Archaeologist Alan Ward reviews the archaeology of the Cathedral Precinct. Featured in The Hidden Treasures, Fresh Expressions Project Archaeology Report, Keevill Heritage 2021.

The area to west and south of the cloister has, unsurprisingly, long been known as 'The Precincts', but in this report to make identification of buildings, both standing and buried, more intelligible the area has been divided up by giving different parts of the access route their own road names. On the west, 'Prior's Gate road', south 'Minor Canon Row road' and further south King's School road and on the east 'Archdeaconry road'. The area of the precincts to the east of the latter has here been called King's Orchard and Deanery Garden their correct titles and the area to the north is that between the church and the High Street, including the site of the Prior's Lodge, usually referred to as the Old Deanery.

The precinct to the north and north-east of the cathedral

Location: Basically we know next to nothing about buildings in these areas. The only building we know anything about, and then not very much, is the Prior's Lodge or as it was called in the seventeenth century the Old Deanery.

Documentary evidence: This northern area, sandwiched between the church and the High Street, is somewhat narrow and was made narrower by there being a precinct boundary wall set back some 4.50m from the High Street frontage. Probably our first document for this area is that of the time of Edward III (Hope 1900, 22-23), for in 1345 he granted a licence to build and crenellate a stone wall from the city East Gate to the Priory Gate of St. William. However, an earlier document (DRc/T 290) dated to between 1283-1291 tells us that,

'Richard de Rof, clerk, brother of Solomon de Roff, clerk to Thomas prior of Rochester gave a piece of land 42 long x 17 wide in the City of Rochester lying under the wall of the priory on the north and the main road on the other side.'

This statement is implying there is land, presumably without a building (or buildings), between the High Street and an already existing north wall of the priory. The presence of St. William's Gate by the late thirteenth century and the mention of this wall suggest, hardly surprisingly, that the precinct was shut off from the town by a wall prior to that constructed in 1345. Admittedly there is no way of knowing whether or not the circuit was ever complete.

Of St. William's Gate, other than its position, we know nothing. It may have simply been a gated gap in the wall. Another gate may have existed further to the east next to Phelip's Lodge (82 High Street) where a narrow alley still exists.

Chertsey's, College or more correctly Cemetery Gate at the north-west extremity of the precinct is a fifteenth century structure still standing to full height. As far as AW knows this building has never been the subject of a building survey. Presumably it always had a timber framed, rather than stone built, chamber above the gate passage. There is obviously no point in having a gate without it being connected to a wall or, as now, buildings. Initially a precinct wall is assumed, and this would have extended southwards and eastwards. Admittedly there is no scar of a contemporary boundary wall on the south side of the gate and even the scars of the timber framed buildings that existed down to the late nineteenth century have weathered away. Presumably there was a gate here from at least c.1225 when North Gate (formerly Pump Lane) was created (Ward 2005, 2011 in preparation b and c) and the present writer suggested such was there from c.1090 (Ward 2017), only a few months later to find there had been a gate, probably on this spot, since before 762 (Brooks 2006, 11).

There is of course no point in having a gate unless it allows access to a road, alley or courtyard. As the west front of the cathedral is just 50m or so to the south it is a reasonable deduction that there was a road leading up to the main door from the late eleventh century. It may not have been present before that date, for the simple reason the west front of the Anglo-Saxon cathedral was further to the west adjacent to the main north to south route, Doddingherne Lane, through the town.

We have two documents which mention land belonging to Geoffrey de Cok (Geoffrey the Cook) and being given to the priory. From the measurements given it would appear that two separate plots are involved. One messuage was 52ft x 18ft and lay in Doddingherne (DRc/T 278 dated to 1215-1226) and the other concerned land,

'lying the length of the burial ground near the road, having in longitude from Godfrey Cocs great stone house E-W 68ft. and in (a measurement may have been left out here by AW) latitude from the road to the priory burial ground' (DRc/T 289 dated to 1225-1239).

We can see for the latter, and perhaps deduce for the former, the general area as to where these plots were situated. However, from the measurements given neither plot of land can be accurately positioned. Hope regarded the former as being between the Cemetery Gate and St. William's Gate (1900, 25), unfortunately no reason for this positioning is given.

The area to the north of the nave was always the lay cemetery whilst that to the east of the Sextry or Deanery Gate was the monks cemetery. The lay cemetery, on land with the name of Green ChurchHaw, supposedly took up nearly all the land as far as the High Street, including the site of the early fifteenth century Church of St. Nicholas. From west to east it extended from Doddingherne lane to a wall extending from the north-east corner of the north transept up to St. William's Gate (Hope, 1900, 23). As it can be shown (Ward 2011a, 2015, 2017) that Doddingherne Lane was situated some 15m further to the west than realised by Hope and, indeed everybody else, including AW, until recently and as no burials have ever been found in all the service trenches adjacent to the King's Head Hotel, the supposition that the medieval cemetery stretched us far as the lane should be dismissed. It is also unlikely the cemetery reached as far as the High Street for the frontage was a far too valuable rent commodity.

Within the precinct, 'Gundulf's Tower' has been discussed at considerable length (see above, pages 16-19 and 26-27) and not a great deal more is stated here. It date of construction varies depending on whom your read. McAleer's date of c.1070 was always pretty ridiculous and the present writer has shown that it is later than the crypt (Ward 2002). A date somewhere in the first two decades of the twelfth century appears to be that favoured by Tim Tatton-Brown and Colin Flight and there is no reason to doubt that. The present writer is pretty sure that he has seen shelly mortar used as its bonding matrix, but Tim Tatton-Brown thought not (pers. com. 2016). If shelly mortar was present that would tend to reinforce an early twelfth century date.

An excavation in 1973-74 in the garden to the east of the fifteenth century Sextry Gate may have found the floor of a Roman building or evidence of earlier medieval buildings.

Phelip's Lodge: As far as the writer is aware Phelip's Lodge is the only building to have had detailed building surveys undertaken (Austin 1997; Bacchus 2010) and both agree on the fifteenth century date for the timber framed structure hidden behind post-medieval refacings.

Archaeological evidence: Very little meaningful excavation has taken place in the northern precinct area. In 1887 part of the precinct wall was observed (Arnold 1889, 201) on the lawned area now taken up by the war memorial and again in 1894, a few metres to the west (Livett 1895, 57). The east end of the wall up against the Roman town wall, was seen in 1969 (Harrison 1972, 122). Assuming these are both parts of the 1345 wall there must be a dog-leg somewhere along its course and as David Bacchus points out the wall must pass to the north of Phelip's Lodge (2010, 214). It is just possible that a c.3m length of Ragstone wall forming part of the west wall of 78 High Street, represents that dog-leg. That it is of Ragstone, whilst the rest of this wall is of brick, certainly creates suspicion that there is medieval masonry here. Another stretch of wall was apparently seen in the arches of the cellar of 86 High Street in 1959 or 1960 (Chaplain 1961, lxxiii). The arches must be to the rear of the cellar and hence the walling seen is assumed to be on the line of that seen in 19769 coming off the east face of the town wall. The rear of the cellar is (probably) several metres to the south of the wall line seen in 1887 and 1894.

In 1990 the present writer observed trenches being dug for services in the yard immediately north of the St. Andrew's Centre and in 1992-93 in the rear garden of Phelip's Lodge (Ward 1996, Figures 6 to 8 in that report). A short length of wall over 0.50m wide was seen on the east side of the trench in the garden of Phelip's Lodge and this same wall has been picked up again in a 2017 project undertaken by Graham Keevill to the rear of 84 High Street. That this wall is medieval there can be no doubt and it has to be earlier than the fifteenth century timber framed building. An architectural fragment of fourteenth of fifteenth century date was removed from the wall in 1990 and the wall cut through soil producing pottery of the period c.1275 to c.1350. From this soil, fragments of Purbeck and Onyx Marble and at least one fragment of green 'serpentine' marble were also recovered. Admittedly a fragment of brick was also recovered, but this could easily have been intrusive from a higher level. What is regarded as an internal partition wall came off at a right angle to the west. Of this wall a length of c.3m survived.

Below the modern southern boundary of Phelip's Lodge a post-medieval stone lined cess pit was discovered, but other than a few bricks used in its construction no dating evidence was found.

The 1990 a 2.80m deep soakaway trench was dug in the St. Andrew's Centre car park. Natural Brickearth was reached at 5.65m above Ordnance Datum with a pit cut descending a further 0.40m. The base of the pit was not seen. From its fill late Iron Age and early Roman pottery was recovered. The pit was overlain by dark brown clayey loam top-soil the top of which was at about 6.70m above O.D. Two layers of dark earth were seen above, only distinguished by there being a stake hole 0.50m deep infilled with purple ash cutting the lower layer. The upper layer of dark earth was sealed by a 0.10m thick deposit of levelled, compacted gravel presumably forming a courtyard surface. This was almost certainly associated with a wall of Ragstone with some flint and chalk bonded by a sandy mortar. This stonework survived for a depth of 1.57m, the lower 0.60m being below the gravel. There was no obvious cut going through the gravel and if that surface and the masonry go together that means that about 1m of upstanding wall survived. However, the masonry could have been trench built, the cut through the gravel not creating any damage. The impression gained in regard this undoubted medieval masonry was that it was either free-standing or a buttress to a wall situated to the south or west.

A 0.40m wide north to south aligned wall on the east side of the trench was presumably associated with the Old Deanery, possibly its east wall as shown on Hope's plan, Plate V (Fig. 6). A further wall, also perhaps is forming part of the Old Deanery was observed by Arthur Harrison in 1983 when a drain trench was dug (Harrison 1985, 265-266). This wall, again 0.40m wide, was cut through by the drain trench, Due to what is considered to be its narrow width it was regarded as a garden boundary wall. It may of course have supported a timber framed building, but its position 2,15m in front of the Deanery (now the St. Andrew's Centre) would suggest it is to the south of the Old Deanery structures as depicted by Hope on his Plate V.

