The Earliest English Church? A Reconsideration of the Chapel of St Pancras at St Augustine’s Abbey, Canterbury

Abstract

This paper presents an archaeological reconsideration of the chapel of St Pancras at St Augustine’s Abbey, Canterbury, using published and archival sources from over a century of excavation and archaeological survey. The evidence considered, including previously unpublished elevations and geophysical survey, sheds new light on the chapel’s structure and its context, of which it presents the fullest account yet published. Using all available sources, it is argued that the earliest phase of the ruinous building neither dates from the Roman period, nor from the mid- to late 7th century, as previously argued by other scholars, but from between 597 and 609. The same evidence supports the interpretation that the structure was rebuilt in the 7th and 8th centuries. As such, the first phase of St Pancras may be both the oldest ‘Anglo-Saxon’ ecclesiastical building visible today, and founded, consecrated and used by Saint Augustine of Canterbury himself. It provides, therefore, unique testimony to a pivotal moment in British history and offers previously unrecognized material evidence for the foundation of the English Church.

Keywords:

• ‘Anglo-Saxon’
• church
• St Augustine’s Abbey
• Canterbury

Saint Augustine’s Abbey is located just to the east of Canterbury’s Roman town wall, between the modern streets of Longport on its south, Monastery Street on its west and Havelock Street and North Holmes Road on its north. The complex, partly excavated in sporadic seasons throughout the 20th century, is today a popular English Heritage tourist attraction. It forms part of the Canterbury World Heritage site, along with Canterbury Cathedral and nearby St Martin’s church. The abbey was founded by St Augustine, probably at some point between the return of its first abbot, Peter, from Rome in 601/2 and the death of St Augustine in 604.¹ As the first monastery established in an ‘Anglo-Saxon’ kingdom it was central to the introduction of Christianity to the ‘Anglo-Saxon’ kingdoms in the 7th century.² The later medieval abbey is well understood in both archaeological and architectural terms, but its ‘Anglo-Saxon’ predecessor is far less so, current knowledge about the complex largely deriving from pre-Second World War excavations and limited written sources.³

One of the first structures of St Augustine’s original complex was the church of Sts Peter and Paul (hereafter, Sts Peter and Paul), probably begun in 609, judging from textual references, but consecrated by Augustine’s immediate successor, Laurence.⁴ Sts Peter and Paul was the most important of a group of three churches forming the ‘Anglo-Saxon’ monastery’s liturgical core. The others were the church of St Mary — axially aligned with Sts Peter and Paul and just to its east — and then, on a different alignment, St Pancras, the principal subject of this paper (Fig. 1). Perhaps because it is by far the best preserved of the three churches, St Pancras is frequently cited in studies of ‘Anglo-Saxon’ architecture and history, as crucial — along with Sts Peter and Paul — to understanding both the Augustinian mission and the origins of English ecclesiastical architecture.⁵ However, the dating and interpretation of its first two constructional phases — those relevant to such questions — have been disputed among scholars for over a century.

Fig. 1. A general view of the chapel of St Pancras at St Augustine’s Abbey, Canterbury

© The author

Published discussions of what took place at the site immediately before the foundation of the monastery are also, at best, brief and general.⁶ Consequently, far less is known about the origins of St Augustine’s Abbey and its immediate physical context than might be assumed for such a famous monastery. This has wide-ranging implications for the broader questions concerning which the site is often used as evidence. Before discussing St Pancras, this paper therefore examines what was at the future site of St Augustine’s Abbey immediately before the foundation of the monastery, and its relationship with the church of St Martin c. 250 m to its east.

St Martin’s church

Bede, writing in 731, despite drawing on the authoritative testimony of the then-abbot of St Augustine’s, tells us nothing about what was on the site used for the abbey. But he does claim in his Historia Ecclesiastica gentis anglorum that nearby St Martin’s church (Fig. 2) was constructed in the Roman period and used by the Frankish queen of Kent, Bertha, from the time she arrived to marry King Aethelberht, until the arrival of the papal missionaries led by Augustine in 597.⁷ This is all that Bede, or any other near-contemporary textual source, has to say about what was immediately outside of the Roman walls of Canterbury before the foundation of the monastery. It is usually assumed that the current site of St Martin’s parish church was where Bertha worshipped, according to Bede, but Charles Thomas argued that it was instead St Pancras.⁸ This immediately highlights the significance of understanding the date and character of the earliest building at St Martin’s for St Pancras, but as we shall see, St Martin’s may also be relevant in other ways to understanding the construction history of St Pancras.

Fig. 2. The present church of St Martin’s, Canterbury

© The author

It might be hoped that archaeology could resolve the question of whether Bertha’s church was dedicated to St Martin, by showing that one or the other of these locations had no building that could even have been used as a church in the late 6th century. While this might anyway be difficult to demonstrate without excavation on a scale impossible at St Martin’s, the question is rendered more difficult to resolve by the complicated structural history of the standing building today used as St Martin’s church.

The second phase (Phase 2) of St Martin’s, as known through structural analysis of the standing fabric of the church and limited excavation, is itself ‘Anglo-Saxon’ in date.⁹ But debate has continued since the 19th century over whether the earliest phase of St Martin’s — a rectilinear, almost square, brick-built structure measuring 5.2 m × 4.3 m, with small corner buttresses and an opus signinum floor — is Roman period or post-Roman.¹⁰ It had a square-headed doorway (1.82 m high × 1.01 m wide) with door jambs incorporating ragstone internally, and a green limestone lintel, although another — round-headed — doorway was later inserted into it. There was a small subsidiary room on its south.¹¹

The limited amount of excavation at St Martin’s, in 1896 and 1954,¹² has made using excavated evidence to date and interpret the first phase (Phase 1) structure problematical.¹³ This has led to a wide range of dates and interpretations for the structure being proposed. For example, although Paul Bennett argued for a date for its initial construction of c. 580, and saw it as a chapel built for Bertha and those accompanying her from Frankia, Alec Detsicas interpreted the first phase of the structure as a possible Roman mausoleum, and Tim Tatton-Brown dated it to the Late Roman period, with Phase 2 belonging to the 7th century.¹⁴

Tatton-Brown’s dating of the first phase of St Martin’s was accepted by Kevin Blockley in his unpublished MPhil thesis.¹⁵ Blockley also considered it a Roman mausoleum, and argued that in the late 6th century it was used by Bertha’s Frankish entourage as a church, prior to the Augustinian mission.¹⁶ Harold and Joan Taylor, in their classic three-volume study of ‘Anglo-Saxon’ architecture, considered it ‘sub-Roman or early Anglo-Saxon rather than of Roman construction’, but nonetheless ‘part of the fabric of the church in which Queen Bertha worshipped before the coming of Augustine’.¹⁷ That is, opinions range from the Phase 1 structure being a Roman or sub-Roman mausoleum to a late-6th-century — in effect Merovingian — church built for Bertha and her retinue.

