Excavations at St Martin's House: archaeological investigations in the vicinity of St Martin’s Church, Canterbury

EXCAVATIONS AT ST MARTIN’S HOUSE: ARCHAEOLOGICAL INVESTIGATIONS IN THE VICINITY OF ST MARTIN’S CHURCH, CANTERBURY

christopher sparey-green

St Martin’s Church, a kilometre east of the centre of Canterbury, is well known for its association with the mission of Saint Augustine and the conversion of Anglo-Saxon England to Christianity, its reported use by Queen Bertha identifying it as the earliest recorded church in the country. Archaeological sites in its vicinity are therefore of the highest importance, whether throwing light on its origins or its later development. The site lies below a small spur on the east of the Stour valley overlooking the city (Plate I). The geology is Thanet Beds overlain here by at least 2m of brickearth. On the higher ground immediately to the east this geology is overlain by Woolwich Beds and River Terrace Gravels, being the source of the springs which drain through the churchyard and the grounds of Glebe House, south of the church.

The present investigations preceded the creation of a car park for Canterbury Christ Church University in the garden and orchard north of St Martin’s House (fancifully renamed St Martin’s Priory in the last century) and east of the extended churchyard of St Martin’s (centred at TR 1591 5776) (Figs 1a and 1b). A limited evaluation excavation undertaken in December 1999 identified post-medieval garden trenches but did not extend far enough to reveal the full complexity of the site (Pratt 2000). This became apparent during the watching brief maintained over site reduction in May to September 2001, and limited area excavation was then undertaken with the permission of the contractors, Kier.

The following report describes the result of this latter phase of work. (Contexts from the evaluation are referred to by context number preceded by the letter E). The site is described as if aligned due north and south, the true axis slightly east of north. Data from specialist reports by John Cotter on the pottery, Ian Riddler on the iron objects and bone comb, Robin Bendrey on the animal bones, Rebecca Nicholson on the fish remains and Susan Pringle on the Roman ceramic building materials are summarised in this report; the full reports are held in the site archive.

Prior to excavation and recording, a terrace approximately 65m long north-south by 35m wide and up to 2m deep had been created by the contractors, the long side aligned approximately with the hillside. The formation level for the car park was set on a gentle slope not conforming with the natural slope, resulting in the truncation of features on the eastern uphill side and the retention of intact stratigraphy downhill. The reduction took place in two phases, the features recorded in the early stages observed in plan at approximately 33.60-34.00m aod, later features at a surface approximately 0.3m deeper. Due to the circumstances of the watching brief, not all features were levelled to od.

results of the excavation and watching brief

Five phases of past activity were recognised. These are described below:

Phase 1 the earliest phase of activity was represented by the curving southern edge of a cut running into the northern section of the excavation area [130] (Figs 2 and ³). Only the southern 0.5m of this cut was visible in plan, which was about 2m broad and 0.45m deep where it entered the section, the profile gently sloping to a rounded bottom. This may have been the southern edge of a circular depression around 7m in diameter. Its primary fill of redeposited brickearth (71) was around 0.20m thick, above which were three clearly defined deposits (68-70) of fine grey-brown or grey silt containing rare pebbles and flecks of charcoal. This feature was sealed by an orange-brown clayey silt (67), only visible in the eastern 9m of the section (it had been truncated by machine elsewhere), thickening towards the north-eastern corner of the excavation area where it was around 0.4m thick.

Phase 2 downhill, west of feature 130, was an irregular, trench-like feature [131], which was only partially exposed, but at least 0.8m wide and 0.3m deep with steep sides and a flat base. Its fill could not be differentiated from an extensive overlying deposit of grey-brown clayey silt and frequent sub-angular pebbles with a high moisture content up to 0.9m thick (66), which also overlay the Phase 1 deposit (67). Visible along the northern section, this was traced for 25m along the western side of the site (Fig. 2). Apart from a single medieval sherd the only finds retrieved from the surface of this deposit were two unabraded fragments of Roman tegulae, one in Canterbury fabric 10 the other in fabric 2, its surface bearing a double concentric signature (Harrison 1998). The former at least had been re-used, suggesting deposition in the later Roman or early post-Roman period.

Overlying this deposit was a grey-brown clayey silt (65), 0.20m thick, containing frequent charcoal flecks and occasional gravel and clay lenses. This deposit may have conformed in extent with layer 66, but had been truncated.

In the area of layer 66 the eastern side of a steep sided feature [127] at least 0.2m deep was recorded. Its relationship with layer 66 or the truncated layer 65 could not be determined. It was filled with grey brown silt (126).

Phase 3 Anglo-Saxon burials within the central eastern and south-eastern areas of the site nine sub-rectangular features were recorded cutting the brickearth (Table 1). These features appear to have been graves from their shape and the few finds they contained.

These nine rectangular features ran in an approximate row down the centre of the site, the individual graves varying in alignment between 105 and 150 degrees, with one [87] running roughly north-south at 197 degrees. The more complete features measured between 0.8m and 2.2m long by 0.35m to 0.7m wide and 0.04m to 0.23m deep. The sides of the cuts were generally steep, the bases set at a gradient slightly less than that of the actual hillside. The fills were largely sterile other than a fraction of gravel in the greyish-brownish clay silt, the gravel differentiating this from the natural and suggesting the features had been cut through an overlying deposit similar to 66. Only two graves produced traces of human bone. In Grave 57 fragments of human bone and teeth were identified in the bulk sample residues, whilst in Grave 55 was a pale stain from the decay of the skull at the west end. Grave 55 also produced a single rim sherd of sandy ware, either Mid-Saxon fabric MLS2 or Early Saxon fabric EMS1 of the fifth or sixth century whilst Grave 57 produced a heavily corroded whittle-tang knife of Anglo-Saxon type from the mid-point of the south side of the cut, a position suggestive of a knife placed at the waist as a grave good.

