Roman and Medieval Development of a Canterbury Suburban Area: Excavations at Land Adjoining No. 10 Wincheap

Roman and Medieval development of a Canterbury suburban area: Excavations at land adjoining No. 10 Wincheap

grant shand and alison hicks

with Sheila Sweetinburgh and Luke Barber, Robin Bendrey, Lynne Bevan,

Malcolm Lyne and John Shepherd

In 2002 the Canterbury Archaeological Trust carried out excavation on land adjacent to 10 Wincheap, Canterbury (TR 61449 15735) in advance of redevelopment (Plate I, Fig. 1). The site had remained as vacant ground since 1963 following demolition of the standing properties to make way for a roundabout on Canterbury’s new ring road. Evaluation trenches cut in 2001 had revealed the potential for significant archaeological remains, leading to excavation in September 2002. This revealed archaeological horizons down to the foundation level of the proposed development. A small slot (A) cut in the southern corner of the site examined a lower sequence of features and deposits. Prehistoric, Roman and medieval activity was identified, although the occupation sequence was not continuous.

Nothing was formerly known of prehistoric activity in this area of Canterbury, perhaps in part because few opportunities have arisen to explore these lower levels. Roman occupation is better understood. The site lies c.65m south-west of Roman Worthgate, where the Roman road from Lympne and Wealden areas entered the town. Metallings thought to have formed part of the road were discovered in Gordon Road, c.50m south of the present site (Rady 1999). A burial ground bordered the east side of the road, inhumations and cremations being excavated within the vicinity of Station Road East and Gordon Road (Anderson and Rady 1990; Bennett 1991; Jenkins 1952).

Wincheap became a substantial suburb during the Anglo-Saxon period, and continued to flourish throughout the medieval period. Canterbury Castle, of Norman origin, lies close to the site, just inside the city walls, and may have been one factor influencing development of the area. Evidence of medieval occupation is primarily provided by the documentary sources, which suggest that by c.1200 plots of ground had been carved out of farmland lying between Wincheap Street and the river to the north-west. Dwellings and workshops flanked the road, and gardens lay behind. These properties were rebuilt and re-occupied over the centuries in what appears to have been a largely uninterrupted sequence of occupation, though access through Worthgate was blocked in 1548 and the route from Wincheap into the town diverted to a new gate (Wincheap Gate). Wincheap Street remained a major thoroughfare and retained properties along its borders.

Properties standing on or close to the site prior to 1963 included The Cedars, a large late eighteenth-/early nineteenth-century structure, to the north of the excavated area, and Georgian/Regency style properties, perhaps with earlier origins, on the site itself. These were demolished to make way for the ring road.

This report provides a summary of the site history, a more detailed discussion of the archaeological remains, the documentary evidence and an overview of the finds assemblages. All the material from the excavations has been catalogued and copies of the archive texts and finds catalogues, including those for the tile not reported here, are available for consultation. During stratigraphic analysis, the contexts recorded during excavation (prefixed C) were formed into groups of related activities, and it is these groups (prefixed G) that form the basis for the narrative which follows. Where more specific context information is required, context numbers appear in the text.

summary of the site history

Prehistoric activity was identified in the form of pits and a sequence of soils. Although sparse, its presence is of note since prehistoric occupation has only rarely been attested in Canterbury, and never before in this area of Wincheap.

Roman occupation, dating from as early as the second half of the first century ad and continuing into the middle of the third, was no doubt stimulated by the proximity of the town of Durovernum Cantiacorum and the presence of the major road out of Worthgate. Although burial activity might have been expected given the position of the site, instead metallings and spreads of occupation material, together with butchery evidence, suggest the presence of road-side yards used as work areas. Intercutting features, including pits and post-holes, were indicative of associated activities, although structural occupation was only very tentatively suggested. There was little evidence of Roman activity after the middle of the third century ad.

Evidence of Anglo-Saxon activity was entirely absent from the site. This part of Wincheap, at least, appears to have remained open land until at least the eleventh century.

Medieval occupation was influenced by the growing importance of Wincheap Street. Pits and post-holes, perhaps cut as early as the middle of the eleventh century, suggest open ground but nearby occupation. By the thirteenth/fourteenth century, if not earlier, properties probably fronted the adjacent length of Wincheap Street. Ground to the rear, accommodating flint metallings, pits and soils containing occupation debris, perhaps formed an associated plot or garden. Animals could have been kept, and slaughtered, here as suggested by the animal bone evidence.

The earliest property on the site was perhaps erected during the sixteenth century, probably fronting Wincheap Street. An external yard was bounded by a wall, partly enclosing what was thought to have been a garden. After later renovations, the building was demolished, perhaps during the eighteenth century. In the later eighteenth/nineteenth century, new properties were erected which stood until 1963.

the excavated evidence

Evidence for Prehistoric activity was identified in Slot A, cut slightly deeper than the development formation levels in order to investigate early remains lying above the natural subsoil (Fig. 2). Cutting the Head Brickearth were two small pits (G2), one identified only in section, infilled with soils providing little indication of their function. The excavated feature [C179] contained a single sherd of what was thought to have been middle Iron Age, or possibly earlier, pottery. A flint waste-flake, of Bronze Age date, was also recovered.

Overlying the pits was a sequence of soils flecked with chalk, charcoal and daub [G3], recorded only in section. They were thought to have represented open ground, which could have been used for agriculture, either as arable or grazing land, or been empty and largely devoid of activity. Three sherds from the uppermost horizon were of middle Iron Age or earlier date whilst a fourth was dated 50 bc-ad 50.

The earliest Roman activity [G6] (Fig. 3) was identified within Slot A, where a deposit of fine flint metalling overlay the earlier soils. Upon the metalling were layers of what appeared to be occupation material, comprising thin silt and clay deposits containing inclusions of ash, charcoal, oyster shell and daub. The metalling was too insubstantial to have represented a road or track, but perhaps formed the surface of a yard or work area. To the north-west lay a sequence of soils flecked with charcoal, shell, daub and animal bone [G4]. Excavated within a narrow slot, their stratigraphic relationships with the metalled surface were unknown. If contemporary, as the levels tentatively suggest, they perhaps indicated open ground to the rear of the work area. The pottery recovered suggests that this occupation could be dated to the second half of the first century, the soils to the rear perhaps forming shortly after ad 70. Other sherds of mid to late first-century pottery, recovered as residual material in later contexts, provide additional evidence of early Roman occupation, at least within the vicinity.

Cutting the occupation deposits above the metalling was a post-hole [G7.1] (Fig. 4), the fill of which included pottery fragments dated ad 120-160. Overlying were further spreads of silty clay [G7.2; not illustrated], excavated solely within Slot A, containing occupation material including abundant charcoal fragments, common bone and oyster shell and a dolphin type brooch of probable first-century ad date. Their presence suggests continued use of the ground, either as a work area or for the disposal of refuse.

Subsequent activity saw the laying of an extensive cobbled surface [G8, some identified only in section], different in character from the earlier metalling since it comprised large closely-set flint nodules. The deposit formed a substantial, hard-wearing surface thought to have been used as a yard. Overlying were spreads of occupation debris [G10, G13.1, G14, G15, G16.1, G25.1; most not illustrated], including patches of burnt clay, lenses of ash and charcoal and small patches of sand, attesting to varied use of the surface, perhaps over a period of time. One of the deposits [G10] comprised a sequence of clay and occupation lenses suggestive of flooring which had been successively patched, and may have indicated the location of a specific activity, or even an overlying structure of which all other trace had been removed. Pottery recovered from the cobbled surface was dated ad 90-130, although the underlying post-hole [G7.1] indicates that it could not have been laid until at least ad 120. Sherds recovered from the overlying occupation material, including the clay and occupation sequence to the south [G10], suggested use of the surface throughout the second half of the second century ad.

Cutting the cobbled surface was a number of features, including a north-west/south-east aligned linear cut, 0.30m deep, infilled with soils containing domestic debris [G9]. The feature could have originally accommodated an installation in use with the cobbled surface, later removed and the hole backfilled with unwanted material. Also cutting the cobbling was a small pit [G5.2] and two post-holes [G12], one of which cut the infilled linear feature. Two further pits, possibly contemporary, lay towards the north [G5.1, G19.1]. Both were infilled with soils but few inclusions, providing little indication of their original function.

