered from the fill of the earliest ditch indicates that the structure could date no earlier than the late second century. Three wells were discovered during the course of the excavation (G.9A, G.19A, G.19B). Whilst none were fully excavated, two (G.19A, G.19B) were taken to depths of between 2.20-2.40 m., where the current water table was reached (8.11 m. O.D.). One well (G.19A), perhaps the earliest of the three, was roughly circular in plan, and broad at the surface where the sides had clearly eroded to become shallow and undulatory. At a depth of 1.80 m., however, the shaft, whilst still circular, narrowed and the sides dropped almost vertically. No evidence of an original lining was visible. The well was infilled with lenses of erosion material, interspersed with bands of humic material containing bone, charcoal, daub and pottery dated to the late second to third century. Abandoned, and gradually infilling naturally, the feature would have been an ideal location in which to discard refuse generated by the nearby settlement activity. The well was finally deliberately capped with a 0.90 m. deep deposit of sterile clay, and a hearth (discussed below) sited above. 105 ALISON J. HICKS This well was perhaps replaced by that to the west (G.9A), which was only partially excavated due to time constraints. Apparently circular or ovoid in plan, the sides were steeply sloping and slightly undulating to a depth of 1.40 m., after which they broke almost vertically down to an excavated depth of 1.70 m. (8.74 m. O.D.). The well had become gradually infilled with a succession of erosion deposits, intermingled with humic lenses containing bone, charcoal, tile and pottery of late third- to fourth-century date, deliberately discarded within the abandoned feature. The third well lay to the north-east (G. 19B) and was excavated to the depth of the current water table. Circular in plan at its upper levels, the well had undulating sides probably caused by erosion. However, the shaft became square at a depth of 1.60 m. where the sides were highly compacted and vertical, suggesting an original plank lining. A linear slot (C.1003) running up to the edge of the well may have indicated the location of a wood or stone structure at the well head. Two adjacent shallow, irregular features (C.l001, C.l002) were probably formed by spilled water and trampling feet. Infilling the well was a succession of erosion deposits from which occasional fragments of later third- to fourth-century pottery were recovered. The lack of refuse may suggest that the settlement had largely been abandoned by the time this well became redundant. Overlying the capped well (G.19A) was a hearth (G.17), constructed from tiles set within a shallow depression. The tiles were arranged with their flanges outwards to form an enclosed area. Lying upon the tiles was a deposit of silty loam (C.327) heavily impreg- nated with ash and charcoal, which would have formed during use of the hearth. The surrounding ground surface was scorched to a deep red-brown colour and flecked with concentrations of charcoal: material was no doubt raked out still hot from the hearth, scorching the earth and depositing waste material upon its surface which would subsequently have become trampled into the ground. The hearth may be suggestive of small-scale industrial activity, although analysis of samples taken from the fill and surroundings gave no indication of its use. Pottery of late third- to fourth-century date recovered from this feature indicates that it was in use during the later period of the settlement. A number of pits (G.18) were identified, most backfilled with a mixture of occupation debris, including pottery, bone, charcoal, daub and humic material, and naturally formed silt deposits. Unlikely to have been cut for storage, since the ground would have been too wet, it seems likely that these features were rubbish pits, although why these would have been necessary within a settlement surrounded by 106 EXCAVATIONS AT EACH END, ASH, 1992 fields is uncertain. An exception may have been the disposal of industrial waste, possibly considered unsuitable to spread upon agricultural land. However, soil samples analysed from the pits provided no evidence of industrial activity. The bulk of the ceramic material retrieved from the pits was of mid to late second- to third-century date. Only one of the rubbish pits (C.l 030) included pottery of a later date, its range extending up to the fourth century. Interwoven amongst the settlement features described above was a complex of ditches and gullies, all concentrated to the south of the road metalling (G.26, G.27, G.28, G.29, G.30, G.31, G.32, G.33, G.35, G.36, G.37, G.38). Varying greatly in both width (0.18-3.90 m.) and depth (0.10-1.06 m.), some took the form of a single ditch or gully cut which demonstrated no evidence of cleaning or re-cutting, whilst others had clearly been well maintained, having undergone numerous phases of re-cutting, some even to the extent of changing the direction of water flow. In places, channels had also been deliberately blocked by the insertion of tightly packed flints. Complex formations of parallel and intersecting gullies were uncovered, in places up to six gullies abreast and closely set, although to what extent they were in use contemporaneously was not always easy to determine. The gullies varied greatly in profile, from shallow, U-shaped cuts with undulating sides to steep-sided, V-shaped cuts, no doubt depending upon the degree of weathering and re-cutting to which each had been subjected. There was no suggestion that any of the gullies were originally covered. Instead, the evidence indicates that they were open drainage channels, cut with sloping bases along which water would have flowed, rather than sumps or soakaways into which water would have drained and remained static. The principal aim of the channels appears to have been to drain water from settlement areas and direct it away, either into the road ditch to the north or along a series of ditches and gullies to the south. An alternative suggestion, that they were land boundaries which also serviced to drain areas of cultivation, seems implausible. There was no evidence for arable activity in the ground between the drainage features. A suggestion that the features were beam or timber slots can be discounted due to the nature of the cuts. That a system of water drainage was required suggests severe problems with ground water. Why such a complex system came into being over a period of time is open to debate. It is possible that the use of areas between the gullies was changing, and their lines were therefore moved to create space required for other activities or, in the 107 ALISON J. HICKS case of the first southern road ditch, because the channel into which they drained had been re-located. Alternatively, and perhaps a more likely explanation for the series of closely-spaced central gullies, the nature of the subsoil may have caused the gullies to rapidly fill with erosion deposits during periods of bad weather. Anecdotal evidence from local farmers indicates that, today, drainage ditches need to be cleaned out every year to remain effective as water channels. Thus it may be suggested that, in Roman times, if a gully was neglected for a few winters in succession, its original line may have largely disappeared. Under such circumstances, it may have been easier to re-cut a gully line rather than locate and clean the original channel. A combination of both factors is perhaps the most likely explanation for the series of gullies uncovered at Each End. Most of the drainage channels were filled with erosion material; some contained sequences of material indicating a succession of erosion episodes. A large proportion of the pottery recovered was of later second- to third-century date, although fourth-century pottery was also retrieved, particularly from the uppermost deposits. One of the gullies (C.54) contained a complete, articulated dog skeleton, which lay above the primary silty clay deposit covering the gully base. The dog lay upon its side with head, tail and legs fully extended. Above it, deposits containing quantities of charcoal, pot and bone, probably conveniently dumped within a redundant feature, capped the channel. Another of the gullies (C.57) contained an almost complete pottery vessel (C.l55), dated to the mid second to third century, lying above the primary deposit. The vessel fill was intact (C.l56), but analysis provided no trace of any contents suggesting that the pot was simply discarded into the gully because it was broken. The settlement activity: Discussion A sufficient quantity and variety of features was uncovered at Each End to determine that settlement activity was present upon the site. The concentration of many of these, including the structure and yards, wells, hearth and pits, towards the south of the excavation area may indicate that the settlement was concentrated here. However, the location of burials within this part of the site may indicate that this was an area of peripheral settlement. In addition, the complicated sequences of gullies and ditches interspersed with the settlement features indicate that extensive drainage was required for the area to remain suitable for habitation. It therefore seems plausible to suppose that the site excavated at Each End would not have been the primary area chosen for habitation. 108 EXCAVATIONS AT EACH END, ASH, 1992 Instead it is probable that the settlement nucleus was located upon the higher, drier ground to the south-west, west or north-west, beyond the bounds of the excavation area, whilst the occupation activity discovered at Each End may have represented later expansion of the settlement into peripheral and less suitable ground. Anecdotal evidence from members of the Mid Kent Search and Recovery Club supports this. They report that a much larger concentration of metal objects of Roman date has been recovered from the fields on higher ground on the opposite side of the A257 than from the field containing the excavated area. What form this nucleated settlement took, its full size, extent and nature cannot be determined until a much larger proportion of the settlement has been examined. Events at Richborough would undoubtedly have affected the settlement at Each End. Since the sites were connected by road and ferry, trade would clearly have been conducted between the two both in goods and services. The location of the Each End site, on the fringes of the Wantsum Channel and maybe on the closest raised ground feasible for habitation, ensured that its inhabitants reaped the benefits of the nearby coastal zone. The benefits would have included plentiful supplies offish, water fowl and reeds, and trade could easily have taken place not just with Richborough but also with other ports along the coast. No doubt small boats would have been pulled up against the shoreline near Each End, ready to sail long the coastline collecting and transporting cargoes as they went. Settlement at Each End may have begun as early as the late first century B.C. to the mid first century A.D. Two contexts were identified which contained pottery of this date, although this may suggest occupation within the locality rather than at Each End. Activity was certainly present by the late first to early second centuries A.D.; a number of pottery sherds of this date were recovered. It is possible that the construction of the metalled road, thought to have occurred in the second half of the first century A.D., may have provided the major impetus for occupation, whilst trade with the growing port at Richborough would have been an additional spur. The principal date of the Each End settlement, however, was during the late second to third century, the site perhaps being allowed to develop in its own right due to decreased activity at Richborough. Pottery evidence indicates that settlement expansion, or at least development, occurred in the late second century, when activity moved into areas previously used for burial (the burials were broadly mid to late second century in date). The settlement may have seen its major period of expansion during this period. Coin evidence suggests that the later part of the third century saw a period of declining activity at the site. It has been suggested that this 109 ALISON J. HICKS Group 21 .