As far as this writer is aware these are the only observations made in this part of the precinct.

The precinct to the east of the cloister

The east range itself has been mentioned above (pages 58 and 59) and will be returned to in more detail below (pages 83 to 85). For the infirmary range situated further to the east we know very little and even less for anything across what is now a lawned area, the Deanery Gardens.

Documentary evidence: As with so much of the cathedral, documents in relation to the buildings to the east of the cloister are notable by their absence. Hope tells us that whilst the crypt was being built (c.1185) Heymeric of Tonbridge, 'made the cloister towards the infirmary'. Such a date is becoming a bit late for such a buildings and its is perhaps more likely to be a refurbishment or rebuilding of a previously existing infirmary cloister. At about the same time the sacrist Osbern of Sheppey afterwards prior (?1186 to 1199) made a lodging for himself beside the infirmary. This was presumably whilst he was still the sacrist, but this structure was probably that which became the Prior's Lodge and later the Old Deanery.

For a decade or two after 1540 the former infirmary range to the east contained the royal chapel, presumably the old infirmary chapel, and was probably used as the Queen's Lodging, the queen in question being Catherine Howard. However, the information in regard the royal palace, as far as it is known, is mentioned below (pages 81 and 82) in relation to the cloister.

A rere-dorter existed in the area to the south or east of the dorter. We are told in a document of 1177 / 1178 that Prior Silvester, 'removed the privy which formerly adjoined the dorter' (Hope 1900, 45). The question has to be where did he move the privy to? It certainly would not have been too far distant. Perhaps it was built spanning the defensive ditch.

Archaeological evidence: To the east of the dorter range the infirmary was situated. In 1983 Arthur Harrison may have found one of the walls of this building. A Ragstone rubble wall was found 1.40m below the modern ground surface. The wall found was 3.40m long and stood to a height or 2m. Whether this 'height' was all upstanding wall or foundation or a mixture of both we are not told. It was considered possible that this was a wall of the rere-dorter, but subsequent excavations in 2010 by Graham Keevill on the line of the Roman town wall to the north of Harrison trench make it more likely that this was the south wall of one of the infirmary ranges. The 2010 excavation saw four trenches dug, two either side of the Roman town wall. Quite by chance, at the west end of the length of wall in both the north and south trench detail for a door into the building was found (Keevill and Underwood 2010b).

The c.2m high mound situated within the Dean's Garden immediately to the north of the Roman wall is almost certainly a post medieval 'prospect or viewing mound' for the Dean and his guests'.

The precinct to the south of the cloister

The stretch of road between Minor Canon Row and Garth House is referred to here as 'Minor Canon Row Road', officially it is just referred to as The Precinct. This area is dominated by the finding in 1998 of a 59m long medieval building immediately below the southern pavement (Ward 2002, in prep. c). This structure was first observed by Greville Livett in c.1894 who identified fragments of wall at both the east and west end of this area. The latter was seen at the junction of 'Prior's Gate road' and 'Minor Canon Row road' and may be too far to the north, into that junction, to be part of this buildings (Fig. 16). He regarded this masonry as being the north-west corner of a building butting up to the town wall. If that is correct it must have been a structure aligned at a right angle to and joining the long building. There was certainly a building directly opposite coming off of the north-west corner of the Prior's Gate. Tim Tatton-Brown tells us (but with no reference for his statement) this structure was the medieval almonry (1984b, 187) and which, c.1540, became the first King's School. It was demolished c.1840. The length of wall Livett uncovered at the east end of Minor Canon Row he regarded as being a later Norman defensive wall. In fact he had found the south wall of the long structure (Livett 1895, 49).

Documentary evidence: In 1331-42 wide ranging rebuilding work was undertaken within the monastery by Hamo de Hythe (Bishop 1319-1352), most notably the refectory but also the 'long bakehouse' (Hope 1900, 49, 52). The latter structure has never been identified (Hope 1900, 52; TattonBrown 1984b, 187; Harrison and Flight 1986, 18), but it seems a perfectly reasonable deduction to equate this long structure (leaving off the east end, which is later in date) with that building.

A 1588 survey refers to a structure called the Canon Place or the 'long gallery'. At the Dissolution this building had been converted into rooms to house the minor canons of the new Dean and Chapter. In a later survey of 1647 what is almost certainly the same building was called Canon Row when, “all the long row of buildings within the wall, consisting of eighteen several low rooms and five upper ones, in which divers old and decrepit poor people inhabit, that did belong to the cathedral church” (Hope 1900, p.75; M.A.O. DRc/Ac 4/10 pages 61a and 64a; DRc/Esp 1/1 to 1/5). The eighteen rooms at ground level mentioned in this survey may represent the medieval bays of the undercroft and partitions might have been constructed within them.

Both surveys refer to a long building and tell us, or at least imply, that at one time the building had been used to house the minor clergy. The place-name evidence from the sixteenth through to the seventeenth centuries (petty canons houses, Canons Place, Canon Row, Minor Canon Row) in itself does not mean very much, for names can be adopted by later buildings not on the same site as the original. However, a further document, makes the position of the medieval building and its name certain. In 1605 the minor canons houses are described as being to the north of the great wall dividing it from the Hoghawe (M.A.O. DRc/Ele 120/1) and we know that the Hoghawe or Hoghaugh was on the site of the 1842 King's School building and adjacent area. Therefore it must follow that the great wall in question is the town or precinct wall extending east from the Prior's Gate. The only space available would be the site of Minor Canon Row.

In 1698 the Dean and Chapter ordered the demolition of a 'long building' housing the petty canons: “Row of ancient and ..?.. Buildings within the precinct ..?.. Petty Canon row formerly a prebendal house . Old house from time to time for many years ..?.. ..?.. let out by lease many years past very ruinous and not suitable habitation for persons of ability has been let out to very poor and indigent persons or ..?.. ..?.. ..?.. poor and indigent have of their own accord ..?.. out of the neighbouring parishes and ..?.. into ..?.. and lodged themselves in the said old and ..?.. buildings . Dean and Chapter put to very great trouble ..?.. proposed hath been made to the said Dean and Chapter to pull down the said Row of old buildings called petti canon Row and in the (rooms?) thereof to ..?.. or ..?.. to be ..?.. new buildings to be made ..?.. for persons of good ...?... 40/- recompense to the Minor canons for the advantages which the said Minor canons have or might of could have made by their (Rooms?) or tenements. … Consent for pulling down of the said ruinous old buildings and for the erection of new ones in the (Roomes?) thereof shall be build or ..?.. in the place of the old” (Hope 1900, p.75; M.A.O. DRc/Ac 4/10 pages 61a and 64a; Ac 4/11 pages 7a and 8).

We can see that this medieval building was used into the post-medieval period. We know Minor Canon Row as we see it today was constructed in 1721-3 to replace the previous canons residence and there is no reason to doubt that the long medieval building found in 1998 was that to destroyed. Whether all was actually demolished at that time is a debatable point. Certainly the far east end wasn't, that survived until c.1860, although admittedly it may have been rebuilt. It is shown on the precinct survey of 1801 (Fig. 75a) and although an order for demolition was given in 1836 (M.A.O. DRc/Ac 11, page 319) the east end appears on another survey of 1840 (Fig. 75b) and more importantly on a sketch plan of 1859 (see Beale Poste 1859; Fig. 75c). It does not appear on the first edition Ordnance Survey map published in the 1860s and was presumably demolished c.1860 during road widening. Whilst a delay of demolition between 1698 and 1721 can only be suspected, here in the map and documentary evidence a delay between 1836 and c.1860 can actually be seen.

Archaeological evidence: To the south archaeologically we have, rather surprisingly, more information than might be expected. Unfortunately a good deal of that information is ignored in relation to this exercise. Discussing the work that has taken place on the defences is a total nightmare.Here just one point is emphasised. It was the king who controlled the development of the defences whether in the late eleventh, early thirteenth or the fourteenth centuries. In the late eleventh and early thirteenth, at the very least, it was the king, not the priory, who was the lord who controlled the land. Although we have no documents telling us so, it would appear the king had allowed the monastic community to build on the land, that he controlled. As the king controlled the land and wanted defences refurbished or created it may well have been he, not the priory, who drafted in the work force.

We ignore most of the work undertaken or discussed by George Payne, Greville Livett, William St. John Hope, Arthur Harrison, Colin Flight and Alan Ward. The latter has undertaken two recording projects in relation to the defences (Ward 1997b and 1997c). Two points are briefly mentioned.

In the eastern precinct / town wall two arches are situated in the fill of the Deanery Garden Ditch and one very large arch situated across the King's Orchard Ditch. The latter and one of the former appear to have been constructed to allow the discharge of water, in the case of the KOD arch, at the very least, large amounts of water, here perhaps the ingress of tidal water from the creeks. The other two arches, mainly covered by the fill of the DGD, have in the past both regarded as 'arches of construction' (Harrison and Flight 1968). However, the present writer regards just one of these arches to be for that purpose, the other is regarded as the arch for a drain. There is also an infilled arch within the upstanding town wall, about 2.25m (externally) above the grass bank, may merely be some form of relieving arch, but what it relieved is anybody's guess. This arch can also be seen on the inside of the wall about 1m above modern ground level. Conceivably it could be a blocked window.

The second point is in regard the absence of the Roman wall bank and earlier rampart. These earthworks have been levelled for the whole of their length from the cloister area up as far as the south-east corner of the town. It is here suggested that this levelling was undertaken as one operation in the late eleventh century so as to create an open space, not only for the building of Gundulf's cloister and infirmary ranges, but also as a work compound and as an area for the temporary accommodation of the monks whilst the cloister was built. The old Roman wall was left standing.