Sealing the metallings and occupation deposits, and extending across much of the south-eastern area of the site, were mixed soils and clays [G13.2; not illustrated], the nature of which suggested that direct occupation on the site had ceased. The deposits contained flecks of charcoal and daub, lying in abundance towards the far south-east, as well as common flints and rare pottery and bone. A dolphin type brooch, a spoon probe, a ceramic counter and glass vessel fragments were also recovered from these layers. The mottled, patchy appearance of the deposits suggested that they had not been worked as agricultural soils, but perhaps represented demolition horizons, possibly from Roman timber structures tentatively hinted at by lower clay and occupation deposits but of which no direct evidence remained. The latest elements of the pottery assemblage from these deposits were dated ad 170-250.

No late Roman remains were identified on the site. Similarly, the pottery assemblage provides little evidence of occupation after the middle of the third century ad. An exception was material recovered from a single medieval pit [G17.1 (C129)], potentially dating as late as ad 400.

Anglo-Saxon activity was entirely absent from the site. Whilst remains could have been truncated by later occupation, the lack of any finds of Anglo-Saxon date strongly suggests that this part of Wincheap remained open land until at least the eleventh century.

Cutting the underlying Roman horizons was a small number of Medieval features positioned towards the centre of the site. The pottery recovered suggests that they represented activity occurring possibly as early as the mid eleventh century but certainly from the later twelfth. The features comprised five pits and eight post-holes, the latter forming a cluster towards the east and a single isolated feature further west [G17.1, G18.1, G18.2, G19.2] (Fig. 5). The post-holes could be described as forming an approximate L-shape, and may have represented some sort of boundary or fence line. The pits were infilled with soils containing domestic waste suggesting that they were used, at least at the end of their life, as refuse pits.

Sealing the earlier features was a spread of metalling, comprised of fine flint gravel mixed with small fragments of chalk and peg tile [G24] (Fig. 6). It probably formed a yard surface positioned to the rear of a property fronting Wincheap Street to the south-east. Pottery recovered was dated between the early/mid thirteenth and the mid/late fourteenth century. Later activity, spanning the mid/late fourteenth to sixteenth centuries, was represented by a small number of pits [G17.2, G19.3], a shallow linear feature aligned north-west/south-east [G11] and extensive overlying spreads of mixed soil containing inclusions of pottery, brick, peg tile and domestic debris [G13.3, G25.2; not illustrated]. The features and deposits indicate that the metalled surface had gone out of use but the ground continued to be utilised, perhaps forming a plot associated with nearby dwellings or workshops. More unusual finds, including a child’s finger ring, a quatrefoil mount or tack decorated with a heraldic motif, a scabbard chape and a decorated openwork mount, probably represented casual loss during use of the area.

In the Post-Medieval period, perhaps during the sixteenth century, a building was constructed on the site (Fig. 7), possibly an extension to an earlier property which faced onto Wincheap Street. The structural remains formed an internal area to the east, with an associated cellar behind to the west [G20.1a]. Along the west and south sides of the building was a yard, bounded from an area of open ground to the rear by a further length of wall [G20.1b]. The walls were constructed predominantly of flint and chalk, bonded with pale brown mortar, not coursed but faced with neatly laid (unworked) flints. They stood to a maximum height of 0.65m, the deepest length being that which formed the rear of the cellar.

Only sketchy evidence remained of the internal layout of the building, one or more divisions perhaps represented by two stumps of wall of identical construction to the other structural remains [G20.1c]. A length of trench [G20.2], 0.34m deep and infilled with soils containing common flint pebbles, may have originally accommodated an internal feature, later removed. Nearby lay patches of clay flooring, mixed with occupation spreads of silt and charcoal containing fragments of oyster shell [G13.4, G16.2]. No floor remains survived within the cellar.

The yard bordering two sides of the property was probably used for access to the rear of the building and also as a work area. It was surfaced with flint gravel and fragmented peg tile [G23], overlain by a fine deposit of silt thought to have formed during use. Cutting the yard was a flint-lined well (unexcavated). Beyond, at the rear of the property, the open ground was cut by a drain, formed with side walls of bonded chalk blocks and large flints surrounding a peg tile base [G22]. The feature would have channelled surface water from the metalled surface, through a gap in the yard wall, down towards the west. The gap was later blocked by a single piece of chalk. Also cutting the rear ground was a probable cess tank [G21], somewhat crudely constructed with uneven walls of un-bonded flint, chalk and peg tile. The inner area was capped with a deep (0.48m) deposit of clay, perhaps sealing unpleasant contents (not excavated) below. During the medieval period, cess tanks were commonly positioned against the edge of property plots, and it seems that this feature conformed to usual practice if, as seems likely, the site boundary represents continuity of an earlier division.

Later modifications to the property involved the rebuilding of lengths of the cellar walls and part of the wall dividing the cellar from the internal area of the building [G26]. As part of the process, a diagonal cross wall was built to form a small triangular area to the west, probably used for storage, although insufficient height remained to identify a point of access. The new walls were formed mostly of brick but also included rough blocks of chalk, perhaps re-used from the demolition of earlier walling. Within the back (western) wall of the cellar, two brick fireplaces were added. These modifications were, on the evidence of the bricks, probably no earlier than the seventeenth century.

The building appears to have gone out of use, perhaps some time during the eighteenth century. Demolition deposits, comprising fragmented flint, chalk, mortar and tile within a soil matrix, infilled the lower levels of the cellar and also extended across the south-eastern area of the property [G27, G28; not illustrated]. Fragments of late seventeenth- or early eighteenth-century bottle glass were recovered from these deposits. The footprint of the building, and its associated metalled surface, were cut by a number of pits [G29; not illustrated] from which pottery dated as late as ad 1750-75 was recovered. An iron key retrieved from one of the pits, may also have dated to the eighteenth century.

Further properties were subsequently erected, perhaps during the sec-ond half of the eighteenth century, or possibly in the nineteenth. These properties stood on the site until 1963.

the documentary evidence (Sheila Sweetinburgh)

The development of a number of substantial suburbs from the late Anglo-Saxon period onwards is a particularly interesting aspect of Canterbury’s history. Wincheap is one of these suburbs, its growth linked to the presence of at least two markets in the area and probably its proximity to the Norman castle and ancient Worthgate. Another indicator of the importance of this thoroughfare out of Canterbury is the location of St James’ leper hospital. This mid twelfth-century foundation was sited close to the city boundary of Worthgate Ward, a position which would have allowed its inmates to beg for alms from the many travellers who passed by on this major road between Canterbury and Ashford and beyond. The hospital’s patron was Christ Church Priory, a major landholder in Wincheap, many of whose holdings became the property of the Dean and Chapter of Canterbury after the Dissolution in the mid sixteenth century. However, the Dissolution did provide some opportunities for land-hungry members of the gentry and leading Canterbury citizens, and their mid seventeenth-century successors were even more successful in acquiring church lands, including those in Wincheap. With regard to researching the history of the excavation site, these developments have serious repercussions and it has been extremely difficult to trace landownership over time. For the more recent history census returns and directories have been useful but there remain some gaps, especially for the early nineteenth century.

During the post-Conquest period the area between the city defences, Wincheap Street and the river seems primarily to have been large fields, those nearest the river used as water meadows for the pasturing of cattle and sheep. Some of this farmland may already have belonged to Christ Church Priory, while other plots were in the hands of the archbishop, who granted some of his holdings to William Calvel, the borough reeve, c.1100 (Urry 1967, 386-7). Although it is not clear how the land was divided at this time, by 1200 the priory had apparently cut out a series of plots along the length of Wincheap from these large fields. Unfortunately the priory rentals provide little information about the status of these plots but Urry believed the dimensions indicate that many contained a dwelling on the street frontage with garden ground behind (ibid, 187-9). Among the priory’s tenants on the west side of the road at this time were Robert the priest, the widow Godith, Orbert Pret’s widow, Matthew le Wantier of Chilham, Wimarca and Gleduse Fairhegne. These men and women were in the parish of St Mildred, their neighbours across the road in that of St Mary de Castro.

