The Southern Defences of Medieval Rochester

THE SOUTHERN DEFENCES OF MEDIEVAL ROCHESTER COLIN FLIGHT. M.A., and A.C. HARRISON, B.A., F.S.A. In a previous paper, completed in 1968, we described the results of a series of excavations carried out in Rochester, from 1960 onwards. in and around the area of the Deanery Garden. 1 We went on to discuss the implications of this evidence for a long-standing problem in the historical topography of Rochester - that is, the evolution of the Roman and medieval defences surrounding the south-east quarter of the city. By the time that the excavations came to an end, we had reached agreement between ourselves on every important point of interpretation. Before the report was published, however, one of us had started to feel some doubt - a doubt which prompted the inclusion, in that report, of a single cautionary footnote. 2 Over the last few years, we have thought about the problem once again. To some extent, we have had to take account of subsequent investigations, elsewhere within the city, which - directly or indirectly - affect the questions at issue.3 We have also paid closer attention to the written sources relating to the medieval phases in the development of the fortifications. Our main objective, however. has been to make a fresh assessment of the excavated evidence from the Deanery Garden. 1 A.C. Harrison and C. Flight, 'The Roman and medieval Defences of Rochester in the Light of recent Excavations', Arch. Cant., lxxxiii (1968), 55-104. From here onwards, we cite this report as 'Defences'. i 'Defences', 78, n. 41. 􀂓 A.C. Harrison, 'Excavations in Rochester', Arch. Cant., lxxxv (1970), 9.5-112; A.C. Harrison. 'Rochester East Gate, 1969', Arch. Ca11t., lxxxvii (1972), 121-57; C. Flight und A.C. Harrison, 'Rochester Castle, 1976', Arch. Cant., xciv (1978). 27-60; A.C. Harrison and D. Williams, 'Excavations at Prior's Gate House, Rochester, 1976-77', Arch. Cant., xcv (1979), 19-36; A.C. Harrison, 'Rochester, 1974-75', Arch. Cant., xcvii (1981), 95-136. A.C. Harrison, 'Rochester Priory', Arch. Cant., cii (1985). 265-6. 1 C. FLIGHT AND A.C. HARRISON We have little to add to our account of the Roman defences. The earthwork fortifications of Phase 1 (late second century?) are now known to have continued westwards, on the same line as the later wall, at least to a point about 30 m. from Southgate,4 and northwards to a point beyond Eastgate.5 We still know nothing about the ditch or ditches associated with the Phase 2 wall (early third century?); but subsequent excavations have given us a clearer picture of the rampart-bank thrown up against the inner face of the wall, enveloping what remained of the Phase 1 rampart. The Phase 2 rampart-bank had a width at the base of about 12 m., and a height of at least 2.50 m.6 In order to flatten a stretch of the Roman defences, it would thus have been necessary to shift a considerable volume of earth, quite apart from demolishing the wall itself. During the medieval period, the rampart-bank was in fact removed from behind the Roman wall, all the way from the south-east angle to Southgate, though perhaps not all at once; and two stretches of the wall were demolished, more or less completely. Further west, the Roman wall was partly destroyed, partly buried, when the earthworks surrounding the castle site were constructed.7 In consequence, the Roman south wall ceased to be a discernible feature in the topography of the city, except to the eye of an archaeologist. Its alignment was recognised first by Payne, in the 1890s.8 With this we turn to the medieval defences. On several significant points of interpretation, it pains us both to have to report that we are no longer in agreement. One of us (A.C.H.) is still in favour of the interpretation proposed in 1968, subject only to certain changes in detail. The other (C.F.) is now convinced of the need for some drastic revision. Each of us, after failing to convince the other, has had to admit that the evidence is not decisive as it stands. In time, no doubt, the issue will be resolved by further excavation; but of this there seems to be no immediate prospect. We have, therefore, thought it desirable to publish this further paper, re-assessing the evidence, and explaining briefly the source and scope of the disagreement between us. The principal features for which we need to find room in any interpretation are shown in the plan (Fig. 1). For the purposes of this discussion, it will be understood that we have had to refer to the medieval features by names which are merely descriptive, in order 4 Harrison and Williams, Arch. Cant., xcv (1979), 19-20. 5 Harrison, Arch. Cant., lxxxvii (1972), 124-5, 126-7. 6 It is likely, of course, that. the height had been reduced by erosion, the width perhaps increased. 7 Flight and Harrison, Arch. Cant., xciv (1978), 38. 8 G. Payne, 'Roman Rochester', Arch. Cant., xxi (1895), 1-16. 2 Bole.􀀇 >1<11 /Goto Fig. 1. 􀀆t.:ord 11f11 I! JJ I I ji} 111 \ C. FLIGHT AND A.C. HARRISON not to assume or imply the correctness of one particular interpretation. What we refer to here as the Deanery Garden ditch, for instance, is the same feature which, in the original report, was referred to throughout as the 1225 ditch. In fact, the dating of this feature turns out to be of crucial importance. If we could say for certain when the Deanery Garden ditch was dug, most of the other evidence would fall into place accordingly. The crucial question, re-phrased, comes down to this. From the documentary evidence (see below), it is known that a 'great ditch' was dug around the city in 1225. Where is it on the ground? The view we used to share - the view which still seems preferable on balance to one of us - is what we speak of henceforth as scenario A. On this view, the Deanery Garden ditch is identified as part of the 'great ditch', known to have been dug in 1225. The filling of this ditch, and the digging of the King's Orchard ditch further south, are then assumed to have taken place in 1344, in accordance with a scheme for the redevelopment of the defences devised by the prior and convent (see below). A dislocation of this chronology produces an alternative interpretation, scenario B. On this view, it is the King's Orchard ditch which was dug in 1225. The Deanery Garden ditch then has to be regarded as a feature of earlier but otherwise uncertain date, for the digging of which there is no explicit documentary evidence. Correspondingly, the scheme proposed in 1344 is assumed to have been abandoned, more or less. As far as we can see, these are the only possibilities which need to be considered: to that extent at least we still agree. 9 In fact, we agree on almost every point, apart from the crucial question. In some respects, our disagreement coincides with an earlier difference of opinion. For scenario A, the basic assum􀁪tion we make conforms with the interpretation proposed by Livett, 0 following up on Payne's success in tracing out the line of the Roman wall. For scenario B, by contrast, the assumption from which we start is consistent with the alternative interpretation suggested later by Hope.11 We thought we had settled the question in Livett's favour; but now we are not so sure. 9 There can only be two other possibilities - that both ditches were dug in 1225; or that neither of them was. On this basis we could set about constructing scenarios C and D, respectively. Because of the stratigraphic evidence, however, C is out of the question. As for D, we will only say that this seems to us a hypothesis of last resort; and neither of us feels driven to that extremity. 10 G.M. Livett, 'Mediaeval Rochester', Arch. Cam., xxi (1895), 17-72. 11 W.H. St J. Hope, 'The architectural History of the Cathedral Church and Monastery of St. Andrew at Rochester, Part 2', Arch. Cant., xxiv (1900), 1-85. 