The Topography of Roman Canterbury: A brief Re-assessment

THE TOPOGRAPHY OF ROMAN CANTERBURY: A BRIEF RE-ASSESSMENT P. BENNETT The first attempt to plot out archaeological discoveries and assess the size and extent of the Roman town was undertaken by James Pilbrow in 1868.' The observations and records made when he was City Engineer, during the laying of main drainage, consisted of random sightings of Roman masonry, deposits, burials and finds, which were often inconsistently and incompletely plotted and almost invariably misinterpreted. Despite these shortcomings Pilbrow's basic observations have been fruitfully used as supportive evidence by successive archaeologists, including ourselves.2 Pilbrow's limited approach led to a large underestimate of the size and complexity of the Roman town - a view which persisted until the Second World War.3 He believed the town to be centrally located within the medieval enceinte, flanked to the west by the intra-mural river (the eastern arm of the Stour) and contained approximately by the line of Guildhall Street to the north, St. Margaret's Street to the east and Hospital Lane to the south. The publication of the Victoria County History for Kent, Volume III (1932), did little to reinterpret or alter the assumptions of Pilbrow. The authors recognised that the Roman town was much larger than Pilbrow suspected. The eastern line of the medieval city wall was recognised as having a Roman predecessor on the same line, but again the intra-mural river was interpreted as the western boundary of the Roman town. The most important contribution of this volume 1 J. Pilbrow, 'Discoveries made during Excavations at Canterbury in 1868', Archaeologia, xliii (1871), 151-64. 2 Pilbrow's observations have been particularly useful for plotting out the theatre and forum areas. 3 For example, see 'The Summer Meeting at Canterbury, 15th-24th July, 1929', Arch. Journ, lxxxvi (1929), 237. 47 P. BENNETT was the catalogue of Roman finds made before and after Pilbrow, the recognition of the antiquity of the main roads leading to Canterbury and the listing of Roman burials, fragments of buildings and general finds. The first serious archaeological investigations began with the formation of the Canterbury Excavation Committee in 1944, when trenching took place within bomb-damaged street-frontage properties and in a number of back gardens and yards. From 1946 until the late 1950s, numerous spring and summer excavations directed first by Mrs. Audrey Williams and then Professor S.S. Frere, chance finds, observations and minor excavations by Dr. F. Jenkins and Mr. J. Boyle and others, laid down the basic foundation on which an assessment of the Roman town can be made. The wealth of information gleaned from nearly two decades of excavation has, since 1975, been supplemented by numerous large and small scale excavations and observations in advance of redevelopment by the Canterbury Archaeological Trust. These hectic nine years of almost continuous fieldwork have produced a corpus of new information which collectively allows a tentative re-assessment of the Roman town. One of the most useful pieces of work of the past nine years, and something that had not been done before in detail, has been the plotting of all known discoveries on to the First Edition Ordnance Survey 1:500 (1874) maps for Canterbury. These documents, the first triangulated survey of the city, are extremely accurate and have been used by the Trust, mainly by Mr. P. Blockley, to plot out all known Roman finds (walls, streets, burials, etc.,) from excavation and observation. The figures included with this discussion have been based on these old maps and are presented here as the most accurate and up-to-date plans of Roman Canterbury.4 More specifically, the maps have been used to re-plot all elements of the large Roman Theatre and the postulated Forum, including recent discoveries. These buildings will be discussed in more detail below. One of the most important topographical features associated with the development of the Roman town was the River Stour. Current knowledge based on excavations in the Mint Yard5 of the King's 4 The features shown on this plan are not necessarily contemporary. Some buildings had fallen out of use or had been modified before others were built and some roads were abandoned before others. The plan, therefore, shows the disposition of main buildings and streets irrespective of construction date or contemporaneity, to illustrate current knowledge of the salient topographical features of the town. 5 P. Bennett, S. Campbell, J. Rady, and S.S. Frere, 'Excavations at Lanfranc's Dormitory, The Aula Nova and The Almonry Chapel' in The Archaeology of Canterbury, iii (forthcoming). 