Phase 4 Anglo-Saxon pits a scatter of sixteen pits of varying size and shape was also identified in the same area of the site, the majority of these lying uphill or south of the grave row (Table 2). The only stratigraphic relationship noted between the pits and the graves was at the point where pits 52 and 53 intersected with Graves 55 and 57. The fill of pit 52 and that of Grave 55 could not be distinguished and the relationship between them was uncertain. Grave 57 abutted the southern edge of pit 52 and there was no stratigraphic relationship. Pit 53 cut both graves and the fill of pit 52.

The pits can be divided into two groups; a northern group focussed in the central part of the site (41, 44, 46, 52, 53 and 81, probably an overcut of Grave 79) and a southern group in the south-east corner of the excavation area (61, 73/102, 89, 91, 93/105, 95, 108, 111). West of the grave row were two other potential pits that may be related (125 and E31). The profiles of these pits were truncated by both erosion and ground reduction, and most had a homogeneous clean, silty fill containing limited quantities of occupation material and industrial debris, described below.

The eight pits forming the southern group were more complex in their fills and prolific of finds. Pit 73/102 was a neatly-cut rectangular pit, distinguished from the others by a fill of clayey silt with frequent pebbles, (109) at its base (Fig. 3). A layer of occupation material (101) overlay this, sealed in turn by sterile silts (72) and (100). Pits 93/105, 108 and 95 formed an overlapping group, with the former two lying adjacent and perhaps coeval. Pit 93/105 was an L-shaped cut, separated from the deeper 108 by a low bank of what appeared to be in situ brickearth, forming a ‘rim’ up to 0.50m high around the north-western side of that pit. Pit 108 contained slumped clayey silt layers (114) and (120) within a deep shaft, the lower part of which could not be investigated. The upper deposits (104, 117-119) were common to both pits. The relationship of 108 with 95 was uncertain.

The partially exposed cut E31 was possibly a grave from its alignment but its content of animal bone, charcoal and Roman brick suggested it was a domestic pit of more regular shape (Pratt 2000, 30-31). The relationship between pit 125 and the area of gravel (66) on the western side of the site is uncertain; no finds were recovered from its fill, which was rich in animal bone.

The five pits forming the northern group (ignoring feature 81) were generally smaller, shallower hollows containing less pottery and metalworking finds, though pit 53 produced quantities of oyster shell and animal bone.

The range of finds recovered from these pits by hand collection, supplemented by the result of bulk sample processing, comprised mostly animal bone with some pottery and limited amounts of Roman building material and industrial waste. Pits 61, 73/102, 91, 93/105, 95 and 108 produced approximately 1kg of Anglo-Saxon pottery of the late eighth to ninth centuries (Fig. 4). Pit 61 contained sherds of sandy MLS2 and sandy organic-tempered MLS1 while 73/102 yielded MLS2 ware with bossed decoration and a base in MLS7B. Pit 91 produced parts of at least six vessels, comprising at least two jars/cooking pots in sandy MLS2 (Fig. 4.1), a MLS3 body sherd, jar rims in fabric MLS4D, (Fig. 4.3) and MLS6 (Fig. 4.5) and a worn jar base sherd in Ipswich sandy ware (MLS7A). Pit group 105/108/95 produced a range of finds, the upper fill of Pit 93/105 yielding two sherds of MLS2, including an everted cooking pot/jar rim with heavy internal sooting while 108 contained a cooking pot/jar in MLS6 (Fig. 4.4). Pit 95 produced sherds of MLS2 and Ipswich pimply ware (MLS7B). The lower fill of pit 73/102 produced a cooking pot/jar in MLS2 (Fig. 4.2) and a large body sherd of Ipswich pimply ware (MLS7B). The upper fill produced three sherds of a MLS2 jar with bossed decoration, and the base of an Ipswich MLS7B ware jar, possibly from the same vessel as in Pit 95.

Other finds from these pits included a fragmentary double-sided composite comb of middle Anglo-Saxon type from layer 90 in pit 91 (Fig. 5), an iron whittle-tang knife of Anglo-Saxon type from 73/102, an iron strip and sheet fragments from 93/105 and a nail from 95. The comb comprised two antler connecting plates and six tooth segments with three iron rivets from its central section, the fragments comparable with other examples of Middle Saxon date from Canterbury and other sites in southern England. The connecting plates are shallow and rectangular in section, rather than of the more usual D-shape. Evenly-cut saw marks from the cutting of the teeth are visible on both edges of one connecting plate which bears two bands of four vertical incised lines. The comb thus had a display side, as with other Middle Saxon combs from Canterbury. The form of the connecting plates and single-sided decoration is similar to that of combs from excavations at Christ Church College, Canterbury (Richardson forthcoming). Elsewhere, Frisian combs of eighth-century date and the earliest Anglo-Saxon combs with display sides belong to the late seventh or early eighth century.