-. G21 Site plan ® a Group 22 Fig. 5. Cremation burial groups G.21 and G.22. 110 EXCAVATIONS AT EACH END, ASH, 1992 may have been due to the threat of invasion; the occupants may have responded by abandoning their settlement, whereas at Richborough a stone fort was constructed. Pottery dated to the late third to fourth century recovered from a small number of the features suggests that some form of settlement was present during the later Roman period. However, it may have largely represented a time of gradual neglect and abandonment, or perhaps continued contraction of the principal site of nucleated settlement. The burials: The cremations (Figs. 5 and 6) Three groups of cremation burials were uncovered during the course of the excavation. A total of 15 burials contained 49 pottery vessels, many of which were complete (see Pl.l). Appendix A provides a summary of the contents of each grave. Two groups lay to the south of the road metalling and one to the north, the latter separated by the road line from the settlement area. The cremation site to the south-east was the most extensive, consisting of nine burials, that to the south-west the smallest, consisting only of two, whilst that to the PLATE I !;• A selection of the intact pottery vessels recovered from the cremation burials. Ill ALISON J. HICKS north contained four burials. All were truncated to varying degrees, in some cases so severely that only the lower portions of the ceramic grave vessels survived. North of the road metalling, four cremation burials (G.21: S.21, S.22, S.23, S.24; Fig. 5) were discovered within an area of approximately 10 square metres, although the cemetery probably extended further north and west beyond the bounds of the excavation area. One burial (S.21) was unfortunately destroyed by unauthorised metal detector users prior to the start of the excavation so that only a single, fragmented pottery vessel (C.147) remained. No cremated remains or associated grave goods survived. One of the other burials (S.22) also, curiously, contained no cremated human remains, although the cinerary urn and its fill were retrieved. Each of the three un-robbed burials contained, in addition to the cinerary urn, from two to four ceramic grave goods. Other grave goods of note included a copper alloy tumbler lock bolt (SFN.601) and two hobnails corroded to a fragment of mineralised leather (SFN. 1679), part of the sole from a hobnailed shoe or boot. A total of nine cremation burials (G.22: S.4, S.13, S.14, S.15, S.16, S.17, S.18, S.19, S.26; Fig. 5) was recovered from the area lying to the east of the site, south of the road metalling. The area was bounded upon all four sides by ditches and gullies, leaving a rectangular area of c. 11 m. by 13 m. in which this burial group was located. As recov ered, it appears that the ditches and gullies, yielding late second- to thirdcentury pottery, respected the edges of the cremations, which were dated to the mid to late second century. This may imply that the cremations were marked above ground in some way. Alternatively, it is possible that the ditches and gullies had cut away further cremation burials, the remains of which, therefore, no longer existed. However, since no burials were identified on the opposite sides of these bounding ditches, this is thought to be unlikely. Interestingly, each of the burial cuts within this group had a deposit of earth lying upon the base, evidently deliberately laid. It was invariably flecked with tiny fragments of charcoal, daub, chalk and tiny flint inclusions, and thus could not have formed by erosion from the natural subsoil into which the burials cut. Even the amphora burial (described below) had earth lying within the ceramic base which could not possibly have formed by anything other than deliberate deposition. Within the burials, the grave goods sat directly above this earth deposit, often pushed a short way into it. It therefore appears to have been laid as bedding, to ensure that the ceramic vessels stood upright, and remained so whilst the burials were backfilled. Neither of the other cremation burial groups had similar deposits. 112 EXCAVATIONS AT EACH END, ASH, 1992 The apparent wealth of the burials within this group varied enormously. One (S.13) comprised solely of the ceramic cinerary urn; at the opposite extreme, a burial (S.4) was excavated which contained the cinerary urn plus nine ceramic grave goods. The remains of a pair of hobnailed boots (SFN.634) were also recovered from this grave. Two other burials (S.18, S.19) were found to contain remnants of shoes or boots; one (S.18) contained three burnt hobnails within the cinerary urn, suggesting that the corpse was shod when placed on the funeral pyre. The other burials each contained between one and three ceramic vessels, one of which was an oil lamp (from S.19). Of particular interest was the amphora burial (S.18). The cut had been lined using the lower half of a large amphora (South Spanish Dressel 20), into which an earth bedding (discussed above) had been laid. The grave goods rested upon this deposit. They comprised the cinerary urn, two ceramic vessels, a glass goblet containing fragments of iron sheeting, and a number of iron hobnails from shoes or boots. A second deposit of earth (C.24) had then been laid, sealing the goods, and the upper portion of the amphora (C.67) positioned above. Complete with two handles and the rim, this portion had fortunately slumped down into the body of the amphora, protecting the grave and its contents from the effects of later truncation. One burial (S.26) (not illustrated), containing four vessels, was uncovered by mechanical digger during levelling work prior to road construction. A bucket scoop lifted a complete cremation burial; ceramic vessels were seen by the archaeologist conducting the watching brief literally hanging out of the side of the bucket. Thanks to the immediate co-operation of the machine driver, the bucket was carefully lowered to the ground and the excavation of its contents undertaken. This ensured that the complete burial was recovered, although it could not be illustrated due to its understandably disturbed nature. The third cremation group (G.23; Fig.6), containing two burials (S.5, S.10), lay to the south. Further burials may, however, have been located beyond the bounds of the excavation area. The earliest grave (S.10) contained no ceramic vessels. Instead the cremated remains had been enclosed within a wooden box. The base of the cut was lined with iron nails, indicating the original outline of the box, now decayed. The nails surrounded a patch of light grey-brown, moderately compacted silt (C.231) containing frequent human bone fragments. The cremated remains may originally have been placed within a leather or cloth bag, of which no evidence survived. No grave goods were included within the cut. However, a number of pig and bird bone fragments, interspersed with the human bone, suggests 113 ALISON J. HICKS that offerings were made upon the funerary pyre and that the soul of the deceased was not neglected. Capping the burial was an earth deposit (C.232); occasional iron nails were recovered from this deposit, probably originating from the wooden box and indicating that post-depositional mixing had occurred. It has been suggested that the presence of a wooden box, in preference to a ceramic cinerary urn, may indicate the burial of a person from a low status or poor family. However, there is no reason to suppose that a wooden box would be cheaper or easier to obtain than a pottery vessel. With one exception, all the cinerary urns exhibited characteristics of a type known as Canterbury or 'transitional Native Coarse Ware', common throughout East Kent in the mid second to third centuries. Thus the box burial may, in fact, have reflected a burial of higher status. The other burial within this area (S.5) cut through the edge of the box burial. It contained only the ceramic cinerary urn. The fill of the vessel (C. 130) was intact but curiously contained no cremated human bone fragments although unburnt shrew and mole bones were recovered. No other grave goods were recovered from the burial. The cremations: Discussion The cremation burials date between the mid to late second century. Thus they were contemporary with similar burial rites fashionable in Roman Britain at this time. All of the burials contained a cinerary urn except one which contained a wooden box. In four of the burial urns, fragments of cremated mammal bone were recovered (see Appendix A). Pig was the only identifiable species, whilst bird bones were recovered from the box burial. Pieces of meat are likely to have been burnt with the body on the funerary pyre as offerings to feed the soul on its journey to the afterlife (Alcock 1980). It is probable that in this way some would have become intermingled with the human remains. In addition, unburnt bones, of mole and shrew, were recovered from two of the cremation vessels. It would appear that the rodents either fell into the vessels before soil had slumped down into them and so were unable to escape, or they used the vessels as nesting sites. Curiously, two of the cinerary urns were found to contain no cremated human remains, although one of these contained unburnt rodent bones. If nesting within the vessel, rodents may well have disturbed the deposits, yet it seems likely that post-depositional mixing, by rodents and also earthworms, would cause all the bone to disappear. Is it perhaps more likely that, for some reason, the cinerary 114 EXCAVATIONS AT EACH END, ASH, 1992 urns were buried with no bone inside them? Other examples of cremation burials containing no bone are known in Britain. At the Anglo-Saxon cemetery at Spong Hill, at least one complete urn (no. 1881) was found to be devoid of bone (McKinley 1994).3 It is perhaps possible that the cremations represented symbolic burials for bodies which were missing, for example having been lost at sea. Most of the burials contained grave goods in addition to the cinerary urn. Many would have contained food and drink to provide nourishment for the soul's journey to the underworld (Alcock 1980). Examination of the earth fills of the grave goods provided no evidence of their contents, but this was only to be expected, as they were probably perishable. One burial also contained an oil lamp. Interestingly, a cremation recovered from Warwick Square, London, containing the remains of an adult and a child, had been buried with two oil lamps; clearly it was thought that each individual would require a lamp to light their way into the next world (Merrifield 1987). Hobnails were also recovered from four of the burials, indicating that shoes or boots were buried with the grave. The practice derives from the Celtic custom of providing the deceased with footwear to ensure safe passage into the next world (Van Gennep 1977). One cremation vessel contained burnt hobnails suggesting the deceased may have been burnt wearing boots or shoes. Evidence that any of the burials once had markers was tenuous, being provided solely by the location of the most easterly burial group within an area surrounded by gullies and ditches which appear to respect their positions. Widespread truncation which had occurred on site could have removed all traces of markers. It is known that at least some cremation burials were marked above ground during the Roman period. At Caerleon a cremation dating to the first century had a lead pipe leading to the surface so that libations could be poured into the grave (Merrifield 1987,197, PI. 22), evidence that individual graves were visited and as such must have been visible. The siting of cremation burials surrounded by settlement features is somewhat unusual, since it was more common during the Roman period to position burial grounds beyond the occupation zones. It is possible, however, that gradual settlement expansion down onto the lower ground forced the abandonment of the cemetery areas. The cremations date between the mid to late second century, whilst is was the late second to third century which saw the major period of settlement, which perhaps, moved at that time into areas formerly used for burial. Of the 2,384 cremations recovered at Spong Hill, 29 contained no bone. However, most of these were not complete vessels. 