The Roman and Medieval Defences of Rochester in Light of Recent Excavations

Aiver Medwo y Cost le Presumed Site of Roman South Gate .,,,, ROMAN RO C & MEDIEVAL H if" S TE Based on G.M. /J,ivett. Arch. Cont., xxi (teps). 50 R 􀀂l ---------- -- Metres% 50 I 100 I Yards Fxo. 1. Presumed "Site of Roman North Gate Remains of Rom'on East Gate [f.U􀀇 p. 55 THE ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER IN THE LIGHT OF RECENT EXCAVATIONS* By A. C. HARRISON, B.A., and CoLIN FLIGHT, B.A. THIS is an account of a series of excavations in Rochester, jointly undertaken by the writers between 1960 and 1966; but, in the absence of Mr. Colin Flight abroad, it is only right that the responsibility for any shortcomings should rest with the first-named, who prepared it for publication. The account of the first excavation near Northgate (Northgate A) is by Mr. P. J. Tester, F.S.A., who directed it. Initially, there was no intention to challenge the accepted account of the history and development of the defences of the city; in fa.et, the original excavation was merely an attempt to date the Roman wall, a problem which still lacks a definite solution. Each excavation, however, posed questions demanding further enquiry and made inevitable a reconsideration of all the relevant evidence, both documentary and archooological. For the Roman period, our findings are additions to the accepted account, but our view of the sequence of events for the medieval periods differs radically from that of our predecessors. The Roman defences of Rochester were last treated systematically in 18951 when George Payne, for the first time, correctly identified portions of the city wall as Roman and traced its circuit on the south, east and north sides of the city (the exact line on the west side facing the Medway was then, and still is, unknown);2 in the same publication Canon G. M. Livett described the development of the medieval defences for which there are two main sources of documentary evidence ;3 firstly, a series of entries in the Close Rolls of 1225-27 describing payments made for work in Rochester which included a new ditch and, secondly, documents relating to the grant to the Priory by Edward IlI in 1344 of that part of the ditch extending from the Prier's Gate to the east gate. Briefly, Canon Livett considered that the Roman wall was intact to the end of the Saxon period and underwent several extensions starting from the time of Gundulph (1077-1108). The position of the walls • This pa.per ba,a been printed with the aid of a grant from the Council for British Archroology. 1 His account was to some extent supplemented by R. E. M. (now Sir Mortimer) Wheeler in V.O.H. Kent, iii, 80 and 88. 1 It bas been a. convention, since at least the seventh century, to regard the city as ba.ving its long axis due east-west. 3 These a.re given in full by St. John Hope, A.rch. Oant., xxiv (1900). 55 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER and ditches mentioned below is given in Fig. 1 and, diagrammatically, in Fig. 2. These were: (a) A rectangular enclosure round the old Bishop's Pa.lace whose north wall coincides with the Roman city wall, which Canon Livett ascribed to Gundulph or his successor, Bishop Ralph (1108-14), and considered that the small length of much-patched wall at the junction of St. Margaret Street and Boley Hill Street, site of the second South Gate, belonged to this enclosure. (b) A long narrow strip running from the south-east corner of (a) to a few feet south of the Roman south-east angle. Canon Livett considered that Prior's Gate, or rather its twelfth-century predecessor, as the present gate can hardly be earlier than the late-fourteenth century and may be later, belonged to this wall, ascribed to Bishop Ernulf (1114-23). (c) Following the capture of Rochester in 1215 by King John, the great ditch mentioned by the Rochester chronicler was begun in 1225;4 according to Canon Livett this ditch ran south from Eastgate, turned west near the Roman south-east angle and continued just to south of the wall mentioned in (b) above. At the west end, he considered it swung further to south to enclose Boley Hill and, spanned by a bridge or causeway, to give access both to the second Southgate and to Prior's Gate. (d) Canon Livett thought that in 1344 this ditch was filled in and replaced by a new one further south and a wall was built along its northern lip as far as Prior's Gate, which he considered was rebuilt then. The east wall of the city was e:x:tended to meet this new wall which Canon Livett claimed to have 'traced throughout its course', but in fact the tracing was done with the probe. (e) At some unknown later date, another wall was built enclosing a further area to south, including the ditch mentioned in (d) above, and the east wall was again e:x:tended with a bastion at the south-east corner. This account was in pa.rt challenged by St. John Hope who accepted the early Norman and later Norman extensions and agreed that the documentary evidence shows a ditch dug in 1225. He identified this, however, with the one visible in the Deanery orchard and considered that the wall along its lip was built at the same time, although not specifically mentioned in the Close Roll entries. This enabled hlm to equate the still existing wall near the Vines with that authorized in 1344, but he had to admit that the provisions of the charter were not fully carried out, for neither was the eastern part of the earlier ditch filled in nor was there any trace of the new ditch. The position of these walls and ditches is shown in Fig. 1 and, diagrammatically, in Fig. 2. ' 2 Cott. M.S. Nero D.2. 132, cited by St. John Hope. 56 ROCHESTER N N C D Reconstructed North-South Section Alter G.M. livett { 􀀄- r L,, _ ,􀀅 Romo"n Woll ,. After W.H. St.John Hope I L., 7' ., ...:.il Romon Woll A.C.H. mens. ltr /0 * .. 20 I 9 6 3 5 E Sookowoy -􀀄- 'Ernu/1/011' Woll ( J ,,L....,J Lute- Norman Wal/- JO 40 Grovel 50 ,I 1225 Ditch I Feet H 16th Centu ry:':,,;;----1.μ,,i.J.J-'A Pit 1225 Ditch 0 s FIG. 2. D e a n e r y G Romon PU G a r cl e n ,. ( 1 ..L....,􀀋...,..J 1225 Wc,/1 10 IS Metres s s [face p. 50 ROCHESTER 􀁑 Rom o n 􀁒 Mediev al D Modern New Deanery S 0 /0 20 Fee t I 9 6 .J .. S b._. I --. . I lcioy Romporl Gorden Woll Inner race, excavated by - Llvett,/89S 1 ,,... "\ oloss House ' J 􀀉 / Deep Sookowoy 16th Crmtury Pit Conons'Houses .JO 40 I O s Metres 10 FIG. 3. f 1nn<1r L.I' . no PosW0 ·,note Approx, .- K\ .,. .- .\- -\ p\ \ ,_ \ - ':7 0\ I \_.􀀘 [facep. 57 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER TB:E ExoA VATIONS Culling A. (Fig. 3 shows the positions of all these cuttings, except Northgate A and B.) In 1961 Cutting A (Section e-f) was excavated against the inner face of the Roman wall which, though completely buried internally, here survives to a height of about 16 ft. On the other side of the Roman wall the upper part of its facing is visible but, as the lowest few courses have fallen away, the underpinning is modern. At the bottom of the cutting (Fig. 4) a layer of dark earth, 5 in. thick, overlaid clean orange brick-earth. At the extreme western end of the cutting was a small round pit cut into the brick-earth and containing an urn in white-slipped red ware (Fig. 15, no. 56). There was no sign of anything else in the pit nor of anything in the urn. The pottery from the dark earth (Layer 9) was of the third quarter of the second century A.D. and also contained a fragment of a pseudo-Venus figurine.° Flints and earth formed the foundation of the remains of the Phase I rampart consisting of alternate layers of blue clay and dirty yellow sand (Layer 7). The rampart had survived only to a maximum height of 3 ft. 6 in. and, on the evidence of the pottery, cannot be earlier than the third quarter of the second century A.D. There is no more precise terminus ante quem for the rampart than the construction of the wall, though there are indications {e.g. the small amount of silt in the ditch belonging to it),6 that its life may have been fairly short. The foundation of the Roman wall (Phase II) was a layer of concrete some 15 in. thick, though it is possible that deeper digging would have disclosed a foundation trench as well, similar to that in Cuttings B and C. This concrete raft (Fig. 4, Layer 11) was approximately 8 ft. wide; set back 6 in. from its edge, the bottom course of the internal facing projects to form a plinth 6 in. wide. For the next six courses the face is 'battered' until an 8-in. offset reduces the thickness of the wall to 5 ft. 10 in.; the next seven courses are vertical up to where another offset, 12 in. wide, further reduces the thickness to 4 ft. 10 in. Five more courses of the facing are intact; then the core is exposed to a total height of 16 ft., and it is likely that only the foot-walk and the parapet have been lost. The core consists ofragstone rubble liberally grouted with hard gritty brown mortar; no use was made of lacing courses of bonding-tiles. The face is of untrimmed ragstone, coursed and heavily pointed with the same mortar. The pointing of the inner face of the wall was unweathered, showing that it had been protected by earth piled against it from the outset. This wall-bank (Layer 5), covered the face of the wall to a. point just a.hove the upper offset and enclosed the stump of the Phase I rampart. a Mr. Fre.nk Jenkins, M.A., F.S.A., kindly identified tWa fragment as originating in the Allier area. • As shown in Cutting D (Fig. 6), 57 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER CUTTING A 40'6"0.D. . . . . .. . . 3· . . . . . . . . . . . . . . . . 9 1 I I I I I Section e-1 - - ' ROMAN WALL ' SE 'I'., '' :' . . / : . . . . . ' . .. . . - - - - - - _,_ -4 . . ...... // : .' -- - - -..!..--t' ' - - - -------- I ' I I I Flo. 4. Cutting behind Roman wall at its south-east angle. 1. Post-medieval debris. 2. Dark soil a.nd yellowish sand. 8. Dirty brick-earth. 4. Filling of medieval pit. 5. Roman wall-bank. 6. Sandy fill of wall foundation trench, divided by mortar-scatter. 7. Rampa.rt. 8. Layer of flints forming foundation of 7. 9. Secondcentury occupation layer. 10. Pit with vessel no. 56 (fig. 15). 11. Conoret.e raft of flints in hard brown mortar. 