The metalwork comprised a whittle-tang knife with rounded back similar to others from Canterbury from Anglo-Saxon and early medieval contexts. The fragmentary iron objects comprised a strip with rivet hole, a plate with a series of flat-headed discs, probably nail heads, on one side and a strap hinge with an oval terminal and central rivet hole. Bulk samples from Pits 73/102 and 93/105 produced significant amounts of burnt clay, slag and debris from the working of copper alloy and from iron smithing. These two also produced faecal concretions.

Pit 93/105 produced two fragments of Roman brick in Canterbury fabric 1, both with double concentric finger marks on their surface. Other fragments of Roman brick and tile came from pits 44 and 46.

Of the 796 fragments of animal bone from both pit groups the greatest quantity, by number of fragments (NISP), occurred in the southern group, in 61 (82 fragments), 73/102 (295 fragments) and 93/105 (136 fragments). In the northern group pit 44 produced 158 fragments; a large number of fragments were also observed in the side of pit 125 but were not retrieved. Treated as a single group, the commonest species were sheep and cattle with smaller quantities of pig, horse (fragments from a single skull) and single fragments of dog and roe deer. The cattle bones were the most frequent by number of fragments and were the main producers of meat from their larger body mass but the higher minimum number of elements for sheep indicate a greater number were culled. The range of elements from the cattle skeleton suggested local butchery but the number of sheep mandibles indicated some differential primary butchery, although most fragments derived from one context. All elements of the cattle carcass were represented, the animals slaughtered beyond the age of a useful life for dairy products, breeding or traction. The sheep mandibles suggested culling while immature or, again, beyond their use for wool or milk production or for breeding. The lack of neonatal bones may suggest this was not a breeding but a consumer site, supplied with poorer quality meat.

Bird bones were recovered from eight of the thirteen bulk samples and seven samples produced a range of fish, the majority from the overlapping pits 52 and 53 with smaller quantities from pits 73/102 and 93/105 to the south. The freshwater fish was mostly eel but there was also one fragment of salmonid, the marine species mostly flatfish, but also herring, gadid sp., with single finds of dogfish/ray, whiting, gurnard and garfish. The range of species was typical of medieval deposits in Canterbury and represented the produce available in the local fishmarket.

No structural features were identified other than four possible post-holes [75, 77, 83 and 85] around 0.3-0.4m in diameter and 0.1-0.15m deep (although post-hole 83 was just 0.07m deep). Three of these post-holes [75, 77 and 85] were in alignment, perhaps representing a fence line running north-south in the area of the potential boundary between the burial area and that used for domestic pit digging.

Phase 5 Post-medieval activity no features could certainly be identified from the late Anglo-Saxon or medieval periods. Features 129 and 134 in the northern section of the site had been cut from the recent garden levels and continued below the floor of the excavation. Both features were circular shafts approximately a metre in diameter, the former lined with brick the latter lined with greenish-grey clay, 133, and was filled with recent debris, 132.

The initial ground reduction had revealed thirty broad, shallow trenches, narrow gullies and pits, features 3-33, in the southern area of the site, these filled with grey-brown sandy silt, chips of brick, pebbles and cinder. These were not planned in detail but two, E21 and E22, recorded in the evaluation are shown on Fig. 2. From their form and content they were probably garden features of the late nineteenth or early twentieth century. There was also one burial of a domestic cat. An almost complete South Italian red figure lekthyos of the Hellenistic period, from a layer of ash and humic soil, 64, at the northern end of the site, was undoubtedly a modern import lost, hidden or discarded for reasons now irrecoverable (Pratt 2000, 25-26). No trace was seen of a Second World War bunker reported to lie in this area; irregularities in the lawn uphill suggest it may remain in situ.

Observation in St Martin’s Lane, close to Glebe House, encountered ground water feeding the small stream running from the garden west into an existing culvert in North Holmes Road. Service trenches in St Martin’s Lane opposite the entrance to Priory Cottages revealed a brick foundation which crossed the trench on an east-west alignment before turning to continue south for at least 20m. This comprised at least eight courses of un-frogged red-brown bricks and Caen stone blocks cemented with yellow sandy mortar over a foundation of ragstone blocks. The corner was buttressed in crimson-red bricks cemented with white mortar, the differences in materials suggesting a seventeenth- or eighteenth-century date for the wall and a later date for the buttress.

discussion

Within the area investigated, four phases of activity from the Roman and Anglo-Saxon periods can be recognised with a fifth phase from the post-medieval and recent gardening activity. No prehistoric features or redeposited finds of this period were noted but early prehistoric activity occurred on the lower slopes, in the area of Canterbury Christ Church University (Hicks forthcoming).

Phase 1: the earliest feature on the present site was the undated hollow [130] at the northern end of the site. The form of the feature, as far as it could be observed, was as a shallow, pond-like hollow, the fill a water-lain silt. This feature is best considered in conjunction with the more extensive deposits of Phase 2.

Phase 2: consisted of the trench-like feature 131 and its possible continuation to the south-west, 127, these perhaps sections of a more substantial cut running approximately along the contour and filled with deposits 65 and 66. These deposits extended uphill to seal feature 130, the observed dampness of the soil suggesting the retention of groundwater from a source nearby. Natural springs are known in Querns Place, close to the 40m contour above the site, so a pond-like hollow or channel could have been supplied from there.