115 ALISON J. HICKS A number of theories can be proposed for the three separate burial groups, though without wider knowledge of the surrounding unexcavated areas, these must remain speculative. The groups may have represented different periods of time, although the pottery dating is, unfortunately, insufficiently precise to determine whether any group largely pre-dates another. If this suggestion is correct the north-western group (G.21) would appear to be the latest, the other groups having been overtaken by occupation activity and therefore been abandoned at this time. Another suggestion is that the burial areas represent different spatial, social or family groups, although each group contained a diversity of apparent wealth, and the burial group to the south-west (containing an inhumation (discussed below), a box burial and a cremation containing a ceramic urn) clearly showed significant variations in burial practice. The burials: The inhumation (Fig. 6) A single inhumation burial (G.20A) was uncovered within the excavation area, adjacent to a number of rubbish pits and post-holes. The grave was of a particularly curious construction. The outer burial cut (C.l032) was rectangular in plan, aligned north-east to south- west and 3 m. long, 1.5 m. broad and 2 m. deep. The sides were largely vertical but stepped inwards towards the base. At the base, an erosion deposit (C.422) of silty clay almost identical to the surrounding subsoil, suggested that a period of time had elapsed between the cutting of the feature and its use as a grave. However, this did not appear to be slumping from the existing cut sides which, over much of their height, were absolutely vertical. It is possible that the material was banked up around the sides of the feature after its cutting and it could perhaps be envisaged that heavy storms may have washed the material into the cut. Alternatively, the upper portions of the cut, once demonstrating the effects of slumping and erosion may have been removed by truncation. After the deposition of this erosion material, no attempt appears to have been made to clean out the original cut. Instead, a narrow shaft (C.l 128) was cut into its centre, 2.05 m. long but only 0.50 m. wide, into which was inserted a wooden coffin; dark brown lines of staining (C.406), indicating the original coffin position, were visible within the narrow shaft. Lying at the base of the coffin were the human remains (C.408), in a poor state of preservation but identifiable as those of a woman, estimated to be of 35 to 45 years of age. The body lay supine, with the head aligned towards the south-west. The grave shaft was backfilled with a deposit of green-grey silty 116 EXCAVATIONS AT EACH END, ASH, 1992 C405 Site plan G23 & G20A Section +10.00m.OD Group 20A cut C1032 C420/ 423 fSp* cut C1128 Coffin C406 Group 23 Fig. 6. Inhumation burial G.20A and cremation burial group G.23 117 ALISON J. HICKS sand (C.420/423) and the entire outer cut infilled with a sequence of silty clay deposits (C.214, C.419, C.421), probably representing material upcast from the digging of the original hole. As the coffin began to decay and collapse under the weight of material above, the upper infilling material slumped down to partially fill the void, so becoming less compacted within a shaft-like area directly above the coffin (C.405). This burial was much deeper and broader than was common for Roman inhumations, but it is not known why such effort was expended upon this single burial. What seems probable, however, is that it was the grave of a person of high status. However, it is curious that the inhumation was located within an area of peripheral settlement activity, containing rubbish pits, post-holes and, to the northwest, a hearth, when Roman burials were usually sited outside the area of the settlement. It was originally thought that the burial may have pre-dated this occupation, the settlement having spread to incorporate an area used for burials. However, pottery evidence suggests that the occupation activity and the inhumation burial were roughly contemporary, being of late second- to third-century date, and its location therefore remains as enigmatic as its construction. Stratigraphically, the inhumation burial can be seen to post-date one of the cremations, whilst the pottery recovered from the inhumation (late second to third century) is broadly later than that from the cremations (mid to late second century). It can perhaps, therefore, be tentatively suggested that this burial represents the beginning of the change in Roman burial practice from cremation to inhumation. Late/post-Roman activity (not illustrated) Only sparse evidence of late and post-Roman occupation was uncovered upon the site. A deposit identified as demolition debris (G.71) was recovered from the area adjacent to the wooden structure. This consisted of an irregular spread of flint, chalk, tile and sandstone fragments which lay above the western yard surface. The deposit may have formed when the structure was demolished or fell into decay. Unfortunately, no dating material was recovered from the layer. Above the road metalling, deposits of silty clay (G.70) formed, probably due to a combination of natural silting and erosion processes, and continuous trampling across the road surface. This build-up may have occurred both during the lifetime of the surface and in its primary stages of abandonment. Subsequently, a dark soil (S.3), up to 0.30 m. deep in places, formed above the line of the road 118 EXCAVATIONS AT EACH END, ASH, 1992 and ditches, indicating that the road had ceased to be maintained by this time. However, although the metalling was now covered and the road ditches abandoned, it should not be supposed that the route was no longer used. A large quantity of fourth-century coins recovered from the dark soil horizon suggest activity in the later Roman period. It seems probable, therefore, that, having become well established as a roadway during the previous centuries, it would continue to be used as a trackway when the administration involved in its maintenance was no longer present. No features specifically indicative of late Roman occupation were identified at Each End, although fourth-century pottery recovered from a number of features suggests some activity was present. However, extensive truncation had taken place across the site, no doubt due to later agricultural practices which may have removed evidence of late Roman activity. Although the road and ditches were neglected, local settlement may have continued. Alternatively, the late Roman period may have seen continued contraction of the nucleated settlement site, which perhaps never recovered from its decline in the late third century. Materials of late-Roman date retrieved during the excavation may have derived solely from use of the surviving trackway. Fragments of Anglo-Saxon pottery recovered from the layers of colluvium (G.72) and topsoil (G.73) are probably stray finds, possibly from field-manuring, and are not necessarily indicative of post-Roman occupation. THE HUMAN REMAINS Trevor Anderson THE CREMATED MATERIAL Introduction Seventeen pottery vessels, with their contents intact, were forwarded to the osteologist for excavation. In three cases (vessels C.37 (S.15), C.260 (S.4), C.471 (S.26)), the urns were sufficiently complete to permit examination by computerised tomography (Anderson and Fell 1994)4 before excavation. An advanced form of radiography where sections or slices can be obtained without damaging the object. The tomographs were kindly provided by Jon Billings and Colin Fell of Kent and Canterbury Hospital. 119 ALISON J. HICKS Eleven vessels were found to contain cremated remains (Appendix A). An additional cremation was discovered in a poorly preserved wooden container. This is discussed separately below. The condition and preservation of the material Bone condition The fact that all the bones were hard and solid, did not leave a mark when scratched and were not calcined, suggests that the firing temperature was between 400-750° (Hoick 1987, 131-2). Experimental work, however, suggests that calcination may occur at a lower temperature in an oxygen-rich environment (Parker 1985). In vessel C.43 (S.13, 11.6 per cent of the bones were only lightly burnt which argues for either a low firing temperature or a very short period of cremation. In two vessels (C.78 (S.14) and C.471 (S.26)) the recovery of intact teeth suggests that the maximum temperature was probably less than 500°C.5 Colour The majority of the fragments (89.7 per cent, with a range from 75.0 - 97.8 per cent) were pale cream or off-white in colour, which suggests that most of the bones were burnt in a well-oxygenated environment. All the vessels, however, contained sooty, dull-grey blue/black bone fragments. This would argue that approximately 10 per cent of the material was burnt in a reducing environment. In all vessels, it was noticed that the trabeculae were frequently blue whilst the external surface was white. This supports the view that a lack of oxygen will result in a blue coloration. Other factors, however, such as liquefaction of the marrow, will cause localised temperature reduction, which will influence bone colour (Reverte 1987). In vessel C.78 (S.14) the majority of the fragments were off-white but a few ribs and various vertebrae were blue externally. These bones are well-protected by soft tissue and would tend to be the last skeletal elements directly exposed to the flames of the pyre. This suggests that the cremation was probably of rather short duration, with the centre of the body not fully burnt. Also it is a possibility that the body in question was rather thin, since a lack of fatty tissue would tend to slow down the cremation process. In vessel C.70 (S.17) the majority of the fragments were off-white but larger portions of the lower leg and ankle were blue. There is very little soft tissue over the lower leg, which means that normally these bones are quickly burnt. The fact that the rest of the body has been consumed points to the interpretation that the heat of the pyre was confined to the upper body, with the lower legs being burnt in a less intense heat, possibly a smouldering fire. It is possible that a sudden heavy downpour reduced the efficiency of the 5 Work by the Japanese forensic odontologists, Furhata and Yamamoto, suggest that tooth eGoldstein 1973,23).namel will be completely exfoliated at c. 500°C (quoted in Buikstra and Goldstein 1973, 23). 