58 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER In Eagle Court the much-weathered external facing of squared ragstone blocks regularly coursed is intact above the level of the lower internal offset. George Payne observed7 that the lowest courses of the facing extended back into the core. If this were the case (and nothing of the sort was to be seen internally), they must have formed ashlar bonding-courses and cannot be the foundation, as he supposed, because they are approximately 4 ft. above the actual base of the wall. The south-east angle forms a regular curve with an external radius of about 32 ft. and lacked indications of either an external bastion or of an internal tower. The construction trench between the Phase I rampart and the wall was filled with layers of brick-earth, separated by a thin spread of mortar-scatter resulting from the building of the wall (Layer 6). A large medieval pit had been dug through both the wall-bank and the remains of the Phase I rampart (Layer 4). The top of the overlying deposit of dirty brick-earth (Layer 3) is at about 30½ ft. above O.D. against the wall and has a very slight slope to the west. A similar levelling is apparent inside the wall in the Deanery Garden (Cutting C) at about 29½ ft. above O.D. and is probably to be connected with the back-filling of the 1225 Ditch. If that is so, it is probable that this surface also represents the same mid-fourteenth-century horizon, of which there are other indications to the west and south. Layer 3 was overlaid by a midden of dark earth streaked horizontally with yellow sand {Layer 2); it contained pottery of c. A.D. 1300, like the preceding layer, and including a grey-ware jug (Fig. 16, no. 75). There was also part of a louver (see Appendix I), and a small stone mould (see Appendix II) which point to a fourteenth-century dating for this layer. Deanery Garden, 1962-65. With the object of discovering whether the Phase I rampart found at the south-east corner continued to the west, excavations were begun in 1962 among the trees at the end of the Old Deanery lawn (TQ 74366838). Cutting B was sited 2 ft. to the west of the point where George Payne dug in 1894, and proved to be directly south of the line of the Roman wall (Fig. 3). Gutting B (Section c-d). (a) The earliest features were the truncated remains of a comple.x: of nine pits and four gullies, with very little datable material (e.g. Fig. 13, nos. 20 and 21}. So little was left of the pits and gullies of this ea.rliest period, amounting to no more than a few inches in each case, that no firm conclusions as to the nature of this OCQμpation could be reached, except that there was some unidentified activity during the late.-fust o r early-second century A.D. (b} Approximately 3 ft. 6 in. from the south end of the cutting the lip of a ditch rumung east-west-was traced; it was filled by precisely 1 Arch. Oant., xxi (1896). 59 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER SOUTH CUTTING 8 Sttf/011 c-d NORTH 2 4---------···- 10 SCALE 11'1 FEET Flo. 5. Cutting B, across line of south Roman wall in Deanery Garden. 1. George Payne'e excavation, 1894. 2. Tudor debris. 3. Mortar back-fill of wall-robbers' trench. 4. Brown soil with tiles and debris. 5. Medieval domestic rubbish and Roman mortar rubble. 6. Medieval filling of brown soil and rubble. 7. Medieval accumulation of dark soil and rubble, 8. Rema.ins of Roman wall resting on foundation of flints and brick-earth. 9, Ditch of pre-wall defences, back-filled with clay over silt. 10. Filling of first-century gully (G.l). the sa.me mixture of blue clay and dirty sand noted in the construction of the Phase I rampart in Cutting A, though the proportion of clay was higher (La.yer 9). Trees prevented further extension to the south (but see Cutting D, below). ( c) At the north end of the cutting was the south face of the Rom an wall, here demolished down to the concr ete foundation, though a very large lump of the core rested upon this. This concrete raft was 1 ft. 3 in. thick and rested upon a foundation-trench 2 ft. 8 in. deep, alternately filled with rammed flints and brick-earth. (For a complete section, see Cutting C.) 60 SOUTH CUTTING D 2 1---------------- ---··-··· --·--- ----1 ---- --··-· -·-· .. --····--·---l 1-----------= 6 =-- ----·-· ··- - ·-·-·-----·-···----1 ---····· ·--·--·· UN EXCAVATED Section a.I, .Jl'B"QD. ' ' CUTTING C ·Fxo. 6. Cutting C and Cutting D, across line of Roman wall in Deanery Garden. 1. George Payne'a excavation, 1894. 2. Post-medieval build-up of tiles and building debris. 3. Mortar and tile back-fill of wall.robbers' trench. 4. As for 2. 5. Soil and rubble. 6. Medieval filling of soil with tiles, rubble and domestic refuse. 7. Mortar r ubble. 8. Medieval turf.line. 9. :Medieval filling of soil and chalk rubble. 10. Flinty gravel. 11. Remains of Roman wall. 12. Foundation of wall composed of layers of flints and brick-earth. 13. Olay rampart (Phase I). 14. Gravel rampart material (Phase I). 15. Sandy rampart material (Pha􀂗e I). 16. Phase I ditch backfilled with clay over silt. 17. Flint foundation of rampart. 18. First• to secondcentury occupation level. / / 4 SCALE IN FEET NORTH 1 I. I [facep. 01 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER (d) All other Roman levels had been removed by an excavation which extended from the face of the wall to the southern limit of the cutting. The excavation had been back-filled by a mass of material containing medieval pottery. Four layers were distinguishable, but the pottery from Layers 4-6 does not differ appreciably in date, which is considered to be c. A.D. 1300 (Figs. 16-17, nos. 79-94). This large medieval excavation is fairly confidently to be identified with the ditch known from documentary evidence to have been dug in 12258 and which the monks of the priory obtained leave to fill in 1344,9 Layer 7 may represent the surface of the 1225 Ditch while it was still open. On the surface of Layer 4 a small drain, made of ragstone and floored with roofing-tiles of unusually large size, 11 in. X 7½ in., crossed the cutting obliquely near its centre and discharged into a circular soak-away pit, 2 ft. 6 in. in diameter, filled with stones and broken roofing-tiles as well as numerous fragments of a large grey cooking-pot with slight shell-filling (Fig. 17, no. 95). Ai, this pot was clearly not a rubbish survival, it may indicate that this common thirteenth-century form continued in use until the middle of the fourteenth century. (e) Above the medieval filling and directly below the topsoil was Layer 1 consisting almost entirely of building debris and architectural fragments with pieces of 'billet' ornamentation paralleled on the Norman west front of the Cathedral (Fig. 18, no. 7). This layer, which has been found to extend for a considerable distance and to contain sixteenth-century material, is considered to represent a levelling of the area after the demolition of monastic buildings, either in 1541-42, when the cloisters were being converted into a house for Henry VIII or, more probably, after 1558 when Lord Cobham sold the site of the Priory to the Dean and Chapter. Cuttings O and D. (Section a-b.) With the double purpose of obtaining a continuous section across the wall and the Phase I rampart, if present, and of investigating the ditch discovered at the south end of Cutting B, Cuttings C and D were opened 2 ft. to the west of B and overlapping it at either end with a 6-ft. baulk between them (Fig. 6). (a) The earliest period noted in Cutting B wM represented in Cutting C by the large Gully G.l which ran obliquely the length of the cutting, passed beneath the wall and through Cutting B in a southerly direction. It was 5 ft. wide and 2 ft. 6 in. deep, had a flat bottom and sloped southwards fairly steeply. Pottery from the filling was scarce, except for a few Roman sherds suggestive of mid-first-century date. (b) The filling of the gully was covered by Layer 18 which may be identified with the top-soil in the middle of the second century as it 8 Arch. Oant., xxi (1895),· 51, and xxiv (1900), 12 ff. 0 Arch. Oant., xxiv (1900), 16 ff. 61 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER contained Antonine pottery, and above this was the gravel foundation (Layer 17), of the Phase I rampart (Layer 13). The rampart itself, which was of the same construction as previously noted in Cutting A (see p. 57 above) survived to a height of 2 ft. 6 in. The bands of clay which were thick and close together at the south end tapered off to the north. Evidently, the clay formed a facing or revetment of the bank and was bonded back into it. The top of the surviving portion of the rampart was more or less flat with only a thin layer of ruhbly earth separating it from the sixteenth-century debris above and there was no trace of a wall-bank overlying the cut-back rampart as in Cutting A. (c) In Cutting D the profile of the ditch located at the south end of Cutting B was exposed (Layer 16); it was V-shaped in section with its deepest point about 8 ft. 6 in. below the estimated Roman ground level. It had, of course, been truncated by the medieval ditch, but, if the sides are projected to the level of the original surface, a width of about 20-22 ft. seems likely. This agrees fairly closely with the dimensions of the first-century ditch at Verulamium which was found to be approximately 9 ft. 6 in. deep and 19 ft. wide at one place1° and 10 ft. deep and 22 ft. wide at another.n A small amount of fairly clean silt (Layer 19) had accumulated in the bottom. Above this silt the ditch was filled with material identical with that of the Phase I rampart. The conclusion seems inescapable that this filling was derived from cutting back the rampart when the wall was built and that the ditch it was used to fill belonged to the same defensive system. Presumably it was thought to be too close to the wall for re-use in Phase II. (d) (Plate I). A complete section was out across the wall foundation, which was all that remained at this point (Layers 11 and 12). About 15 in. of the south edge of the concrete raft had been eroded M a result of the medieval ditch being dug right up to its face. Immediately above the concrete raft was a quantity of mortar debris (Layer 3) which probably represents the remains of the robber-trench mostly obliterated by George Payne's excavation. (e) The medieval layers to the south of the wall presented much the same picture as in Cutting B except that the filling of the ditch was more homogeneous and the line representing the weathering of the surface of the ditch more clearly defined. Again, at this level quantities of Roman mortar suggestive of demolition had been dumped. A large oval pit 4 ft. 6 in. X 3 ft. 6 in. x 8 ft. 6 in. had been dug from the level of the turf-line through the lower medieval layer and the filled Phase I ditch into the subsoil. The filling of dark earth produced a bone comb of early medieval type (Fig. 18, no. 2). From the earth at the top of 10 S.S. Frere, 'Excavations at Verulamium, 1955' ,Antiq. Journ., xxxvi (I 956), 4. 11 Antiq. Joum., xli (1961), 80. ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER the pit, however, under a clay seal, came a few fragments of glazed pottery identical with those from the layer above it and indicating that the sealing of the pit and the filling of the ditch were probably contemporary. The medieval pottery from Cutting D was similar to that from Cutting B (Fig. 17, nos. 96-107). (f) The layer of sixteenth-century building debris was continuous across Cuttings C and D except where interrupted by George Payne's excavation. The results of these excavations therefore were, firstly, that the presence of the Phase I rampart and ditch on the south side of the Roman town was established and, secondly, that the medieval excavation south of the wall was shown to extend for at least 20 ft. OuUing E (Section g-h). The above results immediately called in question the identification and dating of the wall whioh both Canon Livett and St. John Hope show as running from Prior's Gate to the north-east corner of the Deanery Garden, described as the 'later Norman wall' and attributed to Bishop Ernulf (1114-24) (Fig. 3). Cutting E (Fig. 7) was therefore made across the known line of this wall which S CUTTING E Svction g-h N 1 WALL ' FEET Fta. 7. Cutting E, through Livett's 'Ernulnan' wall. 1. Broken roof-tiles. 2. Sandy soil with tiles and chalk. 3. Flints, chalk and tiles. 4. Sandy soil with tiles and flint rubble. 5. Sandy soil with chalk. 6. Coarse gravel. 7. Medium gravel. 8. Coarse gravel. 63 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER was found to exist to a height of 3 ft. with a 4-in. projection at the bottom above a foundation layer of large lumps of ragstone and flint. (These were not laid obliquely as described by Canon Livett.)12 Its width was 2 ft. 8 in. and its construction was of fairly small pieces of ragstone, roughly coursed and with the mortar joints set with small pieces of stone in the manner known as 'garretting'. The face of the cutting to the south of the wall disclosed an important fact. Under the usual sixteenth-century layer of building debris, five layers of earth and gravel sloping at about 45° from south to north were distinguishable. The three lowest layers all ran underneath the foundation course of the wall, which was therefore very clearly built on and into 'made' soil. Although this tipped material of clean gravel differed from that in Cuttings Band D, it seemed certain that it formed part of the same process of back-filling and, as this could not be earlier than c. A.D. 1300 on the evidence of the pottery and was probably later than A.D. 1344 on the documentary evidence, it was clear that the wall could not possibly be earlier than the fourteenth century. This conclusion was supported by an examination of the structure itself: not only did it seem inadequate for a defensive wall, being narrow and with very slight foundations, but the masonry itself was quite unlike any known Norman work in Rochester. Finally, the practice of garretting does not seem to be usual before the sixteenth century which makes it probable that this wall was erected in connection with the conversion of the Priory in 1541-42. Cuttings F and G-H (Figs. 8 and 9). The identification of the medieval ditch and the doubt cast upon the 'later Norman wall' CUTTING F )' ·c ,, ! I I 􀀋 · ·--- ' '\ ' ' 􀀊 GAROEN PATH I -·- --· - 􀀍 - I I 1 ·-· -􀀆- 2 I􀀃. =-, --- Ill Sect/of/ i-i .. - 􀀌 I􀀇 FIG. 8. Cutting F, through channel under path in Deanery Gorden. I. Topsoil. 2. Soil containing rubble and tiles. 3. Dark soil containing rubble, tiles and mortar. 11 Arch. Oant., xxi (1895), 50. 64 SOUTH If) () )> rI'll z-n I'll C CUTTING H d .... • I ••••• • • •• •'1 . . . • · 1 /. NORTH 3 Loyers 5-7, /225 Ditch Filling . : : -: . :_:-:. ·.: .· .· : :_. ·_. :- ·_. . ----- --- ----- CUTTING G Section k-1 Fro. 9. Cutting G and Cutting R, through filling of medieval ditch and Roman pit. 1. Tiles, etc., filling pit or re-cut channel. 2. Silting of channel. 3. Tudor debris. 4. Dirty brickearth. 5. Si.ndy gravel. 6. Coarse sand. 7. Dirty brick-earth. 8. Silt. 9. Silted gravel and sand. Natural deposits: a. Brick-earth. b. Yellow melt-mud. c. Sandy gravel. d. Sand and pebbles. ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER suggested that the next point to be investigated should be the southern extent of the ditch and the wall presumed to lie to the south of the former beneath the grass path in the Deanery kitchen-garden (see p. 56, para. (d) and Fig. 2). For this latter purpose Cutting F (section i-j) was cut across this path and it wa-s found that this overlaid a channel cut into undisturbed brick-earth. This channel was 12 ft. wide and 3 ft. 3 in. deep with sloping sides and was filled with building debris, mostly roofing-tile.13 There were only two pieces of pottery recovered, both probably of Tudor date. Of the wall there was no trace whatever and the channel cannot reasonably be interpreted as a robber-trench. Cutting G (Fig. 9, Section k-1) was sited with the intention of finding the lip of the medieval ditch, but fell short of it. At this point the layer of sixteenth-century building debris (Layer 3) was much thicker and should be regarded as part of a. pit; it contained a. sixteenth-century German brass jetton having the arms of Wertheim.14 Beneath this was the same clean gravel noted in Cutting E with tip-lines again running from south to north and followed by dirty brick-earth. One significant point was noted: the natural geological sequence here is reddish brickearth and then, after a narrow band of pale yellow melt-water mud,16 a sandy gravel. In the :filling of all the southern portion of the medieval ditch the order is reversed with brick-earth at the bottom, indicating that the material was dumped in the order in which it was dug out from another deep excavation. This can hardly have been other than the ditch still visible in the orchard to the south, which Canon Livett, correctly it would seem, considered to be the one authorized in 1344, whereas St. John Hope identified it with that of 1225 (see Fig. 2). Cutting H was then made as an extension to the south of Cutting G and presented a rather complex picture. The earliest feature is a deep excavation reaching no less than 19 ft. below the surface and containing some Antonine pottery at its lowest level. The south end of the section had been distur bed by the channel noted in Cutting F, here 10 ft. wide and 3 ft. 6 in. deep (Layer 2) which had apparently been recut before being finally filled in with tiles and debris (Layer I). This channel ma.de it impossible to define the exact lip of the 1225 Ditch at this point. The sequence appears to have been as follows: a deep excavation of unknown e;xtent was dug in the Antonina period or later. It is possible that this may represent the southern lip of the ditch dug in front of the Roman wall, if this was steep-sided and with a flat bottom, contrasting with the V-shaped profile of the ditch associated with the Phase I 1s It is suggested that this was what Canon Livett felt with his probe (Arch. Gant., xxi (1895), 62. l& Miss M. Archibald of the British Museum kindly provided this identification. l5 Apparently a sludge-deposit of the kind described by M. P. Kerney, in 'Late-glacial Deposits on the Chalk of South-east England, Phil. Trans. lfoJt. Soc. L to the end of the third century any final judgement must be reserved. The table on the facing page summarizes the evidence available. There is some slight evidence of a third period of the defences (Phase III) near the north-west angle. In 188936 according to George Payne the town wall was exposed at a distance of about 100 ft. from the river and is described as being 7 ft. thick, of rags tone in hard pink mortar with a double bonding-course of tile. A. A. Arnold in 188737 described a similar 33 Arch. Gant., xxvii (1905), lxix•lxx. 8' A. L. F. Rivet, Tawn and Oount1-y in Roman Britain, London, 1958, 90-92. On the other hand! Professor Frere has already suggested a. date still later 'in the first. half_of the third century, perhaps under Alexander Severua or Gordian III'. (Brua,n,nia, 263.) 35 For Reculver see J.R.S., li (1961), 191, and lv (1965}, 220· Antiq Journ xli (1981), 224-8. ' ·• For Canterbury, see S.S. Frere, Roman Oanterbury 3rd ed. p 10 36 Arch. Oant., xxi (1896), 8. ' ' · · 31 Arck. Gant., xviii (1889), 1.94. 76 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER DATlNQ EVIDENCE FOlt ROll!AN DEFENCES Coarse wa.Tea Samian Suggested Cutting Context (Figs. (Fig. 15) da.ting 13-14) PHASE I Z6 50-75 Earlier deposits C Gully l sealed below Layer 18 22-24 150-175 ra.mpaTt A Layer 10 66 75-100 Layer 9 40-42 150-175 B Pits in Layer 10 20-21 80-120 Foundation layer of C La.yer 17 19 150-175 earth and flint.a A La.yer 8 27-30 170-190 Body of rampart C Layer 13 16-17 150-175 Layers 14-15 31-35 150-175 A Layer 7 36-39 5 170-190 PRAsE n Earlier deposits, B Layer 16 18 170-190 Northgate Wall-trench B Layer 8 25 170-190 Body of wall-bank A Layer 6 46-47 170-190 Layer 5 43-45 170+ Northgate B Layer 12 1-15 1-4 170-190 fragment of wall in the same area cut through by a. drain as 'undoubtedly Roman' whloh in context means almost certainly that the mortar was pink. Thls seems to suggest a. wall very different from that found elsewhere, and closer, for example, to the late third-century work at Richborough.88 0. Saxon. The survival of the Roma.n defences is attested by the description of the city as a 'castellum' in charters of the seventh-ninth centuries and by reference to the walls and gates in the same documents and in accounts of the unsuccessful siege by the Danes in A.P. 885. D. Norman. The circuit of the Roman wall continued in being during the Norman period, though the building of the Castle towards the end of the eleventh century must have involved alterations in the south-west corner, as the defences of the castle seem always to have been independent of those of the oity. A length of the Roman wall near Southgate, and possibly part of the west wall as well, must therefore have been demolished when the oastle-ditoh was dug. It will be remembered that both Canon Livett and St. John Hope considered that two extensions were ma.de to the south during this period, but neither of these now seems supported by the evidence. The suggestion of an early-Norman rectangular extension enclosing the 38 J.P. Buiihe-Fox, Exca.vationa of the Roman Fort a.t RichboroughIV, Oxford, 1949, 81. 