Ground water has been observed in St Martin’s Churchyard immediately downhill and a stream still runs through the grounds of Glebe House to the south-west at approximately 30m aod. This area was known as Conduit Meadow in the seventeenth and eighteenth centuries when it was the origin of a pipeline feeding the city (Rady 1987, 124-7). This source presumably feeds the large pond (the ‘Court Sole’) 250m south-west, in the grounds of the present Barton Court Grammar School. An aquifer on this line could have served as a gravity-fed supply to the Roman town, entering at a relatively high point on the eastern perimeter near Burgate, although no structures have, as yet, been identified in the St Martin’s area which could have been part of an aqueduct, distribution centre (castellum divisorum) or reservoir (Bennett 1989, 127). A section of brick and concrete channel 450m to the north-west has, however, been identified as a Roman conduit with a source in the Scotland Hills, close to the later St Augustine’s Conduit House, 300m to the north (Bennett 1990; Jarman 1997; Hicks forthcoming). The former structure adjoined a large pond-like hollow, similar to the silted hollow [130] seen on the present site (Bennett 1990, 8).

A water source here, in the area of St Martin’s Church, is also likely to have been an important factor in the sanctification of the site as a Christian focus, the water supply allowing use for baptismal purposes, an activity that Bede mentions as having been carried on by Augustine at this church (Colgrave and Mynors 1969, 157). The site’s relationship with this famous church is further discussed below.

The upper levels of these deposits (65 and 66) were truncated, the remainder left in situ unexplored but perhaps originally laid to consolidate damp ground, the Roman tile incorporated in them suggesting a late Roman or early post-Roman date. The few brick fragments did not bear traces of mortar but whether debris from a Roman or later structure is uncertain; the extensive use, or re-use, of Roman brick in the church is clearly relevant to this find. Although a cemetery associated with Roman grave goods of the second or third century is recorded on the higher ground to the east no indication of contemporary burials or funerary structures was noted here, the excavated graves being of a later character (Whiting and Mead 1928).

Phase 3: comprised the nine inhumation burials, set in a row with a range of alignment between 105 and 150 degrees. Skeletal remains were almost entirely absent apart from traces of decayed bone and human tooth crowns and the only accompaniment was the single whittle-tang knife from Grave 57. The human bone showed that at least two of the rectangular cuts had contained burials, the remains consistent with the soil conditions for, although animal bone did survive in the deeper pits of a slightly later date, it was often in a very decayed state and possibly only the final surviving fraction of a larger mass. Human bone laid in a shallow grave in this brickearth subsoil would have preferentially dissolved. The larger cuts were comparable in shape to adult graves while the smallest, 87, might well have held an infant. The shallowness of the features should be noted, implying that others may have been lost through erosion or were unrecognised in the area of 66 and the disturbed ground in the western part of the site. The fills were notable for the lack of occupation debris or Roman tile, material that occurred in significant amounts in the features of Phase 4. None produced closely datable finds or intersected Phase 2 features but the form of the graves and general arrangement is similar to that in early post-Roman cemeteries where extended inhumations are arranged in approximate rows but with frequent variations in alignment.

The presence of a knife in Grave 57, though it could not be dated beyond ‘Anglo-Saxon’ is perhaps suggestive that this grave at least did not post-date the seventh century. The position of the blade in the grave suggests it was deliberately placed as a grave good, a practice that recent research indicates did not continue into the eighth century (Hines and Bayliss 2013, 464-473).

Anglo-Saxon graves of the late sixth or seventh centuries are exemplified at Eccles and Holborough in the Medway valley. At the former site approximately 200 inhumations lay in a close-set group outside the presumably ruined (but re-used?) eastern extension of the main wing of the villa, the graves oriented and mostly without grave goods (Shaw 1994). There was at least one case (J43) of a knife as the only grave good, as here. At Holborough a similar rite was employed but with greater dispersal of the thirty-seven oriented graves and a higher proportion of grave goods, although two each contained little more than a knife and a copper alloy pin (Evison 1956, 118-119, 122-123, graves 6 and 15). Most recently, and at a somewhat earlier date, the Buckland, Dover, cemetery produced several instances of simple graves containing a knife as sole grave good, placed in the region of the pelvis or torso (Parfitt and Anderson 2012, 180-181, graves 321, 330, 386b). The type of whittle-tang knife was one of the most common items at Buckland, whether single or associated with other finds (Riddler 2012, 164-171).

These burials may have lain on the north-east of a larger cemetery close to St Martin’s church, the alignment between 105 and 150 degrees somewhat at variance to the 97 degrees of the earliest building. Their presence here, some 100m east of the church, might imply a seventh-century burial ground of some size, potentially containing thousands of graves, but the lack of grave outlines downhill suggests this was an isolated row. The western side of the site, however, was not available for detailed investigation and graves there may have been missed. The disposition of the graves in an irregular arc is probably accidental rather that reflecting the perimeter of some focus uphill. Early Anglo-Saxon burials may have existed nearby however, since iron spear-heads of Saxon type were found on St Martin’s Hill in the nineteenth century (Brent 1879, 42) and another grave with a spearhead is reported lower down the hillside beside St Pancras church (Sam Moorhead, pers. comm.).

The identification of early graves here is important for the understanding of the origins and early history of the church. The latter question requires detailed analysis of the fabric and will not be addressed here; the earliest structure is at present variously interpreted as a Roman domestic building, fountain or shrine to a water deity, or a mausoleum (Jenkins 1965; Tatton-Brown 1980; 1994, 214-218; Taylor 1998 and most recently by Ward 2004). For the first time, however, early burials have been identified here, close to the reported findspot of a group of eight gold ornaments known as the ‘St Martin’s hoard’, similar to ornaments interred with rich burials of the sixth or seventh century ad. These early graves place the ‘hoard’ in a new light.