120 EXCAVATIONS AT EACH END, ASH, 1992 cremation process.6 The upper and central region of the body, fuelled by the ignited body fats, would continue to be consumed, quite probably being stoked vigorously, and thus fed with oxygen, by the attendants. The lower legs, however, peripheral to the centre of the cremation, would tend to burn fitfully in such circumstances. Size In the sample as a whole, almost half the bones (by weight) were between 20-50 mm.; just over a quarter between 10-20 mm.; 15.8 per cent less than 10 mm. and bones over 50 mm. accounted for only 11.9 per cent of the total. In two vessels (C.142 (S.24) and C.148 (S.23)) no bones over 50 mm. were found: these two urns contained a higher than average frequency of fragments under 20 mm. The fact that these two vessels were incomplete when excavated, with only the base of C.142 recovered, appears to have artificially biased the results for these vessels. In the more complete urns, there was little variation in mean bone size. This suggests that a similar method of cremation and collection was practised throughout the sample. Bones that fractured during the cremation would have their broken ends exposed to the heat of the pyre and they would display a uniform white colour. The fact that many of the fragments are blue internally and white externally argues that the fractures occurred towards the end of the cremation process, perhaps as a consequence of the pyre collapsing, or else they were broken during collection. The latter view is supported by the fact that even in present day cremations, bone shafts up to 25 cm. in length are recovered, prior to crushing in the cremulator (McKinley 1989). Also archaeological evidence has demonstrated that the bones of cremated bodies are relatively intact as long as the burnt remains are left in situ, and are not collected (Buikstra and Goldstein 1973, Fig. 16). Weight The weight of human bone recovered from the incomplete cremation vessels is of limited value for assessing the percentage of the skeleton which was deposited in the vessel. In the eleven cremation vessels a total of 10,417 gms. of human bone was recovered. From the four complete urns, however, the mean bone weight is 1,317 gms., which range from 859 to 1,959 gms.. In modern crematoria somewhere between 1,600-3,600 gms. of bone is recovered from an adult body (McKinley 1989). This suggests that, in our sample, approximately half the bones were recovered and deposited in the cremation vessel.7 In larger towns, most cremations probably took place in permanent structures such as the one found at Carnuntum, Germany, as illustrated by Wenham (1968, Fig. 8). In rural areas, an unenclosed pyre, open to the elements would probably be the normal method of cremation. 7In three complete vessels, C.30(S.18),C.37(S. 15) andC.471 (S.26) (mean height 19 cm.), it was possible to establish a soil:bone ratio; the mean was 1:19.8. Obviously, the weight of the soil will vary due to both its composition and the shape of the urn. As a rough guide, however, this suggests that the ratio should be in the region of 1:40, if all the bones from a single individual had been deposited in the vessel. 121 ALISON J. HICKS This raises the question of what happened to the rest of the cremated bones. There is a sufficient number of bones to rule out the possibility of a token deposit. Clearly, the smaller fragments may have been lost or overlooked during collection, but these would not account for half the body weight. Even more enigmatic, is the fact that two apparent cremation vessels (C. 131 (S.5), C.l65 (S.22)) did not contain any cremated bone.8 Location In three complete vessels (C.30 (S.18), C.37 (S.15), C.471 (S.26)) it was possible to sub-divide the soil matrix into five layers of equal depth. Examination of the bone weight from each layer showed that the majority of the bone fragments were located in the lower portion of the vessel, with the upper layers being largely devoid of bone fragments. The mean bone:soil ratio for the lower two layers was 1:28.8 and 1:30.3, respectively; the same ratio for the upper two layers was 1:4.5 and l:12.7.Tomographicexamination of two of these vessels demonstrated the exact 3-D location of individual cremated fragments (Pis. II and III). The evidence clearly indicates that the vessels, when deposited, were less than half full of cremated bone. The vessels presumably filled with soil during the years after deposition; there is no evidence to suggest that cremation vessels were deliberately filled with soil prior to burial. No distinct pattern of bone deposition was recognised in the three complete vessels which could be divided into layers. The various skeletal elements, such as skull, vertebrae and upper or lower limbs, were not confined to individual layers. In vessel C.87 (S.19) evidence of earthworm activity was noted. In this and in vessel C.1319 (S.5) unburnt shrew and mole bones were recovered10 (PL IV). Clearly their activity disturbed the cremated fragments. In three vessels, cross-layer joins of the bone fragments was possible. In vessel C.30 (S.18) distorted skull fragments, which must have shattered during the cremation, recovered from the three lower layers of soil, were re-united. In vessel C.37 (S.15) two fragments of left fifth metatarsal could be joined. In vessel C.43 (S.13) two fragments of patella from different layers were found to join. Our findings suggest that post-depositional infill of soil, coupled with earthworm activity and the presence of burrowing animals, will have destroyed any information regarding the original location of the cremated human bone fragments. 8 At Spong Hill the smallest deposit of cremated bone was 117.2 gms.; at St. Stephen's, St. Albans, the figure was 71.8 gms. (McKinley 1993). In both cases the cremations were undisturbed and were identified as the remains of a single adult. 9 This vessel did not contain human bone. 10 It is possible that these animals had fallen into the cremation vessels prior to their infill with soil: if the containers had been full of soil the animals would have been able to burrow their way out. Alternatively, the animals may have been using the vessels for nesting. 122 EXCAVATIONS AT EACH END, ASH, 1992 PLATE II Vessel C.37: CT scan, demonstrating the exact location of cremated bone fragments. Skeletal elements present In all the cremations many of the fragments could not be firmly identified. On average just over half (50.7 per cent) of the bones were classed as unidentified fragments. At best almost three-quarters of the bones were identified (vessel C.260 (S.4)). In incomplete cremations, containing minute fragments, the percentage of firmly identified skeletal elements fell as low as 9.1 per cent (vessel C.148 (S.23)). Examination of the identified fragments showed that all skeletal elements, skull; axial; upper; and lower limbs, were represented. It has been established that, in a complete body, these four elements account for 18.2 per cent; 23.1 per cent; 20.6 per cent; 38.1 per cent of the skeleton, respectively (McKinley 1989). In the Each End cremations the mean of the four elements, being 22.0 per cent; 21.6 per cent; 21.8 per cent; 34.6 per cent, equate closely with these figures. In the sample as a whole, the skull and the arm bones are slightly 123 ALISON J. HICKS PLATE III A \ *\W Vessel C.471: CT scan, demonstrating the exact location of cremated bone fragments. over-represented, whilst the lower limbs and the remaining axial elements are slightly under-represented. Bearing in mind that the majority of vessels are incomplete and that over half the fragments were not firmly identified means that no definite conclusion can be drawn from this finding. There is no evidence, however, that certain elements were noticeably over- or underrepresented. There is evidence for an inverse relationship between skull and axial elements and also between upper and lower limb bones. Cremations with many skull fragments have very few other axial elements and vessels with arm bones over-represented have relatively few leg bone fragments. Since many diaphyseal fragments can not be readily identified, an apparent discrepancy between upper and lower limbs may be illusory. 124 EXCAVATIONS AT EACH END, ASH, 1992 PLATE IV Unburnt shrew (mandible) and mole (humerus) bones from vessel C.l31; evidence of post-depositional contamination. Scale: c. 3:1. The inverse relationship, noted in four vessels, between the weight of cranial and axial fragments is very interesting. There is some evidence that cranial fragments are over represented in vessels with a high proportion of small fragments or an absence of larger bones (C.37 (S.15), C.142 (S.24)). In vessels with larger and lightly-burnt bones, vertebrae are over-represented and cranial fragments are less frequent (C.30 (S.18), C.43 (S.13), C.70 (S.17)). The reason for this is that in badly fragmented remains, small pieces of cranium will still be recognised due to their distinctive sutural lines and vascular impressions; in lightly burnt remains, even a few large portions of vertebrae will outweigh minute cranial fragments. The occasional wrist or ankle bone or phalanx was noticed but, in general, the smaller bones of the hands and feet were rarely recovered. This may be the result of inefficient collection. It should be borne in mind, however, that hands crossed over the body are normally protected by 'the heat shadow' and are well-represented (Hoick 1987, 160). Thus, the absence of hand bones supports the view that the bodies were cremated with arms by their side." Ageing and sexing Ageing and sexing of cremated remains is notoriously difficult. The standard 11 In modern forensic work, hand bones are rarely recovered, if the arms are outstretched, since the adequate supply of oxygen will quickly consume them. 125 ALISON J. HICKS ageing and sexing criteria are not normally available, being either absent, fragmented or distorted by the cremation processes. However, juvenile individuals can be separated from full-grown adults when it can be demonstrated that epiphyseal fusion is incomplete. From the minimum number of twelve individuals (including the casket/box burial) all were full-grown adults. The presence of an unfused clavicle and a fused femoral head suggests that the female in vessel C.30 (S.18) was a young adult aged c. 18-23 years. The degree of suture closure in the unsexed remains from vessel C.260 (S.4) suggest an age of over 30 years. The other remains could not be accurately aged, except to say they were full-grown. In six cases (vessel C.30 (S.18), C.37 (S.15), C.43 (S.13), C.78 (S.14), C.471 (S.26) and in the wooden box S.10) the remains were assessed as female. These diagnoses were based on the width of the greater sciatic notch; the pre-auricular sulcus; the frontal bone; the size of the glenoid