77 ROMAN AND MEDIEVAL DEFENOES OF ROCHESTER Bishop's Palace depends upon two points: firstly, the identification of a small piece of much-patched wall as being of that date (it will be shown that it is more probably of the fourteenth century) and secondly, upon the existence of three documentary references to the Bishop's Palace;39 the first dates from the time of Gundulph, the second relates to a re-building by Gilbert de Glanville soon after 1185 and the third occurs in a document of 1459 which Bishop Lowe dated 'from his new palace at Rochester'. It is only in the last case that there is a clear correspondence between the document and the existing building located on the line of the Roman wall and the reference to the 'new palace' could be taken to mean that it was built here for the first time: Gundulph's and Gilbert's palace may well have been elsewhere. The 'later-Norman extension' ascribed to Bishop Ernulfwas bounded by the wall investigated in Cutting E. This excavation showed conclusively that this wall must be later than the fourteenth century and the style of the wall suggests the sixteenth. In his account of this extension Canon Livett shows a drawing4° of what he takes to be the continuation of this wall at the east end of Minor Ca.non Row. It is, however, very different in character and the sketch suggests that it Wa.':! sunk into the filling of the 1225 Ditch. E. Thirteenth century. The 1225 Ditch referred to by the Rochester chronicler and in the Close Roll entries may with some confidence be identified with the great excavation shown to extend from the south face of the Roman wall to a point about 100 ft. further south.41 This ditch seems to have had sloping sides and to have been of considerable depth in the centre (the bottom was not reached even in the deep soak-a.way, which reached a depth of 24 ft. from the present surface). It is not quite established whether the 1225 Ditch continued to follo,v the line of the Roman wall all the way to Southgate, leaving the Refectory (and possibly the Palace) projecting into it like towers, but there are two indications that it may have done: (I) Canon Livett's drawn section41a of the wall he found at the east end of Minor Canon Row shows the sloping side of an earlier ditch running beneath the wall with its lip a.bout 8 ft. from it which can hardly be a foundation-trench as he suggests. Unfortunately, since the orientation is not given, this could either be the inner lip of a ditch 87 ft. from the Roman wall or the outer lip of one 106 ft. from it. In the so Arch. Oant., xx:iv (1900), 60. ,o Arch. Gant., Jed (1896), Plate I. 41 But one note of caution should be sounded. We have, like previous writers taken it for granted throughout that the ditch back.filled in 1344 was the same􀃿 􀄀he d.it<:h dug_in_ 1225. This seems only common sense but, in fact, direct evidence 1s lacking. Suni larly, although the Deanery garden ditch can confidently be equat ed, in view of the pottery evidence, with that be.ck-filled in 1344 the date when it was first dug ce,nnot be closely .fixed. ' 41􀄁 Arch. Cant., x.xi (1895), Plate I. 78 P1.,\TE J [foe, p. ,􀀈 l?LATE HA Pholo: R. G. Foorcl A. 'Arch of Construction' in fourteenU1-century Eust \Vall. PLATE llB Pholh: C. R. F/ig/1/ .B. Roman North 'Wall showing Conduit. PLA'l'E Ill Photo: I' . .J. '/'tsler A. :Sqt1t.ll'ecl fttcing Stone􀁫 on outside of' Ro111un Wull. Photo: P . ./, Test,r B. Coursed Rubble l'ocing on tho inside of tho Roman Wall. PLATE IV Photo: R. G. Foord A. Copy in Lead of sixteenth-century Jetton. .{,·􀀈'- . I ',') ,' , .•, (' /, B. Stone Mould for casting Buckle and Trinket. ROMAN AND MEDIEVAL DE.FENCES OF ROCHESTER latter case it can reasonably be equated with the south edge of the 1225 Ditch and this is supported by the fact that the filling, shown as gravel and brick-earth, is the same as was found further east. (2) The heavy buttressing of the south side of the Refectory, where no less than seven buttresses project for some 6 ft., can hardly be part of Bishop Ernulf's original building. Perhaps they date from Bishop Haroo de Hethe's rebuilding in 1336, when the ditch would still have been open. (Indeed it seems quite possible that the rebuilding was necessary because the Refectory was tending to slip into the ditch.) On this assumption the 1225 Ditch would link up at Southgate with the ditch a.round the Castle. The street leading to St. Margaret's must have been carried across it by a causeway, or possibly a bridge, and it then swerv-ed sharply eastwards around Boley Hill, an earthwork south of the Castle which, on present evidence, seems also of thirteenthcentury da.te.42 If our interpretation of the medieval wall underlying the pavement at Northgate is correct, then this gate was rebuilt during the second half of the century. Otherwise there is no reason to suppose that there were any other alterations to the rest of the circuit. F. Fourteenth century. It is this period that presents the greatest problems. It is common ground that in 1344 the monks were granted a charter empowering them to fill in the 1225 Ditch from Eastgate 'to the gate of the said Prior' and to take over its site: our excavations have shown the ditoh was in fact so filled (see pp. 61-2 above). The width known from documentary evidence agrees well enough with that of the ditch outside the east wall and the length with the distance a.long the Roman wall from Eastga.te to Southgate: the 'gate of the said Prior' may be Southgate or a gate nearby leading into the preoinct (it cannot have been on the site of the present Prior's Gate, as both Canon Livett and St. John Hope supposed).4'3 In return for this concession the monks were to build a new wall and dig a new ditch further south. The ditch which still exists, extending the line of the previous one on the east and curving around a new south-east angle westwards for about 400 ft., can safely be identified with this 1844 Ditch. It is clear, too, that the old ditch was largely backfilled with spoil taken from the new and it is even possible to trace the progress of the work, which proceeded from east to west. The tip-lines near the east wall which slope to the west show that to the south of the old south-east angle a large bank of earth was first thrown across the line of the 1226 Ditch, in order to delimit that part of it which was now obsolete (cf. Cuttings J.Q). After that, filling proceeded directly from the south and the tip-lines slope to the north or north-west. The ground ,􀀢 Arch. Cant., lxxiv (1960), 197-8. 0 .Arch. Cant., xxi (1896), 63, 8Jld .xxiv (1900), 21. 79 ROMA.i.'i AND MEDIEVAL DEFENCES OF ROCHESTER inside the wall was levelled off at the same time, producing the horizon already noted in the Cuttings A and 0, and approximately 250 ft. of the east end of the Roman south wall was demolished, as indicated by the dumps of mortar debris underlying the back-fill in Cuttings Band D. The new wall was evidently intended to run along the strip of undisturbed ground between the two ditches--Qn the line laid down on Canon Livett's plan, parallel with the Roman wall and about 120 ft. from it. This was clearly started from the west end. The wall running westwards from Prior's Gate, intact at least until 1720,44 but now la,rgely demolished except for the fragment mentioned on p. 56 above, is on the line and in all probability belonged to it. Prior's Gate itself is probably later, but the soar left on its eastern. face suggests that the wall continued some way in that direction. Contemporary with this work was the construction of the wall along the south side of the High Street in accordance with the charter of 1345.45 It has been located opposite the choir of the Cathedral in 188746 and again i n 1960 in the cellar of No. 88, High Street.47 It is, however, abundantly clear that the work was never completed. The eastern section of the wall which should have been apparent (a) in the foundation trenches of the new Deanery, (b) in Cutting F, (c) in Cutting H, and (d) in Cuttings L-M, was simply not there. The ditch also seems to have come to a premature end after about 400 ft. just ea.et of the end of :Minor Canon Row (see p. 75, above). It is reasonable to suppose it was the Black Death of 1348-49 that interrupted the work. The next development was the abandonment of the unfinished ditch and wall and the extension of the walled area to the south to include the site of the 1344 Ditch and a strip of land beyond it. The new wall ra.n southwards a.cross the line of both ditches (piers being sunk through the 'fill' as was observed in Cuttings J a.nd -K) for some 325 ft., to a oiroular bastion and then turned westward for about 680 ft. (This is Canon Livett's 'post-1344' Wall.) The west wall of this enclosure was intact in the eighteenth century but its exact line is now lost; it returned to join the unfinished wall previously mentioned a little to the west of the present Prior's Gate which may itself be contemporary with it (Fig. 1). Whether this extension was ever authorized or whether the king ever complained that the monks had failed to carry out in full their share of the undertaking will probably never be known; nor is it possible to assign an exact date to the work which may well be of more than one period, as the east wall at least is certainly a complex structure. There « Thorpe,. Regi8trum Rcffen.se, 552. 45 Arch.. Cant., xxiv (1900}, 23. •• A.rah. Oam., xviii (1889), 201. 􀁫7 Unpublished. Information lcindly supplied by Mr. R. E. Chaplin, B.Sc. 80 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER a.re, however, two indications of a possible date which are worth noting. The first is documentary: John Sheppey, the Prior (1333-51), claimed in a petition to the Pope,48 that among other good works he had 'enclosed the whole (priory) with a strong wall'; thls obviously relates to the two charters of 1344 and 1345 already mentioned, but in view of the established fact that the 1344 Wall was abandoned unfinished, Prior John's claim might mean that the farther 'post 1344' extension was completed by 1350. The seoond po.int is that, as already noted (see pp. 67-8above), the building technique of the eastern city wall where it crosses the thirteenth-century ditch is identical with that of the east curtain-wall of the Castle known to have been built by Prior John of Hartiip in 1368. In view of these points and the probability that the southern defences would not have been left for long in such disarray, it may well be conjectured that this final phase was completed during the third quarter of the fourteenth century. THE FINDS* I. PO'l'T:ElW (i) Romano-British By A. P. DETSICAS, M.A., F.S.A. Abbreviations and References Oamulodunum C. F. C. Hawkes and M. R. Hull, Oamulodunum, Oxford, 1947 Oanterbury Sheppard Frere, 'Canterbury Excavations, Summer, 1946', Arch. Gant., lxviii (1954), 101-43. OGP J. A. Stanfield and Grace Simpson, Central Gauli8h Potter8, London, U)58. Cobham P. J. Tester, 'The Roman Villa in Cobham Park, near Rochester', Arch. Gant., lxxvi (1961), 88- 109. Colchester M. R. Hull, The Roman Potters' Kilns of Colchester, Oxford, 1963. D. J. Dechelette, Les VaseB ceramiques ornes de la Gaule romaine, ii, Paris, 1904. Dover L. Murray Threipland and K. A. Steer, 'Excavations in Dover, 1945-1947', Arch. Oam., l:xiv (1951), 130-49. Gillam J .. P. Gillam, 'Types of Roman Coarse Pottery Vessels in Northern Britain', AA', xxv (1957), 1-40. Greenhithe A. P. Detsioa.s, 'An Iron Age and Rornano-British Site at Stone Castle Quarry, Greenhithe', Arch. Gant., lxxi (1966), 136-90. 