It is difficult to track their successors in the much less detailed later priory rentals. Yet it seems likely that they too would have had dwellings and (work)shops fronting the street. Some may have traded at the neighbouring markets. In the fourteenth century, the timber market may have attracted men such as the woodmongers John Coupere and Godfrid, and carpenters John Plomer and Henry the carpenter, who lived in St Mildred’s parish (Canterbury Cathedral Archives Library [CCAL]: DCc/ Rental 128; Rental 147). Similarly, these men may have traded at the ‘waegn ceap’ or wain (wagon) market, assuming this rather than wine is the origin of the name ‘Wenchiape’ or ‘Wenchepe’, which dates back to the thirteenth century or possibly earlier (CCAL: Lit MS C20, pp. 74-5). There was a cross nearby, Barnacle Cross, just beyond the city ditch on what became known as Wincheap Green (part of the city fee farm), and a cross house though its early history is uncertain (CCAL: PD5, p. 179; Somner 1977, 81) The fifteenth-century testamentary records are equally frustrating regarding the residents of Wincheap because some testators do mention their messuages, tenements and occasionally workshops such as forges, but they do not indicate where in Wincheap these premises were located. For example, in the late fourteenth century Henry Lincoln had several tenements in Wincheap from which he received rents and a century later Thomas Dere lived in a tenement there with his wife and two children (Centre Kentish Studies [CKS]: PRC 32/1, f. 7; 17/3, f. 140).

For the early modern period the Christ Church Priory rentals and those of the Dean and Chapter provide relatively little information beyond the name of the tenant and the rent to be paid. Consequently, it is impossible to trace those occupying the area under discussion. However, among the landholdings of the Dean and Chapter in St Mildred’s parish in 1567 were several tenements, though others held pasture and one held arable land (CCAL: DCc/Rental 7). This mix of dwellings, workshops, gardens and farmland may have changed little over the next century but this is difficult to prove because the Parliamentary Survey provides negligible information about individual properties, and those it does detail are well outside the designated site (CCAL: DCc/Survey 8; Survey 22). Yet it is worth noting that some properties were of a reasonable size, including one just outside the Wincheap Gate, which was said to consist of a hall, kitchen, parlour, buttery, three chambers, garret, woodhouse and a chamber over the same, possibly a reflection of the area’s prosperity (East Kent Archives [EKA]: U88/T18). Nevertheless, St Mildred’s parish as a whole was con-sidered to be poor and the corporation had constructed a few houses close to the city wall and churchyard to be given to poor people who were willing to nurse and to bury plague victims (CCAL: CC/FA19, f. 184v).

However even if the status of the Canterbury end of Wincheap had remained largely unchanged, its topography had altered. In 1548 access to the castle and the city via Worthgate was stopped and instead traffic passed into the city through Wincheap gate, to the east of Wincheap Green, before rejoining Castle Street (Gosling 1825, 25). This situation remained the same for 250 years at which time the corporation negotiated an exchange of lands with Samuel Balderstone thereby allowing the city to extend Castle Street through the castle grounds, city wall and Worthgate, and over the moat to the north end of Wincheap Street (CCAL: CC/Millens/37/2/A; CC/PD5, 178-9, 283). In part this new carriage road was funded by public subscription and to complete the development Worthgate was demolished in 1791. The ward map of Worthgate, dated c.1792, shows the new layout, and as well as the ‘new road’, Wincheap Street and old way via Wincheap Gate, there is a lane running from Wincheap Green westwards outside the city ditch and almost parallel to the city wall (CCAL: Map 4). At a later date this was known as Wincheap Grove. The map is fairly detailed, indicating buildings, gardens and fields but it does not show individual properties clearly. Nevertheless, it would appear from this map that the designated site was to the south of the garden ground of 1 Wincheap and occupied the area of 2 Wincheap (possibly also part of 1) and the gap in between 2 and 3. Though difficult to be sure, the garden for 1 Wincheap looks to have been larger and more formal than those nearby. Behind the line of gardens are several fields, part of the water meadows illustrated in many paintings of the city made at this time.

Within a few years of the drawing of the c.1792 Worthgate Ward map a large house called The Cedars was built on the northern part of the garden ground of 1 Wincheap, that is on the corner of Wincheap Street and Wincheap Grove (post-dating the land tax and window tax assessments). It is not clear from the surviving records who built the house, nor when it was completed but it is marked on the 1851 tithe map (CCAL: TO/C1/7A; 7B). The land behind The Cedars and that of its neighbour is listed as garden ground in the tithe apportionment, and was said to be owned and occupied by Miss Lasaux, suggesting that it might be part of the estate belong to the house. Even though this may be about 50 years after the construction of The Cedars, members of her family had been living in the area during the same period: Mr De Lasaux lived at Wincheap Green in 1817 and according to the list of freemen and electors a Mr Thomas Thorp De Lesaux resided in Castle Street in 1835 and a Mr Thomas De Lesaux in Wincheap (Kentish Gazette: 23/5/1817; published list of freemen and electors, beginning 1835). This Huguenot family had become extremely wealthy and it is feasible that they owned The Cedars.

It ought to be possible to ascertain the occupier of the property in 1841 from the census returns but the absence of either house names or numbers means the returns are difficult to use (The National Archives [TNA]: Census Returns 1841, St Mildred’s parish). The 1851 returns appear to show that the house was unoccupied and it may have been at about this time that the house was converted into a school (TNA: Census Returns 1851, St Mildred’s parish). In 1861 ‘The Cedars’ was a school for young ladies under the governance of Martha Mason, who employed a French mistress and two music teachers, as well as her daughter (TNA: Census Returns 1861, St Mildred’s parish). Within a few years, however, the school was run by Miss Mary Jenkins. The house had a rateable value of £72 in 1865, its neighbour was valued at £24 and the next two houses at £12 and £20 respectively, an indication of its considerable size compared to those in the vicinity (CCAL: CC/DS1/H1/1/8).

Afterwards The Cedars once again became a private residence, first in the hands of Frederick Share (1878), and then Stephen Horton Williamson, the eldest son of Stephen Williamson, the owner of St Mildred’s Tannery. Stephen junior was persuaded by the family to return to Canterbury to help manage the business in the 1880s, and he lived in Wincheap for almost 30 years before finding a house in St Stephen’s on the other side of the city (Canterbury Directory 1878; Bedwell’s Directory of Canterbury 1888; Pike’s Blue Book 1894; Pike’s Blue Book 1904; Kelly’s Directory 1905; Pike’s Blue Book 1911; Canterbury Directory 1915). Whether the house was empty after he left is unclear, but in 1924 A.J. McNicol, sectional engineer of the P.O. Engineering Department, seems to have been there. By 1930 the Yeoman family had acquired the house as their business premises but following nationalisation in 1949 the family firm became part of British Road Services Ltd, continuing to use The Cedars (part of the building seems initially to have been let to Ragusa Asphalt Paving Co Ltd) and the depot behind until the site was compulsorily purchased for the ring road development in the early 1960s (Kelly’s Directory 1930; Canterbury and District Directory 1935; Kelly’s Directory 1949-50; Kelly’s Directory 1958; Kelly’s Directory 1961; Crampton 1993, 2-3).

Several other buildings were purchased by the council and demolished at the same time, including those on which the new development now stands. Of these, the first to the south of The Cedars was ‘in a sort of transitional Georgian-Regency style’, though it may have been a much older timber-framed house like its neighbour at 2 Wincheap (and presumably on the site of earlier buildings), the third being a three-storey Victorian house which was an infill development, the space clearly shown on the Worthgate Ward map. The next four properties were not purchased by the council until 1975. Thereafter they suffered from neglect, squatters and vandalism as various organisations and the council decided their fate (Kentish Gazette 28/7/1978; 19/1/1979. Kentish Herald 24/10/1978). They were finally saved from demolition by the intervention of Coombs Ltd, a local building firm who undertook a very careful restoration of the properties, including a passageway from the street to the back of the premises (Kentish Gazette 14/11/1980; 25/9/1981).