Five coin pendants and a ‘medalet’, a composite disc brooch fragment, and a pendant with a cornelian intaglio make up the ‘hoard’, which is not a hoard in the true sense and more likely, according to Haith (1991, 24) to be the contents of more than one female grave. The exact circumstances of their discovery is unknown, but the earliest accounts in 1844 mention a source in the ‘precinct’ or within the ‘churchyard’ of St Martin’s; there is no direct reference to another findspot such as St Augustine’s Abbey (Roach-Smith 1845a; 1845b; pace Haith 1991, 23). Recovery during grave-digging in the vicinity of the church or in an early stage of the excavations for the rebuilding of the chancel and the vestry on the north side in 1844-5 is thus possible (Tatton-Brown 1994, 218). The latter would give a findspot, east of the early core of the church and 70m downhill of the present site. Three of the coins, however, came into the hands of W.H. Rolfe of Sandwich prior to 1844, too early perhaps to be associated with the building work (Roach Smith 1845b, 187 and 190). The other pieces were then given the same findspot. There may be further objects; one, assigned to the St Martin’s hoard, is now held in the Bibliothèque Nationale, Paris.

The coins in the collection comprise two dated groups, one of the late sixth century and the other of the early seventh, suggesting two contexts of slightly different date. Although Bede does not mention the use of St Martin’s for burial, the reference to Liudhard, bishop to Queen Bertha and the presumed officiant during her use of St Martin’s as a private chapel, makes this site an attractive location for the interment of a medallion in his name, locally manufactured as a personal gift to one of the queen’s entourage or to an early convert buried close to the place of worship (Colgrave and Mynors 1969, 26; Grierson 1952, 42-43). The Frankish origin of other pieces would also be consistent with ownership by members of the royal family, interred at a chapel dedicated to the important Gallic saint.

The hoard find also raises questions about the so-called Queen Bertha’s coffin on the north side of the present chancel. The queen is recorded as buried within the porticus of St Peter and Paul’s church on the abbey site but this stone sarcophagus is of simple tapered pattern with an undecorated, slightly coped lid. On opening the coffin in 1883 the base was shown to have been originally furnished with a cut-out for the head, this subsequently bricked up. The coffin contained fragments of bones and other human remains and was partly filled with rubble (Routledge 1883, 57-58). Routledge later assigned it to the restorer of the church in the twelfth century, set within a pseudo-Norman arch in the later north wall of the chancel (Routledge 1897, 11). The date of introduction for simple coffins with cut-outs for the head is uncertain but in the Merovingian world these were in use by the sixth century. A coffin of this pattern might indeed be of a late Roman or sub-Roman origin, contemporary with an early phase of the church, originally set within or just beyond the east end of the structure and subsequently re-used. This coffin and its interior deserve further study.

Phase 4: comprised fifteen or sixteen pits, presumably an element of the known middle Anglo-Saxon settlement. The pits appeared to fall into two groups, a southern group of eight and a northern group of four or six extending across the site. The few datable finds from the southern group belonged to the late eighth to ninth centuries, the occasional faecal concretions and industrial waste suggesting they had served as cess or rubbish pits although a primary use as quarries for brickearth is possible. One pit, 108, may have originally been a well, the adjacent 93/105 giving access or serving as a receptacle for water drawn from the shaft. Here it should be noted that the church contains an early medieval font now identified as a re-used twelfth-century well-head. Rather than material brought from elsewhere this stone could have been set in a well-head in this area (Tatton-Brown 1980). The later fills appeared to be cess deposits sealed by cleaner deposits before the deposition of the ash and metalworking debris in the uppermost fill (112). Pit 73/102 was similar to a grave but the basal deposit (109) of clayey silt was similar to that in known cess pits, the pebbles possibly placed to aid seepage of the contents.

Finds from six of these pits included pottery, a comb and an iron whittle-tang knife of the eighth or ninth century. As already noted, the fragmentary comb was comparable to other Middle Saxon products from Canterbury and sites in southern England, decorated on one side for display. The similarity of the ceramic content of the southern pits suggests that they were a discrete group of Mid Saxon date. There were no obviously residual sherds nor Late Saxon wares and the widest possible date-range for the pits, indicated by the presence of Ipswich ware and Mid-to-Late Saxon Canterbury sandy ware (MLS2), is c.720-875. Evidence from elsewhere in Canterbury, however, suggests that the main currency here of Ipswich ware, and of MLS2, could be from c.750-875. The range of wares and the typological and decorative traits present is similar to that of Mid Saxon pit groups from nearby excavations at St Martin’s Hill and from St Augustine’s Abbey to the west (Macpherson-Grant 1987). Features on the latter sites were judged to date to c.800-870, and the St Martin’s Hill pit to c.800-875, comparable to that for the pit cluster from the present site. The association here of lesser-understood local and/or regional wares (e.g. MLS4D, MLS6) with established type-fossils of Mid-Saxon date provides a useful extension to the known typology and dating of Mid-Saxon wares in Canterbury.