Of structures in the Deanery Garden / King's Orchard we know very little. In 1959 or 1960 sewer pipes were laid in the Deanery Garden for the New (new) Deanery and several medieval buildings were encountered. A clay floor was found which contained pottery of the eleventh century. A second clay floor in a separate building contained thirteenth century pottery and a mortar floor within a third structure sealed a pit also containing thirteenth century material (Chaplain 1961, lxxiii). Unfortunately that is all that is know of these structures. The plan produced by Harrison and Flight in their 1968 report unfortunately does not extend far enough to the west to show these remains (1968 their Figure 3). That plan does show some chalk foundations within the fill of the Deanery Garden Ditch which may be part of one of these structures (Figure 6). Even if it is not, it would appear that all of these buildings were situated along the line of that infilled ditch. Unfortunately the thirteenth century pottery does not give us a secure terminus ante quem for the ditch infilling, if for no other reason one of the floors contained only eleventh century sherds. Admittedly those sherds my be residual, but we have no way of knowing. Greville Livett also found a building somewhere in this area of the Deanery Garden, for he tells that at a depth of about 0.70m he dug through a rough plaster floor. The demolition debris on its surface suggested the structure was seventeenth century in date (Livett 1895, 50). All the buildings could be of that sort of date. The pottery from the pit is probably 'safe'.

As well as the 'long bakehouse', there was also Ernulf's 'new bakehouse' which in turn tells us there must have been and 'old bakehouse', a brewhouse, stable, guest house, a laundry. Some of these buildings would probably be situated in the Grange Yard on the site of the 1843 King's School buildings. What looks suspiciously like a chapel appears on a not very good sketch of this area drawn by Canon Wheatley in (Wheatley M.A.O. DE53/1/19/7).

The Sole Pond (sole is Kentish for pond so we end up with Pond Pond) on the west side of St. Margaret's Street (Fig. 9) may have been a reservoir, although one suspects only for the Bishop's Palace. Where did the water for the monastic buildings come from. Springs apparently existed at the east end of the Vines and could have been channelled south-westwards, but if so where did the water cross the ditch/s.

We have the additional problems in regard the defensive sequence, mentioned above, but also there was a garden wall constructed sometime in the late eleventh or early twelfth centuries, “Odo, Bishop of Bayeux, gave to the church of St. Andrew and our monks to make there their garden beside the wall outside the gate towards the south part of the city outwards, which they have now enclosed with a wall on every side. And for these three acres of land” (Hope 1900, 5). This wall is believed to have eventually enclosed the vineyard and was completely rebuilt in 1384-1385 (Tatton-Brown 2006, 37 note 74). Below the west garden wall of the former Archdeaconry House fronting onto St. Margaret's Street there appears to be an earlier Ragstone wall surviving for a height of about 0.40m. This may in turn have joined with the wall, still standing 4m or so high. forming the boundary between Archdeaconry House and the King's School. These may be the last parts of that vineyard wall to survive.

In 1998 service trenching at the west end of Minor Canon Row resulted in a small fragment of medieval walling being seen (Ward 1998a). Later in the same year he whole of the road surface was taken up. At a depth varying between 0.25m and 0.45m down from that surface, below the line of the south kerb, a medieval building was observed for practically the whole length of the road (Ward 2002). To the north of the kerb line only compacted soil and rubble and flints were seen, this material formed the make-up deposit for an earlier road surface, or perhaps even a badly made road surface itself. What was probably a medieval road surface was seen at the base of a c.0.50m deep service trench in front of No. 7 Minor Canon Row in 2010 (Ward 2010). In 1998, at the east end of the road, a concentration of brick culverts and two parallel brick walls, 0.50m apart, again in front of No.7 Minor Canon Row were observed. The two latter walls are assumed to have formed a drain or sewer and originally to have been capped with flat slabs.

Of the medieval building the wall and buttresses for most of its course were made of Ragstone rubble with occasional angular flints and a few peg-tiles. The face of the wall was faced with a mixture of flints, Ragstone and occasionally other types of stone The buttresses abutted the wall face and had greensand ashlar blocks at their corners, but occasionally bricks were also used. At the east end of the structure, chalk was the predominant building material and here also the whole width of the structure was revealed, 8m internally and c.10m externally. The wall seen by Greville Livett in this area way back in 1893 was part of the south wall of this building not, as he thought the supposed later Norman (Ernulfian) defensive wall.

More or less in the centre of this structure was a 1.50m wide doorway Excavation down to the threshold produced large numbers of medieval floor tiles from within the soil removed. The Ragstone jambs along with iron pivots for supporting a two-leafed door remained in situ. The long broach stop carved on the jambs tend to be of mid-fourteenth century or later date (pers. com. Tim Tatton-Brown; pers. com. Rupert Austin architectural draughtsman for C.A.T; unfortunately Margaret Wood states fifteenth century (Wood 1983, 410). A mid-fourteenth century date would of course be acceptable for the building work undertaken by Hamo de Hythe.

The threshold itself was at depth of 1.08m below the modern kerb and almost certainly represents the external medieval ground level. An internal step down shows that the floor level of the structure was lower still. At the east end, between the main north and south walls, partition walls and clay floors of different dates were seen. Artefact dating evidence was sparse, but enough, along with the overall stratigraphy and the map evidence showed that three phases of occupation could be discerned. The medieval phase and two post-medieval (c.1540 to c.1750) phases, the last continuing into the early modern period (c.1750-c.1875). The map evidence tells us that the east end was not pulled down until c.1860.

Below the lowest clay floor, sandy clay and clayey loam deposits continued down for a depth of c.4.00m with the base of the trench at c.6.50m. Man created deposits were still continuing downwards and there can be no doubt that this material infills the Deanery Garden Ditch. Just one artefact was recovered, one floor tile of Flemish manufacture datable to between the very late fourteenth through to the mid/late sixteenth century. That does not help us date the infilling of the Deanery Garden Ditch, but a date in the late fourteenth or early fifteenth century for the construction of the east end of this long building would be fine.

Just one area produced a copious amount of pottery. On the north side of the south wall an external chalk lined post-medieval cess pit was excavated and produced a dated jug of 1664, but with an overall ceramic date range pointing to its being backfilled in the early eighteenth century. A very rare teapot (without its lid) manufactured by John Dwight and dating to about 1700-1710 was also found (Ward 2001, 41, information from John Cotter).

Further south opposite the door into Mackean House, the foundation of a wall was observed. and found to extend the whole width of the road. At its west end a distinct kink could be seen in the line of this wall and archaeological cleaning showed that the longer eastern portion was later in date than that to the west. The earlier part of the foundation continues the line of the wall that at one time existed to the south of Minor Canon Row. The scar for this wall can be seen in the east wall of Priors Gate and was also found in a small excavation in the garden of No. 7 Minor Canon Row (Ward 2011b). A wall coming off at a right angle to this wall may be that of a medieval building (Fig. 76). The foundation across the road may be that of gate blocking, whether the wall continued eastward or turned to the south to join with the south-east corner of the east range is an unknown quantity.

The precinct to the west of the cloister including the Bishop's Palace

Location: Much of this area is taken up by the access road, 'Prior's Gate road'. and the site of the medieval Bishop's Palace, of which only a small part survives. That the garden, or more correctly the court, of the palace had different layouts at different times in the medieval period is shown to us by the traces of archaeology recovered from the various observations and excavations (Fig. 76). None of those archaeological projects, individually, tell us a great deal, either about the Roman (Fig. 77a) or medieval periods. However, by bringing the information, collected over a hundred or more years, together we can, with great difficulty, begin to see the complex nature of the structural build-up of this area in the period c.1100 to c.1800. This area is supposedly the site of Gundulf's cloister, by looking at the archaeological remains and topography we can see that it was not (see below, pages 67 to 69).

Documentary evidence: The documents in relation to this area are few and far between and none are of much use in attempting to understand the archaeological remains. We are told that Bishop Gundulf (Bishop 1077-1108) built a residence for himself and his successors (Flight 1997, 185, 222), but we are not told where it was situated. This building was burnt down in 1137 (ibid, 203). It was presumably rebuilt and then damaged again, for Bishop Gilbert de Glanville (Bishop 1185-1214) rebuilt the palace after a fire, presumably the conflagration of 1179 (Flight 1997, 221). However, yet again we are not told where it was situated. Colin Flight would have the pre-Bishop Gilbert palace situated to the east of the cathedral on the site of, or adjacent to, what is now the St. Andrew's Centre. This interpretation is based on the archaeological work carried out in 1978 on the site of the fifteenth century palace, the results there showing that

'it seems fairly certain, from archaeological evidence (Harrison and Williams 1980) that the residence built in the eleventh century - the aula (= hall) mentioned by Gundulf's biographer - could not have stood on this site. Apparently it must have been located somewhere else; and the only available site for it would be somewhere east of the church', (Flight 1997, 222).

Whilst the palace may have been elsewhere, that is certainly not shown by that excavation. The trenches excavated did find the east wing with of what is regarded as the fifteenth century palace, but no mention is made of what was below its mortar and chalk floor. That apparent failure to find an earlier building could be illusory for no section drawings were drawn, or at least none published, for us to examine, from Trench L or Trench G (Harrison and Williams 1979, 24). Certainly earlier structures were present immediately to the east, although, with the exception of a mortar floor of, possibly, a timber framed building in Trench H, none appear to reach as far westward as the excavated trenches (ibid., 1979, their Figure 1). Also of course any late eleventh century palace may not have been in the area excavated, it may have been under the standing south block or up against the western boundary wall, or projecting eastwards below the grassed area of College Green (Fig, 77b). At the end of the day we just do not know.

About the definite earlier structures to the east of the fifteenth century remains found we know very little due to a lack of section drawings and a failure to dig exploratory sondages at strategic points (Harrison and Williams 1979; Guinness 2005).