Of those properties that were demolished in 1963 (1, 2 and 3 Wincheap), the first and largest had been the residence of Robert Fill, a retired wine merchant, and his wife for much of the second half of the nineteenth century before it was apparently briefly owned by George Mundie M.D. and then Major T.H. Jones in the 1890s (TNA: Census Returns 1861, 1871 St Mildred’s parish. Bedwell’s Directory of Canterbury 1888; Pike’s Blue Book 1894). James Dadds, a local builder, acquired the house about the turn of the century but by 1911 Charles and George Yeoman were living there (Pike’s Blue Book 1904, Pike’s Blue Book 1911). The house remained in their family until it was purchased by the council in the early 1960s, George having moved to 3 Wincheap by 1930 and 1 Wincheap eventually sold to the council by Walter Yeoman, who, with his brothers William and Edward, had inherited the place from his father and uncle (CCAL: CC/H169/9/1388/1). The house in between, that is 2 Wincheap, had been the home of Barlett Chambers and his family, and then Miss Brooks in the late nineteenth century; and the Misses Godden in the early twentieth century (TNA: Census Returns 1861, 1871 St Mildred’s parish. Bedwell’s Directory of Canterbury 1888; Pike’s Blue Book 1894; Pike’s Blue Book 1905; Pike’s Blue Book 1911; Pike’s Blue Book 1924; Kelly’s Directory 1930; Canterbury and District Directory 1935). Its last known resident was James Russell who was living there in 1958 (Kelly’s Directory 1958). The home of George Yeoman until his death, 3 Wincheap was said to be ‘new built’ in 1861 and a decade later it was the home of a retired farmer, George Morgan, and his daughter (TNA: Census Returns 1861, 1871 St Mildred’s parish). Thereafter it passed through a succession of hands, its last known occupier being David Morgan in 1958 (Kelly’s Directory 1958).

The demolition of these and a considerable number of other buildings in the area had allowed the council to construct the ring road, the Wincheap roundabout and a subway from the castle to the west side of Wincheap. Habitat had also built its large store nearby but much of the site of 1, 2 and 3 Wincheap was still unused. There were plans in 1985 to develop the site to provide a public open space with seating and landscaping but this was not completed and the area effectively remained undeveloped until the present scheme was undertaken (Canterbury Extra 25/11/1988).

the finds

the pottery

The Prehistoric and Roman pottery (Malcolm Lyne) (Fig. 8)

The site yielded 2,646 sherds (46,812g) of prehistoric and Roman pottery in sixty-three contexts. Of this material 1,287 sherds (24,961g) were either unstratified or residual in post-Roman features. Apart from a tiny amount of prehistoric material from the earliest contexts and a handful of fourth-century sherds from post-Roman ones, the pottery all falls within the period c.ad 43-250.

One pit [G2: C179] yielded a single abraded flake of indeterminate Prehistoric pottery with silt and up to 1.00mm crushed-flint filler: it is probably middle Iron Age in date but could be earlier. The pre-Roman soil sealing this pit [G3] yielded a further four abraded prehistoric sherds, comprising three similar in character to that from Pit C179 and a rim sherd from a combed bead-rim jar of Thompson Class C3 in coarse ‘Belgic’ grog-tempered ware (1982, c.50 bc-ad 50).

Assemblage 1 [from sequence of soils G4: C167, C168, C169]: much of the early Roman pottery was residual in later Roman contexts but the G4 contexts pro-duced a less disturbed assemblage. The 141 sherds (2,028g) have a predom-inance of ‘Belgic’ grog-tempered wares in coarse fabric B2 (63% by sherd count). These sherds include rim fragments from at least three combed storage-jars with everted rims, and three bead-rim jars of Thompson’s Classes B5-5 (late Iron Age- ad 70), C1-4 (late Iron Age- ad 100) and C4 (c. ad 30-100). Romanised Canterbury products are poorly represented by seven jar fragments in sandy grey fabric R5 and eighteen sherds from the following vessel;

1. Collared flagon in rough oxidised sandy fabric R6.1. External rim diameter 70mm. Context 168.

Other wares include eight sherds from a warped jar in Stuppington Lane fabric BER1 (Bennett et al. 1981, fig 6-45, c.ad 50-80), a fragment from a Central Gaulish roughcast beaker (c.ad 50-120) and two sherds from a South Gaulish Samian Dr 18 dish (c.ad 70-90). Nine sherds in North Kent Fineware fabric R16 include a rim fragment from a biconical beaker of uncertain type (c.ad 43-130) and there is a large sherd from a lid-seated jar of Monaghan type 3L0.5 (1987, c.ad 70-150) in white-slipped oxidised Hoo fabric R18.1.

The poor representation by Canterbury products, but their presence in all three deposits suggests deposition soon after ad 70: a BB2 basal sherd from an open form (c.ad 110-250) present in the assemblage from the uppermost dump [C167] is probably intrusive.

Assemblage 2 [from fill of post-hole G7.1: C165]: this feature yielded ten sherds (141g) of largely undiagnostic pottery but including the following;

2. Undecorated bead-rim dish in handmade BB1 fabric. External rim diameter 160mm. Similar to example from Redcliff, Dorset (Lyne 2002, fig. 10-52) but unusually high-fired grey. c.ad 120-160.

The deposit overlying the post-hole [G7.2: C159] produced an assemblage (130 sherds, 1,632g) of residual pre-Flavian pottery with ‘Belgic’ grog-tempered sherds making up more than three-quarters of the material.

Assemblage 3 [from cobbled surface G8: C139]: the eight sherds from this cobbled surface are largely undiagnostic but include three fresh fragments from a biconical beaker of Monaghan class 2G1 in grey North Kent Fineware fabric (1987, c.ad 90-130). This vessel suggests an early second-century date for this metalling.

Assemblage 4 [from the fill of linear feature G9: C142]: this feature yielded ten fresh sherds from the following vessel;

3. Carinated bowl of Monaghan type 4G1.1 in fine grey North Kent Fineware fabric R16. External rim diameter 130mm. c.ad 80-120.

The presence of this vessel suggests that the feature was broadly contemporary with cobbled surface G8.

Assemblage 5 [from clay floor G10: C90]: the twenty-two sherds (119g) from the clay floor span the period c.ad 70-200 and comprise twelve fresh sherds from a beaker in Central Gaulish Black Colour-coat fabric R35 (c.ad 150-200), five sherds from open and closed forms in grey North Kent Fineware fabric R16, an abraded fragment in ‘Belgic’ grog-tempered ware and three sherds from the following two vessels;

4. Moulded neck flagon in hard blue-grey ‘North Gaulish’ Canterbury ware variant fired pink with discontinuous external cream slip. External rim diameter 60mm. Paralleled in form at the Area II kiln at St Stephens Road (Jenkins 1956, fig. 8-7). c.ad 50-80.

5. Necked jar in sandy grey fabric R5. External rim diameter 120mm. c.ad 70-175.

Assemblage 6 [from soil deposit G13.3: C63]: the assemblage, although residual within a medieval context, comprised 228 sherds (3,101g) of Roman material. The most significant component by sherd count is North Kent Fineware: vessels in this fabric include examples of Monaghan’s beaker form 2A5 (c.ad 150-190), jar forms 4A2 (c.ad 110-200) and 4H2.1 (c.ad 70-130) and dish forms 5B4 (c.ad 70-130) and 7A3.2 (c.ad 50-150). ‘Belgic’ grog-tempered wares, represented by sherds from at least three necked-jars (c.ad 70-150), account for a fifth of the sherds and ‘Native Coarse Ware’ (c.ad 170-250) for another 13%.

The thirty-two sherds in sandy grey Canterbury fabric R5 include fragments from the following vessels;

6. Lid-seated jar. External rim diameter 160mm. Paralleled in Burgate Street (Wilson 1987, fig. 114-567). c.ad 140-170.

7. Necked-jar. External rim diameter 140mm. c.ad 70-175.

8. Reeded-rim bowl with undercut rim. External rim diameter 260mm. Paralleled off Canterbury Lane (ibid, fig. 90-209). c.ad 130-170.