4 THE SOUTHERN DEFENCES OF MEDIEVAL ROCHESTER We plan to proceed as follows. First, we comment on the evidence from documentary sources, without as yet attempting to correlate it closely with the evidence on the ground. Second, we re-assess the results of the Deanery Garden excavations, making some agreed changes in the interpretation, but avoiding any mention of the documentary evidence. Third, we discuss an area to the west of the Deanery Garden, still very largely unexplored, where the sequence of events, on any interpretation, must have been somewhat different. Last, we deal in turn with the two possible scenarios, pointing out where each seems stronger or weaker than the other. It is only in this final section that the disagreements between us become significant. (1) Documentary Evidence From the records of the monastic community attached to the cathedral church, we can see that the monks were frequently engaged in transactions concerned with the acquisition of extra land, not only inside the city, but also beyond the wall towards the south. The earliest such transaction seems to have been the one involving an area of three acres, just outside the wall, given to the church by Odo, Earl of Kent - presumably between 1075 and 1082 - to provide the space needed for the monks' garden. 12 After 1088, the king was induced to issue a charter confirming the previous grant. Thus, we learn - what it would in any case be safe to assume - that the monks had surrounded their garden with a wall.13 During the twelfth century, further parcels of land outside the city wall continued to be acquired. The existence of a boundary wall enclosing the monastic precinct is mentioned now and then, but without any clue which would help us to fix its position on the map. 14 By this time, the monks were starting to put up buildings beyond the line of the wall. 15 The laying-out of the new cloister, the credit for which is ascribed to Bishop Ernulf (1115-24), may mark the begin- 12 The original grant by Odo was apparently not recorded in writing. Though the monks did not arrive till 1083, preparations for their introduction are likely to have been under way for some years previously (perhaps since soon after the death or Bishop Siward in 1075). Odo was thrown into prison in 1082, and not released again till 1087. 13 Two documents connected with the confirmation of Odo's grant by William II, in or soon after 1088, are printed, from the Textus roffensis, by Hope, Arch. Cam., xxiv (1900), 4-5. We note that the Textus is now accessible in a fascimile edition. 14 Hope, Arch. Cant., xxiv (1900), 5-6, 8-9. 15 The priory buildings are fully described by Hope, Arch. Cant., xxiv (1900), 23-66. Very little survives above ground, but of this there is a useful account in J. Newman, The Buildings of England: West Kent and the Weald (Harmondsworth, 1980), 485-8. 5 C. FLIGHT AND A.C. HARRISON ning of this process. 16 Till then, it seems, the monks had been housed in wooden buildings, awkwardly arranged on a less than adequate site - north of the wall, south of the new cathedral.17 Among the new buildings for which they had Ernulf to thank, the refectory is mentioned explicitly; and this, we know, stood up against the Roman wall, on the outside (see below). Indeed, a long stretch of the Roman wall owes its survival to the fact that it became incorporated into the monastic buildings, during the twelfth century. Some other structures, mentioned in documents of this period, were certainly located outside the Roman wall, though their sites are not known exactly. In time, the monks' attempts to expand their precinct southwards brought them into collision with the policies of central government. There is no word of any such conflict during the twelfth century; nor do we know of any attempt to repair or strengthen the defences of the city, in any part of the circuit, except on one occasion.18 During the 1220s, however, large amounts of money were invested in the defences of the castle and city. A series of entries in the Close Rolls of Henry III, mostly concerned with the funding of the work, help us also to understand the policy which lay behind it. 19 The castle, besieged and badly damaged in 1215, was brought back into repair. More than that, it seems to have become the centre of an integrated system of fortifications surrounding the city as a whole. Work on the castle began in 1221; in 1225, a ditch began to be dug around the city, under the supervision of 'three upright and lawworthy men' chosen from among the citizens. To the east, and perhaps to the north as well, the ditch followed the line of the Roman 16 Hope, Arch. Cant., xxiv (1900), 7, citing the list of benefactions preserved in an early thirteenth-century manuscript (British Library, MS. Cotton Vespasian A xxii). In a document drawn up in 1203-05, the monks themselves draw attention to their buildings, as proof of the generosity with which they had been treated by Ernulf - by contrast with the malevolence of Bishop Gilbert. Just look at our buildings, they say, erected from the foundations, on a larger scale than before, through Ernulfs efforts, and at his expense (Testantur hoe officinae nostrae, per eum et eius sumptibus ampliatae et a furuiamentis aedificatae): J. Moule, Gilbert Glanvill, Bishop of Rochester, 1185-1214 (M.A. thesis, Manchester, 1954), 241, from Canterbury, Dean and Chapter, Ch. Ant. R 70a. 17 Gundulf is said to have completed all the necessary buildings 'so far as the capacity of the site would allow': Textus roffensis, f. 172. If these buildings had been of stone - whether they were south of the nave, or on roughly the same site as the later cloister - their foundations would almost certainly have been discovered, where they adj6o ined the church, in the course of the underpinning carried out in the 1870s. Pipe Roll, S Richard I (II93), 166, where the sheriff accounts for the money he has spent in 'strengthening the city of Rochester'. The sum in question, however, is less than £10. 19 All the relevant entries are printed by Hope, Arch. Cant., xxiv (1900), 12-14. 6 THE SOUTHERN DEFENCES OF MEDIEVAL ROCHESTER wall;20 the chief uncertainty arises towards the south. How much work was carried out on the walls themselves, the sources fail to reveal. A passing reference to the construction of lime-kilns need not imply very much.21 Some later sources (see below) mention the existence of towers: perhaps, this is when they were built. 22 In 1227, for the first time ever, the citizens were granted a charter allowing them some measure of municipal autonomy.23 On any interpretation, the new ditch must have ploughed its way through the middle of the monastic precinct. It is not surprising, therefore, that the priory's annalist should have thought it worth recording the fact that the digging of 'a great ditch around the city' was started in 1225.24 That same year, Hartlip church was granted to the monks, no doubt as compensation for the land which had had to be requisitioned. 25 No further developments affecting the defences of the city are recorded till 1344. Before this, beginning in 1331, the monastic buildings had undergone some fairly extensive repair and reconstruction, subsidised by the bishop, Hamo de Hethe (1319-52)- who had himself been prior of Rochester before his appointment to the bishopric.26 In 1344, the prior and convent applied to the king, Edward III, for permission to reconstruct one stretch of the fortifications. The specific proposals they made can be summarised as follows: (i) to fill in a section of 'the king's ditch' extending from Eastgate as far as 'the gate of the said prior'; (ii) to dig a new ditch, 20 Its outer edge seems clearly marked by the streets which circumscribe the Roman wall, at a distance from it of roughly 20 m. Seen in this context, the excavations at Northgate fail to prove that the ditch did not exist, as was suggested by one of us: Arch. Cant., xcvii (1981), 99. They could equally well be thought to indicate that the berm was very wide - not less than about 9 m. 21 We agree that this passage need not carry the significance attached to it by Hope, Arch. Cant., xxiv (1900), 15. 22 If towers were part of the plan, perhaps this would help to explain the width of the berm (see above, n. 20). 