4; Cat. Papal Pet. 1, 192 and .217, cited by R. C. Fowler, V.O.B. Kent, ii 123 The flnda have been deposited at the Eastgate House Museum, Ro􀂔hes􀂕r. 81 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER J.W. Leicester Lu. Lulling stone o. Ospringe Rich.borough I-IV Southwark Su:arling Winchester AA' Arch.Oant. P. J. Tester a.nd J. E, L. Caiger, 'Excavations on the Site of a Romano-British Sett.lement in. Joyden's Wood, near Bex ley', Arch. Oant., Ixviii (1954), 167-83. K. M. Kenyon, Excavaftion8 at the J e'UR'Y Wall Site,• Leicester, Oxford, 1948. W. Ludowici, Katalog V, Stempel, Nannen, und Bilder romischer Topf er aus meinen A usgralJungen in Rheinzabern, 1901-1914, Jockgrim, 1927. G. W. Mea.tes, 'The Lullingstone Roman Villa', Arch. Oant., lxvi (1953), 15-36. F. Oswald, Index of Figure-types on TerraSigiUata, Liverpool, 1936-37. W. Whiting, W. Hawley and T. May, Report on the ]jJxcavaftion of the Roman Cemetery at Ospringe, Kent, 0:lrl'ord, 1931. J. :P. Bushe-Fox, Excavafli.ons of the Roman Fort at Richborough, Reports I-IV, Oxford, 1926-49. K. M. Kenyon, Excavations in Southwark, 1959. J. :P. Bushe-Fox, Excavation of the Late-Celtic Urn,. field at Su:arling, Kent, Oxford, 1925. B. Cunliffe, Winchester Excavations, 1949-1960, Winchester, 1964. Archreologia Aeliaria, Fourth Series. Archc.eologia Oantia!na. Coarse Wares {Figs. 13-16) The coarse pottery discussed in this report is a small portion of the wares recovered in the various cuttings, and a still smaller selection is illustrated; wherever possible reference is being made to such wares in published reports. Pa.ate is used to denote the core of the vessel, and fa.bric to describe the finished vessel as for texture and colour. Northgate B, Section o-p, Layer 12. 1. Dish in grey-brown fabric and grey paste, burnished (Oolchester 40B, post A.D. 120; Greenhithe 189, A.D. 120-150; Ospringe 28; Leicester, fig. 20, no. 3, Antonine). 2. Dish in grey fabric and light grey paste (Oolche.ster 38, .A..D. 120- 400; Southwark, fig. 15, no. 3, Antonine). 3. Dish in grey fabric and sandy paste, rimless, but with a. cordon above the lattice decoration (Gillam 234, A..D. 140-200; Greenhithe 168, A.D. 120-150). 4. Dish in grey fabric and light grey paste; smaller but closely similar to no. 1, above. 5. Dish in grey fabric and paste, burnished (Oo"lchester 40A, mthout wavy-line decoration, poat .A.D. 120; Ospringe 47, etc.; J. W., fig. 5, no. 29, Antonine; Lullingstone 115, Antonina). 82 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER FIG, 13. (¼) 83 ROMAN AND MEDmVAL DEFENCES OF ROCHESTER FIG. 14. (¼) ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER I (½) r1 ,r ' i,,r - jl , r I l 1' I\ 1 I I t-􀀃 I􀀊 \-i1J 􀀉 :: :: '􀀊',' 1I! • (d; 60- . . , ii• '. (¼) FIG. 15. 85 L-U ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER 6. Dish in light brown fa.bric and grey paste, burnished ( Ospringe, 304, 364; J. W., fig. 5, no. 30, .Antonina). 7. Dish in grey-brown fabric and grey paste, burnished (Southwark, fig. 24, no. 8; Lei.cester, fig. 20, no. 1, post A.D. 150; Oobham, fig. 5, no. 26, post A.D. 150). 8. Dish in grey fa.bric and paste, similar to nos. 1 and 4, above. 9. Dish in grey-black, smooth fabric and grey pa-ate, with traces of polishing, imitating samian Forms 18 or Lu. Sb (Colchester 310, post A.D. 140; Riclworough III, 235, A.D. 80-120; Leicester, fig. 38, no. 2, Trajanic). 10. Necked jar in grey fabric and grey, sandy paste, burnished around the rim and neck (Southwark, fig. 16, no. 1, second century A..D.). 11. Jar in grey fabric and light grey, shell-gritted paste, with a pronounced cordon round the neck of the vessel; this may well belong to the series of cordoned jars discussed in Greenhithe (q.v.) and dated to the second century A.D. 12. Bead-rim jar in grey sandy fabric and paste (Greenhithe 219, 222, A.D. 120-150; Southwark, fig. 17, no. 14, second century A.D.). 13. Dish in grey-black, smooth fabric and grey paste, with a rather flattened rim and traces of polishing, probably in imitation of samian Form 27 (Riclworough Ill, 226, A.D. 80-120), although the outline of the vessel suggests a shape less rounded than the samian prototype. 14. Necked jar in grey fa.bric and paste, burnished (cf. Ospringe 57, ll0). 15. Necked jar in grey-black fabric and grey paste (of. Ospringe 6, 51, etc.) . .Also from this layer: Dish in grey fabric and pa.ate, burnished (Ookhester 40A or B, post A.J>. 120); another dish, closely similar to the preceding vessel; cavetto-riro jar in grey fabric and paste, with bur• nished rim (Gillam 116, A.D. 125-150; Colchester 278, .A.D. 100-200; G-reenhithe 126, A.D. 120-150); another cavetto-rim jar, similar to the previous one; dish in light-brown fabric and paste, burnished (Cokhester 38, post A.D. 120; Southwark, fig. 15, no. 6, Antonine); 'poppy-head' beaker in grey fabric and paste (Greenhithe 192, A.D. 120-150); cordoned jar in grey fabric and burnished rim (probably Greenhithe 234, A.D. 120-150); cavetto-rim jar in light grey fabric and burnished rim {similar to Greenhithe 233, A.D. 120-150); jar in grey fabric with a rolled rim {similar to Ookhester 268A, A.D. 100-250); dish with recurved rim in briok-red fabric and grey paste (Canterbury, fig. 8, nos. 68 and 79, A.D. 100-150); fragments of a colour-coated, folded beaker in cream paste and grey-black slip, with faint traces ofrouletting (probably similar to Colchester 406, late-second to early-third century A.D.); a. fragment from the base of another colour-coated beaker in light-brown fabric and white paste (post A.D. 140); large cavetto-rim jar with a wide, burnished 86 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER band below the neck a,nd burnished lattice decoration (Cokhe,st,e1· 279B, A.D. 200-350; Greenhithe 233, A.D. 120-150). Chronologically, the most significant factor is the presence in this deposit of sherds from two colour-coated vessels, providing a, secure terminus post quem certainly not earlier than A.D. 140, scarcely earlier than A.D. 150.48°' The majority of the pottery found in this la.yer consists of vessels in the tYPical Antonine black-burnished fabric, whether as dishes or jars, pa.raUeled by the Colchester and Greenhithe vessels; the latter, though dated in their own context to c. A.D. 150, undoubtedly remained in use much later, certainly at least as late as c. A.D. 170. It would seem, therefore, almost certain that, on balance, this layer cannot be dated much earlier than the last quarter of the second century A.D., a dating further supported by the samian ware (Fig. 15, nos. 1-4) found in the same context. Dating: c. A.D. 170-190. Gutting C, Section a-b, Layer 13. 16. Small jar in grey fabric and paste, polished and decorated with dots applied en barbotine; no exact parallel was found as to the shape of this vessel, but vessels with similar decoration are well known (Colchester 122, A.D. 150-350; Greenhithe 142, A.D. 90-120; RichJJorough III, 278, A.D. 70-100; Lullingstone 137, .Antonine; Leicester, fig. 27, no. 23, A.D. 80-120, and fig. 42, no. 45, A.D. 125-130). 17. Dish in grey fabric and paste, burnished (Colchester 37, A.D. 70- 170; Greenhithe 169, A.D. 120-150). Also from this layer: Fragment ofrough-ca.st beaker (post A.D. 150); cordoned ja.r, with burnished lattice decoration below the cordon, probably similar to the Greenl􀅧ithe (q.v.) series of such vessels (lateAntonine). Dating: c • .A.D. 150-175. Northgate B, Section o-p, Layer 16. 18. Dish in grey fabric and paste, burnished (G7eenhithe 180, A.D. 120-150; Colchester 87, .A..D. 70-170). Also from this layer: Oavetto-rim jar in grey fabric and paste, burnished (of. Gillam 140, A.D. 180-270; Oolchest,e,r 278, A.D. 100-200; Southwark, fig. 21, no. 24, Trajan-Ha.dria.nic). The pottery found in this la.yer falls within the same chronological limits as that recovered in Layer 12. Dating: c. A.D. 170-190. Gutting 0, Section a-b, Layer 1'1. 19. Dish in grey fabric and paste, burnished (Greenhithe 182, .A..D. 120-150; Oolchest,er 37, A.D. 70-170). ,11a It is worth noting that a ooin of Marcus Aw·elius was found in the wallbank nearby in an excavation in 1967. 87 lO ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER Also from this layer: Rim-fragment from a small-necked jar in grey fabric and paste (of. Southwark, fig. 22, no. 7, A.D. 100-160); various scraps of storage jars in shell-gritted paste (of. Greenhithe, 202, 204-5, A.D. 120-150). Dating: c. A.D. 150-175. Cutting B, Section c-d, Pit intru.sive in Layer 10. 20. Cordoned jar in grey fabric and paste, with a cordon immediately below the rim and another round the neck; burnished decoration of trellis (cf. Oankrbury, fig. 12, no. Ill, without decoration and rim missing, A.D. 100-120; Richborou-gh II, A.D. 80-120, and III, 274:, A.D. 80-120, but with different rims). 21. Jar in brown fabric and paste, faintly rilled (Leicester, fig. 35, no. 18, Flavian). Dating: e. A.D. 80-120. Cutting 0, Section a-b, Layer 18. 22. Cordoned jar in grey-brown fa.brio and grey, sandy paste (Greenhithe 155-6, 158, etc. A.D. 120-150). 23. Bead-rim je.r in grey fabric an.d paste, sandy (Greenhithe 216, A.D. 120-150). 24. Flagon in cream fabric and grey paste, with white slip on external surface (cf. Ospringe 522). Also from this layer: Fragments of shell-gritted storage jars (second century A.D.). Dating: o. A.D. 150-175. Cutting B, Section c-d, Layer 8. 25. Bead-rim jar in grey fabric and shell-gritted paste (Greenhithe 88, 99 and 138, A.D. 120-150). Though the material recovered in this la.yer was slight, it is quite clear from the evidence of other layers that no chronological distinction can be established between those deposits and this layer. Dating: c. A.D. 170-190. Gutting 0, early Gully Filling continuous with Gutting B, Section c-il,, Layer 10. 26. Storage jar in light brown fabric and paste, with a. double cordon below the rim; probably hand-made (Leicester, fig. 36, no. 26, Claudius-Nero). The filling of this gully supports the evidence mentioned above (p. 59) of first-century A.D. activity, but the material available is too scant to allow for more than a genera.Ida.ting of c. A.D. 50-75. Cutting A, Section e1, Layer 8. 27. Mortarium fragment49 in fine cream fabric with thick pink core; •0 Notes on the mortaria. were kindly provided by :Mrs. K. F. Hartley, B.A. 88 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER a few flint grits survive. The broken herringbone stamp can be identified as one used at the Colchester kilns, and as one of the Colchester dies most commonly attested. on other sites, particularly Antonina sites in Scotland (Ookhester, fig. 60, no. 30, and pp. 114-16; ..A.D. 150-190). 