48 THE TOPOGRAPHY OF ROMAN CANTERBURY • Spot heights from 1st edition Ordnance Survey (1873) All levels in leet n Li Fig. 2. Contour Map of the City of Canterbury. 49 P. BENNETT School (near the North Gate), at various sites close to the city wall between the West Gate and the North Gate6 and at the Poor Priests' Hospital,7 Stour Street, suggest that at least a branch of the Roman river may have followed the line of the present intra-mural stream. From the late first or early second century onwards, attempts may have been made to divert the course of the stream or at least restrict the area of flood plain. These attempts perhaps culminated in diverting the river north and west of the town when the defensive walls were built in the late third century. A recent examination of the spot heights recorded on the 1:500 1874 Ordnance Survey maps (by Mr. J. Rady) has enabled a thorough contour survey of the late nineteenth-century city to be made (Fig. 2). This map clearly shows the line of the intra-mural stream and its attendant flood plain and may in some respects reflect the situation prevailing in Roman times. Since the early Roman period there has been a considerable rise in sea level (and hence an associated rise in river level, perhaps by as much as c. 10-15 ft.), and it is now considered possible that the river broke through the walls and moved back to an earlier line in the flood plain (the line of the present intra-mural stream) in the sixth or seventh century. Although much evidence exists for some form of Flavian occupation in the area later contained by the town walls, a growing body of evidence now supports the theory that it was not until the early second century that the topography of the Roman town was really established. The recently excavated elements of the public bathhouse, opposite St. Margaret's Church and the newly discovered Temple Precinct west of the Theatre," together with new north-east to south-west aligned streets also found on these sites, were constructed in the early second century. Current knowledge from excavations and observations has comprehensively modified the regular chequer-board street grid originally postulated9 by Professor Frere. The network of streets plotted out on Fig. 1, though regular in many respects, does suggest that the development of the grid was piecemeal, with modifications to the street pattern occurring through 6 Interim reports on these excavations will appear in the next issue of this journal (ci, 1984). 7 P. Bennett, 'Excavations in the Castle Street, Stour Street Areas', in The Archaeology of Canterbury, vi (forthcoming). * For a reasonably up-to-date plan of these areas see 'Interim Report o Excavations in 1981 by the Canterbury Archaeological Trust', Arch. Cant., xcvii (1981), Fig. 2. '' For Professor Frere's current views on the street-grid, see S.S. Frere and S. Stow, 'Excavations in the St. George's Street and Burgate Street Areas', in The Archaeology of Canterbury, vii, 334-5 and Fig. 144 on p. 336; Britannia, i (1970), 84, Fig. 1. 50 *< BRICK, TILE AND POTTERY KILNS "liii'lHIIiilHliiiimMMiniiti. NORTHGATE .' WESTGATE OUfcNINGAU FORUM CANTERBURY C.75-C.450 LONDON GATE / SILVER HOARDIS ' IBB URIED c.410l = WORTHGATE / /OO KEY: POTTERY KILNS TILE KILNS ROMAN BURIAL MOUNDS CREMATIONS INHUMATIONS • 0 / ' / / V-S,>«V / / \ BATHS />• .* \X. / / ////Terns X^W / / \ / DRAIN-,'** " XSy > / / / / PRECINCTS S/V ">•/ . \.* .• y . -Tt ''* >/ —~ ; //^/ //'s^^Mty /' *•"*' ^ BURG / y V / S H R I N E X>i - .•' „„„,., • ^ S S ^ -«*. // x PORTICO BURGATE DRAN BATH HOUQSE --W/ . , • •M> ^x v •> /X "Vx THEATRE ,%*;'-% ' / / v RIDNGATE LATER ST. MARTIN'S CHURCH ^^^'''- ^C '"'"HHIH^ 1000 FEET 300 METRES Fig. 1. Plan of Roman Canterbury c. A.D. 75-450 (based on 1:2500 First Edition Ordnance Survey map (1874)). THE TOPOGRAPHY OF ROMAN CANTERBURY time and radical changes (particularly between Burgate and Ridingate) occurring in the late third century when the town wall was built. Turning to a discussion of Fig. 1, one is immediately struck by the enormous gaps in our knowledge of the Roman town. The line of the late third-century town wall with gates and external streets leading to the Saxon Shore forts and to London, established by Professor Frere and confirmed by numerous excavations by the Trust10 are the most dominant features. Only three intra-mural streets aligned north-west to south-east (the Riding Gate - London Gate street, the West Gate - Burgate street and a street located equidistant between them) indicate any regularity in plan, whilst only two others, one south-east of the Theatre and a second south-east of the St. George's Street bath-house, have been found roughly at right angles to the three streets. Elsewhere, the known streets seem to follow differing and seemingly random alignments. Two streets, aligned roughly northeast to south-west, may have been