9. Reeded-rim bowl. External rim diameter 180mm. Paralleled off Canterbury Lane (ibid, fig. 90-211). c.ad 130-200.

Vessels in other fabrics include two ‘pie-dishes’ of Monaghan class 5C3 (c.ad 150-250), a dish of type 5E3.1 (c.ad 130-230) and a class 5F dish (c.ad 130-230) in BB2 fabric R14, as well as a lid-seated jar of type 3L1.1 (c.ad 150-200) in Thameside greyware fabric LR2.1. Samian sherds include fragments from a Dr.31 platter (c.ad 150-200) and a Dr. 45 mortarium (c.ad 170-200).

Assemblage 7 [from flooring G13.4: C49; and G16.2: C51]: the assemblage, although residual in medieval contexts, comprised 142 sherds (2,101g) and constitutes an assemblage too small for any kind of meaningful quantification, particularly as the material from C49 includes thirty-five sherds from the same vessel. The make-up of the material is similar to that from C63, but includes a large sherd from the following vessel;

10. Hammerhead mortarium in white Rhenish fabric R64. External rim diameter 260mm. Paralleled on many sites, including St Magnus House Quay (Richardson 1986, fig. 1.80). c.ad 170-250.

There is little ceramic evidence for occupation on the site after the mid third century and what there is comes from residual pottery assemblages in post-Roman features. These include two cooking-pot rim fragments in late Roman grog-tempered fabric LR1.1 (c.ad 250/70-400+) from medieval metalling G3: C52 and G24: C71.

Medieval pit G17.1: C129 yielded one fragment each from a beaded and flanged bowl of Lyne type 7A.12 (1994, c.ad 370-400+) and a convex-sided dish of type 7A.16 (c.ad 370-400+) in similar fabric, as well as a girth-cordoned jar sherd in imitative Alice Holt greyware fabric LR5.1 (c.ad 270-400) and two abraded bowl sherds in Oxfordshire Red Colour-coat fabric LR10 (c.ad 240-400). Why all this late Roman material should be concentrated in one pit is a mystery, but one should not overlook the possibility that soil was brought in from elsewhere in Canterbury during the medieval period. If one discounts the late sherds from Pit C129, then there is no firm evidence for occupation on the site much after ad 250.

The post-Roman pottery (Luke Barber)

The excavations produced 880 sherds of post-Roman pottery, weighing just in excess of 19kg, from forty-four individually numbered contexts. The overall assemblage is of variable condition with a great range of sherd sizes, from small (< 30mm across) to large (> 100mm across). Most of the pottery is in good condition with only low to moderate signs of abrasion though this is probably due to the well-fired nature of most of the fabrics. The deposits on site appear to have suffered quite extensive mixing in many places as intrusive material is often low/moderate while residuality appears to be high/very high. Indeed the stratigraphic position of some contexts shows their ceramics assemblages to be totally composed of residual pieces. Although the high degree of residuality may be expected in the twenty-one pottery-producing levelling and metalling layers it is also apparent in the majority of the fifteen pit fills to produce pottery. This residual element consists of both Roman and medieval material. By far the largest two groups on site are from C63 [G13.3] and C60 [G25.2] which produced 121 (2,115g) and 102 (2,930g) sherds respectively. However, these deposits may contain as much as 62% and 57% residual medieval sherds. The majority of contexts produced under twenty sherds, often making ceramic dating difficult. The assemblage has been fully quantified by fabric on an Excel database with an overview summary presented here.

The post-Roman assemblage contains a wide chronological range of material, the earliest of which may be of mid eleventh-/twelfth-century date. Although the High Medieval period is better represented, perhaps the most intense period of domestic refuse disposal is the late medieval/early post-medieval (Transitional) period. Although significant quantities of post-medieval pottery are present the late post-medieval period is, with the exception of a group from an unstratified evaluation context, poorly represented. The assemblage is characterised in Table 1 in order to demonstrate the quantities involved by period. This table is based on the date of the pottery itself, whether intrusive, residual or contemporary with the context in which it was found. The divisions are based on the Canterbury Archaeological Trust post-Roman fabric types (prefixed EM, early medieval; M, medieval; LM, late medieval; PM, post-medieval and LPM, late post-medieval). This gives a fairly accurate breakdown though it should be noted that some fabrics can cross some of the chronological boundaries given.

The degree of residuality at the site precludes describing the assemblages of the different periods using specific well-sealed context groups. As such a much more general approach has been taken giving a chronological overview of the pottery regardless of its context.

The majority of the Early Medieval assemblage consists of small to medium sherds with moderate signs of abrasion. There are few feature sherds to help refine dating in most instances and although some material may be as early as the mid eleventh century all could be of later twelfth-century date. Although small quantities were recovered from a few pits in G17.1, G18.1 and G19.2 much of the assemblage was residual in later deposits. Canterbury Sandy Ware (EM1: 17/234g) cooking pots dominate, though there are a few later twelfth- to early thirteenth-century shell-dusted (EM1.1: 9/195g) cooking pots and bowls and a single bodysherd from a Brittoncourt Farm type (EM1.BCR: 1/5g) rouletted jug [C78, fill of pit C79; G18.1]. Other vessels include cooking pots in shelly (EM2: 11/180g) and sand and shell (EM3: 5/90g) tempered wares. Too little material is present to draw firm conclusions but the material clearly demonstrates some activity in the immediate environs from the later twelfth century at the very latest.

There is a larger group of wares of the High Medieval period, fairly typically dominated by Tyler Hill products (M1: 223/3,794g) (Fig. 9). The majority of these are well fired and probably toward the end of this period (where they merge into LM1, see below). However, some definite early/mid thirteenth-century vessels are represented, most notably the lower two-thirds from an oxidised medium fired sparsely glazed jug in pit C107 [fill C106; G17.1], suggesting continuation from the previous period. A range of M1 cooking pots and bowls, mainly with flat-topped, or slightly concave club rims are present. A single heavily abraded London ware jug handle (M5: 1/55g), belonging to this period was residual in pit C66 [G29] and a green glazed jug bodysherd with pressed-moulded stamped decoration in Flemish Highly Decorated Ware (M14: 1/55g) was recovered from unstratified deposits. Overall it would appear that low levels of activity continued until the later thirteenth century when there was an increasing amount of refuse deposited through to the mid fourteenth century. Some typical vessel forms are given in the catalogue.

1. Bowl with horizontal-topped club rim. M1.1. Dull brown orange throughout with external sooting. Pit C107, fill C106 [G17.1].

2. Lower portion of jug with thumbed base and thin patchy external green glaze on upper body. Mid grey core with brown orange surfaces. M1. Pit C107, fill C106 [G17.1].

3. Cooking pot with lid-seated stabbed rim and applied oblique thumbed strips. Externally sooted. Dull brown core with brown grey surfaces. A well fired vessel of probable mid/later fourteenth-century date. M1. Layer C63 [G13.3]

4. Bowl with horizontal-topped stabbed rim. Mid grey core, dull orange brown surfaces. Exterior sooted. M1. Layer C63 [G13.3].

5. Bowl with wide flat-topped stabbed rim. Spots of internal glaze. Mid grey core, dull orange margins and brown grey surfaces. Of similar type/date to No. 3. M1. Layer C63 [G13.3].

The Late Medieval/Transitional period produced the most pottery certainly sug-gesting refuse disposal was continuing to increase in the later fourteenth and fifteenth centuries (Fig. 10). The pottery is dominated by cooking pots, bowls, jugs and pitchers in hard-fired Late Tyler Hill ware (LM1: 184/2705g). These develop seamlessly from the earlier M1 Tyler Hill products making precise dating difficult. However, their high firing and utilitarian nature (notably lacking decoration) suggest a later fourteenth- to mid/late fifteenth-century date for most.

Occupation appears to have continued without any obvious break throughout the fifteenth and into the middle of the sixteenth century. During this time more refined hard-fired earthenwares begin to dominate, such as Canterbury-type Fine Earthenware (LM2: 81/1930g) and Canterbury Transitional Sandy (LM1.2: 39/865g), as well as a number of more poorly represented wares probably of general Wealden origin (e.g. LM4: 15/540g and LM17A & B: combined 6/85g). These local earthenwares are present in a range of oxidised and reduced jars (often lid-seated), bowls and pitchers, again with very little decoration save internal glazing, usually on the bases, with external glazing being sparse.