28. Dish in grey fabric and paste, burnished and probably decorated with a lattice which has now worn off (cf. Greenhithe 176, A.D. 120-150; J. W., fig. 5, no. 28, Antonina). 29. Dish in grey fabric and paste (cf. Greenhithe 181, A.D. 120-150; Colchester 37, A.D. 70-170; Gillam 219, A.D. 125-150; Leicester, fig. 19, no. 3, A.D. 125/30-200; Southwark, fig. 15, no, 2, second century A.D.). 30. Jar in grey fabric and paste, burnished. Also from this layer: Fragments of three storage vessels in shellgritted paste (Greenhithe 202 and 205, ..A.D. 120-150); fragment of a cavetto-rim in grey fabric and paste, burnished over rim (Colchester 278, second century .A..n.). Dating: c. A.D. 170-190. Culling 0, Section a-b, Layers 14 and 15. 31. Dish in grey-brown, burnished fabric and grey paste (GreenMthe 187, 239, A.D. 120-150; Gillam 310, .A..D. 170-210). 32. Dish in grey fabric and light grey sandy paste, burnished (Cokhester 37, A.D. 70-170; Greenhithe, 180-1, A.D. 120-150). 33. Jar in grey, smooth fabric and grey paste, rouletted decoration (late-Antonine). 34. Jar in brown fabric and grey, sandy paste, with traces of burnishing on the high shoulder and rim, rilled (close to Southwark, fig. 20, no. 3, A.D. 100-120). 35. Jar in brown fabric and paste, with some shell-grit (near Greenhithe 118, A.D. 120-150). Also from this layer: Carinated beaker in grey, smooth fabric and paste (cf. Riclworough III, 291-2, .A..D. 80-120); rim fragment of 'poppy. head' beaker, Antonine; storage jar in light brown, shell-gritted paste (Southwark, fig. 18, no 2, second century A.D.); jar in brown fabric and shell-gritted paste (Antonine). Dating: c. A.D. 150-175. Gutting A, Section e-f, Layer 7. 36. Mortarium fragment in very hard, dark pink-brown fabric with near black core, closely reminiscent of tile in texture. This piece is certainly not from any of the major potteries producing mortaria, but there is some evidence that such a fabric could have been made in Kent and origin there seems most likely. A Flavian date would not be impossible for this piece though it could be second-century. 37. Dish in grey fabric and paste, burnished (Greenhithe 182, A.D. 120-150; Ookhester 37, .A..D. 70-170). 89 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER 38. Dish in grey fabric and paste, burnished (Southwark, fig. 15, no .. 5, late-second to early-third century A.D.; 08pringe 5, etc.; Oolchester 38, post A.D. 120; J. W., fig. 5, no. 31, .Antorune). 39. Dish in grey fabric and paste, burnished (Ookhester 37, .A.D. 70- 170; Greenhithe 58, A.D. 150-200; Riolworough Ill, 46, Antonine). Also from this layer: Fragment from a storage jar in brown, shellgritted paste, with thumb-nail decoration (Greenhitke 202, A.D. 120- 150). Dating: c . .A.D. 170-190. Gutting A, Section e-f, Layer 9. 40. Jar in grey fa.brio and paste with shell-grit, probably handmade (of. Colchester, fig. 73, no. 22). 41. Dish in dark grey fabric and grey sandy paste, burnished on the inner and outer surfaces (of. Greenhitke 170, A.D. 120-150; Gillam 327, A..D. 130-180; Leicesur, fig. 49, no. 5, to .A.D. 220). 42. Probably a dish, in drab grey fabric and light grey paste, with rouletting (of. Richborough m, 212, A.D. 75-100). Also from this layer: Fragments from a storage jar in light brown fabric and shell-gritted paste (second century A.D.). Dating of this layer can only be very appro;ximate as the pottery deposited in it is not only very fragmentary but also covers a. fairly wide range of time; this is particularly true of the samian ware. On balance, a date in the third quarter of the second century .A.D. is not unlikely. Gutting A, Section e-f, Layer 5. 43. Necked jar in grey fabric and paste (Soutliwark, fig. 16, no. 3, second century A.n.; Leicester, fig. 38, no. 20, Trajanic; Winchester, fig. 17, no. 29, A.D. 43-140). 44. Cordoned jar in grey sandy fabric and paste, with one cordon below the rim. 45. Jar in grey-brown fa.bric and brown, sandy paste. Also from this layer: Dish in dark grey fabric and paste, burnished (Greenhithe 184, A.D. 120-150; Colchester 37, A.D. 70-170); a. fragment from the base of a vessel in shell-gritted paste (second century A.D.). Dating: post c. A.D. 170. Gutting A, Section e-f, Layer 6. 46. Jar in light brown fabric and grey, sandy paste, with a rolled rim and traces of soot internally; perhaps hand-made (cf. Leicester, fig. 24, no. 20, A.D. 95-200 and later). 47. Mortarium fragment in fine, yellowish cream fabric with whitish trituration grit. This fabric and grit was used both in the extensive potteries at Colchester, and at Canterbury in Kent. The form is certainly second-century, probably earlier than A.D. 170. 90 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER Also from this layer: Dish in grey fabric, burnished (Colchester 37, A.D. 70-170; Greenhithe 180, A.D. 120-150); mortarium fragment in finetextured, pinkish cream fabric with a trace of grey in the core-this is a difficult mortarium to assess; its place of manufacture is uncertain but Kent would be a possibility. It was probably made within the period A.D. 90-150. Dating: c. A.D. 170-190. Gutting G-H, Section k-l, Layer 8. 48. Carinated beaker in grey, smooth fabric and paste (Rickborough, I, 75-77, first-second century A.D.; Greenhitke 144, A.D. 120-150). 49. Dish in dark grey fabric and sandy paste, burnished (Greenhithe 187, A.D. 120-150; Gillam 310, A.D. 170-210; Cobham, fig. 3, no. 14, A.D. 70-180; Dover, fig. 9, no. 8, late-first to second century A.D.). 50. Dish in grey, smooth fabric and paste, probably in imitation of samian Form Lu. Sb (RicnborOU{lh m, 235, A.D. 80-120; Colchester 310, post A.D. 140). 51. Dish in grey smooth fabric and paste, imitating samian Form Lu. Sb, a.s above. 52. Dish in grey smooth fa.brio, burnished, imitation of samian Form Lu. Sb (RickborOU{lh III, 235, A.D. 80-120; Colchester 310, post A.D. 140; Leicester, fig. 38, no. 2, Trajanic). 53. Dish in brown fabric and paste, rather coarse and with a groove at the top of the vessel: probably hand-made. 54. Bead-rim jar in grey fabric and sandy paste (Greenhithe 222, A.D. 120-150), 55. Bead-rim jar in reddish brown fabric and grey paste with shellgrit (Southwark, fig. 17, no. 15, second century A.D.). Also from this layer: Fragment from a carinated beaker, probably similar to no. 48, above (second century A.D.); jar in smooth grey fabric and paste (cf. SC>Uthwark, fig. 16, no. 3, second sentury A,D.). Dating: post c. A.D. 170. Cutting A, Section e-f, Layer 10. 56. Necked jar in dull red paste, with yellowish cream slip. This vessel can best be described as a. debased copy of a pedestal urn, and its ancestry could be traced back to Swarling 3 and 9, except that the foot of those vessels is rather different from that of the Rochester jar. The shape of this vessel approximates Camulodunum 108 Aa and Bb, reechoed in Colchester, :fig. 75, no. 10; its paste, however, is identical with that often used in the manufacture of flagons (cf. Greenhithe 143, A.D. 120-150). Typologically, this jar would seem to be one of the earliest vessels recovered, and a date in the last quarter of the first century A..D. would not be improbable. 91 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER Samian Ware (Fig. 15). (a) Plain Forms. The great majority of the sherds found were too fragmentary for the purpose of close dating but, in general, agreed broadly with the coarse wares. Form 33 predominates in the series of cups, Form 27 is represented by few certain fragments, and the early Form 24 is totally lacking; likewise, in the plate forms, Form 18 is not represented, and most pieces belong to Forms 18/31 or 31, with a fair proportion of the late Form 3 l(Sa). Only one fragmentary stamp was found on a cup of Form 33, and this probably reads (PISTI)LLIM. South Gaulish samian is entirely absent. (b) Decorated Forms. Here, too, no South Gaulish samian was present. Forms 29 and 30 are absent, except for one or two scraps which may belong to the latter form. All the sherds described below were found in Northgate B, Section o-p, Layer 12, except for no. 5, found in Cutting A, Section e-f, Layer 7. I. Form 37. East Gaulish, in fair condition but with glaze almost completely worn; in the style of HELENIVS of West􀄛rndorf (K. Kiss, A Westerndorfi Terra-Sigillata. Gyar, in Archaeologiai Ertesito, Serie III, 1946-48, Budapest, 1948). The design is enclosed by a double basal line over which have been impressed several spiral decorative details (Kiss, OJJ· cit., pl. VI/64), also known in use by this potter in lieu of the ovolo (Kiss, op. cit., pl. XV/I, 3). The double half-medallion (Kiss, op. cit., pl. VI/80) contains one figure-type, Putto to left (0.459 = Lu. V M288) and outside it a free-style incision in the mould simulates vine tendrils. Dating: c. A.D. 175-200. 2. Form 37. Central Gaulish. Fair glaze. The decoration is badly blurred as this bowl was probably made in a worn mould. The ovolo is rather large and rounded, without a central projection, has a corded tongue ending in a badly-blurred rosette and is enclosed by a mediumsized bead-row border. The remnants of the decoration preserved show a vertical bead-row border terminated by another blurred rosette, which divides the decoration into panels: to left, Bird to right (0.2317), though this figure-type is of such poor relief that it looks rather smaller than Oswald's bird; to right, part of a basket with fruit (D.1069), which has not been completely restored in the drawing as it could be 0INNAMVS's variant (OGP, pl. 158/20} of this ornament rather than Deohelette's detail. The ovolo is ADVOCISVS's no. 1 (OGP, p. 205, :fig. 33); this potter ha.s used a bird similar to, though larger than the figure-type on this sherd (cf. OGP, pl. 112/3, D.1010 = 0.2316). Dating: c. A.D. 160-190. 3. Form 37. Central Gaulish. Rather worn. A remnant of a doublebordered ovolo with a corded tongue ending almost straight. The ovolo roulette obviously completed its circumference of the mould at this 92 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER point and careless application has resulted in superimpression. This is . CINNAMVS's ovolo no. 5 (OGP, p. 267, fig. 47). Dating: c. A.D. 145-190. 4. Form 37. Central Gaulish. In fair condition. A fragment from the upper part of the decoration, with a double-bordered ovolo which has a corded tongue ending in a swollen tip; this is probably the ovolo no. 1 ofIANVARIS II (OGP, p. 213, fig. 34). Dating: c. A.D. 150-180. 