23 Calendar of Charter Rolls, i, 64. The original charter survives, in duplicate, among the muniments of the city: P.H. Bartlett, The City of Rochester Charters (Rochester, 1961), 18-20, with a facsimile. We wonder whether this charter can be taken to fix the date of the city seal - a very handsome piece of work, inscribed with the proud (but alas ungrammatical) legend: + SIGILLUM CIVIUM ROFENSIS. A date of 'about 1210' is quoted by Bartlett (ibid., 5), on the authority of the British Museum. 24 Hope, Arch. Cant., xxiv (1900), 12, from a fourteenth-century manuscript (Cotton Nero D ii). This is Rochester's copy of the composite chronicle known as the Flores historiarum. 25 The likely connection was pointed out by Hope, Arch. Cant., xxiv (1900), 16. 26 The relevant passages are printed by Hope, Arch. Cant., xxiv (1900), 49-50, from a contemporary record of the bishop's movements and business transactions (Cotton Faustina B v). Repairs to the church were also carried out. 7 C. FLIGHT AND A.C. HARRISON further out from the wall, on their own land; and (iii) to build a new wall, 16 ft. high and 5 ft. thick, along the inner edge of the new ditch.27 After this had been done, the land enclosed by the new wall was to pass into their possession. By implication, they would then be free, if they wished, to demolish the existing wall - which, as they assured the king, had already been allowed to fall into disrepair. The monks, it is clear, were prepared to go to considerable expense for the sake of additional space inside the wall: they were even willing to make themselves responsible in perpetuity for maintaining this sector of the defences. Three documents relate to this transaction: a writ issued by the king, ordering the constable of the castle to hold an inquiry, in front of a jury of citizens; the report of this inquiry, recommending approval of the scheme;28 and the letters patent, authorising the monks to go ahead.29 It will be seen that none of these documents amounts to proof that the work was ever actually carried out. On any interpretation, in fact, we have to assume that the scheme was modified, either before or during its execution. This is obvious at once, from the fact that the stretch of ditch between Eastgate and the Roman south-east angle was certainly not filled in - as it certainly ought to have been, on the original plan. A fourth document exists which may cast some light on this problem. Five or six years later, the prior who had sponsored the scheme, John of Sheppey (1333-50), submitted a petition to the pope, asking to be allowed to retire from office.30 He gave an account of the many improvements for which he had been responsible, including the fact that he had 'caused the priory to be surrounded by a strong wall'. The petition was approved in February 1350. By this 27 Exact figures are quoted for the width and length of the section of ditch which the monks are permitted to fill. The width (87½ ft.) poses no problem; the length (905½ ft.) is a more interesting piece of information, because it ought to help us in fixing the site of 'the gate of the said prior'. Sadly, because we have not been told quite where the measurement was made (along the wall? along the outer edge of the ditch?), we are still prevented from coming to a definite conclusion. 28 Hope, Arch. Cant., xxiv (1900), 16-19, from the originals in the Public Record Office. The writ is dated 28 April, 1344, and the inquiry was held on 'Wednesday the vigil of the Lord's ascension' - that is, on 12 May. 29 Hope, Arch. Cant., xxiv (1900), 19-21, from the original, preserved among the muniments of the Dean and Chapter. See also Calendar of Patent Rolls, Edward Ill, 1343-5, 262. The Charter is dated 23 April, yet cites the fact that un inquiry has been held. There is some contradiction here which we cannot explain. 30 (Ed.} W.H. Bliss, Calendar of Entries in the papal Registers relating to Great Britain and ire/and: Petitions to the Pope, i, 192. Attention was first drawn to this source by R.C. Fowler, V.C.H. Kent, ii, 123. 8 THE SOUTHERN DEFENCES OF MEDIEVAL ROCHESTER time, therefore, it is possible that the monks had carried out at least some part of the plan proposed to the king. In addition - or even instead - the prior's statement could be taken to refer to the construction of a crenellated wall along the northern boundary of the precinct, between Eastgate and 'the gate of St. William. '31 The proposal to build such a wall, approved by the king in August 1345, seems certain to have been carried out according to plan. Though nothing is left above ground, a wall following this line has been located at several points, most recently close to Eastgate.32 Having thus taken steps to isolate themselves more thoroughly from the outside world, and from the rest of the city, the monks seem also to have been concerned to demonstrate their usefulness to the crown. In the 1360s, when the outer defences of the castle began to be extensively reconstructed, it was the prior, John of Hartlip (1361-80), who took responsibility for superintending the work. Similarly, in the 1390s, another prior John - the second John of Sheppey (1380-1419) - was made a member of the management committee 'apRointed to repair the walls, gates and turrets of the city of Rochester'. 3 The committee was headed by the constable of the castle, Sir William Arundell, who engaged to pay a proportion of the cost himself.34 Money was contributed also by the king, the Duke of Lancaster, and 'certain other lords'; apparently, the citizens had no share in the scheme. Here at least, the documentary evidence ( of which we were till recently unaware) can be correlated well with the evidence actually existing on the ground. We can now give a definite date for the reconstruction of Eastgate ( demolished no later than the beginning of the eighteenth century, but partly excavated in 1969), for the surviving stretch of wall to the north (retaining some of the original 31 Calendar of Patent Rolls, Edward lll, 1343-S, 539. The text was printed by J. Thorpe, Registrum roffense (London, 1769), 552, ostensibly from the original. 32 Harrison, Arch. Cam., lxxxvii (1972), 122-3. On the other hand, we can take it for granted that the monastery would always have been walled off from the rest or the city; and there are casual references to a precinct wall (not necessarily on the same line as the wall built in 1345) in a number of thirteenth-century documents: Hope. Arch. Cant., xxiv (1900), 26-7. 3·1 Calendar of Pate/I/ Rolls, Richard II. /396-9, 137. This document is a writ or aid, issued on 6 March, 1397, authorising the impressment of the workmen required. 34 Calendar of PatenJ Rolls, Henry IV, 1399-1401, 379. Arundell had promised to pay the contractors a sum of £88, 'for the making of a parcel of a wall around the town of Rochester'. Arter his death (in August 1400), there was some dissension over the raising of the money, and this document, dated 10 November, 1400, was intended to settle the matter. 9 C. FLIGHT AND A.C. HARRISON crenellation), and for the bastion at the north-east angle of the city.35 Without much doubt, the intention had been to- reconstruct the defences in their entirety - a plan which followed on, perhaps, from the rebuilding of the bridge in the 1380s - but there seems no evidence of any attempt to persevere with the work, after 1400. By the sixteenth century, the defences were derelict. This is clear from Leland's description, dating from the 1530s.36 Eastgate was mostly still intact,37 but the other gates had gone. Leland remembered seeing 'six or seven towers' - six or seven which 'yet remain', he says, as if implying that he thought there had once been more. From then on, the story is one of continuing dilapidation, punctuated by piecemeal demolition. (2) Archaeological Evidence: Deanery Garden Sector Turning to the medieval features revealed by the excavations in the Deanery Garden, it seems best if we discuss them first in sequence from north to south, rather than in chronological order. We begin, therefore, by noting the existence of a broad, flatbottomed gully, dug alongside the outer face of the Roman wall, to a depth which resulted in the partial exposure of the ·wall's foundations. 