Also in the west gable of the standing south range there what appears, to be a Romanesque window at first floor level. Whilst it could be reset, it is just as likely that the wall around has been refaced. Having said that, Livett states that window and another in the south wall,

'which from a distance look like Norman windows, are in the style of the fifteenth century, their heads being four-centred' (Livett 1895, 43).

These windows need to be looked at again by a building specialist. If we take Livett's view of the window as being correct, the end product of this exercise still shows us that the reasoning for placing the eleventh century palace elsewhere is by no means secure and it may have been in its late medieval position from the outset

We can go further, the implication is that the structure burnt down was that re-erected. However, we know from documentary sources that Prior Osbern (Prior 1189-93),

'made a lodging for himself next to the infirmary' (Flight 1997, 222)

In other words a lodge was constructed to the north of the infirmary in the area of what is now taken up by the St. Andrew's Centre (the eighteenth century New Deanery). Colin Flight would have it, that this was the time, c.1190, that the palace was created on what we know to have been the site of the late medieval Bishops Palace (Flight ibid. and his Fig. 26). This, however, is merely a surmise, there is no documentary evidence to that effect, and just as the excavation of 1976-77 did not show the presence of a c.1100 palace, nor did it show the presence of a c.1200 palace either. It is here suggested that the latter did exist on this site, but so did the earlier structure.

Our next documentary reference along, with the archaeological evidence discovered by Harrison and Williams, is enough to show that the palace was on the present site in the fifteenth century, for Bishop Lowe (Bishop 1444-1467) headed a document stating it had been written, 'in our new palace at Rochester' in 1459 (Harrison and Williams 1979, 26). Which is all well and good, but of course still doesn't tell us where the 'old palace' was situated. It could have been on the same site and been repaired or rebuilt or it could be on a different site. Overall, due to our knowing that the prior had his lodge to the east of the cathedral from c.1190 and as we know Bishop Glanville rebuilt a residence for himself at about the same time, it is just as reasonable to assume that the palace was situated to the south of the nave, rather than to the east of the cathedral, from the outset.

Much of the south range still stands and has usually been regarded as being of mid-fifteenth century date. However, Patricia Clarke suggests a thirteenth or fourteenth century date is more likely (Clarke 2014). She specifically points out three windows of earlier architectural type (ibid., 6 and 7). By itself such earlier architecture should not be regarded as conclusive for the dating of structure. Whilst the north ('public') face is pretty non-descript, the south face is much more interesting and she points out that in that face (at least) one of the two windows here is not designed to 'fit in' with the building as in its later medieval design. The third window is high up within the east gable (Plates 80 and 81). Archaeological / architectural plans and elevations as well as the relevant photographs would have greatly helped in confirming her (almost certainly correct) ideas. For the sake of simplicity the traditional date of the mid-fifteenth century for the east wing has been used in the phase plans. They were hard enough to produce on that assumption alone and with the number of structures to fit in to the courtyard they will become harder still to phase once an earlier date (if correct) for the south wing is confirmed. Perhaps even harder than can be imagined for she tell us (ibid., 33, note 12) that the cellars were definitely cellars from the outset and not the ground floor as had been previously (no reference) suggested. She tells us they could only have been the ground floor if they had been constructed in a 'pit'. The whole point is, of course, they were (see below, page 74). The south wing was constructed along the line of the Deanery Garden Ditch. Until such time as an archaeological trench is dug up against the south face we have no-way of knowing whether they cut into the infill of that ditch or whether that material abuts the wall. If it was a ground floor, in theory windows should be visible internally, but that face has almost certainly been re-pointed, rendered and white-washed several times.

The east range found by Harrison and Williams was still standing in 1719 (Fig. 78), but had been demolished by the time the Baker's map of Rochester was published in 1772 (Fig. 9). The west range was still shown on Sale's map of 1816 (Fig. 29) but, presumably because it was no longer considered to be part of the precinct, was omitted from the 1801 survey (Fig. 75a). Even the south range was only shown by a dashed line and then omitted altogether from the survey of 1840 (Fig. 75b). From the 1820s, and perhaps earlier, the palace had been leased to Mr. Twopenny and his mother (Livett 1895, 44). The office of Mr Twopenney, at the far north end of the west range, is marked, but not shaded on both maps. It seems likely that the southern part of the west range was demolished sometime after 1816 and certainly by the time the first large scale Ordnance Survey map of Rochester was produced in the 1860s.

Archaeological excavations and observations: There have been a considerable number of observations made from the 1880s onwards which will be mentioned as we progress. Four larger projects (two as watching briefs') have taken place in the garden of the Bishop's Palace or immediately to the east within the roadway (Figure 76). The first of these four projects was in 1976-77 on the site directed by Arthur Harrison and D. Williams (1979, 19-36), the second in 1998 in the adjacent roadway (Ward 1999b, 2002, in prep. b), the third, directed by Abby Guinness in 2005 was back into the garden between the 1976-77 excavations and the garden wall bounding the road (Guinness 2005) and lastly back in the roadway (Ward 2007, in prep. b). The 1998 and 2007 'excavations' merely saw the then tarmac road surface and its bedding removed with cleaning and recording then taking place. Considerable structural evidence was revealed which although not dated in any detail is relatively easy to understand. The two excavations within the garden revealed complex structural remains, a good deal of which is difficult to put together in a coherent form. Only a few of the structural components have stratigraphic relationships one with another and there is only a limited amount of meaningful artefact evidence. The failure to search for the limits of wall lines and the relationships of layers to those walls makes analysis difficult. All it would have needed was a few small trenches ('sondages') placed at strategic points to look for these stratigraphic relationships. In the 1976-1977 project much needed, section drawings of all the trench sides should have been project undertaken. What we can say however, without any doubt whatsoever, is that the structural sequence within this area is far more complex than is generally thought. We can also state that the often repeated statement (Hope 1898, his Plate 1, 1900, 6; Livett 1895, 39, 47; Palmer 1899, 56, Fairweather 1929, 192; Tatton-Brown 1984b, 186, Flight 1997, 149) that Gundulf's cloister, whether of timber or masonry, was to the south of the nave is wrong. Only Philip McAleer made the brave guess that the cloister had always been in the position that we see it today (1993, 13, 1999, 44).

Roman: Although much mutilated the Roman town wall of the third century was observed in three of the projects. In that of 1976-77 the late second century earthwork defensive rampart was also revealed (Fig.77a). This extended backwards (north) of the wall for a width of at least sixteen feet (see below). A coin, a forgery, of Julia Domna (wife of the emperor Septimus Severus (reigned 196-211) dating to c.204 was found in the remains of the earthen bank contemporary with the wall. This bank had been placed on top of the late second century earthwork defences, but whether the coin was found in the area levelled in the late eleventh / early twelfth centuries or in situ material we are not told. From their section drawing it would appear that the whole of that wall bank, and perhaps part of the second century rampart, had been levelled (Harrison and Williams 1979 their Fig. 2, here redrawn in a simplified format as Figure 79a). There is certainly no difference shown within their section drawing of any levelled wall bank and any in situ material. The coin could, therefore, be within Roman soil re-deposited in the late eleventh or early twelfth century, or in other words it becomes residual and may have little bearing on the date the wall bank was constructed. If all of this soil is disturbed we cannot even be sure that the coin came from the wall bank, it may have been intrusive into the earlier rampart. We just do not know how much of bank and rampart have been destroyed and intermingled. The sketch profile does try to give some hint of the scale of destruction (Fig. 79b). Even if we could be sure the coin was in situ within undisturbed material and whilst it would give a terminus post quem of 204 that can hardly be considered to supply a secure date for the construction of the wall which may not all have been of one phase anyway (Ward 2011a).

The Roman building found below the south and west walls of the cathedral has been mentioned above (page 10) and what we know of it is not repeated here. However, in 1998 the present writer observed a compacted chalk foundation (Building 1) immediately to the west of the south turret of the cathedral west front. At the time he regarded it as being of Norman date, but it may be that this structure was also Roman and formed part of the building below the cathedral south wall. This foundation is, however, discussed in a bit more detail in the medieval section below. Graham Keevill has also seen masonry in this area, about 1m down from the modern road surface (pers. com). Just to the south in 1900 and 3.60m below the floor of what is now the Ship Safe Training Group Ltd a human skeleton was found and has been mentioned above (page 9)

Medieval: In the late nineteenth century Canon Greville Livett made several observations along 'Prior's Gate Road' (Figures 6 and 74; Livett 1895, 71 and his plan of Rochester). He points out that in a Gas Company trench, which would probably have been c.1.20m to c.1.50m deep, a three foot wide wall was seen running,

‘south and making right angles with the Saxon church'.

This wall (Wall P) seems to have been traced for a length of about 2m. Whilst this structure could be part of the medieval gate known to be at this position, Livett regarded this wall as being pre-Norman, which perhaps hints that it was found low down within the trench. This wall was almost certainly seen again in 1998 and due to its depth was regarded as Roman. If that surmise is correct then it could easily be part of the Roman building found below the south wall of the cathedral nave.

However, there is another possibly, for when the Anglo-Saxon church was demolished c.1090 another building was also destroyed. The late twelfth century historian Eadmer tells us that Lanfranc,

‘demolished the ancient church of the bishopric together with the adjoining building', (Flight 1997, 173 note 9).

We know nothing more about that building, but if it was the wall seen by Livett and as implied by Eadmer, it was obviously a masonry structure and there must be more walling to go with it.

Further south, another wall (Wall Q) ran parallel to what is here called 'Prior's Gate road' and seems to have been traced for a length of c.6m. This wall was also seen in 1998 and found to be a stone lined drain. This drain followed the line of the road and was traced intermittently for nearly 20m.