The northern pit group consisted of smaller, shallow scoops containing few finds other than the quantities of animal bones and oyster shell in pits 53 and 125. These pits may have lain on the edge of a settlement uphill and east of the church and its burial zone, the intersecting of the northern pit group and burials occurring on the perhaps poorly-defined interface between the two land-uses. Post-holes 75, 77 and 85, set in an approximate north-south line, did not appear to form part of a building but could have formed part of a boundary delimiting the pit-group on its downhill side but overlying the main group of burials. A boundary here, approximately 80m from the church’s east end, would correspond approximately to the distance between the south side of the church and the northernmost pits in the excavations at Glebe House, south of the church, suggesting the radius of the burial area around it (Rady 1987, fig. 2). The form of the settlement is unknown and, with the limited area of excavation in both areas, no structures have been traced in plan, but any buildings are likely to have lain nearer the line of the Roman road approximating to the course of St Martin’s Hill. The Roman tile and brick recovered from this phase was concentrated in the southern pit complex, the small quantity not indicative of any major structure nearby, the lack of mortar debris again suggesting this was residual, although noteworthy in view of the quantities of Roman tile re-used in the early fabric of the church.

Occupation of this date demonstrates that settlement around St Martin’s Church not only bordered the Richborough road south of the church, but also extended uphill and north-east. Activity of some form extended over a much larger area along the north side of the abbey and west to the line of Broad Street and the edge of the defences. Areas nearer the city, in the grounds of the present Christ Church College, also produced numerous pits rich in ironworking debris (Jarman 1997; Houliston 1999). An industrial zone here might conflict with the image of a religious centre and possible royal palace but this strategically important industry may have been carried out by the monastic foundation under royal control, in an area where the prevailing west wind would have carried pollution from such activity away from the city. That this was an important area for ironworking is corroborated by a charter of ad 689 given by King Oswine to St Peter’s Minster (St Augustine’s, Canterbury) and Abbot Hadrian, granting rights to extract iron ore (Sawyer Charter S12). The dating obtained at St Martin’s complements previous evidence for this settlement and adds to our knowledge of the scale and nature of activity around the church in the Middle Anglo-Saxon period.

This settlement lay within a more extensive wic or trading settlement that extended as far as Fordwich. Whether this wic was comparable to those at Southampton or London, for instance, is doubtful, recent investigations in the Christ Church College area having produced little evidence for exotic, traded items but the range of features and the more common finds are similar (Houliston 1999, 2; Vince 1990; Holder et al. 2000). The ville of St Martin’s appears to have remained in ecclesiastical hands during the later Anglo-Saxon period (Sparks 1980, 21), as recorded in Domesday (DB 2.24; VCH III, 213; Tatton-Brown 1987, 26) and Domesday Monachorum also listed it under the archiepiscopal manor of Westgate (VCH III, 258). How this extra-mural area fared in the time of Viking incursions in the late ninth century is not clear but partial abandonment might be suggested by the absence of tenth- and eleventh-century occupation on the present site and in the area to the south of the church (Rady 1987, 134). Documentary references from the ninth to the eleventh centuries, however, show that, even if settlement was limited, the church continued in importance as a place of worship (Brooks 1984, 34). By the early thirteenth century the western limit of the ville was the eastern boundary wall of St Augustine’s Abbey and by the early fourteenth century the community at St Martin’s had its own school (Sparks and Tatton-Brown 1987, 208).

The present investigations did not reveal any medieval activity, the result perhaps of later erosion or an indication that settlement was then clustered along the roadside. The earliest maps of Canterbury, dating to the sixteenth century, provide little detail but one early perspective map looking east shows the church and buildings fronting onto St Martin’s Hill, set between the precinct of St Augustine’s Abbey and the King’s Park on the high ground beyond (CALC Map 49). Another, from the opposite direction, shows the area between Canterbury and Fordwich with St Martin’s and St Austin’s separated from the King’s Park by the park pale (CALC Map 57).

W. and H. Doidge’s more detailed map of 1752 confirms the presence of the church within an enclosure smaller than the churchyard of the nineteenth century. The range of houses on the road front are shown with an isolated block to the east which is presumably the present building of St Martin’s House, later renamed as ‘Priory’. A building is shown in the north-west corner of the grounds, close to the churchyard, and two smaller blocks are shown to its north. The brick structures observed in St Martin’s Lane may be part of one of these buildings or the garden wall north-west of the main house and Glebe House, as shown on the Doidge plan and a City Terrier of 1792. The area north of the house and north-east of the church, approximating that of the present site, was orchard and is still shown as open ground with scattered trees on the 1872 Ordnance Survey map. The trenches and pits from the fifth and final phase of activity belonged to gardening activities within this area and, of the two wells recorded, one is shown on the Ordnance Survey plan of 1872. Their presence confirms the continued extraction of water from the natural aquifers here until recent times. The churchyard was extended northwards in the Victorian period and again eastward up to the boundary of the present site in 1935.

To return, finally, to the most significant find on this site, the tenuous remains of burials close to this World Heritage Site require this area to receive close monitoring of any further ground disturbance within the present St Martin’s graveyard or its vicinity. The burial here of early Christian converts of the late sixth or early seventh century ad would be of the greatest significance for understanding the early history of the church but until more certain evidence can be recovered for such activity this remains only one interpretation. Future observation might provide some confirmatory evidence for an early cemetery and any work on the church should be directed to testing the extent and nature of the earliest structure, testing the hypothesis of an origin as a Roman funerary monument or hydraulic structure, identifying the earliest conversion of the building and its subsequent development into the present church.

acknowledgements

Sections of this report have drawn on research carried out by Simon Pratt as part of his evaluation work in 1999 and much of the documentary evidence given here is from that source. The author is very grateful to Peter Atkinson for preparing the site drawings; the finds drawings were prepared by Dominique Bacon and Barbara McNee. He would like to thank the contractors, Kier and Co., for providing access to the site during construction of the new car park and also Canterbury Christ Church University for funding the archaeological work.

bibliography

Bennett, P., 1989, ‘Canterbury’, in V.A. Maxfield (ed.), The Saxon Shore: A Handbook , Exeter University Press, 118-129.