38 We used to believe that this gully represented the innermost part of the Deanery Garden ditch; but now, for a number of reasons, we agree that it ought to have been treated as a separate feature. Instead of strengthening the defences, it tended rather to weaken them, by undermining the face of the Roman wall; and the filling consisted of rubble and refuse, more or less horizontally bedded - quite different, that is, from the clean spoil, tipped in from the east or south, which forms the filling of the Deanery Garden ditch. We are inclined to interpret this gully as a drainage-ditch, meant to carry rainwater from the buildings around the cloister, and eventually discharging into the city ditch beyond the south-east angle of the Roman wall. 39 On this view, the gully must surely be later than the 35 Harrison, Arch. Cant., lxxxvii (1972), 130-1. The dating suggested there - late fourteenth-century - was based on considerations of style alone. For directing our attention to the documentary evidence, as cited in the last two notes, we have to thank H.L. Turner, Town Defences in England and Wales (London, 1970), 158. 36 (Ed.) L.T. Smith The Itinerary of John Leland (5 vols., London, 1907-10), iv, 45. 37 It is featured in a drawing of Rochester made by the herald William Smith in 1588 (and reproduced in Arch. Cant., vi (1866), facing p. 54). 38 'Defences', 61, 62-3. Though we cannot be certain, we continue to regard this gully as a linear feature, parallel with the Roman wall. 39 We take it that the water would have been flushed through the undervault of the latrine, presumably sited to the south of the dormitory, outside the Roman wall. There was certainly some sort of building on this site: Harrison, Arch. Cant., cii (1985), 266. 10 THE SOUTHERN DEFENCES OF MEDIEVAL ROCHESTER Deanery Garden ditch: it may well have been the decision to fill up that ditch which necessitated some reorganisation of the drainage system. The succession of deposits forming the fill of this gully is also in need of some comment. Above layers of dark silt and rubble (Cutting B, layers 6-7), we found a scatter of mortar debris (layer 5), derived without doubt from the demolition (partial or complete) of this particular stretch of the Roman wall. On the other hand, the quantity of debris is comparatively small: the bulk of it, we infer, must have been collected and removed from the site. Hence, it appears that the gully continued· to serve a purpose, even after the Roman wall had been (or begun to be) demolished. Some time later, however, the gully was deliberately filled with a mass of loose rubble and broken tiles (layer 4), evidently generated by reconstruction somewhere within the precinct. The whole area seems to have been levelled off at this stage,40 but still remained clear of buildings. No structures were found in any of our cuttings, though a soakaway pit, sunk into the surface of layer 4, appears to indicate the existence of a building close by. A trench dug obliquely across the middle of the Deanery Garden, when the new Deanery was built, intersected three.walls, probably medieval, about 28 m. to the west of Cuttings C-D.41 At the time of the Dissolution, it seems, the easternmost part of the precinct was occupied only by gardens.42 We ought to make it clear that the main assemblage of medieval pottery, stated to have come from the filling of the Deanery Garden ditch, comes in fact very largely from the filling of this gully. 43 It includes several cooking-pots with horizontally flanged rims, of a type which is thought not to have come into use till late in the thirteenth century. How long it persisted seems to be more uncertain; but fragments of one such pot were found in the filling of the soakaway pit just mentioned; and this pit, whichever scenario is right, cannot be earlier than the middle of the fourteenth century. As will hardly need to be said, we regret not having prolonged these cuttings towards the south, far enough to locate the outer edge of this feature, and also (more importantly) the inner edge of the Deanery Garden ditch itself. Cutting E, some 13 m. from the Roman 40 If we are right in thinking that this gully was dug as a drain, it seems necesary to assume that drainage was now diverted once again, presumably towards the south. 41 The features discovered in the digging of this trench were recorded at the time by R.E. Chaplin, who kindly put his notes at our disposal. 42 Hope, Arch. Cant., xxiv (1900), 59. 43 The pottery is described by P.J. Tester, 'Defences', 98-9. 11 C. FLIGHT AND A.C. HARRISON wall, is now the northernmost point at which we can claim to have identified the characteristic filling of that ditch.44 Though there must have been a berm of considerable width - not less than 6 m. - the exact figure is unknown. Further west, however, the refectory projects by more than 10 m. beyond the outer face of the Roman wall (see below); and this may give some clue to the width of the berm, regardless of whether the ditch is thought to have been dug before (scenario B) or after (scenario A) the construction of the new cloister. These changes in the interpretation do not affect our previous conclusions regarding the boundary wall which Livett believed to be of 'later-Norman' date.45 In fact, this wall was built across the filling of the Deanery Garden ditch. It is probably a post-Dissolution feature, though earlier than the blanket of rubble derived from the demolition of the monastic buildings. At all events, it cannot be 'later-Norman'. With the Deanery Garden ditch we can deal quite briefly, setting aside the question of its date. The deepest section was revealed, not by any cutting of ours, but by the digging of a soakaway pit in connection with the redevelopment of the site. 46 This shaft was cut down through the sloping layers of sand and gravel (here about 4 m. thick in all) which form the final filling of the ditch, and then into a deep deposit of dark silt (more than 2.50 m. thick), still without reaching the bottom. In our excavations, we found or came close to the outer edge of the ditch at several points, most notably in cutting H.47 Here again, we think, some reinterpretation seems to be called for. Considering the evidence again, we prefer to abandon the idea that the lower part of this section (layers 8-9) formed the filling of an earlier (Roman?) feature truncated by the digging of the ditch. More probably, these layers represent stages in the silting of the ditch itself, till the moment came when it was at last deliberately filled with spoil (layers 5-7), derived from the new ditch further to the south. On this view, the outer face of the Deanery Garden ditch would originally have been about 34 m. from the Roman wall, roughly 5 m. high, and very nearly vertical. A sheer face was also found in cutting M, one of the line of trenches we dug just inside the existing east wall. 48 The stratigraphy 44 'Defences', 63-4. 45 Livett, Arch. Cant., xxi (1895), 48-50. The wall was first located by Hope in 1886. 46 'Defences', 74. 47 'Defences', 66-7. 48 'Defences', 67-8. 12 THE SOUTHERN DEFENCES OF MEDJEV AL ROCHESTER sectioned by these trenches (Fig. 2) was generally similar to that already revealed in cutting H; but the edge of the ditch had swung outwards here, by about 4 m. From these sections, it is clear that the profile of the ditch had become very considerably degraded, even before it was filled with sand and gravel. Little was found in any of these deposits. The most significant item is a single sherd, of twelfth- or earlef thirteenthcentury date, from the layer of dark silt in cutting M.4 Beyond this, there is a strip of solid ground - solid, that is, except for a single gully, apparently of sixteenth-century or later date. We do not doubt that it was at least intended for a wall to be built along this line - parallel with the Roman wall, beyond the outer edge of the Deanery Garden ditch - as seems for certain to have happened (see below), further to the west. In the Deanery Garden, however, on the alignment indicated by Livett,50 there is no trace of a wall. Our cuttings failed to find it;51 so did the foundation trenches dug for the new Deanery. 