Livett observed three other walls (Wall R, Wall S and Wall T). Wall T crossed the line stone lined drain more or less at a right angle and was aligned on the sooth-west corner of the southern west front turret. Whether that alignments means anything is doubtful if for no other reason that it must be considerably later than the building of the Norman church. Further to the east, other than saying Wall R and Wall S are presumably of medieval date and showing their alignment, not a great deal can be said (Fig. 6). The gate to the north and the assumed gate to the south would have given access to a court rather than an actual lane, but as so often such has been suggested before (Tatton-Brown 1984, 187 and his Figure 6). The lane itself was presumably created after the palace went out of use in the mid-sixteenth century

Of some considerable interest several fragments of Ragstone walling were found immediately south of the nave in 1937. We are told that Prior Radulf (Prior 1208 to 1214) constructed stone buildings within the cemetery (Flight 1996, 222 note 14). We are not told where those buildings were and it would be extremely difficult to show any of these walls represented those buildings, but such should be borne in mind in any further excavation in the area. One of these walls was noted as being bonded with 'shelly mortar' (Cobb 1938, his plan, here redrawn as Figure 11). Whilst such mortar can be late medieval or even post-medieval, its occurrence in a Rochester context does hint that for this masonry we are looking at an early to mid Norman building. There is only one way to find out, re-excavate the area.

The late second century rampart found in the garden of what is now The Deanery (but formerly Prior's Gate House) in 1976-77, extended northwards for a width of at least 6m. Once the third century masonry wall was built a higher backing bank was constructed. Both bank and rampart were eventually to be partly levelled off. This levelling process was presumably undertaken at the time of the construction of a Norman building in the early or mid twelfth century (see below), for obvious reasons it cannot have been later. This levelling presumably extended the soil of the bank / rampart backwards to the north for a width of at least 15m, more or less the whole length of the garden and, judging by the land profile as it exists today to the east, perhaps far beyond. Whilst no accurate level above Ordnance Survey Datum is given the top-soil was (and presumably still is) at approximately 39 feet above O.D.. This equates to c.12.20m. The top of the levelled Roman bank is 1.50m below giving a height above Ordnance Datum of about 10.70m. This is c.1.70m higher than the concrete path immediately south of the south wall of the nave. The latter is c.1m above the Roman building found by Irvine and shown by Hope on his Plate II. The level of the concrete is approximately the level from which the Norman nave wall was constructed in the late eleventh century. Whilst all of these measurements are approximations they will not be too far adrift from reality and the reader should now begin to see why they need to be mentioned. What we end up with from the south wall of the nave to the land level inside of the standing boundary wall of the Bishop's Palace is a rise of c.3m. If we make an assumption in regard the extent of the levelling out of the wall bank and earlier rampart we can, perhaps, see the approximate slope of the land c.1100 (Fig. 79b).

We cannot say for sure when this levelling of the wall-bank took place. It could, conceivably have taken place c.1090, on the other hand it may have still been present c.1120, we have no-way of knowing. The earlier rampart still survived for a width of 6m to the north (Figures 77a and 79a) and the wallbank would be wider still. If our assumption in regard its levelled extent is correct, when standing it would have projected back from the wall for about 10m and then when destroyed by about 20m. This would allow a measurement of about 30m (c.90ft) between the south wall of the cathedral and the base of the levelled bank. This could allow part of the cloister to be built at this level. At higher level, on top of the levelled bank a further part of the cloister could be built. The cloister would have to be constructed at two different levels. Perhaps unorthodox, but whether the buildings were of timber or in stone, structurally it could be done. If the cloister was only taken back as far as the 'cliff face' it would be about 35% less in north to south width than the cloister area to the south of the quire. Even if we were to include the area of the west range as we see it today, it would also be slightly shorter in a west to east measurement as well. Such an area, even if not regarded as being too small for a monastic cloister 90ft (c.30m) x 190ft (c.60m), would certainly be regarded as being out of proportion. Only if the cloister could be brought forward to the old Roman town wall would it have a more square 'look', the look we usually associated with a monastic cloister.

That this was not done is shown to us by a combination of the archeological stratigraphy produced in the 1976-1977 excavation undertaken in the front garden of Prior's Gate House and that undertaken within the cloister garth in 2014-2016.

In the latter area we have deep gravel foundations and some remnants of masonry pre-dating the standing cloister East Range (see above, pages 55 to 57). It seems reasonable to assume that if there were early foundation trenches within what had been the Prior's Gate House garden (now the new, new, new, new Deanery garden) they would be dug and infilled in the same way. There is no sign of such within the section drawing of the 1976-77 excavation. There may be the temptation to say the monks ranges were of timber and the foundations found in the garth were for a masonry built infirmary ranges. Such a discrepancy seems unlikely, if for no other reason that the monastic ranges would (probably) have precedence, but such does remain a possibility.

Other than three small post-holes cutting into the levelled Roman wall bank each no more than 0.15m in diameter, there is no sign of a timber structure. The cloister ranges would be large buildings and would use large timbers, post-holes at least 0.40m in diameter would be present. The buildings would have earth of perhaps mortar floors. Admittedly, a grey shelly soil overlain by charcoal might represent such a floor, but if it does it is not associated with any structural evidence for walling whether timber or stone. It could be argued that the walls rose from horizontal sleeper beams set into the ground and which would be extremely difficult to see in narrow trenches, but why would such a method be used when gravel foundations over 0.40m deep with masonry walls above was being used in early Norman structures below the garth lawn. The present writer is unable to accept any of these ideas.

Philip McAleer does also point out that there is no hint that any doors passed through the south wall of the nave (1999, 44). Admittedly half the south wall has gone and most of the rest has been refaced. However, Irvine makes no mention of mortar of compacted earth floors for a walkway being present, prior to underpin trenches being dug along the length of the south wall.

If the garth as we have it today, to the east, was terraced for what some might argue were the infirmary buildings, why not terrace the area to the south of the nave as well? This it especially true due to the fact we can see the Deanery Garden area was also levelled. The labour was there, hundreds of peasants could be drafted in as necessary. Also if those infirmary buildings were made of stone, which they were, why should the monks cloister buildings be of wood? After much thought the present writer favours the foundations found within the cloister garth as representing those of the early Norman monks cloister rather than the infirmary range and the total lack of evidence for a cloister to the south of the nave is explained by it never having been there in the first place. Also, of course, the objection to the Bishop's Palace being on this site from the outset is removed.

There is no evidence, either documentary or archaeological, that the late eleventh century cloister was built to the south of the nave, whereas we do now have evidence for structures below the garth lawn to the south of the quire. We also have the very obvious fact that the latter area is that which has been levelled off.

Norman building: Let's start our exploration of the various buildings found in 1976-77 and 2005 with the Norman building (Building 2) made from Ragstone and bonded by shelly mortar. In their interpretation of this structure Harrison and Williams put forward the idea that it could be a 'porch or stairway leading up to a first floor hall', as exists at Canterbury in the building known as the Aula Nova (= new hall or new residence) or alternatively it could be 'part of the original Bishop's Palace' (Harrison and Williams 1979, 25). The present writer, for two reasons, thinks the first interpretation is unlikely. First, even though the hall itself may have been at first floor level there should still be signs of a floor in the lower room. Such a deposit should have appeared in Trench C (Fig. 79a). Perhaps it did, but is so it is not mentioned in the text.

Secondly, some form of structural element, either foundation trenches for stone walls or post-holes for a timber building, should appear. Admittedly there is a mortar filled wall robber trench (their context 29) in Trench D, but this is too far away from the conjectured porch to be the north wall of such a hall and too close for it to be the south. Other than that feature there is nothing. This robber trench is almost certainly a continuation of a wall found in their Trench P which they regarded, without explanation, as being 'relatively modern (ibid., 24). The 'Stone Foundation' noted in Trench F (their Figure 2, here Fig. 79a) is that of the 'blocking wall'.

Building 2 appears to be free-standing. If we look at this Norman building we should ask ourselves, 'What do we see?' If looking at a plan of this building for the first time without anything else around and not knowing anything of the area the answer would (should) be, a gate-house. Measuring approximately 10m by 6m with walls 1m wide and a door (= gate) 2m wide such an interpretation seems perfectly reasonable. However, as with most of the structural evidence from the 1976-77 excavation there are problems with such an interpretation. First, a gate-house does need a boundary going off to either side. There is no sign of such in the sides of Trench F or Trench E. Admittedly a boundary may not necessarily be of masonry, it could be of timber or nothing more than a hedge but, due to the impressive nature of the conjectured gate-house, such seems unlikely. As with the south wall of this postulated gate-house the boundary wall may have been 'robbed'.'

If it was a gate-house it should obviously lead into a courtyard with buildings. There appears to be no sign of a gravel or cobbled courtyard to either north (in Trench I), or south (within Trench L) of this postulated gate-house or, more correctly, not one at the correct level. Cobbles do exist immediately to the north, but these are stated as being 5ft below the modern ground surface. This appears to be too low to correspond with a threshold of the suggested gate. Judging by what we see in their Plate II (1979 facing page 36) the gravel of the gate-passage and the stone threshold appear to be not even 2ft down from the modern ground surface. This would obviously leave too much of a 'step' between the cobbles and the gate-passage. At most there should only be a discrepancy between the surface within the gate-passage and any cobbling to the north of about 0.15m. The absence of all round section drawings, so that we can see the levels of the various layers, here makes itself felt. Be that as it may, the interpretation put forward by the site directors is itself open to doubt. They proposed this cobbled surface created during the levelling of the wall bank. Ignoring the fact that this interpretation is based on seeing cobbles over an area of just c.1.70 x. c.0.25m in extent, for two reasons this seems to be incorrect. First, why would the work force bother? Secondly, if it was to do with the removal of the wall bank such a surface should appear in the section drawing of Trench F. It doesn't. There is only one way to find out what is going on, re-excavate the area.