5. Form 37. Central Gaulish. In good condition. In the style of CINNAf.fVS, with his ovolo no. 3 (OGP, p. 267, fig. 47), enclosed by a slight wavy-line border, and part of one figure-type, Bear to left (D.775 = 0.1619) which has been often recorded in this potter's work (OGP, pls. 157 /7 and 159/33). Dating: c. A.D. 145-190. (ii) Medieval (fig. 15) By G. C. DUNNING, B.Sc., D.Lit., F.S.A. Northgate B, Section o-p. F01J,ndation of Medieval Wall. The pottery consists of sherds of decorated jugs of fine quality sandy ware, belonging to five pots as follows: 57. Rim, neck and body sherds of light red ware, partly grey in the core. The surface is light red with patchy green glaze on the neck and body. Decorated on neck and body with broad, flat, white strips, diamond-rouletted. The strips are vertical on the neck, curved and branching on the body. The inner bevel of the rim has a white slip. 58. Rim and neck of jug, similar to no. 57 of grey ware with light red surface. White slip covers the outside, the bevel of the rim and the inside of the neck for about I inch. The neck is partly green glazed. Decorated on neck with strips, one diamond-rouletted between two with more irregular marks. A body sherd probably belongs to this jug. It has a vertical diamondrouletted strip and green glaze overall. 59. Two body sherds of grey ware, with grey to buff surface inside. Decorated with plain, narrow, applied strips or ribs. The larger sherd haa four strips, two vertical and two curving to the right, suggesting panels with festoons separated by vertical lines. 60. Body sherd from below handle. Grey ware, thicker than the rest and with harsher surface. The inside is light red, and the outside has a thin white slip. Decorated with vertical diamond-rouletted strips. The glaze is overall, thicker on the strips, green mottled with darker green and brown. 61. Body sherd of thin ware, grey core and inside, thin white slip outside. Vertical diamond-rouletted strip, and streaky green glaze overall. 93 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER DISCUSSION The pottery forms a consistent group comprising two main types of jugs: (1) Nos. 57-58. Tall slender jug!!, about 13-16 in. high. The neck a-nd body are decorated with rouletted strips, either of white clay or coloured by glaze. Complete jugs of this type, variously decorated with linear and curvilinear patterns, have been found in London.5° (2) Noa. 59-61. Shorter jugs, a.bout 12 in. high, with ovoid body. The diamond-rouletted strips on Nos. 59-60 are ty:pical of London,51 where the narrow strips on no. 61 also occur.62 The festoon or arcade motif is also known in other techniques in the, London region.53 The kilns which existed primarily to supply London with pottery were situated in East Surrey, where waste heaps of the industry are extensive. Earlswood has produced a waster with stamped and diamond-rouletted decoration,M and the Rochester pottery was probably made here. The products of this kiln and of the workshop of Limpsfield,65 where the pottery is more archaic in character, have been found at half a dozen sites in north-west Kent.66 Further east in Kent the finds are more sca.ttered, at Rochester, Canterbury and Stonar.li7 The evidence points to a. considerable trade in East Surrey pottery to places along the coastal part of north Kent during the second half of the thirteenth century. Cuttings A-D. (iii) Medieval (figs. 16 and 17) By P. J. TESTER, F.S.A.. The medieval pottery from these excavations is here figured and described mainly for the evidence it affords as to the age of the contexts from which it was recovered. Most of the material is very fragmentary and only a small proportion lends itself to illustration. The main groups contain examples of the well-known type of cooking pot with straighttopped flange of marked projection, which is generally accepted as beginning in this area towards the end of the thirteenth century and continuing into the fourteenth. The few earlier sherds found in association must, therefore, be regarded as rubbish survivals. Probably both ao London M.tllleum, Medieval, Catalogue, 214, pl. bd. ai e.g. London Museum Acc. N os. 17129 and 26665. a2 Guildha.U Museum Aco. No. 17728. Rackham, Medieval E ngli8h Pottery, pl. 48. 63 Med. ATch.., v, 270, fig. 72, 1. 54 Su'M'ey AToh. Ooll., xxxvii (1927), 245. 5 6 C.B.A., Exhibition of Medieval Pottery (1964), 5, No. 6. 88 Arch. Oant., lxxii (1958), 31-9. n Deal Castle Museum. 94 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER l- - - --1 􀀂41 ,64 ( ) )66 ) '6S I r ( 4\: 􀀉68 􀀊69 4',o ,, \,2 I .. ,.1 i \ } ( 11\,s r· -----··i l \ .,BI 1 ==;-i "Bl T. () fl􀀂 D '9S 8iib! FIG. 16. (¼) 95 ROl\1AN AND MEDIEVAL DEFENCES OF ROCHESTER /-, l8 I I 􀀆---- t T, 􀀁 -:i --i===--=======;::,( £ ;,-------,;::::==...--==-------=---_,,( \94 I / L _____J ======1 196 ? 17 I \ 1---: --._,----_ -_ ---,-,c ,:· · ·7 ;-------.-􀀐-=::::--=====;::.,􀀑 "! F 􀀃 - \. I ; 􀀂-- ( t I f t r-"7 ). I rt::- FIG, 17. (¼) 96 !.✓JI. ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER the medieval bank behind the city wall in Cutting A and the filling of the medieval ditch in the Deanery Garden were composed to some degree of surface scraping from within the area of the defences which would account for the mixture of sherds of different ages. In support of this it may be mentioned that Romano-British sherds occurred as survivals in these medieval layers in some quantity. M:r. J. G. Hurst and Mr. S. E. Rigold have kindly discussed the dating of Nos. 62-71 and No. 85 and their helpful comments have been incorporated in these notes. Gutting A (fig. 5), Swtion e-f. Layer 3. 62, 63, 64 and 66. Rim sherds of cooking pots, probably globular and thrown on a slow wheel. Short everted necks and plain rims. No. 62 has a facetted body indicative of knife-tooling. The ware is hard, sandy and grey, and akin in form to East .Anglian 'Early Medieval' ware68 but in some ways closer to Thetford ware in fabric. Evolved wares of this general type last well into the twelfth century in East Anglia and are normal c. 1140 at South Mimm.,, Herta. They do not occur at Eynsford Castle, but cf. possible parallels from Canterbury.so Probably eleventh century, perhaps the first half. No. 63 is in similar ware with a straight neck, perhaps approaching the sandy wares that are usual at Dover in the second half of the twelfth century instead of the shell-filled pottery (S.E.R.). 65. Everted rim, grey ware, lighter on the outside, with some shell filling. Can be paralleled in a mid-twelfth-century conte.xt at Eynsford (S.E.R.). 67, 68, 69 and 70. Cooking pot rims of coarse, heavily shell-filled ware, grey in the fracture but reddish on the surface in some cases. Clubbed rims reminiscent of early twelfth-century form from Eynsford (S.E.R.). 71. Pink-buff cooking pot with simple rim. Possibly twelfth century, but this type persisted. 72. Recessed rim. Grey-buff ware with finely ground shell. An unusual form, possibly a pot lid. 73, 74, 76 and 77. Flanged cooking pot rims. 73 and 74 are greybuff with very small amount of shell, while 76 and 77 are darker with much more shell filling. Similar rim forms occurred at Canterbury in the late-thirteenth century60 and also at Joyden's Wood61 where Dr. G. C. Dunning has dated them c. 1280-1320. The form may have persisted well into the fourteenth century. 68 G. C. Dunning in Medieval Archaeology, iii, 44. 68 Aroh. CanJ., bcviii (1954), 129. Go Arch. Cant., lxviii (1954), 133-4. 01 Arch. Cant., lxxii (1958), 32. 97 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER Layer 2. 75. Unglazed grey ware jug with elongated stabs in handle and combed decoration on body. The context of this vessel is considered by the excavators to be probably fourteenth century. Gutting B (fig. 6), Section c-d. Layer 7. 78. Cooking pot with simple everted rim, similar in colour and shell content to 73. Thirteenth-century and possibly a little earlier typologically than flat-topped rims. 80. Rim with square moulded section. Grey ware with very slight shell content. Layer 6. 79. Cooking pot rim. Grey ware with some shell. 81. Reddish on surface with scarcely any shell. 82. Undercut rim. Grey ware with only a few particles of shell visible on surface. 83 and 84. Rims of jugs in hard grey ware. No trace of glaze. 85. Rim of jug in red clay, decorated with applied spots of cream slip and glazed. Mr. Hurst observes that it is copied from a Rouen jug which would, however, have had a white body-not red. Not much before 1240 in France, but }.fr. Rigold has an imitation from Eynsford which cannot be a great deal later. 86. Base of jug in ware similar to 85. Green glaze under base, and patches of brownish glaze down to irregular line shown on drawing. 87. Jug in dull red clay with decoration of narrow horizontal bands of buff slip. Traces of eroded glaze. Layer 5. 88. Rim of baluster jug. Buff surface with patch of glaze. A wellknown thirteenth-century form. 89, 90 and 91. Cooking pot rims in grey to buff ware with small amount of shell filling. Layer 4. 92. Rim of cooking pot with everted rim and slight bead. Finger-tip impressions on inside of rim. Brown surface with grey core and some shell. Apparently a rubbish survival of twelfth- or early thirteenthcentury type. 93. Flanged rim in grey ware with slight shell content. 94. Rim of unglazed jug in hard grey ware. 98 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER Pit dug int,o Layer 4 from its surfew e 95. Rim and sagging base of cooking pot. Grey ware with very little shell. If, as the excavators suggest, Layer 4 is due to the 1344 back-filling of the 1225 Ditch, this pot (represented by a large number of closely associated sherds, and clearly not a rubbish survival} cannot be earlier than about mid-fourteenth century, and is a useful indication of the possible persistence of this common form and fabric. Cutting D (fig 7), Section a-b. The following sherds came from the medieval ditch filling and are generally similar to the assemblage from the equivalent layers in Cutting B. 96-101. Cooking pot rims of similar ware to 73-82, and 89-91. No. 100 is in particularly thin hard ware and :is probably a late representative of the type. 102-104. Rims of unglazed grey ware jugs. No. 103 has a slight pinched lip to facilitate pouring. 105. Rim of straight-necked jug in grey ware with buff slip on both inner a.nd outer surfaces. 106. Rim of unglazed grey ware jug with traces of strap handle. 107. Rim of jug in grey clay covered with cream slip. Combed decoration and extensive green glaze on exterior. Cutting M (Lower Filling of 1225 Ditch}. 108. Simple averted rim in grey ware with considerable filling of finely pounded shell. Finger-tip decoration on rim similar to No. 92. Probably twelfth or early thirteenth century. n. coms .AND ToKENs 1. Cutting A, Layer 2. A 'Short Cross' penny of King John, struck in Lincoln, 1205. Obverse: Class Va. Reverse: Class Vb, i. reads TOM.AS. ON. NICCQ.62 2. Gutting B, Layer 6. An Irish penny of King John, struck in Dublin, 1205-1210. Reverse reads ROBEC/RDON/DIVEC.62 3. Gutting G, Layer 3. A German jetton of the sixteenth century. The coat of arms on the reverse appears to be that of Wertheim. 