The suggestion that it was built as a Roman mausoleum is supported by the similarity of its plan to other Roman mausolea in Kent, such as Stone-by-Faversham.¹⁸ If it was a mausoleum perhaps it was prominently sited on the hill to enhance its visibility, and, therefore, monumentality. A Late Roman date might be supported by the presence of reused Roman brick, although the majority of the brick had no evidence of reuse.¹⁹

Interpreting the structure as a Late Roman mausoleum leaves open the question of whether it already had Christian associations in the 4th or 5th century. It might have been a pagan or even wholly secular structure, but the mausoleum itself could have been associated with Christian burial, as were the mausolea at Poundbury outside Dorchester in Dorset.²⁰

Nevertheless, the discovery at the site of a coin, reused as a pendant, with an inscription to Liudhard, Bertha’s chaplain, and four other coin-pendants, probably from at least two high-status graves, strengthens the idea that it was the place mentioned by Bede.²¹ Whatever the date of the Phase 1 structure at St Martin’s it was probably used as their church by Bertha and other Christians in the Kentish court prior to 597. If so, its small size suggests that they were very few in number. This interpretation may be supported by the fact that the structure attracted further ‘Anglo-Saxon’ burials, as is shown by nine burials just to its east, perhaps from the late 6th to 7th century.²² It was also remodelled in the 7th century into a structure more confidently identifiable as a church, its walls of reused Roman brick in pink mortar.²³

The most plausible interpretation of the first phase of St Martin’s is, therefore, that it was a Late Roman mausoleum, employed in the late 6th century as a church for Bertha and the few other Christians in the court before the Mission. The surrounding area is devoid of evidence for Roman-period settlement or even burial,²⁴ so it is most likely that this mausoleum stood in open countryside, located on St Martin’s Hill because of its visibility rather than because of an earlier cemetery.

the site of st augustine’s abbey before 597

Unlike the area around St Martin’s, the zone around St Augustine’s Abbey was used for Roman-period inhumation and cremation burials, the ‘East B’ cemetery.²⁵ Late Roman inhumation burials are reported on the west of the St Augustine’s Abbey site, near the 7th-century church of Sts Peter and Paul.²⁶ In addition, where excavation reached the underlying geology of light brown brickearth overlying hard grey gravel that is actually below Sts Peter and Paul, this was directly overlaid by Late Roman tip deposits.²⁷ A gully (c. 0.28 m deep × at least c. 0.53 m wide), possibly a foundation-trench, which may also belong to the Late Roman period, is the only feature which could even be structural. If so, whether this represents a building or another structure, and whether it was in a field, cemetery or settlement, is uncertain. Likewise, although much unstratified Late Roman material — including pottery and coins — has been found elsewhere at the site, this may be no more than the consequence of tipping, such as that found at Sts Peter and Paul. That is, there is no confirmed Late Roman occupation anywhere within the area contained within the Abbey precinct.

The nature of the Roman deposits beneath St Pancras in particular was identified by Blockley, based on Frank Jenkins’ excavation records.²⁸ The only Roman features beneath, or immediately adjacent, to the later church were pits and gullies, cutting ambiguously termed ‘occupation soil’ containing artefacts from the 1st to 4th centuries. What may be an east–west Roman road was found in Jenkins’ 45m-long trench south of the church, but again there was no trace of either burial or settlement during the Roman period. Across the whole excavated area within and surrounding St Pancras the latest dated Roman-period features were the latest of the pits, for one of which there is a radiocarbon date of 248–564 at 95.4% probability.²⁹

The interpretation that the area later encompassed by the precinct of St Augustine’s was devoid of Roman buildings is supported by a geophysical survey of the whole of the English Heritage site, undertaken on behalf of English Heritage in 2009 by Claire Graham and Richard Fleming for Stratascan, and — with the permission of English Heritage — published here for the first time (Fig. 3). Although Graham and Fleming identified some possible rectilinear structures, none of these features need be pre-medieval in date. A rectilinear anomaly south of the eastern part of the medieval nave and west of its south transept is probably the Kent and Canterbury Hospital. Founded in 1793, this was rebuilt successively and is visible in this location on the OS 25-inch map revised in 1906. It was moved to a different site in 1937, and the remaining buildings were demolished.³⁰

Fig. 3. Resistance survey of St Augustine’s Abbey

© Historic England. Geophysical survey data conducted by Stratascan for English Heritage

Consequently, the geophysical survey offers nothing to contradict the impression from excavation that in the 4th to 6th centuries the site of later St Augustine’s Abbey was open land, possibly traversed by an east–west Roman road. There was an extensive Roman-period inhumation cemetery to the north and west of the site, but most of the land later used as the monastic precinct was probably outside of its burial area and used for rubbish tips.

It has long been known that pre-7th-century ‘Anglo-Saxon’ burials also occurred within the Roman cemetery area east of the town walls in general, and these have been also reported just south-east of St Pancras.³¹ However, in 2019 the grave of a woman in her twenties was found during rescue excavation just north of St Pancras church.³² This burial was accompanied by a silver, garnet-inlaid, Kentish disc brooch, a necklace of amber and glass beads, a copper alloy belt buckle, a copper alloy bracelet and an iron knife. These artefacts suggest a high-status burial dating between 580 and 600 — exactly in the period in which the Merovingian entourage of Bertha, and then the Augustinian mission, arrived in Canterbury.³³ It is uncertain at present whether this was an isolated burial or part of a larger cemetery.

The religious identity of the woman is also unknown. Although it is often assumed in British archaeology that accompanied burials must have been pagan, contemporary Merovingian Frankish Christians were buried clothed and with accompanying artefacts, and there is no reason to correlate furnished inhumations with pagan burials in this period.³⁴ Given the Frankish contacts of late-6th-century Kent, and the uncertainty over whether the burial dates to before or after the Augustinian mission of 597, it is equally possible from present evidence that this woman was a Christian or pagan.³⁵ This, of course, highlights the question of the date of the nearby church of St Pancras.

the chapel of st pancras

The Chapel of St Pancras lies in the east of the St Augustine’s Abbey site, where it is today a highly visible ruin (Fig. 1). The building has a nave 12.97 m long × 8.11 m wide, with an arcaded wall dividing this nave from an eastern apse.³⁶ The apse has a small rectilinear room (porticus) on its south, and wall stubs which may be traces of another to the north, with two rectilinear porticus on either side of the nave. There is also a rectilinear western porch standing 3.35 m high, with a western doorway 1.93 m wide and 2.35 m high into the nave.³⁷ The porch, like the porticus, has a straight joint between it and the earliest phase of the nave wall (Period 1). Since the work of Routledge and Hope, the two earliest excavators of the chapel (see below), these have been understood as additions to the building, an interpretation confirmed by excavation (Period 2).³⁸ The standing walls of the chancel and chancel arch are later medieval in date, differing in construction from pre-Norman walls at the site by their use of stone rather than exclusively Roman brick.³⁹