first laid out in Flavian times (the small section of street located in the south west quarter of the city and the road separating the Temple Precinct and the Theatre). These streets may hark back to a brief military activity and the position of a fort or fortlet established just after the conquest to guard the Stour crossing." The line of these streets together with a similarly aligned minor road later covered by Temple courtyard metallings,12 the line of Watling Street and an early road which ran from Burgate to West Gate, (from the early fort and military stores depot at Richborough) may therefore have been established before the town was laid out in the early second century. The line of the modern London Road, north-west of the city, has recently been proved by excavation to be of Roman origin.13 It has been suggested that the line of this street, extended eastwards, may indicate the existence of another early Roman focus (perhaps a fort or settlement) on the north-west side of the Stour, in the area of the west railway station. Recent excavations in the insula containing the public baths, the north-west corner of the Theatre insula, the insula south-west of the Theatre'4 and the Temple insula15 have provided overwhelming '" For example, see S.S. Frere, S. Stow, and P. Bennett, 'Excavations on the Roman and medieval Defences of Canterbury', part xi, 'The Church Lane Excavations', in The Archaeology of Canterbury, ii, 82-5 and Fig. 45. " P. Bennett, S.S. Frere, and S. Stow, 'Excavations at Canterbury Castle', in The Archaeology of Canterbury, i, 25-30. 12 The Archaeology of Canterbury, vi (forthcoming). 13 P. Bennett, S.S. Frere, J. Rady and S. Stow 'Canterbury Excavations: Intra Mural and Extra Mural Sites', in The Archaeology of Canterbury, viii (forthcoming). 14 M. Day, K. and P. Blockley, S.S. Frere, and S. Stow, 'The Marlowe Car Park Excavations and associated Areas' in The Archaeology of Canterbury, v (forthcoming). 15 The Archaeology of Canterbury, vi (forthcoming). 51 P. BENNETT evidence for a hiatus in occupation in the early second century. The sparse occupation levels, indicating a possible continuity of postconquest occupation east of the intra-mural stream, were sealed over by dumped deposits and levels associated with major building activity and the formation of a network of interconnecting streets. At the centre of this new street pattern were the public buildings, the Theatre, the Temple Precinct, the Public Baths and the Forum. Space does not allow for a full discussion of all of them and, therefore, the Public baths and the Temple areas, both early secondcentury developments, which have been discussed in interim in earlier issues of this journal,16 will not be considered here. New information has, however, come to light for both the Theatre and the Forum areas, which justifies some further discussion. Excavations at 77-79 Castle Street17 in 1977 revealed large robber trenches undoubtedly associated with the second-phase masonry theatre. The addition of these new elements has considerably changed the postulated theatre plan published by Professor Frere in 1970.'8 A complete replotting of the published elements of the Theatre has been attempted on Fig. 3, with the First Edition 1:500 (1874) Ordnance Survey map, used as a base. Three recent observations made during main service trenching in Castle Street have also been incorporated in the plan. The replotting of the salient elements of the Theatre, largely the work of Mr. J. Rady, has thrown up some inconsistencies in Professor Frere's original plan, which seem to have been based on an inferior, more recent survey map of the area. The slightly amended alignments of some of the walls and the inclusion in the plan of newly discovered fabric and robber trenches, have been supplemented with a tentative reconstruction of the internal arrangements of the theatre. Also shown on the figure is the major Roman street, aligned north-east to south-west, located on nos. 77-79 Castle Street, and the Portico, stylobate drain and courtyard of the Temple Precinct. Two large connecting robber pits, located up against the road frontage, south of the outer curving wall, also shown on the plan, were possibly for a load-bearing structure, perhaps an arch or building flanking the south-west side of the postulated stage area and attached to the south corner of the Theatre. The internal arrangements, based on structural elements located by Professor Frere and James Pilbrow, and in service trenches in Castle Street, have been re-interpreted in the light of a very general comparison with the '"Arch. Cant., xcii-xcix (1977-83). " The Archaeology of Canterbury, vi (forthcoming). '" S.S. Frere, 'The Roman Theatre at Canterbury', in Britannia, i (1970), 83-112. 