Excavations in Rochester

EXCAVATIONS IN ROCHESTER By A. C. HARRISON, B.A., F.S.A. TH:m limited excavations here described were carried out between 1961 and 1969 by the Archreological Society of Sir Joseph Williamson's Mathematical School, Roches.ter. The fust waa conducted jointly with Colin Flight, B.A., who shares the responsibility for it only excepting any errors that there may be in its presentation. The help of all those who took part is gratefully acknowledged, and I am especially indebted to Mr. I. J. Bissett for drawing the coarse ware pottery, to Mr. T. Ithell, B.Eng., for help with surveying, to Mr. R. G. Foord for undertaking the photography and supplying the prints, and to Dr. G. C. Dunning, B.Sc., D.Lit., F.S.A., for his comments on the polychrome jug. Above all, my thanks are due to Messrs. A. P. Detsicas, M.A., F.S.A., andP. J. Tester, F.S.A., who have not only aasumed responsibility for the sections of this paper that bear their names but have given invaluable help and support at all times. 1. THE OLD CoitN ExoHA.NGE, Roo:aESTER During rebuilding operations in 1961 an excavation, limited both in time and scope, was carried out in a cellar at the rear of the Old Corn Exchange (TQ 74306865). A trench 15 ft. by 6 ft. was cut across the width of the cellar, parallel to the High Street and at a distance of 52 ft. from it, the cellar floor being 6 ft. 6 in. below street level. Inevitably, the structural remains were fragmentary but the Roman levels produced useful groups of stratified pottery and there was medieval materiaJ of considerable interest. The Stratification (Fig. 1). Natural soil, a yellow clay, was only reached at the western end of the section, the centre being obscured by masonry, Roman and medieval, and the eastern end by two deep pits, neither of which could be completely excavated. (a) Where undisturbed, the Roman levels consisted of a series of floors containing occupation materiaJ of the fust and second centuries. These were not associated with any structuraJ remains, except for two small stake-holes, though it is possible that Layer 11 may be contemporary with the short length of Roman wall as it contained lumps of the same hard brown mortar. This wall, of ragstone rubble and flints on a foundation of unmortared flints overlaid Pit 1, which is dated c. A.D. 150 and which provides a terminus post quern. 95 EXCAVATIONS IN ROCHESTER (b) Pit 2, which mutilated the Roman wall and was dug into the filling of Pit 1, was filled with occupation earth, greenish-black in colour, which produced pottery of the twelfth century. Its top had been truncated by the modern cellar and it was not possible to determine its depth. (c) The medieval remains comprised part of a room at the western end of our trench, against the north wall of which had been dug a rectangular pit, approximately 3 ft. by 2 ft. 6 in., and a corridor 4 ft. wide which descended 16 in. by three wide steps to a further room just outside the northern limit of the excavation. If there had been another room further east, its floor had been entirely destroyed by the modern cellar. Originally, the western room had been connected with the corridor by a doorway which was subsequently blocked. These medieval walls were built of flints set in a soft creamy white mortar containing crushed shell and the eastern wall of the stairway rested on top of the Roman wall, the west side of which had been cut back and refaced. The pit was lined with flint and chalk set in the same mortar and was thinly floored with mortar. It appeared to be an original feature of the room and is interpreted as a latrine. The filling, which was mostly greenish sand, contained part of a green-glazed jug (Fig. 6D), a fragment of a glass bowl or bottle, an iron adze-head (Fig. 6} and almost all the fragments of a jug of polychrome ware, which dates the :filling to c. 1300 A.D., when presumably the latrine went out of use. The building was therefore at least of thirteenth-centUI'Y date and, as its ea.st wall was sunk into the filling of Pit 2, cannot be earlier than c. HOO A.D. The base of a jug in reddish-buff ware with a, white slip externally found in the dark earth layer above the corridor may indicate that it was out of use by the fifteenth century. Fig. l shows these structures in plan. (d) The two upper floors belong to a brick-walled cellar apparently part of the Corn Exchange built in 1706. 2. TRE COM:Il10N, ROOHlllSTER. This excavation was carried out, in advance of redevelopment, on the line of the Roman wall approximately 48 yards from its presumed north-west corner and 137 yards from the site of Northgate A.1 Fig. 2 shows a. section across the wall at this point. The wall here survives to a. height of 9 ft., nearly 6 ft. of the internal facing being intact (Plate I, .A), but only three courses of squared ragstone remaining on the outside (Plate II, B). Its construction was exactly the same as described elsewhere, 2 with one noteworthy exception: there was no foundation- 1 A.'!'ch. Oant., lxxxiii (1968), 55, fig, 1. a Ibid., 75-7. 96 :::::::::e;c=;;::;;;L􀀇􀀈,L--., i k••++S +-L•♦+ Foundotlon of 1 I unmort1Jr,d Ft/n),, i I I I I I , _________ , I O 1,..-,---- __, I. Cellar Floor and Brick Rubble Build-up. 2, Earlier Floor, 3, Black Occupation Soil. 4, Light-brown Loa.m. 5, Light-brown Loam, 6. Brown Loam. '1. Green Clay. 8. Black and brown Earth and .Ash. : Strtps i s I Fut i.' I M,tttl FIG, 1. 9. Black Occupation Soil, 10, Burnt red Sand. 11. Green Sand with brown Mortar. 12, Floor--orange Loam. 13, Blaok-Earth and greenish Sand. 14. Floor-orange Loam and concreted Gravel. 