The ruin was partly reconstructed by the Ministry of Works in 1962, when the upper parts of the Period 1 walls (the uppermost c. 0.3 m in the nave according to Blockley), were rebuilt using white mortar, but it is possible to identify these reconstructions in the present structure using earlier records and data from the 1970s excavations at the site.⁴⁰ The latter task was largely accomplished by Blockley in his Master’s thesis (2000), although — as will be seen later — his work requires some revisions in both its description of the structures and their dating.⁴¹

There are two additional sources of information about St Pancras. The first consists of photogrammetric elevations of the standing walls of the building, created by Caroline Atkins for English Heritage in 1998. These are published here for the first time, with permission of English Heritage (Figs 4–14). When looking at these elevations it must be remembered that they were taken after the 1962 reconstruction of the upper few centimetres of the nave walls, so the uppermost courses shown in the nave are 20th-century rebuilding rather than earlier walls. For the first two phases of the structure, those considered here, Atkins’ elevations confirm the existence of the straight joints mentioned below and the remarkably uniform character of the brick walls belonging to those phases.

Fig 4. St Pancras Chapel, chancel, east wall. Elevations of the chapel of St Pancras. Figures 4–7 and 12–14 are published here in their original form and with their original captions to render them of maximum use to other scholars. Figures 8–11 have had to be cropped and modified due to practicalities of publication, but the elevations depicted are unaltered. The complete original versions of all the drawings are available online

© Historic England

Fig. 5. St Pancras Chapel, chancel, south wall.

Fig. 6. St Pancras Chapel, chancel, south wall and south-east buttress.

Fig. 7. St Pancras Chapel, chancel, north wall.

Fig. 8. St Pancras Chapel, north wall and north-east buttress (stipple indicates plaster base/render).

Fig. 9. St Pancras Chapel, north wall of nave.

Fig. 10. St Pancras Chapel, south wall of nave.

Fig. 11. St Pancras Chapel, south chapel (south porticus).

Fig. 12. St Pancras Chapel, nave, west wall.

Fig. 13. St Pancras Chapel, northern addition.

Fig. 14. St Pancras Chapel, nave, south-east buttress.

The second new source of information is detailed recording and photographic record of the surviving structure by the author in 2018–19. It should also be noted that, while visible and accessible since the 20th century, the recording undertaken by Atkins in 1998 and the author in 2018–19 were the only two previous occasions on which any sort of detailed archaeological records of the whole structure, including the elevations of its walls, are known to have been made. The significance of both Atkins’ photogrammetric survey and the 2018–19 recording, apart from creating a better record, is that they confirm the structural observations made by earlier scholars. That is, in so far as they can be evaluated without further excavation, the 20th-century descriptions of what was found during excavation may be considered reliable.

Consequently, the most important source of information for understanding the earliest structures at St Pancras remain the series of excavations at the site in the 19th and 20th centuries. Excavations took place in 1881, 1900, 1972 and 1974–75, but none has been published in more than the most summary form.⁴² The records of most of this work were collated and examined by Blockley in his Master’s thesis. The earliest excavation was by Charles Routledge, published in 1882, with subsequent work in 1900–01 by St John Hope, published in 1902, and the latest was work by Jenkins in the 1970s — a 45 m-long trench to the south of the chancel, mentioned above, and twenty-one separate trenches within the church itself, published in brief by him in 1976.⁴³

Jenkins’ records were formed into a structural narrative first by the excavator and later, in much more detail, by Blockley, who also collated and studied all the site archive for the 1970s excavation at St Pancras as this existed in the 1990s.⁴⁴ Jenkins identified four structural phases (Periods 1–4, with Period 1 the earliest), and this terminology for the building will be used here.⁴⁵

Combining these records, it is possible to give an outline stratigraphical description of St Pancras. Although here the emphasis will be on the origins and ‘Anglo-Saxon’ development of the building (Periods 1–2, Fig. 15), in order to understand this it is first necessary to note that the earliest deposits within and around it were badly disturbed by post-Norman intrusions. There were numerous Norman and later burials (from Periods 3–4) and the upper deposits inside the church were adversely affected by its use as a post-medieval farmhouse and pigsty. This was especially so in the chancel of the standing ruin, where there had been a 910 mm-deep chalk-filled hollow constructed for the pigs.⁴⁶

Fig. 15. Revised phase plans of St Pancras Periods 1 and 2. On both plans, black indicates newly constructed walls and columns. Period 1a is indicated in grey on the Period 1 plan. Walls indicated in white on the Period 2 plan are those retained from Period 1. Broken lines indicate possible Period 2 walls, where the existence of these is a matter of interpretation rather than unambiguous evidence.

© Author, drawn for this paper by Petra Dark

One feature post-dating the Norman period is especially significant for the dating of the earlier history of the building. This is a 14th- or 15th-century (Period 4) tiled floor found in all areas of the church at c. 360 mm deep, forming a terminus ante quem for what was sealed below.⁴⁷ Burials were found c. 380 mm below this floor, with stone slabs surrounding a dug grave and overlaid by chamfered limestone slabs, typologically dated by Blockley to the 12th century.⁴⁸

Below both this floor and the burials was a layer of brickearth between c. 180 mm to c. 200 mm deep, and below that a white ‘concrete’ (mortar) floor c. 150 mm deep, with a pinkish or reddened surface, extending across the interior.⁴⁹ Excavation proceeded no lower than this floor across most of the interior of the building, although in places it reached the underlying Roman deposits.⁵⁰ Unless very shallow (the intervening brickearth layer was only c. 200 mm deep at most), the medieval burials must therefore have cut through the mortar floor.

The Period 1 building

Combining my own observations in 2018–19, Routledge and Hope’s published accounts, Blockley’s work on Jenkins’ excavation archive and Jenkins’ publications — some of which are absent from Blockley’s bibliography, and so presumably unseen by him — it is possible to produce the first accurate description of the original (Period 1) structure.⁵¹ This consisted of a rectilinear room 12.87 m long × 8.11 m wide, surviving up to c. 0.76 m high, with a western doorway 2.3 m wide.⁵² The interior surface of its wall showed traces of smooth white plaster up to 130 mm above its original clay floor (see below).