One of the questions we should be asking ourselves is, were there buildings to north or south of this conjectured entrance? Or to put it another way, where were the buildings to which this conjectured gate led? The 1937 plan (Fig. 11) does hint that there was a Norman building to the north, but the 2005 excavation hints of two Norman buildings to the south (see below, pages 74 and 75). If access was from south to north there should be some sign of gravel access routeway within Trench L, and perhaps other trenches. Other than a chalk and mortar floor, originally tiled, for the east wing of the Bishop's Palace, no such surface is mentioned. To the east we do have a road, still is use. We know this road was present in the eighteenth century and, by implication, due to the presence of the Almoner's Gate at one end and the Prior's Gate (and an earlier gate adjacent to Building 3) at the other, there was access to the area in the medieval period. Admittedly this access was more probably into the bishop's court than along a planned lane. That access would, however, have required an opening through the old Roman town wall. The suggested internal turret (Building 3; Figures 76 and 77b, Plan 2) would stand to the west side of the opening which would probably be nothing more than a gap closed by a door, but perhaps with an arch above.

If the identification of this structure (Building 3) as an internal turret is correct, these are normally of Roman date. However, in 1974 a Norman turret, although in the report called a 'guardroom', was found on the north side of the city projecting internally and in association with twelfth century pottery (Harrison 1981, 105). The distinctive shelly mortar used in Building 3 would also suggest a Norman date for this structure. The existence of a gate, first suggested by Arthur Harrison and Colin Flight (1986, 21), presumably just a simple arched opening, immediately to the east seems likely. Such a gate with its attendant road would line up with the later Prior's Gate, and provide access into the bishop's court.

If Building 3 was an internal turret that would imply the Roman wall to the west would still be standing. The suggested gate through the town wall would allow access into the court, there would be the west range of the cloister on the east. If we assume the Norman building does represent another gate there would, presumably, be a boundary wall attached to that building and would (presumably) block off any access between itself and Building 3. From the outer gate the visitor would walk north, with a wall on their left, then turn to the west and then south to pass through the inner gate. Again there is only one way to find out if such ideas are correct (or not).

Between the suggested inner gate and the Roman wall there is only a 7m width. It is unlikely, therefore, that the hall and solar ranges for a conjectured late eleventh century bishop's residence would be built directly in front (south) of that gate. We should remember however, that there is no reason why such buildings should be in front of such a gate, they could be to the west or even the north. The bishops residence could therefore, have been built up against the Doddingherne Lane boundary wall with the hall aligned north to south up against that wall and the solar facing south and east (Figure 79b).

Alternatively the Roman wall could, as it eventually was, have been demolished to the west of Building 3, the internal turret, or it could have been utilised, just as in the refectory to the east. The bishops residence could have been constructed to the south of the old town wall. However, whilst a building was placed here in the fifteenth century (assuming the window in the west gable is that late) such would be unlikely in the late eleventh or twelfth centuries, if for no other reason that it would be built either in the still open Deanery Garden Ditch or if it had been infilled the material would still be loose and unstable. The suggestion of a thirteenth or fourteenth century date by Patricia Clarke, as mentioned above (page 69), seem far more reasonable and such would allow (if of the later thirteenth century or later) to allow some compaction of the ditch infill to take place. However, with infill slumping the cellars could still originally have been a ground floor.

The western boundary wall would be along the line of what is now, very largely, early nineteenth century walling fronting onto Boley Hill (Doddingherne Lane). To the south of South Gate House parts of a medieval wall, assumed to be of the early thirteenth century, still exist, and the old Roman South Gate was still in use attached to another remnant of the old town wall perhaps down to the early thirteenth century. Also at the far north end of this boundary wall some medieval masonry still survives. Two stones of a quoin can be seen. The present writer has always assumed they represented a building earlier than the brickwork above. There are also indications that the quoin stones were themselves later in date than a flint wall (Building 5). Today only two courses of that wall can be seen stretching for a length of about 3m, but twenty years ago a third course could also be observed. The two surviving courses have the hint of herring-bone masonry. Whilst not a good example of such coursing, some of the flints can definitely be seen to be set in opposite slanting directions. Such coursing would suggest a date of pre-1200 and probably pre-1150. A blocked door and threshold of a later structural phase can also be seen (Building 6). This herring-bone wall would have overlaid the compacted chalk foundation mentioned above (Building 1). The latter, at one time thought to be Norman in date is now considered more likely to be Roman. That this foundation steps out at least c.1.50m from the line of the precinct wall would represent a major change in the property boundary which the present writer just doesn't think the medieval Church would accept. In addition, despite what Hope shows on his precinct plan (Fig. 6), there was no sign of an earlier north to south boundary wall, which would have had to exist to the west and which (assuming it came to the west of the footpath) would have been seen in 1998.

Later buildings: The excavation of Harrison and Williams within the front garden of the Bishop's Palace or, more correctly, the eighteenth century building known as Prior's Gate House and now called The Deanery has been the most interesting that has taken place in the precinct. As well as the Roman rampart and wall it revealed a hitherto unknown masonry Norman building dating to the first decades of the twelfth century (pre-1140). The fifteenth century east wing of the Bishop's Palace was also revealed. Problems in relation to the understanding of the function of the Norman building are considerable and have been briefly mentioned above. The east wing of the palace is shown on the 1719 print of Rochester (Fig. 80) which appears in the History of Kent by John Harris. It had been demolished by the time Baker's map of Rochester (Fig. 9) was published (Denne 1772).

Whilst both the Norman building and the east wing of the palace were excavated over a large enough area to create their plan, very great problems arise in understanding the other fragments of walling found (Figures 80 and 81). Six walls were observed:

Wall A. Described as a 'blocking wall' which crossed (i.e. cut) an earlier wall (Wall B).

Wall B. This wall had an associated chalk floor containing fourteenth century pottery.

Wall C. A foundation which overlapped the north wall of the Norman building, described by the excavators as being '3ft thick'. This presumably means its width rather than its vertical measurement. This wall is not visible on their plan (see below).

Wall D. A wall of Ragstone and flint in Trench P. This wall is described by the excavators as 'relatively modern', but no explanation as to why is given in the text and it is not shown on their Figure 3, or if it is, then it is not clearly shown

Wall E. A foundation in Trench N, also described as 'relatively modern'. Again no explanation for this assertion is given, but it does appear in their Figure 3. As far as this writer can judge this wall forms part of the west wall of Building 3, the internal turret, the south end being cut away for the insertion of a soakaway. (The site directors do, quite rightly, have a ? against that interpretation on their Figure 3. The effort involved for a pit to be dug here would be excessive). Admittedly the wall could be aligned east to west and form part of a completely separate structure, but on the basis of the evidence that we actually have a Norman date for this wall seems the most valid interpretation.

Wall F. A wall on top of the Roman wall in Trench N was built partly of eighteenth century bricks.

In addition

Wall G. This wall is still there on site today in the form of the boundary wall. The Ragstone and flint wall long the line of 'Prior's Gate road' stands about 2m high. At one time it had a large fragment of opus signinum concrete (c.0.15 x c.0.10m) near the wall top. but this disappeared after the 2005 wall re-instatement project. This wall may date to after 1840 for a cathedral precinct plan of that date shows the northern part of this wall is at a definite angle rather than the curve that we see today. The north wall facing onto Collage Green is made from late sixteenth or seventeenth century bricks and a date of c.1600 may well be applicable.

The 2005 excavation then revealed a further six walls all of which are probably of medieval date (CAT context numbers 219 and 220 and 236 to 239; here Walls H to M). Two of these walls (Wall H and Wall M) are probably the same as walls (Wall O and Wall N) found in 1998 and 2007.

All the walls found due to mortar colours, alignments or stratigraphic positions, make an already confused situation more confusing.

Assuming that the dates given for the east wing of the Bishop's Palace and for the Norman building are both correct we have to fit all the other walls into the period c.1140 to c.1450. The present writer would be quite happy to see the date of the Norman building pushed back two or three decades to c.1110 or even c.1100. This judgement is based purely on the type of mortar used, which is described as 'very shelly soft yellow mortar' (Harrison and Williams 1979, 22). Whilst no accurate date for this mortar type in Rochester has been definitively shown, a date of c.1110 or c.1100 is just as likely as one of c.1140. However, we should always be careful of archaeological deductions based on mortar type alone, but luckily, here we can see the stratigraphy pushes this building back to the start of the medieval structural development of the site. This factor, along with the mortar type, show us that a date sometime in the early to mid-Norman period for this masonry is assured. Three walls (Walls A, B and D) found in 1976-77 in a section of text coming after the description of the mid-fifteenth century east wing are confidingly described as 'Later Features' in that report (ibid. 24). Whilst there are no stratigraphic relationships between the masonry of the fifteenth century Bishop's Palace and these three walls, at least two, Walls A and B, and probably all three, are earlier, not later, than the east wing. This has to be so, for three reasons. First, Wall B, is associated with a chalk floor containing fourteenth century pottery, in other words circa one hundred years earlier than the east wing of the palace. Whilst in itself, this is not a conclusive reason for it being earlier, for it could still be a 'later feature', it is overlain by the so called 'Blocking Wall' (Wall A). This wall has to be in use when the Norman building was still standing, for it meets the north-east corner of the latter and does not appear in any of the trench sides (Trench C, L or I) excavated within the east wing of the later Bishop's Palace. Lastly, the Norman building, and presumably the 'blocking wall' have to be demolished at the time of (or before) the building of the (supposed) fifteenth century east wing.

Let's move on to the other short lengths of walling found in 1976-77 and see what we can make of them.