This period sees an increase in non-local wares. There are a few Tudor Green (LPM5G: 6/20g) cups represented and a single sherd from a Martincamp flask (LM36: 1/5g), the latter unstratified. Rhenish imports include stoneware from all three of the main producers at this time. A single sherd from a Siegburg jug (LM7: 10g) was recovered from G25.2, C60; three sherds of Langewehe (LM8: 15g) vessel from G25.2, C121 and five sherds of Raeren mug/jug (LM9: 50g) from four separate deposits including C121, G25.2. All in all the imports constitute 2.9% of the late medieval/Transitional assemblage.

6. Bowl with sloping tapering stabbed club rim. Dark grey core, brick red margins and dark grey surfaces. LM1. Layer C64 [G13.3].

7. Bowl with wide concave rim. Spots of external glaze. Mid grey core, dull orange surfaces. LM17B. Layer C64 [G13.3].

8. Jug/pitcher with collared rim. Dull orange throughout. LM17B. Layer C64 [G13.3].

9. Lid-seated jar with thin glaze on internal lid-seating. Dull orange throughout. LM2. Layer C121 [G25.2].

By far the majority of pottery in the Post Medieval period can be placed within the later sixteenth to seventeenth centuries (Fig. 10). The gradual development of the post-medieval redwares (PM1) from the earlier LM1.2 and LM2 local wares makes dating the earlier pieces in this range difficult, particularly without well sealed groups. A wide range of local PM1 (95/3,486g) vessels are represented including jars, pipkins, bowls, plates and chamber pots with internal, or all over, glazes of various shades of red/brown to green. Wealden fine pink-buff earthenwares (PM2.3 & 2.6: 12/665g combined) as well as a few sherds of Calcareous-peppered ware (PM64: 2/30g) are also present. Regional wares include green, but more commonly yellow, glazed Border ware (PM10.1 and 10.2: combined 27/875g). The latter is best represented by a complete bowl from pit C35, fill C32 [G29] similar to Pearce’s No. 52 (1992, 51). Material from London is also represented, including small sherds from tin-glazed ware plates, bowls and jars decorated in a variety of styles (PM9B, PM9P, PM9TB, PM9W: combined 6/45g) and a single sherd from a London stoneware tankard (PM25T: 15g) from pit C39, fill C38 (G29). This, together with two sherds of Staffordshire-type combed slipware press-moulded plate (PM21.3: 30g) and a sprinkling of the PM1 sherds are likely to belong to the early/mid eighteenth century. Imports at this time constitute 5.2% of the post-medieval assemblage by sherd count. They include a single base sherd in Dutch slipware (PM15: 20g) from C231, six sherds from Frechen bottles (PM5: 225g) and a single Westerwald mug sherd (PM6CM: 15g) from unstratified deposits.

G29, consisting of at least six pits mainly of seventeenth-century date, although producing some of the best post-medieval material, demonstrates all too well the level of residuality even in cut contexts. In all G29 produced 126 sherds (4,009g) of which 42.9% can be considered as residual using sherd count. The more fragmented nature of the residual material is notable, as it only constitutes 29.8% of the assemblage by weight. The ‘contemporary’ material is dominated by local earthenwares (PM1, PM2.3, PM 2.6) but London stoneware and tin-glazed ware are apparent (2 sherds) as well as Border ware and Frechen stoneware (4/135g).

(Nos 10-14 all Pit C43, fill C34 (G29), dated mid/later seventeenth century.)

10. Jar with heavy squared club rim. Dull orange with mid/dark grey surfaces. Thin patchy interior glaze. PM1.

11. Base from a tripod pipkin with horizontal looped rod handle. Internal even glaze with patchy thin external glaze. Dull orange throughout.

12. Mug with simple rim and cordons on shoulder. All over dark brown glaze. Dull orange throughout. PM1.

13. Bowl with square club rim. Good internal glaze. Dull orange throughout. PM2.6.

14. Plate with thickened rim. Good internal glaze and dull orange throughout. Some blackening on exterior of rim. PM2.6

The vast majority of the 105 sherds of the Late Post-Medieval period were derived from an evaluation test-pit. This unstratified group consists of seventy-one creamware sherds (LPM11A: 1,887g) from dinner plates, chamber pots, bowls and a tureen as well as twenty-three sherds from early transfer-printed pearlware (LPM12: 97g) tea bowls and saucers with blue Chinese landscape designs. The group probably dates to between 1780/90 and 1800/10. The remainder of the pottery of this period consists of a small background scatter of nineteenth-century material in unstratified deposits or intrusive into earlier contexts.

other finds

The small finds (Lynne Bevan)

The small finds are discussed by material group below. The full catalogue of small finds is held in the archive.

The copper alloy finds included a number of Roman items. These comprised two dolphin brooches, both dating to the first century ad [G13.2, G7.2; catalogue nos 1-2], a spoon probe of possible second-century date [G13.2; catalogue no. 3], a leaded copper alloy fragment, possibly the foot of a stand or casket (unstratified; catalogue no. 4), and three shaft fragments from a pin or implement [G13.3; catalogue no. 5]. Less diagnostic finds comprised a hollow leaded copper alloy fitting [G13.3; catalogue no. 6], five fragments of strip, two fragments of sheet and a fragment of leaded copper alloy strip with a tapering end.

The spoon probe is of particular interest. Such implements were used throughout the Roman period as surgeons’ curettes or sounds, as well as for minor pharmaceutical or toilet purposes, such as extracting cosmetics from containers (Crummy 1983, 60-1). This particularly long example is similar to examples from Colchester, one of which dates to the early second century (Crummy 1983, fig. 65: 1929, 1931, 61).

There was a number of identifiable items of Medieval and Post-Medieval date among the copper alloy small finds, including a sixteenth-century coin (catalogue no. 21) of Charles V (I) of Spain from an unstratified context. Medieval finds included a child’s finger ring [G13.3; catalogue no. 7], with an ovoid bezel decorated with a leaf-shaped motif and punched dots. It is very similar in form and decoration to a ring from London dated to c.1270- c.1350 (Egan and Pritchard 1991, fig. 218:1633, 333-4). A quatrefoil mount or tack [G25.2; catalogue no. 8], decorated with a heraldic motif comprising a shield surmounted with four possible swords on a ground of punched dots, was originally gilded. It may be contemporary with two octofoil-headed mounts from London with similarly robust shanks – interpreted as tacks or nails rather than as dress mounts – dating to c.1330-c.1380 (Egan and Pritchard 1991, fig. 155: 1301, 243). Tacks or nails with similarly ornate heads were possibly used for soft upholstery, since hammering would have damaged their decoration (ibid, fig. 155: 1300, 242).

A scabbard chape [G25.2; catalogue no. 9] with crenellated decoration is similar to other scabbard chapes dated to the thirteenth to fifteenth centuries (Cuddeford 1994, no. 79, 40), and a chape with almost identical cutwork decoration from Sandal Castle has been dated to 1485-c.1600 (Goodall 1983, fig. 1: 77, 232-233). A fourteenth-/fifteenth-century date is probable for the single-looped buckle (unstratified; catalogue no. 10) based upon published parallels (Cuddeford 1994, no. 23, 9-11; Whitehead 1996, no. 57, 19). The buckle is unusual in having the remains of a second loop protruding from the frame (for similar buckles see Egan 1991, no. 472 and a buckle labelled ‘Museum of London Collection’, 102-3).

A simple buckle plate comprising two sheets of copper alloy plate held together with two rivets and two small nails [G31; catalogue no. 11] was probably of post-medieval date, although similar plates were made during the medieval period (e.g. Egan and Pritchard 1991, figs. 72-3, 110-13). A fragment from a skimmer made of copper alloy sheeting, with a flattened, riveted ferrule into which a wooden handle would have fitted, is probably also of post-medieval date [G27; catalogue no. 12]. Although similar to examples from London dating to the early to mid fifteenth century (Egan 1998, fig. 126: 437, 157), an early post-medieval date may be more likely for the skimmer in view of the use of rivets made from rolled sheeting, ‘which seem to be later than the mid-fifteenth century’ (Egan 1998, fig. 126, 155-7).