To the east of this rectilinear room was an apse. This was represented by a curvilinear foundation trench containing a polygonal wall, which could represent an originally curvilinear apse rebuilt to be polygonal, a polygonal apse on a curvilinear foundation or even a wall with a polygonal external face and curvilinear inner one.⁵³ Both Jenkins and Blockley suggest that there was a doorway in the south of the apse, blocked in Period 2, but the same space is shown by Hope as a later grave.⁵⁴

There were traces of a buttress, 0.53 m wide, 0.28 m from the junction between the apse and the south-east corner of the main rectilinear room; evidence for a buttress in the equivalent position on the north of the apse was noted by Jenkins in the 1970s, but its dimensions are unrecorded.⁵⁵ Jenkins observed that a short north–south wall, secondary to the initial construction (Period 1a), but prior to Period 2, projected 3.66 m from the south buttress of the apse, but the reason for this feature is unknown.⁵⁶

The side and rear wall of the main rectilinear room had small external rectilinear corner buttresses 350 mm × 530 mm. Blockley gives the buttresses on the west as projecting 900 mm, two of these were stratigraphically sealed by the south wall and north wall of the western porch of the Period 2 building respectively.⁵⁷ Three buttresses on the south wall were spaced at 3.96 m, the westernmost projecting 560 mm × 350 mm wide and the other two 800 mm × 350 mm wide.⁵⁸ There seems to have been only one, centrally placed, buttress on the north wall, measuring 530 mm × 300 mm, and this was earlier than the north porticus of the present ruin.⁵⁹

All of these Period 1 features are built of Roman brick, most showing indications of reuse, with a few lumps of opus signinum incorporated into the construction. The bricks were set in yellow mortar but all of these walls had un-mortared foundations, c. 760 mm wide and usually c. 560 mm deep, constructed of four or five courses of flint nodules with Roman brick and opus signinum inclusions.⁶⁰ The variability in the coursing of these foundations was because they were deeper where they overlay the relatively softer fill of underlying Roman-period features.⁶¹ A capping of Roman brick or tile — projecting c. 150 mm from the wall-face — overlay the top of the flint foundations. This structure had a light-yellow loamy clay floor 8–130 mm thick, directly overlying a soil layer containing Roman-period artefacts.⁶² Both the offset at the base of the Period 1 walls and a thin layer of yellow mortar outside them presumably represent evidence for their construction.

Four clay-filled postholes were sealed beneath the Period 1 clay floor of the building.⁶³ These were assigned by Blockley to the construction of the structure, and this is the most plausible explanation.⁶⁴ However, it is just possible that they represent an earlier, post-built, structure on the site of, and of only slightly smaller dimension than, the main rectangular room of the structure. If so, and their proximity to the interior of the brick-built walls suggests their interpretation as for scaffolding rather than a pre-existing structure, there is nothing to indicate whether this putative structure was domestic or ritual, pagan or Christian.

Together this enables the Period 1 structure to be reconstructed as an east–west apsed basilica built of reused red Roman brick in yellow mortar. The walls had small rectilinear decorative buttresses. The interior had a light-yellow clay floor and walls covered with smooth white plaster. No artefacts were certainly associated with this structure, but at least three east–west inhumation burials were cut by the Period 2 south porticus and one by the Period 2 apse porticus along with other, undated, burials found within and around the structure, any or all of which could belong to this phase.⁶⁵ The burials, orientation and plan of the structure all support the interpretation — shared by the excavators and by Thomas, Dorothy Watts and Blockley — that the Period 1 structure was a church. This interpretation is also accepted here.⁶⁶

The 1900 excavation in the east of the nave found a fallen arch built entirely of Roman brick, with yellow mortar and white plaster on its surface identical to that used in Period 1.⁶⁷ On structural grounds it is reasonable, therefore, to date the fallen arch to Period 1, but as in 1900 the excavation had stopped at the mortar floor inside the nave, the fallen arch must have lain on that floor.⁶⁸ However, a layer of 230 mm of soil separated this mortar floor from the earlier Period 1 floor. Thus, this arch must have been retained into Period 2 or later.

The burial of a child containing a sceat — the coin was recently numismatically dated to the early 8th century — was cut into a soil layer between the south wall of the nave, and the Period 2 south porticus (see below), and overlaid by construction debris from that porticus.⁶⁹ This burial must have occurred prior to the construction of the Period 2 porticus, giving it a terminus post quem of the early 8th century. The south-east corner of the Period 2 south porticus cut a grave containing human bone with a radiocarbon date of 672–1121 cal AD at 95% probability. This burial gives a terminus post quem for the Period 2 porticus, here no earlier than the late 7th century. These two burials could be interpreted as evidence that the use of the Period 1 church extended into the 8th century, that burial occurred adjacent to the ruinous Period 1 church prior to Period 2, or that the Period 2 structure chancel and nave were constructed prior to the two dated porticus.

The Period 2 church

The design in the next structural phase (Period 2) was much more complex, eventually at least.⁷⁰ Porticus were added to the south side of the chancel and to both the north and south sides of the nave, and a porch to the west end of the nave; the door to the nave was narrowed to 1.98 m by the addition of two brick-built jambs.⁷¹ Jenkins plausibly ascribed a 150 mm-thick mortar floor — almost certainly that found by Routledge in the 19th century — to this Period 2, noting that its relationship to the plaster on the interior of the Period 1 walls suggested that it post-dated them.

The junction of the chancel and apse was elaborated with a triple arcade with four columns, a feature paralleled at ‘Anglo-Saxon’ churches elsewhere in Kent. Hope reasonably calculated the height of this arcade from its surviving remnants as c. 4.72 m.⁷² Later, but within Period 2, the side arches of the arcade were blocked with reused Roman bricks.⁷³ The retention of the Period 1 chancel arch (see above) necessitates that the arcade was inserted while it was supported, presumably by timber scaffolding — a considerable effort, betokening the builders’ concern to keep the arch intact.⁷⁴

No pre-Norman worked stone, other than that from the columns of the arcade, was found when the church was excavated. The existing column is of limestone from the Paris basin, probably newly imported rather than reused from a Roman structure.⁷⁵ The rubble overlying the Period 1 and 2 structures comprises only a large quantity of reused Roman brick, often with pink mortar adhering to it, suggesting that the upper parts of the walls were wholly of this material.⁷⁶

All the Period 2 walls, including the porticus and porch, were constructed of reused Roman bricks in white mortar, unlike the yellow mortar of Period 1.⁷⁷ The Period 1 walls in the nave and apse were rebuilt above the lowest 910 mm, with the bricks of the porch being bonded into the western nave wall above that level.⁷⁸ Rebuilding the Period 1 walls so extensively, and from different heights, is consistent with the interpretation that Periods 1 and 2 were discontinuous, as argued by Jenkins.⁷⁹ Jenkins also noted burials belonging to Period 2 in the chancel and to the south of the church but no further details of these burials survive.⁸⁰

Dating Period 2

The two burials already mentioned — one radiocarbon-dated, the other containing a sceat — support a date in the early 8th century or later for the Period 2 porticus. Nevertheless, the distinctive triple arcade is well paralleled in 7th-century Kentish churches, the latest of which is that at Reculver, probably dating to the late 7th century.⁸¹ Stylistic dating of the column-bass by Dominic Tweddle supports a late-7th-century date, although it would also allow an 8th-century one,⁸² but the analogous arcade at Lyminge is confidently 7th century.⁸³ A 7th-century date for the triple arcade suggests that the chancel and nave of the Period 2 church were rebuilt earlier than the porticus, and also date to the 7th century.