cavity, estimated femoral head diameter and mid shaft femoral circumference, respectively. The other six adults could not be sexed. Given the small sample, there may be no significance in the fact that only females were identified. Non-metric variants The individual in vessel C.70 (S.17) exhibited a minor variation in the path of the occipital artery (exsutural mastoid foramen). The sample, however, was too badly fragmented to permit meaningful study of non-metric variants. Pathology There was very little evidence of pathological lesions on the cremated fragments. The fragmentary nature of cremated remains, as well as the degree of distortion, means that pathological lesions, except vertebral joint degeneration, are rarely encountered. As such it is almost certain that the true prevalence of disease will be under-represented. The adult female in vessel C.43 (S.13) presented with well-defined osteophytes on her lower spine. The individual in vessel C.471 (S.26), an adult female, displayed osteophytic outgrowth on her odontoid peg (second cervical vertebra). The only other possible lesion was noted on an adult female (?) in vessel C.37 (S.15). The medial facet of her right calcaneus (heel bone) exhibited a smooth-edged shallow cavitation (PL V). The oval defect measured 4 x 3 mm. with the greater dimension being mediolateral. The left calcaneus was normal. The defect probably represents an example of osteochondritis dissecans, a term coined in 1887 by Konig (Barrie 1987). The cause was uncertain, although minor trauma was postulated; more recently it has been suggested that localised delay in ossification is the underlying cause (ibid.). One clear finding is that the site of predilection is the medial femoral condyle (knee joint) (c. 90 per cent of cases) and that adolescent males are most frequently affected (Ortner and Putschar 1981, 242). Other sites, including the elbow (Pick 1955), patella (Desai et al. 1987) and the ankle (Bauer et al. 1987; Dashner and Golder 1990; Gerard et al. 1989; Madhok 1987; Nafei et al. 1990), have occasionally been reported in the modern 126 EXCAVATIONS AT EACH END, ASH, 1992 PLATE V Right calcaneus from (individual in) vessel C.37, displaying osteochondritis lesion (arrowed) of the medial facet. Scale: 2:1. medical literature. There is some suggestion that multiple lesions, associated with short stature, are familial (Mubarak and Carroll 1981; Philips and Grubb 1985). In living patients, there is evidence that osteochondritis of the ankle may result in pain and swelling (Madhok 1987); however, it often heals naturally, without further symptoms (Bauer et al. 1987). This finding, coupled with the location and size of the lesion, suggests that osteochondritis of the ankle may be under-represented in clinical practice. In dry bone material, the true incidence of the defect will be recognisable (Wells 1974). Dental pathology Fragmentary teeth were recovered from all the vessels except C.142 (S.24). which was only represented by a fragmentary base. A total of 20 teeth was available for examination.12 The majority were loose root fragments. In two vessels, C.78 (S.14) and C.471 (S.26), a complete tooth was preserved. The 12 The dental remains were examined by Jon Andrews, B.D.S., B.Sc, F.R.S.M. 127 ALISON J. HICKS only evidence of dental pathology was found in an adult female (vessel C.471 (S.26)). Her lower left first premolar displayed evidence of infection and abscess or cyst formation. The small sample means that it was not possible to assess the oral health status of the sample under study. Other finds: Animal bones Small samples of burnt animal bones were intermingled with cremated human remains in four vessels (C.30 (S.18), C.47 (S.16), C.78 (S.14), C.87 (S.19)). The only identified animal was pig ruling out the possibility that domestic pets were being burnt with their dead master or mistress. The mean weight of animal bone was 33 gm.; the maximum weight, 84 gm., was recovered from vessel C.47 (S.16). In the latter, the presence of an unfused femur suggests that the pig was under VA years; unfused metatarsals in vessel C.30 (S.18) argue for slaughter prior to 2 years (Cornwall 1974, 229). Pig teeth and bird bones were recovered from the wooden box S.10. It appears that in Celtic tradition the deceased, for a limited period at least, still required bodily nourishment. This equates with Van Gennep's period of transition, in which the corpse is regarded as an object of fear since it is not yet fully incorporated into the abode of the dead (Van Gennep 1977). Items of food and footwear would be placed with the body so that it would safely reach the next world and would not return to haunt the living. The burning of animal bones on the pyre with the corpse suggests that the Romans were feeding the spirit or soul of the deceased, the ascending smoke symbolising the release of the bodies from their mortal existence and completion of their safe passage into the next world. When present, the small sample of burnt animal bones suggests that only a few bones were placed on the pyre. The age of the slaughtered pigs would be compatible with prime eating quality. It is reasonable to assume, therefore, that the majority of the carcass was used as a meal for the living, whilst inedible portions were burnt with the corpse as a token gesture. In the early Roman period, following the Celtic tradition, a formal meal was occasionally laid out beside the cremation vessel, in which plates, dishes and even cutlery were provided for the deceased (Biddle 1967; Philpott 1991, 197-8 and 237). As time progressed, this practice became infrequent and burnt animal bone fragments are found within the cremation vessel. At Each End, no animal bones, burnt or unburnt, were placed with the associated vessels; unburnt animal bones, shrew and mole, in vessels C.87 (S.19) and C.131 (S.5), and the single unburnt fish bone (vertebral centrum) in vessel C.471 (S.26), represent post-depositional contamination. The box/casket cremation In addition to the eleven cremations deposited in pottery vessels, a single cremation was discovered in a box or casket (S.10). The wood had rotted away but the outline of the container could be identified by the presence of iron nails. The container exhibited characteristics typical of both box and casket burials. The size (length 70 cm.; width 34 cm.; depth 24 cm.), as well 128 EXCAVATIONS AT EACH END, ASH, 1992 as the absence of ornate metal decoration, suggested that it represented a box rather than a casket (Philpott 1991, 12-21). The bone fragments, however, appeared to have been deposited directly in the box and not, as is the usual practice with box burials, inside a pottery or glass container (ibid. 16). It is possible, of course, that they were originally placed in a leather bag or pouch, which would leave no visible trace. Unlike most box and casket burials the example from Each End did not contain any grave goods except for cremated bird and pig bones.13 Over 50 box burials and approximately 40 casket burials are known from Roman Britain (ibid. 12-21). The majority occur in Hertfordshire and Cambridgeshire and both forms are known from the early Roman period but appear to be unknown before the arrival of the Romans. Most casket burials are first or second century; box burials are most frequent in the mid to late second century (ibid.). All the bones were hard and solid and did not leave a mark when scratched. Practically all were white/off-white (98.9 per cent), the remainder being blue/black. By weight, almost a quarter of the fragments were under 10 mm.; just over 8 per cent were larger than 50 mm., the majority of the bones being between 10-50 mm. The total weight of recovered human bone from the box was 1,234 gm. The bone:soil weight ratio was calculated as 1:18.3. Just over 40 per cent of the fragments could be identified. It appears that most elements of the skeleton were represented, but the skull (33.7 per cent) was over-represented, whilst the remainder of the axial skeleton was under-represented (12.9 per cent). Even allowing for differences in weight, the upper limbs were not so well-preserved (12.2 per cent) as the lower. Two teeth were recovered: the root of a lower right canine and the roots and an incomplete crown from an upper right second molar. Examination of the fragments suggests that the box contained only one individual. The remains were those of a fully-grown adult since the end plates of the cervical vertebrae were fused. There was insufficient evidence, however, for a more precise age. The gracile nature of the long bone diaphyses (the femoral circumference is only 69 mm.) supports the view that the remains were female. There was no evidence of either non-metric variants, bone or dental pathology on the available fragments. Pig teeth and lower jaw (4 gm.) and bird bones (10 gm.) were intermingled with the burnt human remains. The presence of two parallel cut-marks on the latter suggest that the flesh may have been removed prior to cremation. As with the other cremations, it appears that the small sample of non-human cremated bone represents a symbolic gesture of feeding the spirit of the deceased, whilst the majority of the animal was eaten by the mourners. The fragments deposited in the box were broadly similar in appearance to However, box burials outside the main distribution have been found without grave-goods; examples are known from Gloucester, York and Cumbria (Philpott 1991, 19, Fig. 2). 129 ALISON J. HICKS those found in the pottery vessels. All the bones were burnt in an oxygen-rich environment: this is most noticeable in those from the box. The maximum temperature, however, by the condition of the bones and the incomplete destruction of the teeth, was probably less than 500°C (Furuhata and Yamamoto, quoted in Buikstra and Goldstein 1973,23). It is possible that the bones were burnt in a fairly fierce fire of short duration. In modern crematoria a cremated adult body yields 1,600-3,600 gm. of bone (McKinley 1989), even allowing for the fact that the remains are extremely gracile, it would appear that less than three-quarters of the bones were collected and deposited in the box. It is possible that the smaller fragments and those that had turned to ash were not recovered. Most of the skeletal elements (skull; axial; upper and lower limbs) were present. The small size of the fragments is probably the reason why the skull was over-represented whilst the rest of the axial skeleton was under-represented. THE INHUMATION One inhumation (G.20A) was discovered. The skeleton was represented by a crushed skull and fragmentary long-bone shafts. The ribs, spine and pelvis, as well as the ends of the long-bones, had not survived. The available bones were very fragile. Fragments of cervical vertebrae in the surrounding grave fill (layers C.406 (S.8), C.415 (S.8)) belonged to the articulated inhumation. Based on the cranial morphology,14 the skeleton appeared to be female. Only two teeth, upper molars, were present. They were not heavily worn which at first glance suggests a young adult. However, the lower teeth had been lost during life; consequently, the degree of attrition on the available teeth may have underestimated the true age of the individual. The fact that the sutures of the vault were not sharply defined argues for an older age. From this evidence, an age estimation of 35-45 is obtained. There was no evidence of disease on the available bones. The fact, however, that the spine and the articular surfaces of the long bones were not recovered means that the severity of in vivo joint degeneration cannot be ascertained. The bones were too fragmentary for detailed metric analysis. Four indices could be calculated and they all fell within the normal range.15 Sixteen cranial and three post-cranial non-metric variants were available for examination. Only one trait was present: the mastoid foramina both present in the temporal. A similar finding was noted on the right side in the cremated remains from vessel C.70 (S.17). The trait is not an uncommon finding and appears to contain a low heritability. As such, it would be incorrect to assume that the two individuals are genetically related. 