Bennett, P., 1990, ‘St Augustine’s Water Supply’, Canterbury’s Archaeology 1988-89, 13.

Brent, J., 1879, Canterbury in the Olden Time, 2nd ed., London.

Brooks, N., 1984, The Early History of the Church of Canterbury, Leicester.

CALC: Cathedral Archives and Library Canterbury.

Colgrave, B. and Mynors, R.A.B. (eds), 1969, Bede’s Ecclesiastical History of the English People, Oxford.

Evison, V.I., 1956, ‘An Anglo-Saxon Cemetery at Holborough, Kent’, Archaeologia Cantiana, lxx, 84-141.

Fletcher, E. and Meates, G.W., 1969, ‘The ruined church of Stone-by-Faversham’, Antiquaries Journal, 49, pt 2, 273-94.

Grierson, P., 1952-4, ‘The Canterbury (St Martin’s) Hoard of Frankish and Anglo-Saxon Coin-Ornaments’, British Numismatic Journal, 27, 39-51.

Haith, C., 1991, ‘The Canterbury St Martin’s “Hoard”’, in L. Webster and J. Backhouse (eds), The Making of England: Anglo-Saxon Art and Culture AD 600-900, London: British Museum, 23-24.

Harrison, L., 1998, ‘The Roman brick and tile’, in A.J. Hicks, ‘Excavations at Each End, Ash, 1992’, Archaeologia Cantiana, cxviii, 91-172.

Hicks, A. forthcoming, Destined to serve? Use of the outer court of St Augustine’s Abbey, Canterbury before, during and after the time of the monks. Canterbury Christ Church Excavations 1983-2007, CAT Occas. Paper series.

Hines, J. and Bayliss, A. (eds), 2013, Anglo-Saxon graves and grave goods of the sixth and seventh centuries AD: A chronological framework, The Society for Medieval Archaeology Monograph, 33, London.

Holder, N., Bowsher, D., Pitt, K. with Blackmore, L., 2000, ‘Across the Saxon Town: three new sites in Lundenwic’, London Archaeologist, 9, 6, 151-59.

Houliston, M., 1999, ‘Christ Church College’, Canterbury’s Archaeology, 1996-1997, 1-4.

Jarman, C., 1997, ‘Christ Church College’, Canterbury’s Archaeology, 1995-1996, 2-4.

Jenkins, F., 1965, ‘St Martin’s Church at Canterbury: a survey of the earliest structural features’, Medieval Archaeology IX, 11-15.

Macpherson-Grant, N., 1987, ‘The Pottery’, in J. Rady, 177-182.

Parfitt, K. and Anderson, T., 2012, Buckland Anglo-Saxon Cemetery, Dover Excavations 1994, The Archaeology of Canterbury New Series, Vol. VI, Canterbury.

Pratt, S., 2000, ‘Archaeological assessment and evaluation in the grounds of St Martin’s Priory, Canterbury’, unpubl. archive report, CAT.

Rady, J., 1987, `Excavations at St Martin’s Hill, Canterbury 1984-85’, Archaeologia Cantiana, civ, 123-218.

Richardson, A., forthcoming, ‘The small finds’, in A. Hicks.

Riddler, I., 2012, ‘Knives’, in K. Parfitt and T. Anderson, 164-171.

Roach-Smith, C., 1845a, ‘Account of the meeting on April 25, 1844, Proceedings of the Numismatic Society’, Numismatic Chronicle, VII, 27-28.

Roach-Smith, C., 1845b, `Merovingian coins etc., discovered at St Martin’s, near Canterbury’, Numismatic Chronicle, VII, 187-191.

Routledge, C.F., 1883, ‘Notes on discoveries at St Martin’s, Canterbury’, Archaeologia Cantiana, xv, 56-58.

Routledge, C.F., 1895, ‘The Baptism of King Ethelbert’, Arch. Cant., xxi, 157-160.

Routledge, C.F., 1897, ‘St Martin’s Church, Canterbury’, Arch. Cant., xxii, 1-28.

Shaw, R., 1994, ‘The Anglo-Saxon Cemetery at Eccles: a Preliminary Report’, Archaeologia Cantiana, cxiv, 165-88.

Sparks, M., 1980, ‘The Medieval Pattern’, in M. Sparks (ed.), The Parish of St Martin and St Paul, Canterbury: Historical Essays in Memory of James Hobbs, Canterbury: Friends of St Martin’s, 21-27.

Sparks, M. and Tatton-Brown, T., 1987, ‘The history of the Ville of St Martin’s, Canterbury’, in J. Rady, 200-213.

Tatton-Brown, T., 1980, ‘The Font in St Martin’s Church’, in M. Sparks (ed.), 19-20.

Tatton-Brown, T., 1987, Canterbury in Domesday Book, CAT.

Tatton-Brown, T., 1994, ‘Kent Churches – Some New Architectural Notes (contd.); Churches in and Around Canterbury’, Archaeologia Cantiana, cxiv, 189-235.

Taylor, M., 1998, ‘Two Relics of English Christianity before the Arrival of St Augustine’, Kent Archaeological Review, 132, 32-38.

Vince, A., 1990, Saxon London: an archaeological investigation, London: Seaby.