Given the care taken to incorporate what remained of the Period 1 structure, especially the chancel arch, it was clearly designed to bring a pre-existing building, the Period 1 church, back into use, rather than as a completely de novo construction. It is therefore logical to propose that the Period 1 church retained or acquired some association which rendered it significant to whoever commissioned its rebuilding in the 7th century.

In sum, the evidence of Periods 1 and 2 strongly supports the interpretation, first proposed by Jenkins, that the Period 1 building was in a disused and dilapidated state when it was comprehensively restored at the start of Period 2, when the triple arcade was added.⁸⁴ Period 2 should therefore be subdivided into Period 2a, the rebuilding of the Period 1 structure with the incorporation of the triple arcade, and Period 2b, involving the blocking of the side arches of the arcade, along with the addition of the south and north porticus and western porch.

Period 2a may be dated to the 7th century until at least the early 8th century because of the typological analogies of the triple arcade and the burial that accompanied the sceat. Period 2b dates to the early 8th century or later, because of the terminus post quem provided by the sceat. If the porticus and western porch were added in the 8th century this would allow the whole of Period 2 to fit within less than a century.

At this point it is possible to return to the most difficult question concerning the Period 1 structure: the question of its date. This is crucial to understanding its significance, both to the origins of St Augustine’s Abbey and more widely.

Dating Period 1

Jenkins offered no date for the Period 1 structure, which he recognized as a church, and Thomas and Watts have both ascribed it to the late 4th or early 5th centuries.⁸⁵ Alternatively, Richard Gem, Eric Cambridge, Blockley and others favoured a date in the 7th century.⁸⁶

Both Jenkins’ phasing and Thomas’ dating of the sequence were completely rejected by Alan Ward in his brief reinterpretation of St Pancras.⁸⁷ Ward, dating the white cement floor to the post-Norman period, failed to take into account Jenkins’ evidence for an earlier clay floor beneath the soil layer separating Periods 1 and 2. Jenkins’ original interpretation therefore stands of Period 1 as discontinuous from, and in a dilapidated state prior to the construction of, the Period 2 additions to the building.

Cambridge considered that the presence of small rectilinear buttresses (his ‘pilaster buttresses’) ruled out a date before the 5th century.⁸⁸ However, such buttresses occur in unambiguously Roman-period contexts in Britain, even in Kent, where they are present on Roman buildings such as those at The Mount, and at Loose Road, both in Maidstone, Folkestone, and at Springhead.⁸⁹ Corner buttresses of this type also occur on Roman-period buildings, as at the Folkstone and Darenth villas.⁹⁰ It is, therefore, incorrect to say that small rectilinear buttresses on the sides or corners of rectilinear structures are a distinctively 5th-century and later feature. If a Late Roman date is credited for Phase 1 of St Martin’s, then, as it also has small rectilinear buttresses, there is a local parallel for this feature before the 5th century.⁹¹

Several scholars have considered the unusual dedication to St Pancras reason on its own to date the Period 1 building. For example, Blockley argued on the basis of the dedication that the structure, which he considered impossible to date on archaeological grounds, was built in the mid-7th century.⁹² In this he seems to have been persuaded by Richard Gem’s observation that the cult of St Pancras was only revived by Pope Honorius after 625.⁹³ Nevertheless, this is invalid as evidence for the date of the Period 2 church, let alone that of Period 1, because no text of the 7th or 8th century records its dedication. It is possible that the dedication to St Pancras dates from the Period 2 rebuilding of the church. Gem’s date of the late 7th century for the Period 1 church might, then, apply to the Period 2a church, rather than that of Period 1.

No dateable objects or architectural stonework were found in association with the Period 1 structure, but it must date to 337 or later, as a coin of Constantius II (337–61) was found in an earlier soil layer. The coin is noted by Blockley who makes no comment on its chronological significance for the Phase 1 structure.⁹⁴ A Late Roman or later date is supported by the use of almost entirely reused Roman brick and opus signinum fragments. The latest dated pit from the preceding phase of pits (see above) has a radiocarbon date of 248–564 cal AD at 95.4% probability, further supporting these artefactual dates. An approximate terminus ante quem may be provided by the architectural analogies of the Period 2a arcade. The Period 1 structure may therefore date to between the late 4th century and the (perhaps late) 7th century, but was probably discontinuous with 7th-century Phase 2a.

Comparison with other churches at St Augustine’s suggests that the Period 1 building belongs to the earlier part of this chronological range. It is very simple in plan — just a small apsed basilica — and has a clay floor and plain plastered walls. Yet even by 609 the Mission was able to begin constructing the church of Sts Peter and Paul, with a more complex plan and cement floor.⁹⁵ Furthermore, unlike Sts Peter and Paul and the adjacent chapel of St Mary, also built in the early 7th century, yellow mortar was used in the walls of the Period 1 building at St Pancras. These considerations suggest that Period 1 was earlier in date than these two other buildings, and they are unlikely to reflect the relative status or function of the buildings given that the chapel of St Mary was constructed in a similar way to the church of Sts Peter and Paul, yet had a different status and function.

The proportions and size of the Period 1 building are comparable to 7th-century churches known from elsewhere in Kent, showing that it was, like those churches, laid out according to a grid using the same units and principles.⁹⁶ However, as it employs what John Blair — who identified the existence of grid-planning in ‘Anglo-Saxon’ churches in general — calls ‘the short-perch module’, unlike the Merovingian ‘long perch module’, it is unlikely to be a Merovingian building, and so probably unlikely to be associated with Bertha’s entourage before 597. In combination, metrological considerations suggest that St Pancras was built by the Augustinian mission after 597, but using an ‘Anglo-Saxon’ module of planning rather than an inherited Merovingian one.

These observations, even taken on their own, support the interpretation that the Period 1 structure was an ‘Anglo-Saxon’, rather than culturally Romano-British or Merovingian, building. They also imply that a date before 597 is implausible, and that the Period 1 church was laid out according to system of measurement at the start of the ‘Anglo-Saxon’ church-building tradition rather than part of the Frankish one.

Nevertheless, if the Period 1 church dates to the 7th century, it might be expected to be aligned with both Sts Peter and Paul and the church of St Mary. These two early-7th-century churches share the same alignment, reflecting the common linear alignment of ‘Anglo-Saxon’ monastic church buildings found elsewhere.⁹⁷ That the Period 1 structure at St Pancras is on a different alignment to these two churches is, therefore, most easily explained if it pre-dates them but was out of use by the time they were constructed. This would, of course, be consistent with evidence of disuse between Period 1 and Period 2 at St Pancras.