4. Found,ation-trench of Canons' Houses in Deanery Garden. Imitation in lead of a jetton (Pl. IV A). The obverse shows a degenerate copy of a late-medieval French jetton's cross, of. Barnard's The Oasting Counter, pl. vi, no. 39, and the reverse a ship which can be paralleled on the Nuremberg jettons of the sixteenth century. Cf. Barnard, op. cit., pl. xxix, no. 9. 63 02 Mr. D. ,T. Bra,nd kindly identified both theae coins. 03 Mr. L. R. A. Grove, B.A., F.S.A., F.R.E.S., kindly provided the information oonta.ined in this note. 99 ROMAN AND :M.EDIEVA L DEFENCES OF ROCHESTER IlI. S:r.u.LL FINDS (Fig. 18). 1. Gutting A, Layer 4. Iron arrow-head, socketed and with an angular blade. This is an early medieval type, very similar to that illustrated in the London Museum Medieval Catalogue, plate XV, no. 1 (Aco. no. 27389). 2. Gutting D. Pit sunk from surface of Layer 8. Handled comb made of bone. The handle is cylindrical in section and decorated with incised lines. The blade is slotted into this and secured by iron rivets, two of which survive. The design is closely similar to the comb illustrated as fig. 30, p. 153, in 'London and the Saxons', London Museum Catalogue, No. 6, 1935, though in that case the handle is of deer-horn. The date is therefore probably late-Saxon or early medieval. 3. Gutting A, Layer 2. Part of Louver. See Appendix I. 4. Gutting A, Layer 2. Stone Mould. See Appendix IT. 5. Northgate B, Layer 5. Bronze handle for chest or cupboard. 6. Northgate B, Layer 5. Iron buckle. 7. Gutting B, Layer J. Fragment ofNorma.n 'billet' ornamentation. The surface of the stone is reddened, probably by heat action, which may point to the fires of 1137 or 1179 which are known to have dama.ged the Priory extensively. MPENDIX I By G. C. DUNNING, B.Sc., D.Lit., F.S.A. Louver (Fig. 18, no. 3). Sherd of grey sandy ware, inside surface light brown to red. The outside is covered by lustrous green glaze mottled with dark brown. The inside has faint wheel-marks, showing that the profile sloped approximately as drawn in the section. At the top is a square-cut edge, roughly horizontal, also glazed. From this edge an applied thumb-pressed strip passes down the side, not quite vertically. The sherd is sufficiently curved to give a diameter of about 20 in. at its lower part. The size and slope, and the cut edge, appear to rule out the sherd being from any recognized form of pot, but all these features are consistent with its being from a roof louver. In that case the out edge is the lower margin of one of the apertures, which are always present on these structures. The diameter is greater than the 14-16 in. at the base of several louvers, and the shape is more openly conica}.04 In these features it is, however, matched by another find from Kent. This comprises large et Louvers of the more usual form v.re described in Arch. Joum., oxvi. 170, figs. 16-17. 100 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER 2 ©-1 4a FIG. 18. (l) 101 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER pieces of a louver found by Professor Sheppard Frere in excavations in St. Margaret's Street, Canterbury, which is provided with large triangular-shaped apertures in the side. On the strength of this analogy the Rochester sherd is identified as part of a large louver of type 1, a separate structure fitted over a hole left in the roof, of the open conical shape forming a variant in north-east Kent. The ware and glaze suggest a date about the turn of the thirteenth and fourteenth centuries. APPENDIX II By BRIAN SPENOER, B.A., F.S.A., F.R.Hist.S. The mould (Pl. IVB and Fig. 18, nos. 4a. and b) is made from a tabular block of dense, homogeneous limestone, about I cm. thick.65 The block has been fractured across its width, perha.ps as the result of thermal shock; a comer is also missing. It is likely, however, that only about a third of the original has been lost. 66 Characteristically, matrices are cut on both sides and the block would have been used interchangeably as one half of two different two-piece moulds. Both sides have been drilled with tapered keying-holes; at least two of these originally contained dowels which would have slotted into corresponding holes in the other sections of the mould to ensure accurate registration. Both sides are also provided with runners through which the molten metal, presumably a lead or copper alloy, would be poured, though the funnel-shaped mouth of one of the runners is missing. One side is designed for casting the two component parts of a small circular buckle (fig. 18, no. 4, which assumes that the matrices in both halves of the mould were identical). The ring (diameter l ·9 cm.) is of round section and has a recess within which the loop of the pin could be fitted. The matrix for the ring is linked by fine incisions to a tiny vent to let out air and gases. The loop of the pin was cast round a core, possibly of wood; a groove to receive this core runs from the edge of the mould and across the top of the pin. Near the point of the pin's matrix are the remains of an iron dowel. Buckles of this type and size appear to belong mainly to the fourteenth century. One specimen has been found in a late-thirteentheo I am indebted to Mr. F. G. Dimes, of the Geological Survey and M.usewn, for examining the stone. He observes that 'in general appearance it me.tches most closely examples of a lithogmphio limestone of Jurassic age from Solnhofen, Munich, Bavaria', thou􀂩h this source has not been definitely established. 66 This supposition 1s based on the evidence of a siroilo.r, complete mould found at Calcott, Readi ng, Berks., and now in private hands. It should be noted, however, that stone moulds were sometimes of considerable length and were designed to cast several small objec􀂪 simultaneously; e.g. Nor/. Arch., ix (1884), 20, and a. mould for casting two rows or'tokens (Norwich Mus. 61.36). 102 ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER century context at Deddington, Oxon.; another comes from the Oxfordshire village-site of Seacourt, deserted in the fourteenth century. 67 Numerous examples were discovered in the mass graves (1361) at Wishy, Gotland, where, in many cases, they were found in pairs, lying in or close to, the pelvis.68 These, and similar discoveries in burials elsewhere, have led to the suggestion that this type of buckle was used for attaching the hose to a kind of suspender-belt known as the brekebelt or brygyrdyL69 Pictorial and written evidence support this belief and indicate that the suspension of the hose in this way was a fourteenth-century fashion lasting until c. 1365. These dates accord well with the date (c. 1300) assigned to the pottery that was found in association with the mould. The mould is thus more closely datable than other known moulds designed for casting small circular buckles.70 The matrix on the other side is for casting a pendant, possibly a harness-fitting. The loop of the pendant was cast round a slender, cylindrical core, again possibly of wood; the channel for this core is damaged but wa..:s apparently cut across the entire width of the mould. 71 The main feature of the pendant is a roundel (diameter l ·9 cm.) on which is delicately incised a grotesque, four-legged animal (fig. 18, no. 4b). The animal's head, legs and tail are arranged in the attitude of the lion passant regardant. But there the likeness to the heraldic lion ends. The waisted body, the scale-like treatment of the chest and hindquarters and the spiky, angular legs give the beast a curiously insectlike appearance. Nevertheless, the nearest analogy to the animal, as well as to the technique of the drawing, is a lamb depicted as the Agnus Dei on certain coin-like tokens of lead or pewter.72 ACKNOWLEDGEMENTS Permission for the excavations was kindly given by the landowners, Mr. John Leonard, the Mayor and Corporation of Rochester, and the Right Reverend R. W. Stannard, D.D., Dean of Rochester, whose G? Med. Arch., iii (1959), fig. 95, no. 2; Oxoniensia, iii (1938), 174. 68 Bengt Thordeman, Amwur from the Battle of Wisby, 1361, i (Stockholm, 1939), 117-18, figs. 117,118, 120 (nos. 1-22). 68 Ant. Journ., xxxvi (1956), 218-21. 70 e.g. from York, Arch. Journ., cxvi (1961), 100, :fig. 28, no. 10; from North Berwick, Proc. Soc. Ant. Scot., xli (1907), 431; from Dunwich (Ipswich Mus. 1920. 74.7). 71 Cf. the same tech nique in a mould for casting a Roman medallion, Kathleen M. Kenyon, ]J)xoavations at the Jewry Wall Site, Leicester (1948), fig. 83, no. 19. The hole to the left of the J?6ndant has been drilled obliquely through the Bide. This was a typical way of mserting and fixing metal dowels; cf. London Mue. mould A.11813. 12 C. Roach Smith, Catalogue ... of London Antiquities (1854), pl. xvi, 5 (found near Aldersgate Street); Oollectanea Antigua, i (1848), pl. xxxi, no. xi (from Bite of New London Bridge); these and other examples are in the Department of Coins and Medals, British Museum. 103 lIA ROMAN AND MEDIEVAL DEFENCES OF ROCHESTER interest and support during our prolonged invasion of his garden are most gratefully acknowledged. The excavation was carried out by the members of Sir Joseph Williamson's Mathematical School under the supervision of the writers, and the sheer volume of work achieved, often by quite young boys, seems worthy of comment. Our thanks are especially due to Mr. Brian Spencer, B.A., F.S.A., F.R.Hist.S., and to Dr. G. C. Dunning, B.Sc., D.Lit., F.S.A., for Appendices I and II; to Mr. R. G. Foord for undertaking the photography and supplying most of the prints; to Mr. I. J. Bissett for drawing most of the pottery, and to Mr. E. R. Swain for drawing the small finds. Above all we wish to record our appreciation of the invaluable help afforded by Mr. P. J. Tester, F.S.A., and Mr. A. P. Detsica.s, M.A., F.S.A., who not only prepared the figures illustrating the paper and reported on the pottery, but also have taken a sustained interest in the project from its inception. Finally, it would be ungracious not to mention with gratitude the work of our predecessors in the field, George Payne, Ca.non Livett, and St. John Hope. Little could have been achieved without their record of evidence now no longer available, and where we have differed from them it has been in the light of farther evidence not available to them. Perhaps we could not do better, in conclusion, than quote Canon Livett's own wise words: 'I do not imagine the subject of the city walls is exhausted; it may be that some of the conclusions herein set forth may have to be abandoned at a further date.' 104