14 Lack of brow ridges; shape of forehead; size of mastoid process and the appearance of the lower jaw. 15 Frontage curve index: 83.3; Brachial index: 85.7 (r); Meric index 74.2 (r); cnemic index 86.0 (r). 130 EXCAVATIONS AT EACH END, ASH, 1992 THE PREHISTORIC POTTERY Nigel Macpherson-Grant The excavation produced 68 sherds of prehistoric pottery (overall weight: 550 gm.). Much of it was residual in Roman contexts, but twenty sherds came from features thought to be of pre-Roman date, nineteen of these from the large pre-Roman ditch (G.3). All sherds are of small to medium size with wear patterns ranging from fresh to fairly heavily abraded. The full range of prehistoric activity represented is difficult to assess; the sherds are generally too small and nondescript to ensure firm attribution in each instance. One body sherd (C.l 20 (G.27)) has gritting suggestive of Late Iron Age ware types; one rim fragment (C.40 (G.22)) could belong anywhere between the Late Bronze Age/Early Iron Age transition to Late Iron Age; two body sherds (C.l67 (G.21), C.282 (G3)) have tempering attributes similar to recognised earlier first millennium B.C. types. A fresh body sherd (C.l (G.73)) and a fairly worn base sherd (C.195 (G.32)) probably represent flint-and-grog-tempered wares of broadly c. A.D. 25-75 date, but they could be earlier (see below). What is indisputable is that is that Early Iron Age to Mid Iron Age activity is represented by four sherds of rusticated pottery (C.2 (G.72), C.42 (G.22), C.195 (G.32), C.333 (G.33)). Though all are fairly worn, the sherd size and degree of wear on those from C.42, C.195 and C.333 suggest derivation, if not directly from earlier (pre-Roman) levels, then from Iron Age occupation in the immediate vicinity. The sherd from C.42 is probably from a heavily rusticated straight-walled bowl similar to a vessel from Highstead Period 3 A (Macpherson-Grant forthcoming). The other three rusticated sherds may be of similar date or more specifically Early to Mid Iron Age (equivalent to Highstead Period 3B). Two sherds from the pedestalled base of a fine ware jar, recovered during the initial programme of trial trenching along the Ash by-pass (side code AB.92, context C.30), tend to confirm a Period 3B-type emphasis for the Each End rusticated sherds. None of the sherds retrieved from the pre-Roman ditch (G.3) are rusticated. Tempering/finishing attributes would not be out of place in either Late Bronze Age/Early Iron Age transition or Early to Mid Iron Age contexts, but, with one exception, most of the sherds are significantly fresher than most of the residual material and in situ occupation of later prehistoric date can be safely assumed. This is reinforced by the twelve sherds (all fairly fresh, some joining) representing one or two vessels from a layer (C.353 (G.3)) in the pre-Roman ditch. The exception is a sherd from a layer (C.279 (G.3)) in the same pre-Roman feature; this is from a thick walled, larger diameter storage jar with overall abrasion significantly heavier than other sherds from layers in the same feature. The ditch either remained open for a considerable period, or incorporates material representing relatively long-term activity in the area. The nature of the infilling deposits suggests the former. The presence of this prehistoric material suggests three possibilities. The first is that all the flint-gritted sherds are derived from in situ occupation of 131 ALISON J. HICKS single period date, equivalent to Highstead Periods 3A-3B. Since grog tempering is a recognised (but minor) attribute of Early Iron Age pottery production, the potential later Iron Age flint-and-grog-tempered sherds could all belong to this scenario. Alternatively, rather longer-term occupation, possibly continuous and equivalent to Highstead Periods 2, 3A and 3B, may be represented at Each End. The third possibility is that multi-period, non-continuous occupation, possibly continuous and equivalent to Highstead Periods 2, 3A and 3B, may be represented at Each End. The third possibility is that multi-period, non-continuous occupation was occurring upon the site, with phases of activity occurring during either of the periods described above but also having a later Late Iron Age element. On balance, the third suggestion is that preferred, with a date of c. 650-450 B.C. proposed for the main phase of pre-Roman activity, whilst the evidence of the Late Iron Age coin and the possibility that some flint-and-grogged sherds are late emphasises the possibility that multi-period (pre-Roman) activity was present. The rusticated sherds confirm the presence of yet another settlement of broadly Early Iron Age to Mid Iron Age date and underlines the growing recognition that Iron Age settlement during the sixth to the fifth century B.C. was regionally far denser than recently suspected. The presence of fabric/ formal attributes similar to Highstead Period 3 A is socially significant, since it further confirms the implications of the Highstead evidence: namely that the ceramic change from Late Bronze Age types to specifically Continentalstyle rusticated Early Iron Age types was more likely to have taken place via peaceful cultural adoption/influence than rapid 'invasion-type' replacement. Although there is a considerable body of indirect evidence supporting this scenario, sites actually exhibiting this potential are rare, so that the present data are a useful contribution. The evidence for immediate pre-Roman activity is slight. Few, if any, of the purely grog-tempered wares appear to represent pre- or Conquest period A.D. native occupation, but the Late Iron Age coin and, less certainly, the possible Late Iron Age 'Belgic'-style sherds do indicate a degree of activity in the area during this time and may act as a potential explanation for the presence of the Roman settlement/activity itself. THE ROMAN POTTERY Andrew Savage INTRODUCTION Excavation yielded a total of 5,305 sherds, weighing c. 66.5 kg. and representing a wide range of local and imported types. Most of the material dated to the mid to late second to third century and much of it consisted of small to medium-sized, often worn sherds. A high degree of chemical weathering had taken place softening sherds and often removing evidence of surface treatment. 132 EXCAVATIONS AT EACH END, ASH, 1992 It would appear, on the basis of sherd size and condition, that most if not all of the pottery represents secondary deposition. Most of the pottery was recovered from stratigraphically isolated features, largely drainage gullies and ditches, but also from a number of pits, wells, etc.. Many of these features are judged to represent peripheral activity relating to a small settlement lying beyond the area of excavation. Ceramically, the most interest feature of the site was the pottery from fourteen cremation burials (G.21, 22, 23), all apparently dating to the mid to late second century and containing a substantial quantity of stamped samian. This material has been described in detail and fully illustrated, and is discussed below. A fifteenth cremation (S.10) yielded no pottery. Only a brief summary of the other pottery from the site is included in this report. Pottery dating for individual features or groups is included in the stratigraphic report, where appropriate. The samian stamps report was written by Brenda Dickinson. The other samian was identified and dated by Maggy Taylor. POTTERY FROM THE CREMATION BURIALS Summary Fourteen cremation burials yielding pottery were excavated from three distinct areas (G. 21, 22, 23). In total, forty-eight pots and one ceramic lamp were recovered. The fabrics present are listed below. Most of the fabrics are local or traded wares whose characteristics are well-known. The local wares are discussed at length by Pollard, in Blockley et al. (1995). See also the Discussion and Catalogue, below. Chemical weathering has severely affected the surfaces of many vessels, especially those in softer fabrics. In the case of some transitional Native Coarse Ware and fine Upchurch-type vessels, decorative motifs have been partially and in some instances possibly completely removed. Six vessels displayed severely truncated profiles. These and two parts of a single Dressel 20 amophora (burial S.18) have not been drawn. Eleven of the burials (S.4, S.14, S.15, S.16, S.17, S.18, S.19, S.22, S.23, S.24, S.26) have been dated to the mid to late second century and three, containing only single coarse ware urns, as broadly second century although they are likely to be of similar date to the others (S.5, S.13, S.21). The importance of these burials lies in the fact that, though small in number, they appear to fall within a relative closely defined date bracket. The dating of the coarse wares, particularly that of the cinerary urns, has been refined by the presence of thirteen samian vessels (c. 26 per cent of the total by vessel count), nine of which are stamped (see samian stamp report (10.3)), and other fine wares. As a group they thus constitute an extremely useful 'dating calibrator' for the mid to late second century in east Kent, a period of ceramic transition (see discussion, below). The assemblage may contribute to future typological studies of coarse grog-tempered ware of the period. 