Drawing all these strands of evidence together, the Period 1 church probably dates to after 597 — because of its metrological characteristics — but before 609 when construction of Sts Peter and Paul began, because of its alignment and the relative simplicity of its construction. Only a narrow date-range of just twelve years between 597 and 609, therefore, accommodates all of these characteristics, placing it at the very start of the ‘Anglo-Saxon’ church-building tradition. A plausible historical context can be found for this. It might be expected that upon Aethelberht’s conversion he would emulate Roman and continental Christian rulers by founding a church — perhaps especially as, in a letter from Gregory the Great, he had been urged to emulate Constantine the Great.⁹⁸

Although, on the basis of Bede’s description of the Conversion, an earlier church could have been built inside the Roman walls, the chronological relationship to the earliest building at St Pancras of both this church, and that archaeologically attested on the site of the present Canterbury Cathedral, is uncertain.⁹⁹ The earliest post-Roman mortared walls so far found at Canterbury Cathedral may be given a terminus post quem by a sherd of pottery dated c. 650–850 found in a mortar layer probably representing their construction.¹⁰⁰ This suggests that these walls belong to a late-7th-century or later building, rather than one of the late 6th or early 7th century. If one discounts this sherd, the only dating evidence for the walls is pottery from the layer cut by the walls, providing a terminus post quem of c. 450.¹⁰¹

Consequently, the Period 1 church may have been built in the earliest years of the Augustinian mission — earlier than any other known ‘Anglo-Saxon’ church. If so, St Pancras is also the earliest church at St Augustine’s, and was necessarily commissioned and consecrated by St Augustine himself. Indeed, the abbey’s monks later perceived it as exactly such a place, associating it with Augustine and the first communion celebrated by him in Kent, although whether there was any basis to this tradition, such as a lost text or inscription, is unknown.¹⁰² If, as Bede suggests, Augustine and the other missionaries initially worshipped at St Martin’s and used it as their base for preaching, then it is possible that, after the conversion of Aethelbert, constructing the Period 1 church at St Pancras was seen by the missionaries as the next logical move.¹⁰³ This dating might also explain both the location of St Pancras itself and subsequently the whole 7th-century monastery, if the latter was located because of the church. In this context, the woman’s burial excavated near St Pancras might either be seen as that of one of Augustine’s early converts, or interpreted as part of a pre-existing cemetery explaining the exact location of the church.

Whichever of these interpretative options is true, the monks presumably used the Period 1 St Pancras building as their church prior to the construction of Sts Peter and Paul. Consequently, the evidence for its disuse prior to the Period 2 remodelling of the building might be explained by the abandonment of St Pancras before the construction in 609 of Sts Peter and Paul. If the Christian use of St Martin’s — whether as a Late Roman, sub-Roman or Merovingian church — pre-dates the Augustinian mission as Bede tells us, then St Pancras may be the earliest church built by that mission — even before the church constructed on the site of the present Canterbury Cathedral.

conclusion

If one accepts this dating and interpretation of Period 1, St Pancras is the earliest post-Roman church visible in England today, the earliest ‘Anglo-Saxon’ church yet known, the earliest part of St Augustine’s Abbey and the only visible building both consecrated and used by Augustine himself. As such, it is uniquely representative of a pivotal moment in British history, one of the most historically important standing buildings in England and a visible testament to the origins of both the ‘Anglo-Saxon’ Church and of English ecclesiastical architecture.

ACKNOWLEDGEMENTS

I am grateful to English Heritage and Historic England for permission to use their illustrations and to inspect and photograph the site in detail, and especially to Roy Porter — who read a first draft of this paper — and Kathryn Bedford, for their practical help and discussions of the site with me. Thanks also to Georgina Morton-Brown, Andrew Richardson and Andy Seaman who discussed the site with me, and to Petra Dark, University of Reading, who drew Figure 15, recalibrated the radiocarbon dates and made many helpful comments on the whole paper. Thanks are due also to the anonymous reviewers of this paper for their support and helpful comments.

Notes

1 The most recent studies to consider the monastic complex as a whole are R. Gameson ed., St Augustine and the Conversion of England (Sutton 1999); R. Gem ed., St Augustine’s Abbey, Canterbury (London 1999).

2 On the kingdom of Kent: S. Brookes and S. Harrington, The Kingdom and people of Kent ad 400–1066: their history and archaeology (Stroud 2010); M. Welch, Anglo-Saxon Kent to ad 800 (Woodbridge 2007). On St Augustine’s Abbey: R. Gem, ‘Reconstructions of St Augustine’s Abbey, Canterbury, in the Anglo-Saxon period’, in St Dunstan and his Times, ed. N. Ramsay, M. Sparks and T. Tatton-Brown (Woodbridge 1992), 57–73; N. Brooks, The Early History of the Church of Canterbury: Christ Church from 597 to 1066 (Leicester 1984); C. R. Peers and A. W. Clapham, ‘St Augustine’s Abbey Church, Canterbury, before the Norman Conquest’, Archaeologia, 77 (1927), 201–17. ‘Anglo-Saxon’ is used in this paper simply as the conventional archaeological term for the period in Kent from the start of the 5th century until the Norman Conquest, while recognizing that identity in 5th- to 7th-century Britain was much more complex than this term has sometimes been taken to suggest. For my usage in this limited way, see K. Dark, Britain and the End of the Roman Empire (Stroud 2000). As there, this restricted use is indicated in the text by placing the term in inverted commas: ‘Anglo-Saxon’.

3 R. Gem, ‘The Anglo-Saxon and Norman churches’, in St Augustine’s Abbey, 90–122, at 110–22; T. Tatton-Brown, ‘The Abbey Precinct, liberty and estate’ in St Augustine’s Abbey, 123–42, at 123–32; T. Tatton-Brown, ‘The Buildings and Topography of St Augustine’s Abbey, Canterbury’, JBAA, 144 (1991), 61–91.

4 Gem, ‘Anglo-Saxon’, 95, 97–101.

5 For the Anglo-Saxon churches and their significance: R. Gem, ‘Anglo-Saxon’; E. Fletcher, ‘Early Kentish Churches’, Med. Archaeol., 9 (1965), 16–31.

6 For example, S. Kelly, ‘The Anglo-Saxon Abbey’, in St Augustine’s, 33–49, at 36.

7 Bede, Historia Ecclesiastica gentis anglorum, 1. 26. The standard edition and translation of Bede's work is B. Colgrave and R. A. B. Mynors, Bede's Ecclesiastical History of the English People (Oxford 1992). Chapter and section numbers cited here are those in that work. On Aethelberht’s kingdom: S. Brookes, Economics and Social Change in Anglo-Saxon Kent, ad 400–900: Landscapes, Communities and Exchange (Oxford 2007); Brookes and Harrington, The Kingdom, 69–92.

8 Gem, ‘Anglo-Saxon’, 93–95; C. Thomas, Christianity in Roman Britain to ad 500 (London 1981), 172–74.

9 Gem, ‘Anglo-Saxon’, 93–94.

10 F. Jenkins, ‘St Martin's Church at Canterbury: A Survey of the Earliest Structural Features’, Med. Archaeol., 9 (1965), 11–15.