52 At the time, we took this to mean that the wall could never have been built. That inference, as we realise now, was not entirely safe. It is possible that Livett, led astray by his probe, drew in the line of this wall too far to the north, perhaps by as much as 5-6 m. 53 In cutting M, as mentioned above, the outer edge of the Deanery Garden ditch occurs at a distance of 38 m. from the line of the Roman wall; and this seems almost to force us to think that Livett was wide of the mark. In that case, we cannot quite be sure that the wall did not exist - in the form of footings, or perhaps of a robber-trench- just outside the area available for excavation. There is a chance, after all, that the wall was duly built, but that this particular stretch of it was later demolished. We have a specific reason for dwelling on this uncertainty: if it could be proved, beyond doubt, that the wall was not completed, scenario B would promptly cease to be tenable. We had no opportunity to investigate the King's Orchard ditch, except at one point, in cutting M, where its inner edge was found, converging with the outer edge of the Deanery Garden ditch. However, the existing profile of this ditch - very wide, but relatively shallow - can hardly have resulted from natural erosion alone. At some stage, we suspect, the original edges were both cut back to a 49 'Defences', Fig. 17, no. 108. so Livett, Arch. Cant., xxi (1895), 62. 51 'Defences', 64-7. 52 'Defences', 74, citing Mr Chaplin's observations. 53 We arrive at this figure by extrapolating the line of the Prior's Gate wall (see below). 13 south+- ·--· ········ ·· ···· • .· . L ditch fill copin.9 \ \ 􀀁 , ' , ' l'=>th-c.uitw·y demolit-ion layer Section ➔noYt"h bric.k wall ·.-.,•.,·.-·:-·: ·:. ·: - :.·:-·:-·: .. :. ·:- · . . -· .. -•· .•· Elevation 10111 Fig. 2. Above: A Section through the Filling of the Deanery Garden Ditch, with the inner slope of the King's Orchard ditch to the south. Below: An Elevation View of the City Wall, from the outside, showing also some of the features revealed by excavation internally. (For the location of cuttings J-Q, see the plan in Arch. Cant., lxxxiii (1968), facing p. 57.) THE SOUTHERN DEFENCES OF MEDIEVAL ROCHESTER slope, and the spoil used for filling the bottom of the ditch. Our section (Fig. 2) shows just the beginning of the inner slope, overlain by a tip of seventeenth/eighteenth-century refuse. 54 Some distance to the west, within the King's School grounds, the construction of a new assembly hall allowed us to make some significant observations. 55 Over most of the site, the holes for the concrete piles which form the foundation of this building were sunk into the earthy filling of the King's Orchard ditch. The southernmost line of holes, however, was dug into solid gravel; and the same was true - or seemed to be true - for the westernmost hole in each of the other three rows. For that reason, we thought it 'virtually certain' that the ditch did not continue beyond this point - a conclusion which seemed to us all the more acceptable because we knew, or thought we knew, that the wall associated with this ditch had also been left unfinished. Here again, considering the circumstances in which these observations were made, there has to be some slight doubt; and the point is of greater importance than we recognised at the time. If we could be certain, absolutely, that the ditch was not completed, scenario B would not be worth considering. It is just possible, however, that the ditch - rather than simply coming to an end - was interrupted at this point, by some sort of causeway.56 In that case, we would have to assume that the further stretch of ditch, to the west of this hypothetical causeway, was filled up completely at some later stage. Whether finished or not, the King's Orchard ditch was eventually superseded by the latest line of medieval defences. This is represented by the wall which begins by running southwards from the Roman south-east angle, to a point lying well beyond the outer edge of the King's Orchard ditch. Here there is a ruined bastion, from which the wall returns towards the west. On the plan (Fig. 1) we call it the Vines wall. The latter stretch, relatively slight in construction, seems likely to be ( or to follow the line of) a pre-existing boundary wall, surrounding that part of the precinct which had lain outside the defences until this time. Besides, there seems no sign of a ditch associated with this wall.57 The only stretch which we had a chance to investigate was that which now delimits the Deanery Garden (Fig. 2). Two of the facts 54 At this point, therefore, the wall - if it existed - would probably have been destroyed. 55 'Defences', 74-5. 5<> Alternatively, there is a chance that the ditch had been filled with gravel, clean and compact enough to give the impression of being undisturbed. 57 We cannot see the 'signs of its having been begun' which Hope believed to exist: Arch. Cant., xxiv (1900), 22. 15 C. FLIGHT AND A.C. HARRISON which we discovered are worth repeating here. 58. First, where the wall is carried across the fiJling of the Deanery Garden ditch, we found that it was supported by wedge-shaped piers of masonry, sunk to some considerable depth, and linked together by crudely constructed arches. The same technique was used in rebuilding the eastern curtain of the castle, during the 1360s;59 but we would not wish to assert a similarity of date, merely on the strength of this analogy. Second, in cutting M, we found a ragged joint in the foundations, apparently proving that the stretch of wall to the south of this point had been built up against the broken end of an earlier wall to the north. We are not sure that any definite conclusion can be drawn from this feature, apart from the obvious fact that it implies the existence of more than one phase of construction. (3) Archaeological Evidence: Western Sector We take it for granted that the Deanery Garden aitch must have continued westwards, in the same relative position, all the way to Southgate. The edge of the ditch, though it has not been traced by excavation, is surely outlined by the medieval wall which once ran southwards from Southgate, for a distance of 36 m., before returning to the east on a line very nearly parallel with that of the Roman wall. We refer to it here as the Prior's Gate wall, in order to preserve neutrality. Except for one very short stretch, no part of the Prior's Gate wall survives above ground, but its line is easily visible, up to a point. Prior's Gate itself is agreed to be a later feature, built or rebuilt in the early fifteenth century;60 whether it occupies the same site as 'the gate of the said prior', known to have existed already in 1344, is a question on which we differ. To the east of this gate, the wall was still standing in the seventeenth century, 61 but seems to have been swept away in the 1720s. Livett claimed to have traced it with his probe, in the gardens behind Minor Canon Row; excavating across this line, about 54 m. from the gate, he and Payne found a robber-trench, or a feature they interpreted as such.62 From this point east, uncertainty prevails. In the Deanery Garden, the wall which ought to continue this line was either never built (in which case we have to prefer scenario A), or else built and later demolished. On either view, we agree, there must have been a time 58 'Defences', 67-8. 59 Flight and Harrison, Arch. Cant., xciv (1978), 32. 60 Hope, Arch. Cant., xxiv (1900), 53-4. 61 Hope, Arch. Cant., xxiv (1900), 75, citing a survey of the buildings taken in 1647. 62 Livett, Arch. Cant., xxi (1895), 62. 16 THE SOUTHERN DEFENCES OF MEDIEVAL ROCHESTER when the Prior's Gate wall did not continue eastwards as far as this. Since it cannot simply have come to an end, we suggest that it must have been made to join up with another wall, running from north to south. A wall which suits this hypothesis does exist, its position being marked on the plan (Fig. 1). In its present form, this wall is admittedly a post-Dissolution feature;63 but we find ourselves obliged to assume that the alignment is medieval. 