52 Observed Jenkins Ibrow Slatters Frere Castle Observed served 1950-51 Frere Observed Fig. 3. A revised Plan of the early third-century Roman Theatre, (based on 1:500 First Edition Ordnance Survey map (1874)). JenkinsV 0%/ /V,/ / // ^fPilbrow °P Frere / / / , s 1951/53 ' ' / Mosaic discovered 1758 Fleur de Lis yard Frere 1955 Sewer ^X. Tunrjel 'V ' 1983 Circular building / (^ discovered 1689 //^ LA- 0, Baths > Temple Precinct Fig. 4. A Plan of the Roman Forum Area (based on 1:500 First Edition Ordnance Survey map (1874)). THE TOPOGRAPHY OF ROMAN CANTERBURY Theatre at Verulamium19 and with particular reference to the radial passage walls used in its construction. The main internal structural features in the Canterbury Theatre may have comprised three pairs of symmetrically placed passage walls, leading from the inner curving wall to the postulated position of a D-shaped orchestra wall. The central pair of radial walls divides the Theatre into two symmetrical halves and between these walls and the radial walls north and south of them are the foundations for a further pair of short passage or stair supporting walls. A curving length of wall (shaped like a flattened apse), located by Professor Frere, joining one of these short passage walls to the inner curving wall, may have been one of a series of similar load-bearing foundations, designed to support upper tiers of banked seats and retain the gravel bank of the earlier theatre. Of orchestra and stage areas, nothing is known and a hypothetical reconstruction of this part of the Theatre has not been attempted. It is hoped that this reinterpretation of the surviving remains of Canterbury's Theatre will stimulate further discussion of this remarkable structure. Furthermore, it is hoped that when time allows, a re-examination of cellars in the area, will produce evidence for a more conclusive plan of the Theatre. Recent observations during the cutting of a sewer tunnel under the High Street, close to the intersection of Guildhall Street with the High Street and at no. 9 High Street,20 have brought to light a number of substantial masonry walls, a great expanse of courtyard metalling and a stone-paved portico and courtyard (under no. 9 High Street). These finds add further weight to Professor Frere's assumption that the Roman Forum was located in this area, in the insula north of the Temple Precinct and west of the public baths. Although insufficient evidence exists for a detailed assessment of the size and disposition of this building complex, a plan (Fig. 4) detailing all the known excavated and observed remains is presented here, together with a tentative reconstruction based on some of the surviving elements.21 Professor Frere postulates that a normal principia-type building did not exist here, since such a building does not conform with the various elements so far discovered and that an irregular plan for the overall complex should therefore be suspected. He has kindly '" J. Wacher, The Towns of Roman Britain, London, 1974, 50. 20 The Archaeology of Canterbury, viii (forthcoming). 21 The observations made by Pilbrow have been particularly useful in this area (op. cit., note 1), as have reports of a mosaic found in 1758 in the County Hotel area, (see Arch. Cant., xv (1883) 338 ff. and colour engraving between pp. 126-7) and a large number of architectural fragments undoubtedly from the forum buildings, found in 1861, opposite All Saints' Church VCH, (Kent) iii 68 (2)). 53 P. BENNETT suggested in correspondence that the area north-west of the High Street and west of no. 9 High Street, may have been the Forum piazza. This suggestion is based on the observations at no. 9 High Street, where a stylobate foundation, flanked to the west by gutter blocks and courtyard paving-slabs and to the east by portico pavingslabs, was found during the construction of new building foundations. Similar paving-slabs were also located by Pilbrow opposite the intersection of White Horse Lane with the High Street. Other paving-slabs, perhaps associated with another part of the structure were observed by Pilbrow at the north end of White Horse Lane. Professor Frere further suggests that the curia may have been located north-west of this, flanked by the walls located by Pilbrow in 1868. The Basilica, he indicates, may have existed south of the High Street, although he admits that the lack of structural evidence under the County Hotel creates interpretational problems. One solution to this may be that the Pilbrow walls turn in front of the County Hotel, creating some sort of porch in front of the Basilica. Professor Frere suggests that the structural elements which he found in the Fleur-de- Lis Yard (part of a hypocausted building)22 may have been part of a range of rooms along the outside of the Basilica as at London or Verulamium, although they could easily