If you have a chalk floor in association with a wall (Wall B) it is a reasonable deduction that the floor will be inside a building (Building 7). Therefore, it follows that there should be another wall parallel to that found. It is also a reasonable deduction that there should be two other walls at right angles joining these two parallel walls together and hence a rectangular building must exist. The width and length of that building are both of course unknown, but we can make a reasonable estimate. To the south-west it cannot extend as far as Trench E of the 1976-77 excavation for otherwise it, or its robber trench, should have appeared in the east face of that trench. No masonry and no sign of a robber trench can be seen in the section drawing, Figure 2, of that report (here Fig 79a). If we are totally objective, it is possible that all trace of either could have been destroyed by later digging, but as the 'displaced wall bank' soil continues from Trench F into Trench E with only a couple of not very large post-holes disturbing its profile such seems unlikely. Therefore, the south wall of this building must exist somewhere to the north of Trench E. If there was a gate next to the Norman tower (Building 3) it seems a reasonable deduction that there was access (whether it is called a road or a courtyard is irrelevant) leading into the bishop's precinct at this point. How far to the north that access actually went we have no-way of knowing, but it seems reasonable to deduce that an open ropeway in between buildings reached as far as the south-west corner of the cathedral where another gate, the Almoner's or Great Gate was situated (Fig. 6). If that is accepted then the east wall of this building cannot extend too far eastwards otherwise it would block this access way, a maximum length of 10m would seem reasonable. It would probably be slightly less, say 8m, in which case a width of 4m or slightly less would seem likely and hence we obtain a building as shown on Figure 83a. The chalk floor of Building 7 contained pottery of the fourteenth century and hence provides a terminus post quem for Wall A and perhaps for Wall B.

As the 'Blocking Wall' (Wall A) cuts Wall B it is a reasonable deduction that the building represented by the latter wall had been demolished at the time of, or before, Wall A was constructed. That Wall A appears to end at the north-east corner of the Norman building (Building 2) would hint, very strongly, that this structure was still standing. There is no hint that Wall A went over a demolished wall of Building 2 and there is no sign of its continuation in Trenches I, C or L, across the east wing. Therefore, it would seem that Building 2, the earliest medieval structure, on the site was still standing in the late fourteenth or early fifteenth century and was, presumably, not demolished until the east wing of the Bishop's Palace was constructed c.1450. The 'Blocking Wall' may be just that, a boundary wall. There appears to have been no contemporary floor surface, or if there was none is mentioned, nor does any wall found in 2005 run parallel or at right angles to it. Based on what we have we have to accept the interpretation offered of it being a 'blocking wall' or perhaps a more correct title might be garden or boundary wall. However, further excavation may lead to revision.

To complicate matters we do however, have a mortar floor to the south of this wall (Building 8). In their Figure 1 of the 1976-77 excavation report Harrison and Williams show such a floor within Trench H. This mortar floor is not mentioned within their text at all! This floor may (or may not) go with the Ragstone and flint wall (Wall D) in Trench P. In their report they describe this as a 'boundary wall of relatively recent date', but because of the lack of section drawings within their report we cannot be sure if the floor went with that wall. We just do not know how deep Trench H was in relation to Trench P and hence do not know the relative depths of wall base and floor. In addition to that, there is no explanation at to why Wall D was regarded as 'relatively modern'. If Wall D and the floor did go together then it will not be a 'boundary wall', it will be the south wall of another building and could easily be of medieval date. Even if that is not the case the presence of the floor tells us that there had to be a building in some form, either masonry or timber-framed, present. A mortar filled 'robber trench' (their context 29) is shown in the east face of Trench D (their Figure 2, here Fig. 79a). One is tempted to say that floor and Wall D and the robber trench all go together and form a building coming off at right angle to the mid-fifteenth century east wing. The problem with that theory is that we have no sign of a floor, north wall or a robber trench for this building in the side of Trench E. The 'stone foundation' shown in Trench F has to be the 'blocking wall' (Wall A) going of at an angle to the northeast so that feature cannot be the north wall of this conjectured structure (Fig. 79a). It is just possible that a wall trench could exist within the 1.25m unexcavated area between Trench E and Trench F, but that (even for AW) is stretching conjecture a bit far. In their plan, what may be a post-hole appears up against the west face of Trench H and it is of course possible the whole of this building, which the floor tells us had to be present, was a timber-framed structure, perhaps partially destroyed by Wall D (of (unexplained) 'relatively recent date') in Trench P. What appear to be further post-holes, at least two and perhaps three, are shown in the section drawing of Trench E. Perhaps significantly a very level grey shelly soil deposit overlain by an equally level layer of charcoal (their context numbers 5 and 6) exist immediately above these possible post-holes. Whether the shelly soil represented an earth floor we are unable to say. The charcoal was presumably cold when deposited for otherwise signs of scorching on the soil surface would surely have been mentioned. (At least one likes to think it would be mentioned, but the more one reads the report the less sure we can be!). There is no sign of the mortar floor in the section drawings of Trench D and Trench E, the east sides of which are only 0.45m away from the west face of Trench H. Such a width is unlikely to contain a north to south aligned masonry wall or its foundation. In this writer's view the presence of a timber-framed structure seems the most likely interpretation of this meagre evidence. If that idea is correct then Wall D may well be a later boundary wall, but this writer is unable to see why it should be 'relatively modern'.

Whether any of these deductions are true can only be found out one way. The areas within the garden need to be re-excavated and the walls and floor surfaces followed so as to reveal their relationships one to another and the adjacent layers. This does of course assume that the layers and walls in question have not been dug away in the eagerness (fanaticism?) of the excavators to uncover Roman deposits. So much more could have been done with this excavation and the subsequent report. AW is quite sure more could be learnt from the archive material, if there is any, but AW is equally sure it would take considerable effort, even if it could be located.

As if the above weren't problem enough, Harrison and Williams tell us there was 'a foundation 3ft thick constructed of Ragstone , chalk and some pieces of fire-reddened Caen stone', (Wall C) overlapping the north side of the Norman building. When first studying the plan this writer thought that they had not shown this foundation. However, after some thought it was realised that what should be the foundation is represented by lightly lined shading. This shading in Trench I, and also in, what is here designated, Trench Ia, looks, to this writer, more like a stone rubble surface. The shading used bears no relationship to the key (nor does the key show the shading of Wall A or Wall B come to that). Their '3ft thick' presumably means 3ft wide. On the east side of Trench I, aligned north to south, we can see an edge to this material, the measurement there does represent 3ft or thereabouts, but we can of course see that it passes beyond the limit of the north side of this trench and, hence, is not its full width. On the west side of Trench I we have another north to south aligned edge, but the east to west measurement between the two edges is the equivalent of 10ft (c.3m). Also of course we can see a northern edge in Trench Ia. The north to south measurement across what is assumed to be the same material in each trench is in the region of 8ft (c.2.60m). Presumably it was not dug away to the east when the stone cobbles '5ft down' were revealed and so we have its limit in that direction, but we can see that the material passes further to the west (as shown in Trench Ia). This is a classic example of why separate plans at different levels need to be drawn, rather than a composite plan or, if not drawn, why a single plan has to have adequate notation. The lack of a section drawing is even more disappointing. If this material does represent a foundation we are looking at a wall almost 12ft wide (not 3ft), of unknown westward extent and which just comes to a stop at its east end. The conclusion has to be that this is not a foundation. The type of shading used (a very minor point) and its extent, at least 12ft east to west and 8ft north to south, suggest this is far more likely to be a rubble pavement outside of the door into the Norman building. We are told that it 'partially' overlaps the Norman building, but unfortunately not which part of that structure. From its position on the plan we could surmise that it overlapped the door passage. Such an overlap is not a problem, the two could still be contemporary. Admittedly, the statement that this wall was apparently constructed, 'At some period after the demolition of the Norman building ...' (ibid., 24), is more of a problem, but the lack of a section drawing showing demolition material of the Norman structure below or being cut by this supposed wall, creates considerable doubt. If it is a wall foundation, then its date has to be sandwiched between the demolition of the Norman building and the building of the east wing or, alternatively, it may represent a separate structure after the building of the latter, but before the ?early seventeenth? century northern garden wall was constructed. Whether it formed a building or a boundary wall or, far more likely, a rubble yard surface in front of, and contemporary with, the Norman building, we also have no-way of knowing. There is of course a way to find out ... shovel, dirt, dig. The only way!

The excavation undertaken in 2005 creates even more problems. Within the archive report (Guinness 2005) there is a considerable amount of information, but there was no attempt to solve any of the problems (even assuming they were identified) created by the earlier excavations. At the end of the day the work is little more than a list of what was found. For this individual the text, drawings and plates were not always easy to relate one to another. (However, that may well be a failing (one of many) on the part of AW). Six potential medieval walls are mentioned.

Wall H (CAT context 219). A length of wall c.2.50m in length and c.0.40m wide in line with Wall O found in 1998 and 2007. However, the masonry was bonded by a completely different type of mortar (see below). The wall is described as being of eleventh century date, albeit with a ? on Figure 2 of the 2007 report.

Wall I (CAT context 220). A very short length of wall coming off at a right angle (to the west) of Wall H.

Wall J (CAT context 236). A wall c.1m in length and c.0.50m wide. This may, and probably did, form the west side of Building 3.

Wall K (CAT context 237). A much robbed fragment of wall abutting Wall J.

Wall L (CAT context 238). A short length of wall abutting Wall J.

Wall M (CAT context 239). Not noted on Figure 2 of the report, but it can be worked out that this must be the wall that abuts Wall J from the east and was probably the same as the northern portion of Wall N found in 1998.