A belt stiffener with a central ribbed zone [G20.1a; catalogue no. 13] was probably contemporary with a stylistically similar sixteenth-century belt stiffener from Norwich (Margeson 1993, fig. 23: 287, 40-41). Two pins recovered (both unstratified; catalogue nos. 14-15), one with a wire-wound spherical head and the other a flat head, date to the sixteenth to seventeenth century (Margeson 1993, fig. 5: 31-35, 12-13) and the seventeenth to nineteenth century respectively (ibid, fig. 5: 45 and 48, 12-13).

Less datable finds included an openwork mount with curved decoration radiating outwards from a central bar [G13.3; catalogue no.16]. While the mount is fairly similar in style to two early to mid fifteenth-century openwork mounts from London (Egan and Pritchard 1991, fig. 129: 1112-1113, 205-6), openwork mounts or plates are also known from the Roman period, for example from the fort at Great Chesters (Allason-Jones 1996, fig. 11: 44-5, 210). Since no convincing parallels have been found, the dating of this item remains uncertain, although a post-Roman, probably later medieval date, is more likely.

Other undiagnostic items comprised an irregularly-shaped ring, a fitting rather than a finger ring [G25.2; catalogue no.17], a disc-shaped object, possibly a button or stud [G19.2; catalogue no.18] and several fragments of copper alloy plate [G29: ?catalogue No. 2]. The curved shape of the plate fragments, which had become fused to other material, including a stone, corroded iron fragments, part of an oyster shell and some possible hearth lining, was suggestive of the curvature of a vessel. Reconstruction of the fragments was not possible, and although they came from a medieval context they could be of Roman date and therefore residual.

Iron small finds were few in number. They comprised a key with an ovoid handle [G29; catalogue no. 19], a probable item of door furniture [G23; catalogue no. 20], a total of twenty-five nails, and two unidentified objects (see archive). Only the key is datable. Its general shape is suggestive of a post-medieval date, possibly as late as the nineteenth century (Cuddeford 1994, nos 9 and 11, 31-2).

Bone items were well-preserved and in a stable condition. A hairpin (unstratified; catalogue no. 21), with two small incised transverse grooves beneath a conical shaped head, is of Roman date, conforming to Crummy’s Type 2 pins which are believed to date from the ‘?Pre-Flavian period onwards’, with a terminal date of c.ad 200 at Colchester (Crummy 1983, fig. 18, 162, 21).

Two tuning pegs (both unstratified; catalogue nos. 22-3) would have been used for tuning stringed instruments such as harps, lyres, lutes or fiddles (Lawson 1990, 713). Both had squared heads, and the complete example (catalogue no. 22) had an octagonal-sectioned shaft with a drilled perforation at the lower end through which the string would have passed. Similar pegs have been recovered from mid fourteenth- to mid sixteenth-century contexts in London (Egan 1998, fig. 218, 286), although the general date range for such pegs across the country spans from the thirteenth to fifteenth centuries (Lawson 1990, 713).

The highly polished pointed shaft from a bone pin or needle [G29; catalogue no. 24] is not closely datable. However, its presence in a post-medieval context is more suggestive of a post-Roman date, although it may be a residual Roman find.

Two ceramic counters were recovered, both of which were made from pottery sherds chipped into a roughly circular shape [G13.2 and unstratified; catalogue nos. 25-26]. Such objects are common finds on many Roman sites, including sites at Colchester, where the smaller versions have been interpreted as gaming tokens used in board games, a possibility supported by the abraded surfaced noted on some examples (Crummy 1983, fig. 94, 93-95). The larger of the two counters (catalogue no. 25) has a polished appearance, perhaps due to this kind of wear, although the fact that neither of the counters has a ground edge implies that they are both unfinished and may not have been used (Crummy 1983, 94). The coarse greyware fabrics, which almost certainly derived from storage jars or pots, are entirely in keeping with the kinds of utilitarian, domestic pottery fabrics used for production of the Colchester counters (ibid, 94).

The glass (John Shepherd)

Thirty-one fragments of glass were retrieved. Seventeen of these are Roman in date and fourteen date from the seventeenth century to the present. A catalogue of the material is held with the archive. The assemblage as a whole is very fragmentary and little can be deduced in any detail about the supply of glass to this site.

The Roman assemblage is made up of just a few vessels. There is an emphasis upon narrow-necked forms but the group is insufficient in size for this to be of any great significance. All the fragments appear to date from the late first and early second century; there are no diagnostic fragments from the mid first century or from the late second century or later.

An amber-coloured body fragment with vertical optic ribs (catalogue no.1) could come from either a bulbous jug (Isings 1957, form 52b) or a bulbous-bodied jar (Isings 1957, form 67c). Although the upper parts of these two functionally-distinct, contemporary forms are different, the body shape is precisely the same, Isings form 67c/52b. It is most likely, in fact, that these two shapes were the products of the same glasshouses (Price 1978, 74).

Ten body fragments (catalogue nos. 3-12) are much more likely, however, to come from a conical jug form (Isings 1957, form 55). These are contemporary with the bulbous forms described above.

The prismatic bottle (Isings 1957, form 50), represented by catalogue nos.13 and 14, is one of the most common forms of the late first through to the third century. These vessels, sturdily built and designed for easy packaging and transporting, were used as in transit and storage containers for a whole range of liquid and viscous foodstuffs and cosmetics. The base design of no. 13, two concentric circles with a central pellet, is one of the more common types although the design itself is not well cut. The precise function of these designs is still not known for sure, but it is more likely that such a mark was intended to be used for identification of the vessel’s owner rather than its maker. It is possible that such vessels were multi-trip containers, similar to milk and soft drink bottles of the modern period, and part of a much larger network of supply that required the purveyor of their contents to recognise their bottles.

The Post-Medieval fragments, excluding two modern window glass fragments, appear to date from the early eighteenth century. The assemblage is dominated, as is so often the case, by fragments of the common ‘English’ wine bottle, ten of which are present. Although fragmentary they would appear to come from more bulbous forms of the late seventeenth or early eighteenth centuries.

One vessel to be noted is a common large inverted baluster stem of a drinking glass. This, in a heavy lead glass common for the period, is contemporary with the bottle fragments and dates to the first half of the eighteenth century. A generic example is illustrated by Charleston (1984, pl. 32, b - right).

The mammal bone (Robin Bendrey)

276 bone fragments were recovered by hand-excavation, of which 160 (58%) have been identified (Table 2). The bones are generally well preserved. Assemblage size is small and interpretations have been made with caution.

Cattle, sheep and pig are the three most common taxa. Cattle is the most common taxa, followed by sheep/goat, then pig. Small numbers of horse and dog bones were also recovered. A bone of wild boar is the only wild animal present. The dominance of cattle, sheep and pig, with small numbers of horse and dog, is what would be expected in small urban assemblages of these dates. The presence of wild boar in the Roman assemblage indicates that hunting was employed to supplement the food derived from farmed animals. In general, the scarcity of wild animal remains in the Roman period may be testament to the efficiency of the agricultural system (Grant 1989, 144).

In the Roman material, elements from the skull are relatively well represented for cattle and sheep, and to a lesser extent so are the metapodials of these taxa. This could suggest some proportion of primary butchery in the assemblage. Waste from horn working is witnessed by cut and saw marks on a number of cattle horn cores within this material. The butchery evidence indicates other possible activities on-going, including meat removal (from a cattle scapula), and carcass division (a cattle femur is chopped through the caput).

The medieval material produced a small number of cattle long bones that had been split longitudinally, indicative of marrow extraction. A cattle femur had been chopped through the caput, indicating the removal of the hind limb during carcass dismemberment. Also, a pig metatarsal iii has transverse cut marks on the proximal diaphysis, perhaps made during removal of the feet from the carcass or skinning.

The post-medieval material is the smallest assemblage. It includes a couple of sheep/goat skull fragments with butchery evidence to indicate that they were cleaved in a sagittal plane, to access the brain.