133 ALISON J. HICKS Discussion Twelve of the burials contained between one and four vessels, typical numbers in second-century cremations (Philpott 1991, 32). Of the remaining two burials, one (S.24) contained five and the other, a richly furnished burial (S.4) which yielded fine ware imports from Gaul, the Rhineland and possibly Colchester, contained the unusually large number of ten, giving a mean figure of 3.5 vessels per burial. It should be noted, however, that burial S.21 was robbed prior to excavation and may have originally contained a larger number of pots. In a sample of 1477 cremations from fifteen cemeteries in south-east England considered by Philpott, only fifteen burials yielded as many as, or more than, ten vessels (ibid. 34). No burial excavated to date at Canterbury has yielded as many. Most of the multiple-vessel interments contained typical variations on the standard jar/flagon/dish/beaker ceramic suites which usually furnished cremations of the period. All but two of the jars are in Transitional Native Coarse Ware, a local (east Kent) coarse grog-tempered fabric. Of the remainder, one (S.15, Fig. 9.21), a double- handled jar/'honeypot', is in a fine buff sandy ware, and the other, ajar/beaker, is in fine grey Upchurch-type ware (S.22, Fig. 7.3). All of the interments appear to date to the mid to late second century, with a possible emphasis in the third quarter. A fifteenth cremation (G.23, S.10), a box-burial, was also recovered. This is not discussed here as it yielded no pottery. Most recorded box-burials date to the mid to late second century (ibid. 17). Although evidence is slight, the interment of fifteen individuals within what seems to have been a short period of time might imply that the cemetery served a small settlement rather than an individual farmstead. It is not possible to suggest, on the basis of the available evidence, that the three groups into which the burials are divided represent a chronologicallybased development. It is interesting, though, that very little of the other pottery recovered from the site pre-dates the grave-groups, the main period of activity being the late second to third century and later. It is possible that the nearby settlement expanded or otherwise developed in the late second century to include an area previously reserved for burial of the dead. It was the Roman custom to bury the dead outside settlement boundaries. None of the burials contain the Canterbury grey sandy ware bowls or jars of late first- to mid second-century date that are widely distributed in east Kent. This constitutes further evidence for the contraction in the Canterbury pottery industry and the distribution of its products which seems to have taken place in the mid second century and which has been observed elsewhere. Canterbury grey wares are, for example, virtually absent from Period II deposits of the Classis Britannica fort at Dover, dated to c. A.D. 155/160-180 (Pollard 1988, 95). Canterbury pink-buff flagon wares however, which continued to be- manufactured to at least the late second century, are present here. They are also present in the mid second- to third-century cemeteries at Cranmer House, Canterbury (Pollard 1987, 284-298) and Ospringe (Whiting et al. 1931). Equally striking is the absence of the BB2 jars and dishes whose incidence 134 EXCAVATIONS AT EACH END, ASH, 1992 in east Kent increased in the later second century as that of Canterbury sandy wares declined (Pollard 1988, 97). Coastal sites such as Richborough and Dover exhibit a higher incidence of BB2 than inland sites such as Canterbury (ibid., 93). Although BB2 jars are generally uncommon at Canterbury in both funerary and occupation deposits, dishes are not and are present in the cemetery at Cranmer House; both jars and dishes are present at Ospringe. The absence of BB2 at Each End is thus worthy of comment. The explanation may involve questions of status and choice of vessel rather than chronology. Although eight of the fourteen burials contain at least one dish as part of the pottery suite, they are all in samian ware. Given our substantial ignorance as to the nature and extent of the settlement with which these cremations are presumably associated and the assemblage-variation between different cemeteries associated with a single occupation-site observed elsewhere (Philpott 1991, 33), questions of settlement status are not addressed here. All of the thirteen samian vessels present are undecorated. Of these, twelve are Central Gaulish and all are dated to the mid to late second century. The earliest is dated Hadrianic/early Antonine and none of the stamped vessels is definitely later than the A.D. 170s (see samian stamp report, p. 136). Although the problems of dating coarse wares on the basis of associated samian which might enjoy a much longer life are well known, the abundant and well-dated samian present here is considered to support the suggestion that few, if any, of the coarse ware vessels are likely to post-date the second century. Similarly, none of the other local imported finewares need be of third-century date and are likely to be earlier. The coarse ware urns are all grog-tempered, in a fabric which is transitional between soft-fired, grog-tempered ware of the first and early second centuries and the much harder-fired Native Coarse Ware of the late second to third centuries. The ware is commonly found in east Kent. Several examples here (e.g. Figs. 8.9, 9.19) display the decorative motifs often associated with it: burnished chevrons on the body beneath a double groove and oblique linear or lattice-burnish on the shoulder. The development of these fabrics and their distribution have been discussed at some length by Pollard (1987, 298; 1988, 98; 1995, 704-5). He considers that the transition was complete 'by the last quarter of the second century, if not somewhat earlier' (ibid., 298). The balance of the fine and coarse ware dating here suggests an emphasis for the dating of the burials in the third quarter of the second century. Ten transitional Native Coarse Ware vessels (nine jars and one bowl) have been illustrated (Figs. 7.4; 8.9, 10; 9.19, 22; 10.24; 11.29; 12.32, 35; 13.39). They demonstrate the substantial degree of detailed typological variation present in this fabric. Given the relatively tight date-range suggested for the burials, it is to be hoped that their illustration may contribute to future studies of the typological development of grog-tempered ware in east Kent in the second and third centuries. Burial S.19 (second half of second century) provides an example (C.92, Fig. 12.38) of the occasional inclusion of lamps in Romano-British cremations. Philpott has suggested, while acknowledging an insufficient body of data, that the practice may have become more common in the mid to 135 ALISON J. HICKS late second century. This example adds to that body of data. Lamps have so far not been recorded in burials (cremation or inhumation) dating later than the early third century (Philpott 1991, 191-92). One amphora burial was recovered (S. 18). Most Romano-British amphora burials are found in south-east England. Most belong to the mid to late second century, although third-century examples are occasionally found (ibid., 22-23). This example adds to the sixty-four noted by Philpott as being recorded in Kent (ibid., 22). Thirty-eight were recorded at Ospringe (Whiting et al. 1931; Pollard 1988, 105); other examples have been excavated by the Canterbury Archaeological Trust at Cranmer House, Canterbury (Pollard 1987) and Crundale (publication forthcoming). Here at Each End, the neck of aDressel 20 south Spanish olive-oil amphora was carefully removed to just below the level of the handle/body junction. This allowed the other vessels associated with the burial to be placed within it. Sometimes, in such cases, the neck was discarded and replaced by another cover; sometimes it was repositioned on the shoulder. In this case it would appear that the latter procedure was followed. However, possibly because the upper body was damaged during the removal of the neck, preventing a secure fit, or as a result of pressure from above, the neck was either placed within or subsequently fell into the body. Another amphora, by its description probably also a Dressel 20, has been previously recorded in a burial at Ash (VCH (1932), Kent, 144; Callender 1965, 27) and two others were recorded at nearby Richborough (Cunliffe 1968, 27). The samian stamps Barbara Dickinson Each entry gives: burial (set) number, context number, potter (i, ii, etc., where homonyms are involved), die number, form, reading of the stamp, published example, pottery or origin, date. (a), (b) and (c) indicate: (a) Stamp attested at the pottery in question; (b) Potter, but not the particular stamp, attested at the pottery in question; (c) Assigned to the pottery on the evidence of fabric, distribution and, or, form. Ligatured letters are underlined. Stamp No. 1. Burial S.17 (C.72). A(e)nisatus la 88/31 IS TVSF (ORL B4, 30, 2) Trier (a). This stamp almost certainly belongs to a potter who also used the name Aenisatus, but on one die only. At some stage in his career he also worked at the Argonne factory of Pont-des-Remes, and at Heiligenberg. There is no site dating for him at all, but his use of this stamp on form 27, which ceased to be made in Central Gaul about A.D. 160 and, as far as can be seen, in East Gaul too, suggests Hadrianic or early-Antonine date. This would be consistent with the form of the Each End vessel. Fig. 11.30. 136 EXCAVATIONS AT EACH END, ASH, 1992 Stamp No. 2. Burial S.4 (C.267). Avitus iv lg 18/31-31 AVITI MA Lezoux (b). Some of this potter's stamps, from other dies, have been noted on samian after c. A.D. 150; they are also common in Antonine Scotland. This evidence, combined with his frequent use of forms 18/31 and 27, suggests that he was at work c. A.D. 125-150. Fig. 8.12. Stamp No. 3. Burial S.4 (C.270). Cerialis ii 4a 33 CERI AL M Lezoux (b). In the early- to mid-Antonine period Cerialis ii was associated with Paullus iv and Cinnamus ii in the production of decorated ware; this occurs frequently in Antonine Scotland and in a group of burnt samian of the 140s at Castleford. This particular stamp, noted only on plain ware, reflects that range, with its use on forms 18/3IR, 27 and Ludowici Tg. c. A.D. 135-170. Fig. 8.15. Stamp No. 4. Burial S.19 (C.91). Osbimanus 7a 31 OSBIM VS Lezoux (b). This stamp has been noted at the Pennine fort of Bainbridge, where it is on a form not current before c. A.D. 160, and elsewhere on the contemporary cup, form 80. However, the form of the Each End vessel suggests that it belongs to the earlier years of the die's use, c. A.D. 150-160, along with dishes of form 18/31R which carry the same stamp. Fig. 12.37. Stamp No. 5. Burial S.26 (C.475). Pugnus ii lc 18/31-31 PVGNIM Lezoux (b). Only two other vessels with this stamp are known to the present writer, both on form 18/31 from a pottery shop at Castleford destroyed by fire in the 140s. Other stamps of Pugnus from this burnt deposit occur on Hadrianic and early- Antonine forms, such as 27 and 81, and one is also on a decorated bowl from a very worn mould in a pit at Alcester, filled in the 150s (forthcoming), c. A.D. 130-150. Fig. 13.41. Stamp No. 6. Burial S.16 (C.49). Sextus v 8b 31 SEXTVS F Lezoux (a). There is no site dating for this particular stamp, but its presence on forms 31R and 79 or Ludowici Tg makes mid to late Antonine use certain. Other stamps of the same potter occur on Hadrian's Wall (Benwell and South Shields) and in a group of late Antonine samian recovered off Pudding Pan Rock, Kent. c. A.D. 160-200. Fig. 10.26. Stamp No. 7. Burial S.4 (C.269). Titurus 2a' 18/31-31 TITVRI M (Walke 1965, no. 370) Lezoux (c). This stamp comes from a broken die, which originally had a frame with ansate ends. Neither version of it has been easy to date so far, but the later one, represented here, occurs on form Ludowici Tg; this suggests that the die was still in use after A.D. 160. The Each End stamp provides the best evidence yet for the early use of the die, in that the dish, on the evidence of its form, is unlikely to be later than c. A.D. 145, and probably belongs to the early 140s. Fig. 8.14. 