64 At some stage, we suggest, a wall was laid out to run southwards from the dormitory, dividing the outer precinct into two unequal parts. The site of the King's School assembly hall, mentioned above, lies just to the east of this line. Without much excavated evidence, we know very little about the pattern of structural and stratigraphic relationships in this western sector. Here, too, a stretch of the Roman wall has been demolished - from the north-west corner of the refectory, westwards as far as Southgate - and this, we think, may prove to be the crucial fact. At the time when the Deanery Garden ditch was dug, the Roman wall, it seems safe to assume, must still have been intact. By dating the demolition of the wall, we therefore ought to obtain a terminus ante quern for the digging of the Deanery Garden ditch; and, ifit turns out that the wall had been swept away before 1225, scenario A would be out of the question at once. For the moment, however, the evidence is less than conclusive. As we have said, the surviving stretch of the Roman wall comes to an end at the north-west corner of the refectory. There may be some significance in the fact that it finishes here, rather than slightly further west, at the south-west corner of the cellarer's building. From the documentary evidence, we know that the refectory was one of the buildings erected by Bishop Ernulf, soon after 1115: he is credited with building the eastern and southern ranges of the cloister, but not the western range. On the other hand, there is a charter of Prior Ralph de Ros, dating from the 1190s, which appears to prove that the whole of the cloister, by this time, had been 'completed in stone'. 65 Given these indications, it might be argued that the cellarer's building, at first presumably of wood, was later replaced by a permanent structure - after Ernulf's time, but before the end of the century. The southernmost part of the cellarer's building, excavated in 63 Livett, Arch. Cant., xxi (1895), 71-2. 64 The same assumption is tacitly made by Hope, Arch. Cant., xxiv (1900), 65-6. 65 Hope, Arch. Cant., xxiv (1900), 30. 17 C. FLIGHT AND A.C. HARRISON 1938, remains more or less exposed.66 Some of the vaulting-shafts are still in place: they have scalloped capitals, of a common twelfthcentury type. Unlike the north wall of the refectory, the south wall of the cellarer's building does not consist of a stretch of the Roman town wall. On the contrary, it seems not to have been built till after the Roman wall had been torn down. Thus, by dating the cellarer's building, we appear to obtain a terminus ante quern for the demolition of the Roman wall, and, by extension, for the Deanery Garden ditch. Unfortunately, the monastic buildings are known to have undergone extensive reconstruction in the 1330s. The refectory is mentioned explicitly.67 How far this altered the picture, we cannot be sure; but without much closer investigation - including further excavation - the evidence seems to be less than completely convincing. A building referred to as the 'long bakehouse' is also said to have been repaired by Bishop Hamo, at the same time as the refectory.68 Its location is unknown, but Hope thought that it might have stood 'upon part of the site now occupied by Minor Canon Row'. He also thought that this bakehouse might be the same building as the one of which casual mention is made, in a few twelfth-century documents.69 If so, we would have to allow for the survival of a twelfth-century building, outside the Roman wall, on the line of the Deanery Garden ditch, as late as the 1330s. It would be hard - if not impossible - to reconcile this fact with scenario A. In fact, quite apart from the doubt regarding its location, we cannot feel sure, by any means, that the building repaired in the fourteenth century was the same bakehouse existing back in the twelfth. Further west, in the area of the bishop's palace, the rampart-bank behind the Roman wall had been levelled off by the twelfth century; and the wall itself had been demolished, down to the same horizon, by the mid fifteenth century at the latest (but possibly long before).70 66 The only account which we have been able to trace occurs in Friends of Rochester Cathedral, Annual Report, iv (1939), 20-2. A brief note of the visible features can be found in Newman, West Kent, 48Cr7. 67 Hope, Arch. Cant., xxiv (1900), 49-50. It is clear nonetheless, that the north wall at least was retained: the entrance doorway, which still survives, is an early thirteenth-century insertion. 68 Hope, Arch. Cant., xxiv (1900), 52. 69 The twelfth-century references are collected by Hope, Arch. Cant., xxiv (1900), 9. He does not explain his suggestion regarding the site of the 'long bakehouse': perhaps he thought it might be the same building as 'the long gallery called the Cannon Place' mentioned in a document of 1588 (ibid., 75). The comments by T. Tatton-Brown, Arch. Cant., c (1984), 18Cr8, do not seem helpful to us. 70 Harrison and Williams, Arch. Cant., xcv (1979), 22-4. 18 THE SOUTHERN DEFENCES OF MEOIEV AL ROCHESTER One range of the medieval palace is still in existence, though very heavily modified. Two small windows and part of a third are the only surviving details: they were thought by Livett to be fifteenth-century insertions.71 The east wing, of which only the foundations remain, can be dated to the fifteenth century with some confidence;72 but the main block may perhaps be an earlier structure, to which this wing was attached - and from which it was afterwards torn away again. The north wall of the surviving block runs just alongside the stump of the Roman wall. Quite certainly, therefore, the Roman wall had been razed to the ground before the palace was built. (In fact, the builders of the palace were apparently not even aware of its existence - or else they would surely have taken advantage of it, as a ready-made foundation.) A date for the construction of the palace would accordingly give us a terminus ante quem for the demolition of this stretch of the Roman wall. Here again, however, we fail to reach a conclusion. We take it for granted that Bishop Gundulf (1077-1108) would have built himself a suitable residence, somewhere near the newly constructed church; but neither of us thinks that the eleventh-century palace stood on this site near Southgate.73 Beyond that point, we begin to disagree. One of us thinks it likely that the palace was built here first by Bishop Gilbert de Glanville (1185-1214).74 The other prefers to regard the whole of the palace - not just the east wing - as fifteenth-century in date, until there is stronger evidence to the contrary. Once again, we have to conclude, the evidence is sadly ambiguous as it stands. (4) Alternative Interpretations We are now ready to turn to a consideration of the rival scenarios, explaining how we would propose to interpret the information presented above, given the initial assumption appropriate in each case. As we indicated at the beginning of this paper, the difference depends on whether we assume that the ditch dug in 1225 is represented on the ground by the Deanery Garden ditch (scenario A) or by the King's Orchard ditch (scenario B). Some of the consequences which follow from this initial choice can be seen by 71 Livett, Arch. Cant., xxi (1895), 43. 72 Hope, Arch. Cant., xxiv (1900), 60; Harrison and Williams, Arch. Cant., xcv (1979), 26. 73 There is, as we said before ('Defences', 77-8), no evidence at all for the 'early-Norman extension' postulated by Livett. 74 With the I.inked suggestion that the old palace, or part of it, was then converted into the prior's lodging, first mentioned at around this time: Hope, Arch. Cant., xxiv (1900), 58. 