Sheep provided the largest number of aged mandibles. In the Roman material two age groups are present (though it is a small sample): there are two ‘subadult’ mandibles, representing animals in their second year, perhaps raised specifically for their meat; and three ‘adult 3’ age group (after Bond and O’Connor 1999, 349) which would have been kept for secondary products, such as wool, milk or breeding, before being culled. Likewise, an adult cattle mandible from a post-medieval context was probably from an animal kept for secondary products. In addition a neonatal sheep femur was recovered from a medieval context and a neonatal cattle tibia from a post-medieval. A Roman cattle distal metacarpal exhibited splayed distal condyles that is probably a work-related change, but may also be age-related.

acknowledgements

The authors would like to thank Mansfield Group Ltd for funding the archaeological works and the architects Lee Evans de Moubray who commissioned the work on behalf of the clients. The illustrations were prepared by Peter Atkinson and Barbara McNee, and the photograph by Andrew Savage. The report was edited by Peter Clark and Jane Elder.

bibliography

Allason-Jones, L., 1996, ‘Museum Note 1996. Roman military and domestic items from Great Chesters’, Archaeologia Aeliana, 5 XXIV, 187-214.

Anderson, I. and Rady, J., 1990, ‘Station Road East’, in Canterbury’s Archaeology 1988-1989, 8-9.

Bennett, P., 1991, ‘Station Road East’, in Canterbury’s Archaeology 1989-1990, 7.

Bennett, P., Macpherson-Grant, N. and Blockley, P., 1981, ‘Four minor sites excavated by the Canterbury Archaeological Trust 1978-1979’, Archaeologia Cantiana, xcvi, 267-89.

Bond, J.N. and O’Connor, T.P., 1999, ‘Bones from medieval deposits at 16-22 Coppergate and other sites in York’, The Archaeology of York, 15 (5), York.

Charleston, R.J., 1984, English Glass and the glass used in England, circa 400-1940, London.

Crampton, P., 1993, Canterbury then and now, Sittingbourne.

Crummy, N., 1983, ‘The Roman small finds from excavations in Colchester 1971-9’, Colchester Archaeological Report, 2, Colchester Archaeological Trust.

Cuddeford, M.J., 1994, Identifying metallic small finds, Anglia Publishing, Ipswich.

Egan, G., 1998, Medieval finds from excavations in London 6: The Medieval household daily living c.1150-c.1450, HMSO, London.

Egan, G. and Pritchard, F., 1991, Medieval finds from excavations in London: 3. Dress accessories c.1150-c.1450, HMSO, London.

Goodall, A.R., 1983, ‘The finds’, in Mayes, P. and Butler, L., Sandal Castle Excavations 1964-1973, Wakefield Historical Publications, 231-60.

Gosling, W., 1825, W., A walk in and about the City of Canterbury: with many observations not … hereto published, Canterbury.

Grant, A., 1989, ‘Animals in Roman Britain’, in M. Todd (ed.), Research in Roman Britain 1960-1989, Britannia Monograph Series 11, 135-46.

Isings, C., 1957, Roman glass from dated finds, Groningen/Djakarta.

Jenkins, F., 1952, ‘Archaeological Notebook. Canterbury, 1949-51’, Archaeologia Cantiana, lxiv, 63.

Jenkins, F., 1956, ‘A Roman tilery and two pottery kilns at Durovernum’, Antiquaries Journal, 36, 40-56.

Lawson, G., 1990, ‘Pieces from stringed instruments’, in Biddle, M. (ed.), Object and Economy in Medieval Winchester, Winchester Studies 7 (2 vols), Oxford, 711-18.

Lyne, M.A.B., 1994 , Late Roman handmade wares in South-East Britain, unpubl. ph.d. thesis, University of Reading.

Lyne, M.A.B., 2002, ‘The Late Iron Age and Romano-British pottery production sites at Redcliff, Arne and Stoborough’, Proceedings of the Dorset Natural History and Archaeological Society, 124, 45-99.

Margeson, S., 1993, Norwich households: The medieval and post-medieval finds from Norwich Survey Excavations 1971-1978, East Anglian Archaeology Report No. 58. The Norwich Survey/Norfolk Museums Service, University of East Anglia.

Monaghan, J., 1987, Upchurch and Thameside Roman pottery. A ceramic typology for northern Kent, first to third centuries A.D., BAR British Series 173.

Pearce, J., 1992, Post-Medieval Pottery in London, 1500-1700, Volume 1: Border Wares, London: HMSO.

Price, J., 1978, ‘Trade in glass’, in Roman shipping and trade: Britain and the Rhine Provinces, by J du Plat Taylor and H. Cleere (eds), CBA Research Report 24, London, 70-8.

Rady, J., 1999, ‘Nos. 7-8 Gordon Road’, in Canterbury’s Archaeology 1996-97, 9.

Richardson, B., 1986, ‘The waterfront group: coarse wares and non-samian fine wares’, in T. Dyson (ed.), The Roman quay at St. Magnus House, London, LAMAS Special Paper 8.

Somner, W., 1977, The Antiquities of Canterbury (first published 1703, second revised edition N. Battely, reprinted with introduction by W. Urry, Wakefield).

Thompson, I., 1982, Grog-tempered ‘Belgic’ pottery of South-eastern England, BAR British Series 108.

Urry, W., 1967, Canterbury under the Angevin kings, London.

Whitehead, R., 1996, Buckles 1250-1800, Greenlight Publishing, Great Britain.

Wilson, M.G., 1987, ‘IIIB: The Pottery’, in Frere, S.S., Bennett, P., Rady, J. and Stow, S., Canterbury Excavations: Intra- and extra-mural sites, 1949-55 and 1980-84, The Archaeology of Canterbury VIII, Maidstone, 195-219.

The Wincheap site during excavation, looking south-east

Plate I

Fig. 1 Site location.

Fig. 2 Prehistoric occupation.

Fig. 3 Roman occupation: 1st century.

Fig. 4 Roman occupation: 2nd-3rd century.

Fig. 5. Medieval occupation: 11th/12th-13th century.

Fig. 6 Medieval occupation: 13th-16th century.

Fig. 7 Post-medieval occupation.

Fig. 8 Roman pottery, Nos 1-10.

Table 1. Characterisation of POST-ROMAN pottery assemblage

Period (centuries)	Quantity/ weight	Average sherd size (g)	No. of different fabric groups	No. of contexts dated to period*
Early medieval Mid 11th - early 13th (EM fabrics)	43 704g	16.4	Local - 5 Regional - 0 Imported - 0	4
High medieval Early/mid 13th - mid/late 14th (M fabrics)	234 4,195g	17.9	Local - 3 Regional - 1 Imported - 1	6
Late medieval/transitional Mid/late 14th - mid 16th (LM fabrics)	344 6,250g	18.2	Local - 8 Regional - 1 Imported - 4	20
Early post-medieval Mid 16th - mid 18th (PM fabrics)	154 5,391g	35.0	Local - 6 Regional - 7 Imported - 3	8
Late post-medieval Mid/late 18th-19th (LPM fabrics)	105 2,568g	24.5	Local - 2 Regional - 5 Imported - 0	1

NB. Totals include all residual/intrusive and unstratified material. ‘Local’ equates to Kentish wares, ‘Regional’ to other English wares. * Excludes unstratified/mixed contexts.

Fig. 9 Medieval pottery, Nos 1-5.

Fig. 10 Late medieval/transitional and post-medieval pottery, Nos 6-9

and 10-14.

Table 2. Distribution of animal bone (number of fragments)

	No.	%	No.	%	No.	%	Total no.
cattle	37	51.4	25	50.0	12	42.9	74
sheep/goat*	22	30.6	12	24.0	13	46.4	47
(sheep)	(2)		(2)		(2)		(6)
pig	13	18.1	13	26.0	3	10.7	30
horse	-		4		1		5
dog	1		-		2		3
wild boar	1		-		-		1
cattle-size	28		28		17		73
sheep-size	20		4		1		25
indeterm.	18		-		1		19
Total	140		86		50		276

* Sheep/goat includes specimens identified to species level.