137 ALISON J. HICKS Stamp No. 8. Burial S.15 (C.39). Verecundus v la 18/31 V RECVNDVSF (Ludowici 1927, 232, d) Blickweiler (a). East Gaulish potters of this name unfortunately still lack convincing dating evidence. There are four examples of this stamp from Rheinzabern, on burnt vessels; some, and probably all, are from the cemetery rather than the kiln-site, so there is no reason to think that he worked there. The other sites on which the stamp has been found are consistent with its origin at Blickweiler. One stamp from another Blickweiler die is also noted from the Trier potteries, but as it is not on a waster, the vessel could equally well have originated at Blickweiler. One of the East Gaulish Verecundi worked successively at several other potteries, but his career seems to have been rather later, and he is unlikely to have been the same man as Verecundus v, who was almost certainly at work in the Hadrianic to early Antonine period, to judge by his use of forms 18/31 and 27. The form of the dish is consistent with such a date. Fig. 9.23. Stamp No. 9. Burial S.18 (C.31). Uxopillus 6c 31 VXXOPILLI Lezoux (a). The unusually deep wall and high kick of this vessel suggest that it is extremely late in the Central Gaulish range of form 31 and this alone is enough to make a late second-century date certain. Uxopillus's other stamps occur in a pit of c. A.D. 150-160 at Alcester (forthcoming), in a group of burnt samian of c. A.D. 170 at Tac (Hungary) and at Benwell and Newstead. His range will be c. A.D. 150-180, but this particular dish almost certainly belongs to the 170s. Fig. 12.34. The catalogue Vessels are listed in Group no. order. Cinerary urns are indicated by an asterisk: (*). Burials north of road (G.21) (S.21, 22, 23, 24) Burial S.21 This burial was robbed. It may therefore originally have contained a larger number of pots. (C.l47) (*) Transitional Native Coarse Ware. Probably necked everted-rim jar. Truncated. Not drawn. Date: Second century. Burial S.22 (C. 161) Samian. Central Gaulish. Curie 15 dish. Antonine. Fig. 7.1. (C.l62) Canterbury fine pink-buff sandy ware. Triangular-lip flagon cf. Kirkman 1940, no. 49. Mid to late second century. Fig. 7.2. (C.165) (*) Fine grey Upchurch-type ware. Everted-rim breaker, cf. Monaghan class 214.3 (Monaghan 1987), which is dated A.D. 90-150. Traces of almost vertical burnished linear decoration beneath shoulder groove. Fig. 7.3. Date: Second half of second century. 138 EXCAVATIONS AT EACH END, ASH, 1992 1 ^ s!m S rp. Ik**-* c•^j*--^r k lll\h'm>Mli !', L 6 ^ ^ 8 Fig. 7. The Roman Pottery: nos. 1-8. Scale 1:4, except stamp at 1:1. Burial S.23 (C.148) (*) Transitional Native Coarse Ware. Necked everted-rim jar. Wiping on lower body. Fig. 7.4. (C. 149) Samian. Central Gaulish. Curie 46 cup. Hadrianic/Antonine. Fig. 7.5. (C.l50) Fine grey Upchurch-type ware. Very small narrow-necked flask, cf. Monaghan class IB. Fig. 7.6. Date: Mid to late second century. 139 ALISON J. HICKS Burial S.24 (C.142) (*) Transitional Native Coarse Ware. Probable necked everted-rim jar. Truncated. Not drawn. (C. 143) Fine grey 'Upchurch' ware. Rouletted, probable bag-beaker. Badly worn and truncated. Cf. Pollard, in Blockley et al (1995), Fig. 308, no. 369 and Fig. 309, no. 376, from deposits dated late second to mid third century. Not drawn. (C.I44) Canterbury pink-buff sandy ware. Probable cupped or ring-neck flagon. Truncated. Not drawn. (C.469) Samian. Central Gaulish. Curie 46 cup. Antonine. Fig. 7.7. (C.470) Samian. Central Gaulish. Curie 23 dish. Antonine. Roughly scratched graffito on exterior. Fig. 7.8. Date: Second half of second century. Burials south of the road-east (G.22) (S.4, 13-19, 26) Burial S.4 (C.260) (*) Transitional Native Coarse Ware. Narrow-necked everted-rim jar, cf. Pollard 1987, fig. 107, no. 3A; cremation dated mid second to third century. Burnished chevron decoration beneath double groove. The shoulder above the groove may originally have also been decorated. Wiping on lower body. Fig. 8.9. (C.262) Transitional Native Coarse Ware. Necked everted- rim jar. Wiping on girth and lower body. Cf. Pollard 1987, fig. 107, no. IA and fig. 109, no. 10A; cremations dated mid second to third century. Fig. 8.10. (C.265) Lower Rhineland fabric 1. Cornice-rimmed, roughcast-decorated beaker, cf. Anderson, Fig. 7, no. 1, which is dated c. A.D. 70-120. Fig. 8.11. (C.267) Samian. Central Caulish. Drag. 18/31-31 dish. c. A.D. 125-150.. Stamp of Avitus. See samian stamp report, No. 2 (10.3). Fig. 8.12. (C.268) Oxidised Upchurch-type ware. Flange-rim segmented bowl, cf. Monaghan class 5B2. ?c. A.D. 90-130. Fig. 8.13. (C.269) Samian. Central Gaulish. Drag. 18/3-31 dish. Probably c. A.D. 140-145. Stamp of Titurus. See samian stamp report, No. 7(10.3). Fig. 8.14. (C.270) Samian. Central Gaulish. Drag. 33 cup. c. A.D. 135-170. Stamp of Cerialis. See samian stamp report, No. 3 (10.3). Fig. 8.15. (C.27I) Canterbury coarse pink-buff sandy ware. Cupped ring-neck flagon, cf. Pollard 1987, fig. 111, no. 19B, dated c. A.D. 130-200/250 in a cremation dated mid second to mid third century. Fig. 8.16. (C.273) Canterbury coarse pink-buff sandy ware. Description and date as for (C.27I). Fig. 8.17. (C.277) Colchester/north Gaulish colour-coated ware. Cornice-rimmed, roughcast-decorated beaker, cf. Anderson 1980, Fig. 13, no. 1, dated c. A.D. 120-150. Fig. 8.18. Date: Mid to late second century. Burial S.13 (C.43) (*) Transitional Native Coarse Ware. Necked everted-rim jar. Vestigial cordon on neck. Burnished chevron decoration beneath double shoulder groove. The shoulder above the groove may originally have also been decorated. Wiping on lower body. Fig. 9.19. Date: Second century. 140 EXCAVATIONS AT EACH END, ASH, 1992 I ' I '•fan-fa/ 1 (@mYEMM> www?jaju^2D gii^K Fig. 8. The Roman Pottery: nos. 9-18. Scale 1:4, except stamps at 1:1. 141 ALISON J. HICKS •.••vtt ev 19 Afr" —• ^H^f=^~ - — ""^ . A^^*\ •• - -.-- —j —/ —-•*=*.-*=*=/ =0^\, ^a\0?®' iMT 22 Fig. 9. The Roman Pottery: nos. 19-23. Scale 1:4, except stamp at 1:1. 142 EXCAVATIONS AT EACH END, ASH, 1992 Burial S.14 (C.78) (*) Transitional Native Coarse Ware. Probably necked everted-rim jar. Truncated. Not drawn. (C.80) Fine grey Upchurch-type ware. Poppyhead beaker, cf. Monaghan class 2A4. Undecorated. Mid to late second century. Fig. 9.20. Date: Mid to late second century. Burial S.15 (C.37) (*) Fine sand-tempered ware; cream-buff surfaces, grey core. Possibly north French. The ware is similar to butt- and pentice-beaker fabrics at Canterbury which have been identified as being of north-west Gaulish origin (Rigby, in Blockley et al. 1995). Double-handled jar/'honeypot'. Vessels of this form are not common in east Kent. Several broad parallels were recovered from Richborough, dated to the late first century (Bushe-Fox 1932, nos. 296, 297, and 1949, no. 389). Fig. 9.21. (C.38) Transitional Native Coarse Ware. Neck everted-rim jar. Wiping on girth and lower body. Cf. Pollard 1987, fig. 109, no. 10A; cremation dated mid second to third century. Fig. 9.22. (C.39) Samian. Central Gaulish. Drag. 18/31 dish. Stamp of Verecundus. See samian stamp report, No. 8 (10.3). Fig. 9.23. Date: Mid to late second century. Burial S.I6 (C.47) (*) Transitional Native Coarse Ware. Narrow-necked everted-rim jar. Burnished chevron decoration beneath double groove. Vertical knifetrimming on lower body. Fig. 10.24. (C.48) Canterbury fine pink-buff sandy ware. Triangular-lip flagon, cf. Kirkman 1940, no. 51. Mid to late second century. Fig. 10.25. (C.49) Samian. Central Gaulish. Drag. 31 dish. Mid to late Antonine. Stamp of Sextus. See samian stamp report, No. 6 (10.3). Fig. 10.26. (C.83) Fine grey Upchurch-type ware. Poppyhead beaker, cf. Monaghan class 2A4. Undecorated. Mid to late second century. Fig. 10.27 Date: Second half of second century. BurialS. 17 (C.68) Fine buff ware (?Upchurch-type); orange-red core. Fine silty micaceous fabric with sparse to moderate, fine to very coarse sub-angular clear and milky quartz, occasional fine to very coarse ferruginous inclusions and sparse fine white mica. Large, tall-necked flagon; single, 3-rib handle. Rim missing. Fig. 11.28. (C.70) (*) Transitional Native Coarse Ware. Necked everted-rim jar. Lattice- burnish on shoulder, it may originally have been decorated beneath the double groove. Wiping on lower body. Cf. Pollard 1987, Fig. 107, no. 2A; cremation dated mid second to third century. Fig. 11.29. (C.72) Samian. East Gaulish. Drag. 18/31 dish. Hadrianic to early Antonine. Stamp of A(e)nisatus. See samian stamp report, No. 1 (10.3). Fig. 11.30. (C.73) Fine grey Upchurch-type ware. Poppyhead beaker; probably undecorated. Probably mid to late second century. Cf. Pollard 1987, Fig. 113, 143 ALISON J. HICKS r n f / ! i .* \ VI" » »r 25 ,* i ' / 24 27 26 ^gfHi K^ y y o i a Fig. 10. The Roman Pottery: nos. 24-27. Scale 1:4, except stamp at 1:1. no. 35B; cremation dated late second to third century. Fig. 11.31. Date: Mid to late second century. Burial S.18 (C.23) South Spanish Dressel 20 amphora (same vessel as (C.67)). Buff fabric. Used as a container for the other vessels in this burial. Truncated; the lower 2/5 of the vessel survives. Not drawn. (C.30) (*) Transitional Native Coarse Ware. Wide-mouthed bowl, cf. Ospringe no. 492 which was accompanied by samian dated A.D. 160-190. Burnished decoration above girth-groove. Fig. 12.32. (C.31) Fine oxidised Upchurch-type ware with cream slip. Cup-mouthed flagon, cf. Monaghan class IE 1.4, dated A.D. 140-190 and Ospringe nos. 124, 164 and 177. Fig. 12.33. (C.32) Samian. Central Gaulish. Drag. 31 dish. Probably c. A.D. 170-180. Stamp of Uxopillus. See samian stamp report, No. 9 (10.3). Fig. 12.34. 144 EXCAVATIONS AT EACH END, ASH, 1992 Fig. 11. The Roman Pottery: nos. 28-31. Scale 1:4, except stamp at 1:1. 145 ALISON J. HICKS KKS/uffl2£P 38 Fig. 12.The Roman Pottery: nos. 32-38. Scale 1:4, except stamps at 1:1. 146 EXCAVATIONS AT EACH END, ASH, 1992 (C.67) South Spanish Dressel 20 amphora (same vessel as (C.23)). Neck and shoulder, used as a 'cover' for (C.30), (C.31) and (C.32). Rim missing. Not drawn. Date: Late second century. Burial S.19 (C.87) (*) Transitional Native Coarse Ware. Narrow-necked everted-rim jar. There may originally have been decoration above and/or below the double shoulder groove. Strongly defined vertical knife-trimming on girth and lower body. Fig. 12.35. (C.89) Canterbury coarse pink-buff sandy ware. Ring-neck flagon. Mid to late second century. Fig. 12.36. (C.91) Samian. Central Gaulish. Drag. 31 dish. c. A.D. 150-160. Stamp of Osbimanus. See samian stamp report, No. 4 (10.3). Fig. 12.37. (C.92) Fine buff ware. Lamp, cf. Loeschke type IX; fine, dense, silty matrix, sparse to moderate, fine- to medium-sized angular and sub-angular, clear and milky quartz; sparse, fine white mica; sparse, fine, rounded and sub-angular ferruginous inclusions. A very similar fabric to (C.68), burial S.17, but finer and denser, containing more ferruginous inclusions, but less coarse quartz. ?Local or Upchurch-type. A lamp of this type was recovered from a Canterbury cremation dated late second to early third century (Pollard 1987, fig. 112, no. 23G). Fig. 12.38. Date: Second half of second century. Burial S.26 (C.471) (*) Transitional Native Coarse Ware. Necked everted-rim jar. Heavily weathered oblique linear burnished decoration is visible on the shoulder; there may also originally have been decoration beneath the double shoulder groove. Fig. 13.39. (C.473) Fine grey Upchurch-type ware. Short-necked flask, cf. Monaghan class IB. Fine 'Upchurch' flasks are usually found in mid/late second- to early fourth-century contexts in Canterbury. Fig. 13.40. (C.475) Samian. Central Gaulish. Drag. 31 dish. c. A.D. 130-150. Stamp of Pugnus. See samian stamp report, No.5 (10.3). Fig. 13.41. (C.476) Fine grey Upchurch-type ware. Everted-rim beaker with latticeburnish decoration on body. The formal and decorative characteristics of this vessel are usually associated with BB2 jars of second-century date. Cf. Wilson, in Blockley et al. (1995) fig. 307, no. F192, from a deposit dated c. A.D. 140-170. Fig. 13.42. Date: Mid to late second century. Burials south of the road-west (G.23) (S.5) Burial S.5 (C.131) (*) Transitional Native Coarse Ware. Probably necked everted-rim jar. Truncated. Not drawn. Date: Second century. 147 ALISON J. HICKS V // y *= t \ \ 40 \ 39 41 7, (\