11 C. F. Routledge, ‘St Martin’s church, Canterbury’, Archaeologia Cantiana, 22 (1897), 1–28; C. F. Routledge, ‘St Martin’s Church, Canterbury’, JBAA, 1st series, 40 (1884), 47–51.

12 Blockley, ‘Anglo-Saxon churches’, 76–78.

13 K. Blockley, ‘The Anglo-Saxon churches of Canterbury archaeologically reconsidered’ (unpublished M.Phil. thesis, University of Durham, 2000), 26, 75, 77–79.

14 P. Bennett, ‘The Archaeology of the Canterbury World Heritage site’, 2019, unpublished lecture as Visiting Professor of Archaeology in the Centre for Kent History and Heritage at Christ Church Canterbury University: https://blogs.canterbury.ac.uk/kenthistory/paul-bennett-canterbury-unesco-and-jayne-wackett/ (accessed January 2021); A. Detsicas, The Cantiaci (Gloucester 1987), 148; T. Tatton-Brown, ‘St Martin’s Church in the sixth and seventh centuries’, in The Parish of St Martin and St Paul, Canterbury: Historical essays in memory of James Hobbs, ed. M. Sparks (Canterbury 1980), 12–18.

15 Blockley, ‘Anglo-Saxon churches’, 77, 128.

16 Ibid., 128.

17 H. M. Taylor and J. Taylor, Anglo-Saxon Architecture, 3 vols (Cambridge 1965–78), I, 143–44.

18 Detsicas, Cantiaci, fig. 19. 102. For the analogy see Thomas, Christianity, fig. 28, 183–85, who notes that the latter was also incorporated into a later church.

19 Jenkins, ‘St Martin’s Church’, 5, 12.

20 C. S. Green, Excavations at Poundbury vol 1: the settlements (Dorchester 1987); D. E. Farwell and T. I. Molleson, Excavations at Poundbury 1966–80: Volume II. The Cemeteries (Dorchester 1993).

21 L. Blackmore, I. Blair, S. Hirst and C. Scull, The Prittlewell Princely Burial: Excavations at Priory Crescent, Southend-on-Sea, Essex, 2003 (London 2019), 150–51, 346–47; M. Werner, ‘The Liudhard medalet’, Anglo-Saxon England, 20 (1991), 27–41; P. Grierson, The Canterbury (St Martin’s) hoard of Frankish and Anglo-Saxon coin-ornaments (London 1959); P. Grierson, ‘The Canterbury (St Martin’s) Hoard of Prankish and Anglo- Saxon Coin-ornaments’, British Numismatic Journal, 27 (1952–54), 39–51.

22 C. Sparey-Green, ‘Excavations at St Martin’s House: archaeological investigations in the vicinity of St Martin’s Church, Canterbury’, Archaeologia Cantiana, 136 (2015), 17–36; C. Sparey-Green, ‘St Martin’s Priory’, Canterbury’s Archaeology 2001–2002 (2003), 18–20; M. Sparks and T. Tatton-Brown, ‘The history of the Ville of St Martin’s, Canterbury’, appendix to J. Rady, ‘Excavations at St Martin’s Hill, Canterbury, 1984–85, Archaeologia Cantiana, 104 (1987), 123–218, at 200–13.

23 Blockley, ‘Anglo-Saxon churches’, 78–79; Jenkins, ‘St Martin’s Church’.

24 Ibid.; Rady, ‘Excavations’.

25 J. Weeks, ‘A review of Canterbury’s Romano-British cemeteries’, Archaeologia Cantiana, 131 (2011), 23–42, at 28. See also M. R. Day, ‘The Roman Period in the Parishes of St Martin and St Paul’, in The Parish of St Martin and St Paul, 5–11, fig. 7.

26 St Augustine, unnumbered anonymous map, plate 2.

27 A. D. Saunders, ‘Excavations in the Church of St Augustine’s Abbey, Canterbury 1955–58’, Med. Archaeol., 22 (1978), 25–63, at 28.

28 Blockley, ‘Anglo-Saxon churches’, 64–65.

29 Blockley, ‘Anglo-Saxon churches’, 65. Recalibrated for this paper by Dr Petra Dark using OxCal v4.4.4 with the IntCal 20 datasets.

30 F. M. Hall, The Kent and Canterbury Hospital 1790–1987 (Canterbury 1987).

31 Augustine, unnumbered anonymous map, plate 2.

32 A. Richardson, ‘Interred before Augustine? Canterbury, Kent’, Current Archaeology, 359 (2020), 16–17.

33 Brookes and Harrington, The Kingdom, 57.

34 Dark, Britain, 78–81; S. Burnell and E. James, ‘The Archaeology of Conversion on the Continent in the Sixth and Seventh Centuries: Some Observations and Comparisons with Anglo-Saxon England’, in St Augustine’s Abbey, 83–106.

35 On Frankish contacts in Kent and more widely, see S. Harrington and M. Welch, The Early Anglo-Saxon Kingdoms of Southern Britain ad 450–650: Beneath the Tribal Hidage (Oxford 2014) 181–205; Brookes and Harrington, The Kingdom, 47–48, 67.

36 Gem, ‘Anglo-Saxon’, 101–03; Taylor and Taylor, Anglo-Saxon Architecture, 146–48.

37 Blockley, ‘Anglo-Saxon churches’, 60.

38 M. Sparks, ‘The recovery and excavation of the St Augustine’s Abbey site, 1844–1947’, Archaeologia Cantiana, 100 (1984), 325–44; D. Sherlock and H. Woods, St Augustine’s Abbey: Report on excavations, 1960–78 (Maidstone 1988); W. H. St John Hope, ‘Recent discoveries in the abbey church of St Austin of Canterbury’, Archaeologia Cantiana, 31 (1917), 1–26; W. H. St John Hope, ‘Recent Discoveries at St Austin’s Abbey, Canterbury’, Archaeologia Cantiana, 31 (1915), 294–97; C. F. Routledge, ‘Roman foundations at St Pancras, Canterbury’, Archaeologia Cantiana, 14 (1882), 103–07.

39 Blockley, ‘Anglo-Saxon churches’, 58.

40 Ibid., 66.

41 Ibid., 25–26; F. Jenkins, ‘Canterbury: St Pancras church’, Med. Archaeol., 20 (1976), 163–64, F. Jenkins, ‘The Church of St Pancras at Canterbury’, Canterbury Archaeology, 76 (1975), 4–5.