Ward, A., 2004, ‘Church Archaeology 410 to 597: the Problems of Continuity’, Archaeologia Cantiana, cxxiv, 375-395.

Whiting, W. and Mead, H.T., 1928, ‘A Roman cemetery at St Martin’s Hill, Canterbury’, Archaeologia Cantiana, xl, 67-78.

Plate I

View looking west across site with feature 46 under excavation and St Martin’s Church and the cathedral beyond. Scale 1m.

Fig. 1 St Martin’s House: site location. Scales a (above): 1:8000, b (below): 1:800.

Fig. 2 St Martin’s House: site plan. Scale 1:250.

Fig. 3 St Martin’s House: plan and sections.

Table 1. List of Graves or Grave-like Features

Feature no.	Length (m)	Width (m)	Depth (m)	Fill no.	Fill	Human remains	Other finds	Alignment (degrees)
48	0.80+	0.40?		49	clay silt		post-med pot?	150
55	0.8	0.35	0.17	54	clay silt gravel	trace	Anglo-Saxon pot?	130
57	1.7	0.7	0.23	56	silt gravel	tooth crowns	Anglo-Saxon knife, post-med brick	122
59	2.2	0.6	0.3	58	silt gravel			125
79	1.27	0.4	0.04	78	silt gravel			106
87	0.9	0.42	0.09	86				197
97	1.10+	0.5	0.1	96	silt clay gravel			105
99	0.60+	0.5	0.2	98	silt clay gravel			105
E31	1.0+	0.75	0.56	E8	clay silt		animal bone, chalk, Roman CBM	105

Table 2. Catalogue of pits and other features

Feature no.	Length (m)	Width (m)	Depth (m)	Shape	Fill no.	Fill	Inclusions
Pits
41	0.79	0.54	0.08	Oval	40	Sand Silt
44	1.4	1.2	0.38	Oval	43	Sand Silt	Animal bone, chalk, Roman CBM
46	1.2	1.1	0.3	Oval	45	Sand Silt	Roman CBM
52	2	1.8	0.15	Sub-rec.	51	Silt	Animal bone, chalk
53	1.3	0.8	0.15	Oval	49, 50	Clay Silt	Animal bone, chalk, oyster
61	1.85	1.57	0.81	Oval	60	Sand Silt	Animal bone, chalk, burnt clay, Anglo-Saxon pottery
73/102	1.4+	0.8	0.65	Sub- Rec.	72, 100, 101, 109	Sand Silt	Animal bone, chalk, burnt clay, pot, Roman CBM, iron knife, slag
81	0.66	0.3	0.06	Circ.	80		?Overcut grave 79
89	0.9	0.46	0.09	Oval	88	Sand Silt
91	1.5	1	0.38	Oval	90	Sand Clay	Chalk, Roman CBM, bone comb
93/105	1.68	1.48	1.1	Sub-rectangular	92, 103-4, 112, 113, 115-9	Sand Silt/ Clay Silt	Animal bone, chalk, burnt clay, pot, Roman CBM, iron object, copper alloy object, slag
95	1.4	0.9	0.80+	Oval	94	Sand Silt/ Clay Silt	Animal bone, chalk, burnt clay, pot, slag
108	1.5	1.2	0.90+	Oval	114, 120–121	Sand Silt/Clay Silt	Animal bone, chalk, pot
111	1.25	0.50+	0.7	?	110	Sand Silt	Chalk
125	1.3+	0.3	1.5	?	122–124	Sand Silt	Animal bone, chalk, oyster
Linear features
E21	1.5+	0.2	0.1	Sub-rec.	4	Clay Silt	chalk, burnt clay, pot, oyster
E22	0.3+	0.2	0.1	Sub-rec.	6	Clay Silt	chalk, burnt clay, oyster
Post-holes/small pits
75		0.4	0.15	Circ.	74	Sand Silt
77		0.38	0.12	Circ.	76	Sand Silt	Roman CBM
83		0.3	0.07	Circ.	82	Sand Silt
85		0.26	0.1	Circ.	84	Silt	Post-med pottery

Table 2 (cont.). Catalogue of pits and other features

Fig. 4 St Martin’s House: Mid to Late Saxon pottery. Scale 1:4.

(Drawn by Dominique Bacon.)

1. (Pit 91, context 90) MLS2. Globular jar with simple, slightly thickened everted rim and rounded, slightly flattened base. The rim and neck bear construction marks, the body elongated diagonal impressions and traces of faint vertical knife-trimming both internally and externally. Sooted externally.

2. (Pit 102, context 101) MLS2. Globular jar with curved neck and beaded/ thickened everted rim and flattened, slightly sagging base. Fairly crude manufacture with external vertical knife-trimming, internal ‘fluting’ in the neck/shoulder area and grooves/traces of coil building below. Some external and internal sooting.

3. (Pit 91, context 90) MLS4D. The everted flanged/beaded rim of a jar with a diameter of c.150-200mm. The internal angle is a common feature of other local wares of this date. The rim apex is sooted.

4. (Pit 108, context 114) MLS6. A jar with wide, squat body and simple everted, slightly beaded rim and sagging base. Irregular oblique burnishing on rim and external and internal surfaces. Sooted externally.

5. (Pit 91, context 90) MLS6. The everted rim of a jar with a diameter of c.200-250mm. Trace of sooting externally.

Fig. 5 St Martin’s House: bone comb. Scale 1:1. (Drawn by Barbara McNee.)