19 ipg r--- l --.... ____ _ --􀀄 I I '-~--------___ ) A -·- .. ··--::•--- B Fig, 3. Hypothetical Plans showing the State of the Defences on the South Side of the City, after 1225, according to scenarios A and B, respectively. Cd = Castle ditch. DGd = Deanery Garden ditch (A only), BHd = Boley Hill ditch, KOd == King's Orchard ditch (B only). Sg = Southgate (alternative sites), Pg = Prior's Gate (alternative sites), Ag = Almonry gate, Sxg = Sextry gate, Eg = Eastgate THE SOUTHERN DEFENCES OF MEDIEVAL ROCHESTER comparing the respective plans (Fig. 3), which show the hypothetical state of the defences, just after 1225.7 This is the point at which the two scenarios diverge to the greatest extent. Scenario A On this view, the Roman wall is assumed to survive, more or less intact, till 1344. During the twelfth century, the monks began to encroach upon the existing fortifications. They removed a considerable stretch of the rampart-bank; they erected buildings of their own up against the wall, externally as well as internally; and they must have cut passages through it, here and there, in order to have access to the outer part of the precinct. Nonetheless, in 1225, when the king's engineers drew up their plans for fortifying the city, the Roman wall was still very largely complete. Whatever damage it had suffered was now made good; and a vast new ditch was dug, paralJel with the wall, but some distance away from it. Because of the width of this berm, the monastic buildings which had been put up against the outer face of the wall did not need to be demolished. They were left to project - 'like towers', as we said before76 - from the newly refurbished wall. Other buildings, further out, are likely to have been destroyed; but the monks presumably erected replacement buildings, sited inside the defences. A gate was inserted in the line of the Roman wall, somewhere to the west of the refectory, probably at this time. 77 It was certainly in existence by 1344, which is when we find it referred to as 'the gate of the said prior'. After 1331, as part of the programme of reconstruction inspired by Bishop Hamo, the refectory was thoroughly rebuilt, but still in the same position as before. By this time, the system of fortifications established in 1225 had already been allowed to fall into disrepair. Some parts of the wall had collapsed, as the monks reported to the king in 1344; and the ditch had reached a badly degraded state. The monks were accordingly given permission to fill in the ditch - and, by implication, to pull down the rest of the wall - provided that they 75 We take it for granted, in either case, that Boley Hill was part of the thirteenth-century fortifications, as we have suggested elsewhere: Flight and Harrison, Arch. Cant., xciv (1978), 39. 76 'Defences', 78. 77 The most likely site, we think, would be just to the west of the refectory, on a line with the existing gate. This is an admissible interpretation of the dimensions quoted in 1344 (see above, n. 27), if we assume that the length was measured along the outer edge of the ditch. 21 C. FLIGHT AND A.C. HARRISON would undertake the construction and maintenance of a new line of defences, outside the existing ditch. For some reason, the work was never properly carried out.78 The Deanery Garden ditch was filled with gravel - not from Eastgate southwards, however, but only from the south-east angle westwards. The King's Orchard ditch began to be dug, the Prior's Gate wall began to be built - but neither seems to have been finished. By 1350, when the prior claimed credit for having surrounded the monastery with a wall, the work had presumably been brought to some temporary conclusion, perhaps with the construction of the north - south wall dividing the outer precinct. The monks never tried to complete the original scheme. Instead, at some later date, they built a wall southwards from the south-east angle, out across the unfinished King's Orchard ditch, to the point where it intersected with an earlier boundary wall. There is no documentary evidence relating to this phase in the sequence of construction. 79 It may perhaps have been connected with the rebuilding of the city wall to the north of Eastgate, at the end of the fourteenth century. Other features mentioned above - the bishop's palace, for example, and the possible drainage-ditch alongside the Roman wall - can all be accommodated well enough within this outline scenario. Scenario B By 1225, in the situation envisaged by scenario B, it was already much too late to think of refurbishing the Roman wall. With buildings abutted against it on either side, the wall had long since ceased to be defensible; and part of it, to the west of the refectory, had probably been demolished. On this reading of the evidence, there would have been no point in digging the Deanery Garden ditch as late as 1225. It has to be earlier than that - earlier even than c. 1120, by which time the Roman wall was already out of commission. On the other hand, we would think it safe to assume that this ditch is a later feature than the ditch around the castle, with which it compares in scale; and the castle ditch, we have argued elsewhere, was probably dug soon after 1066.80 By this 78 Perhaps because of the Black Death, as we suggested before: 'Defences', 80. 79 According to Hope, a strong wall surrounding the vineyard was partly repaired in 1384-5: Arch. Cant., xxiv (1900), 66, apparently citing the cellarer's account for that year (which survives in the form of a transcript made for Thorpe). We doubt whether this is relevant. 80 Flight and Harrison, Arch. Cant., xciv (1978), 38. 22 THE SOUTHERN DEFENCES OF MEDIEVAL ROCHESTER argument, the Deanery Garden ditch would have to be earlier than c. 1120, by some significant margin, but later than c. 1070. There is no mention of a ditch in the documents relating to the parcel of land, 'next to the wall', confirmed by the king to Bishop Gundulf soon after 1088. 81 That transaction was occasioned by the banishment of Earl Odo, following the 'war of Rochester' - that is, the six-week siege of the castle which marked the end of his rebellion. In the course of this siege, the monks would have found themselves uncomfortably close to the firing line; and the church itself is known to have suffered some damage.82 Not much later, the castle was reinforced with a stone-built curtain wall, the construction of which is said to have been paid for, with much reluctance, by Gundulf.83 At around the same time, we suggest, an attempt might have been made to fortify the monastery, on the outward sides at least, by digging the Deanery Garden ditch. Twenty or thirty years later, the monks' priorities had changed. All this time they had had to make do with the wooden buildings put up for them in the 1080s. Under Bishop Ernulf, fine stone buildings began to be erected, outside as well as inside the line of the Roman wall. In the process, a long stretch of this wall became incorporated into the monastic buildings - the refectory to the west, the infirmary complex (or so we suppose) to the east. Its survival was thus guaranteed. A shorter stretch, further west again, is likely to have been demolished at this stage. The Deanery Garden ditch, we think, would also have been treated differently in different sectors. Towards the west, it must have been deliberately filled, so as to provide a site for the kitchen, bakehouse and other buildings associated with the refectory. Further east, however, the ditch remained open, silting up in a gradual way till 1225. The explanation for this must lie, we suspect, in the need to provide for drainage fr9m the monastic buildings, especially from the latrine. Thus, though the wall and the 81 It may be mere coincidence, but the area taken up by the Deanery Garden ditch, from Southgate to Eastgate, amounts to roughly three acres- the same area, that is, as the land given by Odo. 82 In another charter issued at about this time, the king apologises for the damage he had caused, 'albeit unwillingly', while conquering his enemies: Textus roffensis, f. 211. 83 The only basis for this statement is a tendentious piece of narrative occurring in the Textus roffensis, ff. 173-174v, where Gundulf is said to have built the castle 'entirely in stone', at his own expense, in return for the king's agreeing to the grant of Haddenham (which Archbishop Lanfranc had given to the church of Rochester, 'for the sustenance of the monks'). Though the story is told in a style which invites suspicion, we assume it to have some truth. The Haddenham charter (Textus roffensis, ff. 212-213), seemingly genuine, can be dated to the summer or autumn of 1088 - that is, soon after the siege - because of the presence of Prince Henry among the witnesses. 