42 Sparks, ‘The recovery’, fig. 2, 337, 343–44; Blockley ‘Anglo-Saxon churches’, 58–63.

43 Sherlock and Woods, St Augustine’s Abbey, fig. 6.

44 Blockley, ‘Anglo-Saxon churches’, 26, 58–74.

45 Jenkins, ‘Canterbury: St Pancras church’.

46 Routledge, ‘Roman foundations’, 104.

47 W. H. St John Hope, ‘Excavations at St Austin’s Abbey, Canterbury’, Archaeologia Cantiana, 25 (1902), 222–37, at 232.

48 Routledge, ‘Roman foundations’, 105.

49 Ibid., 105; Blockley, ‘Anglo-Saxon churches’, 61.

50 Blockley, ‘Anglo-Saxon churches’, 58, 60; Jenkins, ‘Canterbury: St Pancras church’; Jenkins, ‘The Church of St Pancras’.

51 For Jenkins’ archive, see Blockley, ‘Anglo-Saxon churches’, 63–74.

52 Blockley, ‘Anglo-Saxon churches’, 60, 65.

53 I owe the latter suggestion to Roy Porter, in an email of 17 March 2021.

54 Ibid.; Jenkins, ‘Canterbury: St Pancras church’; Hope, ‘Excavations at St Austin’s Abbey’, main plan.

55 Blockley, ‘Anglo-Saxon churches of Canterbury’, 66.

56 Jenkins, ‘Canterbury: St Pancras church’; Blockley, ‘Anglo-Saxon churches of Canterbury’, 66.

57 Ibid., 65–66; Jenkins, ‘The Church of St Pancras at Canterbury’; Hope, ‘Excavations at St Austin’s Abbey’, 229.

58 Blockley, ‘Anglo-Saxon churches’, 65.

59 Ibid., 65.

60 Ibid., 64–65.

61 Ibid., 64.

62 Ibid., 66.

63 Jenkins, ‘Canterbury: St Pancras church’, 163.

64 Blockley, ‘Anglo-Saxon churches’, 66.

65 Ibid., 68–69.

66 Ibid.; Thomas, Christianity; D. J. Watts, Christians and Pagans in Roman Britain (London 1991).

67 Blockley, ‘Anglo-Saxon churches’, 61.

68 See ibid., 60–61.

69 C. Scull and J. Naylor, ‘Sceattas in Anglo-Saxon Graves’, Med. Archaeol., 60 (2016), 205–41, esp. 209 and 236; Blockley, ‘Anglo-Saxon churches’, 61; Jenkins, ‘Canterbury: St Pancras church’; Jenkins, ‘The Church of St Pancras’.

70 Jenkins, ‘Canterbury: St Pancras church’; Taylor and Taylor, Anglo-Saxon Architecture, I, 146–47.

71 Jenkins, ‘The Church of St Pancras’, 4–5.

72 Hope, ‘Excavations at St Austin’s Abbey’, 228.

73 Blockley, ‘Anglo-Saxon churches’, 60.

74 Jenkins, ‘The Church of St Pancras’.

75 Gem, ‘Anglo-Saxon’, 101.

76 Blockley, ‘Anglo-Saxon churches of Canterbury’, 60.

77 Ibid., 61.

78 Ibid., 61.

79 Jenkins, ‘The Church of St Pancras’.

80 Blockley, ‘Anglo-Saxon churches of Canterbury’, 61.

81 Brookes and Harrington, The Kingdom, 109, fig. 32, 111; E. Cambridge, ‘The architecture of the Augustinian mission. St Augustine’, in St Augustine, 202–36, at 219–22, figs 10.5–9, 10.11–13; Taylor and Taylor, Anglo-Saxon Architecture, I, 503–09, esp. 503, fig. 248 and 506–07.

82 D. Tweddle, ‘The Date and Stylistic Context of the Reculver Fragments’, in Corpus of Anglo-Saxon Stone Sculpture, 4, South East England, ed. D. Tweddle, M. Biddle and B. Kjølbye-Biddle (London 1995), 46–61, 136, figs 59–60; C. R. Peers, ‘On Saxon churches of the St Pancras type’, Archaeol. J., 58 (1901), 402–34.

83 The most up-to-date accounts are G. Thomas, ‘Re-creating Ethelburga’s church’, 2020, https://geopaethas.com/2020/10/07/re-creating-queen-ethelburgas-church/; G. Thomas, ‘Update on reconstructing Ethelburga’s church’, 2020, https://geopaethas.com/2020/07/17/update-on-reconstructing-ethelburgas-church/; G. Thomas, ‘Visualising Ethelburga’s church’, 2020, https://geopaethas.com/2020/07/01/visualising-ethelburgas-church/ (accessed 15 December 2021).

84 Jenkins, ‘The Church of St Pancras’.

85 Ibid.; Jenkins, ‘Canterbury: St Pancras church’. But see Thomas, Christianity, 172–74 and Watts, Christians and Pagans, 13–14, 65, 113.

86 Blockley ‘Anglo-Saxon churches’; Gem, ‘Anglo-Saxon’, 101–03; Cambridge, ‘The architecture’.

87 A. Ward, ‘Church Archaeology 410 to 597: The Problems of Continuity’, Archaeologia Cantiana, 124 (2004), 375–95, at 382–84.

88 Cambridge ‘The architecture’, 212–18, at 232, n. 67.

89 T. F. C. Blagg, ‘Some Roman Architectural Traditions in the Early Saxon Churches of Kent’, in Collectanea Historica: Essays in Memory of Stuart Rigold, ed. A. Detsicas (Maidstone 1981), 50–53, 55, fig. 25; Cantiaci, 136, 150 and figs 32–33.

90 Blagg, ‘Some Roman Architectural Traditions’, 57, fig. 26, 1–2.

91 See Jenkins, ‘St Martin’s Church’, 13.

92 Blockley, ‘Anglo-Saxon churches’, 128, 130, 240.

93 R. Gem, ‘The Anglo-Saxon and Norman churches’, 101.

94 Blockley ‘Anglo-Saxon churches’, 65.

95 Ibid., 140.

96 Ibid., 104. The observations on grid-planning and its contrasting modular systems in this paragraph derive from J. Blair, S. Rippon and C. Smart, Planning in the Early Medieval Landscape (Liverpool 2020), 95–96, 115–23; J. Blair, ‘Grid-planning in Anglo-Saxon settlements: the short perch and the four-perch module’, Anglo-Saxon Studies in Archaeology and History, 18 (2013), 18–61, 23–24, fig. 2.

97 J. Blair, The Church in Anglo-Saxon Society (Oxford 2005), 199–202, fig. 24.

98 R. A. Markus, ‘Augustine and Gregory the Great’ in St Augustine, 41–49, at 43.

99 Bede, Historia Ecclesiastica gentis anglorum, 1. 25.

100 Blockley, ‘Anglo-Saxon churches’, 29–31.

101 Blockley, ‘Anglo-Saxon churches’, 29–31.

102 See Hope, ‘Excavations at St Austin’s Abbey’, 234, on William Thorn’s claim that Augustine celebrated communion at the chapel.

103 Bede, Historia Ecclesiastica gentis anglorum, 1. 26.