None of these masonry elements are easy to understand. Wall H was, at first thought to be a continuation of Wall O found in the roadway in 1998. However, it would appear a completely different type of mortar was used to bond the masonry. In the 1998 excavation the use of a hard chalky offwhite mortar was noted whereas in the 2005 excavation the material bonding the Ragstone and flint rubble was described as, 'dark orange-brown moderate sandy mortar' (Guinness 2005 para 5.3.10). Whilst mortar differences can occur in a very short length of walling the discrepancy here is quite considerable. However, after much thought the evidence of the alignment outweighs the difference in mortar colour and the two walls are here regarded as part of the same structure. They may have been built on different days with different materials. Such an assumption may of course be completely wrong. The statement, 'the wall appeared to predate the previous foundations, located to the north in 1976-77', (ibid. para 6.3.3) and its being given an eleventh century date (albeit with a question mark) on the plan, is unlikely to be correct. Its northern end, just 0.40m below the modern ground surface (ibid. Fig. 2 in that report ) is too close to Wall B and its associated chalk floor. The floor should have appeared in the trench side of the 2005 excavation all we have is a note stating, '1976-77 trial trench'. This is telling us that the floor has either been destroyed by Harrison and Williams or survives at a lower level. One does like to be optimistic and AW has opted for the latter interpretation. Unfortunately because of the lack of section drawings from the 1976-1977 excavation we do not know at what level that wall was found. If it is accepted that Wall B is at a lower level, then Wall H cannot be eleventh century or anywhere near as early. It must be fourteenth century or later. The use of hard chalky mortar in Wall O, here regarded as contemporary with Wall H, would also hint at a late medieval or even post-medieval date rather than one of the late eleventh or twelfth century. (Having said that the dark orange-brown sandy mortar would be fine with that earlier date).

As Building 9 would not have a wall (Wall B) from an earlier structure protruding through its floor, logic tells us that Wall H must either be contemporary with, or later than, that floor. This wall was regarded as either a 'freestanding boundary wall or an internal wall within a larger structure', (ibid. para 6.3.3). One would like to say that it was contemporary with Wall B, but it is no-where near a right angle with that wall, nor with the later Wall A. Admittedly there is a thirty year break between the two plans and different excavators will be using different fixed points and this sort of problem can account for discrepancies in drawings. However, here the difference is so large, twenty degrees, that we should be able to discount such an error.

That another wall, Wall I, came off at a right angle before the north end of Wall H was reached does, suggest that, even if the latter does form a boundary wall, we are looking at a building extending westwards (Building 9). As with Wall B, so this wall also must end before Trench E of the 1976-77 excavation are reached. Also, as no southern wall of this conjectured structure appears in Trench H or Trench P then its west wall must be to the east of those trenches and hence much of the building (assuming such existed) will be in the entrenched part of the garden and at a lower level than the 2005 trench.

Despite the significant discrepancy in mortar type between Wall M and that found by the present writer in 1998 and 2007 (Wall N), compact yellow mortar and shelly mortar respectively, here the two walls are regarded as forming part of the same structure (Building 3). Repairs or alterations may have taken place, indeed we know that they did, for the external face of both the north and east walls were repaired with flints. Rather surprisingly Wall M abuts, rather than bonds with, Wall J, which was bonded by a dark yellow mortar. For the original interpretation of there being a Norman tower at this point to remain 'safe' they should bond. However, even with this (relatively minor) stratigraphic problem it is hard to think of any other valid interpretation for the small square structure that appears here.

Harder to explain, is the presence of Wall K and Wall L. The text of the 2005 report tells us,

'Fragmentary traces of a wall return (237)(Wall K) suggested that the northern end of wall 236 formed the north-eastern corner of a structure', (ibid. para 5.3.2).

Presumably here the two walls (Wall K and Wall M = 236) bonded one with another. It would appear that we have another building to the west of the suggested tower. However, that the two walls together formed the north-east corner of a building, which then has to be situated to the south, is more problematic. More problematic if for no other reason that Wall L would be in the way. Wall L then creates even more problems for in the key of Figure 2 of the 2005 report this wall is shaded as being an eleventh century wall (albeit as with all the purported Norman walls it is shown with a question mark), but in the text it is described as, 'may have been a later foundation' (ibid. para 5.3.7). The foundation abutted Wall J and was bonded with 'a clay-mortar matrix'. Assuming it is a Norman wall it seems unlikely that Wall K could represent a building to the south, up against the Roman town wall, but such a structure (Building 11) whilst quite narrow, c.3m, could be represented by Wall L. If that is the case then Wall L and Wall M do not form the north-east corner angle of a building, but the south-east corner of a building situated to the north (Building 12). As with all the other conjectured buildings we do not know the westward extent of Building 11 or the northern (or western) extent of Building 12. Wall K could be an eastward extension of Wall D, but even if that is correct there still has to be east, north and west wall.

As with the 1976-77 excavation, with only minimal excavation in selected areas, far more could have been achieved with the 2005 project and especially its report. AW is fully aware that there are all sorts of limitations as to what can be done and where with any commercial archaeological project, but very small trenches (a 'sondage') quickly dug, recorded and backfilled before anyone, who might object, knows it has been dug, can tell us so much.

With the two projects undertaken by the present writer in 1998 and 2007 the structural elements are (thankfully) relatively easy to describe, if for no other reason that excavation was limited to removal of the modern tarmac and its underlying brick rubble bedding, a depth of no more than 0.25m. The impressive archaeology revealed was then cleaned and recorded. Fragments of the Roman town wall were found in the 1998 excavation. Extending northwards from the inside face of the Roman wall three fragments of wall (Wall N), consisting of Ragstone rubble bonded by a yellow shelly mortar, were observed. A later flint facing had been inserted along the external face of the wall and was again observed when the north wall was further exposed in 2007. This refacing is not mentioned when a c.1m length, forming the west end of this north wall was found in 2005 (Wall M). A c.1.50m length of the west wall was also observed in the 2005 excavation (Wall J) and a small, but un-noted, portion to the south, probably, in the 1976-77 excavation. This 2m x 1.80m (internally) structure has to be a turret. A wall (Wall O) extending northwards from Wall N is almost certainly a late medieval boundary wall constructed when, at least the east wall, of the suggested Norman turret was still standing.

In the area of the Bishop's Palace, as so often when dealing with the archaeology of Rochester, we are in a bit of a pickle. All we can say for sure in regard this area is that the structural history of the site is far more complex than is shown on any of the excavation plans or reported in any of the articles. We can put forward the following structural elements with a reasonable degree of confidence, but for the most part their extent, function and date are unknown. It is known from documentary references that structures existed on the west side of the garden area, a kitchen, wash-house, prison and 'rooms' are all mentioned (Livett 1895, 44. A corbel still remains on the inside face of the wall immediately north of South Gate House and a brick well was observed by the writer when part of the Roman wall was uncovered in (what was then) a small yard area way back in 1991 (Ward 1991b). Apparently two corbels also exist to the south (Clarke 2014).

Excluding the walls observed by Irvine and Cobb from the south wall of the nave southwards we can probably identified twelve buildings:

Building 1 a chalk foundation, passing to the west of the later monastic boundary wall will, presumably, be earlier than that wall. Now, but not at the time it was first seen, it is regarded as being more likely Roman than Norman in date and may well be associated with the building known to exist below the south wall of the cathedral nave. A wall (Wall P) seen by Greville Livett is also likely to be part of this Roman building, but may possibly be (as he thought) be an Anglo-Saxon structure.

Building 2 is Norman, datable to the earlier decades of the twelfth century and is here considered to be a gate-house.

Building 3 is Norman, datable to the earlier decades of the twelfth century and is here considered to be an internal turret.

Building 4. The northern face of the surviving block of the medieval Bishop's Palace is constructed formed from much reused stonework, but the rear, southern, face (Fig. 82) is much more interesting (and complex). Two (?)original windows survive which Greville Livett tells us are of fifteenth century date (Plates 80 and 81), but have the outward appearance of being Romanesque (see above, page 71). That they are not as early as the twelfth century is (reluctantly) regarded as being more likely, but a date in earlier than the fifteenth century seems certain.

Building 5. Two courses of herringbone masonry in the perimeter wall may represent a Norman building perhaps of the earlier twelfth century (perhaps even late eleventh).

Building 6. Perhaps a quoin, and more certainly a door and threshold, represent a building above the herringbone masonry. This structure could be of any date from the twelfth century down to the eighteenth century.

Building 7 a wall (Wall B) and more especially a chalk floor containing 14th century pottery, indicates a building existed at this point. It must date to before the east wing of the Bishop's Palace (Building 9) was constructed and it seems likely that the twelfth century Norman gate-house, Building 2, was still standing. Wall A cuts this structure, but in itself is regarded as a boundary wall rather than representing a building.

Building 8 is represented by a mortar floor (and possibly Wall D). This building could be contemporary with, date to before or date to after the demolition of the east wing. In other words we do not know its date. At a guess it is contemporary with or slightly later than the construction of the mid-fifteenth century east wing and is so shown on Figure 82a.

Building 10. The east wing of the Bishop's Palace supposedly dating to c.1450.

Building 13. Wall S and Wall R may represent a building, perhaps associated with those found in 1937 in the triangular piece of lawn to the south of the nave.

More problematic

Building 9. Two walls found in 2005 (Wall H and Wall I). In the report of that excavation they were regarded as eleventh century, but as far as this writer can see, for the reasons stated above, they have to be fourteenth or the first half of the fifteenth century. As Wall H continues to the north of Wall I the two may form a narrow room, perhaps a stair.

Building 11. Another wall (Wall L) found in 2005 also regarded as eleventh century in date and parallel to the old Roman town wall presumably forming a Norman or later structure up against that wall.

Building 12. Another wall (Wall K) found in 2005 perhaps representing a Norman or later structure to the north-west of the turret Building 3.

There is only one way to sort out the archaeological problems (basically a total muddle) in relation to these buildings. First, re-excavate all of the trenches previously dug, not necessarily all at the same time. Draw all of the exposed section sides and then expand those trenches to find the extent of individual structures. Such a project would need to be properly funded and need two or three young and enthusiastic (and paid) individuals, as well as a small number of volunteers (no more than four) to undertake the heavy work, some of the on-site recording and washing of pottery. Because of the sheer logistical problems in regard soil movement the excavation process would need to be taken over four or five weeks in summer, spread over a three or four year period.

Visit Rochester Cathedral →

Join the Friends →

Archaeology →

Explore the studies, conservation and excavations revealing the sites 1,400 year history.