be part of a separate building. One final reference to the Forum needs to be made. This is to a large circular Roman structure discovered during building work in the Jewry Lane area in 1689. This curious building, discussed by the late Dr. W. Urry in a recent paper in this journal23 may have been part of a temple or shrine associated with the general building complex, but it may equally have been a laconicum associated with perhaps another public bath-house located in the Forum insula, south of the Basilica. To conclude this brief re-assessment of the topography of the Roman town, a few observations should be made about the extramural area, for which current knowledge is extremely sparse. Large cemeteries flanking most, if not all, of the streets leading out of the Roman town have received comments in many learned journals and books and were first discussed at length in the Victoria County History (Kent), iii. Continuing the policy of plotting information which has substance in fact, only known burials (cremations and inhumations) appear on the figure (Fig. 1). Of the numerous reports of burials found in the environs only those plotted on the map can be proven and provenanced. The greatest number of burials yet found 22 The Archaeology of Canterbury, (forthcoming). * Arch. Cant., xciv (1978), 1. 54 THE TOPOGRAPHY OF ROMAN CANTERBURY on a single site (53 cremation burials from Cranmer House on the London Road, north-west of the town),24 was the product of salvage work during building construction. Other than a small excavation in Lady Wootton's Green (1954), no excavation has yet been conducted specifically in a cemetery area. Most of the burials shown on the figure contain material dating from the early second century onwards and indicate an expanding population from that time onwards. The Romano-British burial mounds flanking the south-east side of the town (and two inside the walled area) have also been discussed on many occasions, but have never been plotted out accurately. Since many of these mounds survived virtually intact up to more recent times, documentary and contemporary pictorial evidence have been employed to locate their positions. None of these mounds has been excavated. A number of burials have recently been located inside the walls,25 in the south-west quarter of the town. These burials date to the late first or early second century and may predate the establishment of the street-grid in this area. A unique double inhumation burial from the same site, containing military equipment dating to the mid-second to mid-third centuries, may testify to a Roman murder. Evidence for industrial activity west of the Roman town, is mainly the product of work done by Professor Frere26 and Dr. Jenkins.27 The north-west extension of the Ridingate - London Gate street outside the town walls, it has been suggested, flanks the south-west side of an industrial suburb, which may have been associated with metalworking and pottery and tile manufacture. There is some evidence to suggest that a network of streets and tracks exists in this area,28 and may have been laid out to give access to the industrial suburb. The recent discovery of a second-century kiln on North Lane,29 (northwest of the walls) and previous discoveries of brick, tile and pottery kilns,30 north of the town indicate a wide distribution of such sites, their location perhaps dictated by the proximity of good quality potting clay (London Clay). Overall, it is hoped that the accurate plotting of old and new finds, will considerably assist and facilitate an understanding of the topography of the Roman town and will finally dispel the myth that our 24 The Archaeology of Canterbury, viii (forthcoming). 25 The Archaeology of Canterbury, i, 33-4 and Fig. 7. 26 The Archaeology of Canterbury, viii (forthcoming). 27 (a) F. Jenkins, 'Two Pottery Kilns and a Tilery of the Roman Period at Canterbury', Arch. Cant., Ixxiv (1960). 151-61; (b) F. Jenkins, 'A Roman Tilery and two Pottery Kilns at Durovernum', Antiq. Journ. xxxvi (1956), 40-56. 28 The Archaeology of Canterbury, viii (forthcoming). 29 Arch. Cant., xciv (1978), 166-8. 3,1 Op. cit., note 27(b). 55 P. BENNETT understanding of the Roman levels is comprehensive and complete. Although an enormous amount of work has been done, and many exciting and important finds have been made, the gaps in our knowledge can be seen, perhaps for the first time, by reference to a single plan.31 31 The figures accompanying this article have been drawn by Paul Blockley and Jonathan Rady. Thanks are offered to them and Tim Tatton-Brown for the considerable assistance they have given me during the writing of this re-assessment. Grateful thanks are also extended to Professors S.S. Frere, J.S. Wacher, and G.D.B. Jones for reading and commenting on the text. 56