23 C. FLIGHT AND A.C. HARRISON ditch had both already lost their defensive function, they had both been partly converted to serve new purposes. In 1225, the monks were obliged to surrender a broad strip of land, beyond the silted-up but still visible ditch, so that the king's engineers could carry out their plans for refortifying the city. This is when the King's Orchard ditch was dug; this is when the Prior's Gate wall was built. We think it has to be assumed - in spite of some indications to the contrary - that the wall and the ditch were both successfully completed according to plan. On this scenario, there is no objection to the view that part of the east wall (as far south as the joint in cutting M) was built in 1225.84 The wall returning to the west, though we failed to find it, must exist, or must once have existed, south of the line where we were looking for it; and 'the gate of the said prior', mentioned in 1344, can be taken to have stood on the same site as the gate now called by that name. 85 Inside the new wall, the Deanery Garden ditch had now been completely filled. Because of this, a new arrangement would have been needed for the disposal of water and sewage; and that may explain why another ditch - relatively small, and not defensive in function - was dug at around this time, alongside the Roman wall, in the strip of relatively solid ground which had once been the berm of the Deanery Garden ditch. After this drainage-ditch had silted up to some extent, a second stretch of the Roman wall was doomed to demolition, westwards from the original south-east angle. A little later, the area behind the Roman wall was finally levelled off. These changes in the topography of the site, we suggest, are likely to be connected with the programme of reconstruction begun in the 1330s. Once they had finished the refurbishment of their buildings, the monks produced an ambitious scheme for realigning the defences of the city. In 1344, after due inquiry, the king approved the plan. What actually happened, however, bore rather little resemblance to what had been proposed. The westernmost section of the King's Orchard ditch - assuming that this had been dug in the first place - was levelled off completely. Further east, the profile of the ditch may have been deliberately altered at this time (or possibly not till later), by cutting back the edges on either side; the corresponding stretch of the Prior's Gate wall was also apparently demolished. A new wall was ·84 Livett and Hope were both brave enough to indicate a conjectural bastion at the new south-east angle. If this existed, it must have been sited a little further south than they supposed; and all trace of it is likely to have been destroyed, in or after the fourteenth century, when the inner edge of the King's Orchard ditch was cut away. 85 This would imply that the length of the ditch, as measured in 1344 (see above, n. r!), was the distance along the. wall. 24 THE SOUTHERN DEFENCES OF MEDIEVAL ROCHESTER built, running southwards from the south-east angle established in 1225, and the defences of the city were thus made good, if only after a fashion. The monks went on to build a dividing wall, southwards from the dormitory, separating the western sector from the more secluded area to the east. We take it that drainage was now diverted southwards, into what was left of the King's Orchard ditch. These rearrangements had all presumably been completed by 1350. At this point, the plan becomes the same that we have postulated for a somewhat later date in scenario A. Conclusion In the nature of the case, we are left with little to say by way of conclusion. It is distressing that the evidence should lend itself, even now, to two such widely divergent interpretations. We are confident, nonetheless, that the issue could be resolved by further excavation; and we do not doubt that it will be, sooner or later. Our disadvantage was that, digging where we could, at the bottom of the Deanery Garden, we were outside the area once covered by medieval buildings. Further west, almost anywhere, the question might be answered very quickly. For the moment, we are left with our two alternative scenarios, neither of which is altogether convincing. We agree in recognising the objections on either side; where we differ is in deciding which objections are of greater weight. Scenario B, despite its plausibility, does not conform with the facts - with the facts, at least, as we understood them to be. It seemed clear to both of us in 1968 that neither the Prior's Gate wall nor the King's Orchard ditch had ever been completed. Unless we were mistaken, scenario B is wrong. Scenario A, on the other hand, though closer to the facts as known (or thought to be known), appears to carry with it some less acceptable consequences. It prevents us, perhaps, from arriving at a satisfactory interpretation, correlating the expansion of the monastery with the evolution of the city's defences. Is it likely, for instance, that the twelfth-century cloister would have been laid out as it was, if the Roman wall had still been regarded as having a defensive function? When the king's engineers set to work in 1225, would they really have been willing to countenance an arrangement in which the monks' refectory ( and perhaps the latrine) were left standing against the city wall, on the outward side? Even if the monks, for their part, were ready to put up with this, some of their other buildings must surely have been destroyed, when the site was cleared for the digging of the ditch. Would it still have been possible to find room for 25 C. FLIGHT AND A.C. HARRISON replacement buildings, within the newly refurbished Roman wall? Finally, with the site as cramped as this, does it make sense for the monks to invest in repairs to the existing buildings, in the 1330s, and only later, in the 1340s, to ask for permission to realign this sector of the city's defences? We can only end by repeating the original question. On the south side of the city, there are these two great medieval ditches: the Deanery Garden ditch, which seems clearly to follow the line of the Roman wall; beyond this, the King's Orchard ditch, which is plainly more recent than the other. Which ditch is the one that was dug in 1225? Though we cannot agree on the answer, we agree in thinking that the question has some importance. The digging of the ditch is part of a cluster of events, during the 1220s, which add up one of the most significant changes in the whole history of Rochester. Before this, the topography of the city had been dominated by the two installations which advertised, in a visible and very solid form, the impact of the Norman conquest - the castle, of course, and the cathedral with the monastery attached to it. By straddling the Roman wall, as if deliberately, both these installations appeared to have breached the city's integrity - the castle at once and violently, the monastery only later, and more by stealth. From 1127, the castle belonged to the archbishop of Canterbury, who already had the bishopric in his pocket; the city itself was administered by the king's officials. In the 1220s, however, the picture changed. Possession of the castle (which John had had to recapture in 1215, after it had been surrendered to the rebel barons) had clearly reverted to the crown, before repairs began in 1221. Over the next few years, the castle became, for the first time, the centre of a system of defences which also encompassed the city. These fortifications would thus have expressed a new configuration of political relationships - a partnership of sorts between the king, represented on the spot by the constable of the castle, and the citizens of Rochester, under the rule of a bailiff elected by themselves. It was the monastery now which found its integrity breached: the monks were left in a visibly ambiguous position, half inside, half outside the city. Considering the importance of these changes, there has to be some interest in trying to understand exactly how they inscribed themselve􀁠 - so to speak - into the topography of Rochester. 26