16. Blo.ok Earth and greenish Sand. 16, Flool'-yellow Loam. (16•5 ft • .A.O.D. at Top of Section.) l i . 17. Black Eiμ-th streaked with orange Loam. i 18-20. Floor􀃚-brown Olay over greenish Se.ndJ 21, Floor m1).oh patched with orange and greylLoam. 22, Greeniah1 Sand. 28. Dirty grey Olay. I , I I I ) I I I I I I I 24. Yellow 0/ay-undiaturbed. (facep, 06 ROCHESTER I I 1. Cla.y with Flint and Tile. 2. Blaok Earth. 3. Clay and Tile. 4. Mortar Scatter. 5. Blaok Loose Earth. 6. Brown Earth. 7. Upper Chalk Floor. 8. Blaok Occiupa.tion Soil. 9. Brown Earth. 10. Reddish Clay. 11. Lower Chalk Floor. AIOdrm Brick W<1I The Co :rq.mo n 1967- 8 North Woll of Romon City z I ., I " I s I f FEET ,. I METRE$ FIG. 2. 12. Dirty grey Olay, 13. Dirty orange Briok-earth. 14, Grey Clay, 16. Rubbish Dump {twentieth Oen· tury), 16. Black Earth (nineteenth Century Pit). 17, Medieval Pit. 18, Olay, 19, Earth and Rubble. 20. Brown Earth. (14·8 ft. A.O.D. at Top of Section.) I 21. Ola􀃕. 22. Chl)lk. 23. Olay. 24. Briok-earth. 26. Bla-0k Occupation Soil. 27. B􀃖t Ola.y. 28. Gr4vel. 30. Black Occupation Soil. 31. Chalk. 32. Bu􀃗nt Clay and Ash. 33. Olay. EXCAVATIONS IN ROCHESTER trench filled with rammed flints below the 8-ft. concrete raft on which the wall stood, and this raft had been laid directly on the existing ground surface. Tke Stratification (Fig. 2). The subsoil in this area is a grey clay and was reached at a point about 11 ft. below the present surface. Above this was an occupation layer (30), which included a hearth, 4 ft. 6 in. in length and oval in outline, surrounded by a kerb of chalk blocks, averaging 6 in. in width. A shallow pit was in the centre of the hearth (32). This layer was sealed by a chalk and gravel floor (34) upon which lay another black occupation layer (26) sealed in its turn by a second floor of chalk and clay (22, 23). Above this was a layer of brown earth (20) and directly upon this rested both the wall and the wall-bank (18), which wa-s largely constructed of clay. All these layers produced pottery, the dating and significance of which are discussed below. Two la:rge medieval pits had been dug deep into the Roman levels, but all the later levels had been removed by a large nineteenth-century pit. In front of the wall the same occupation layers and chalk floors continued for an unlmown distance to the north. Above these were about 3 ft. 6 in. of medieval material, and the layer of clay directly under the modern path contained Tudor pottery. An interesting feature on the inside face of the wall was a series of semi-circular indentations impressed into the concrete of the raft upon which the wall was built (Plate II, A}. There were six of these in a length of 5 ft. and their diameters varied from 6 to 9 in. It seemed evident that these impressions were made by the ends of logs laid at right angles to the line of the wall, ,vhich had been laid in position at the time that the concrete was laid. Evidently, these had then been 1·emoved as no trace of any timber foundation remained under the wall-bank. With the object of discovering whether these indentations continued to the east, another cutting was made inside the wall, parallel to the first and 3 ft. from it. Here, however, the concrete raft was unmarked and, as this section showed substantially the same picture as the first one, it is not illustrated. Discussion. While there was nothing exceptional about the actual structure of the wall at this point, the fact that it stood on top of the existing ground-level without' even the concrete foundation being sunk into it seems noteworthy, and can perhaps be explained by the hypothesis that the ground here was too ,vet for trenches to be dug. The presence of occupation material lying under the wall and extending beyond it is also interesting, as it may indicate that some contraction of the size of the settlement took place when the wall was built, as is known to have been the case elsewhere, for example at Silchester. As no structural remains were found, it is probable that the chalk floors in fact represented a yard at the back of some building closer to the High Street. It 97 EXCAVATIONS IN ROCHESTER is not possible to give any certain explanation of the log-impressions left in the concrete-raft. They can hardly have been 'shuttering' for the concrete itself, as for this purpose they would surely have been laid longitudinally. Perhaps, in view of the suggestion made above that the ground here was swampy, the most likely explanation is that some sort of temporary 'corduroy' roadway was laid here parallel to the wall to facilitate the transport of stone, etc., in carts or sledges. The evidence of the pottery recovered from these sections tends to reinforce rathe􀄈 than modify the suggested dating of the Phase II defences of the city.2 Both the wall-bank itself and the surface upon which it had been piled contained pottery of the last quarter of the second century arid the former also produced a denarius of Marcus . Aurelius issued A.D. 161-165. It would seem, therefore, that the date of the building of the city wall must be later than A.D. 200 but how much later it is not at present possible to say. The considerable amount of second-century pottery contrasts with the virtual absence of any firstcentury material and suggests that this was the most prosperous period of Roman Rochester. 3. MARTIN'S BANK, 47 HIGH STREET, R0OHESTER A small-scale excavation was carried out in the garden behind the bank in advance of building operations by kind permission of the manager, Mr. N. G. Willis. Structural remains were fragmentary but suggested that a wattle-and-daub first-century hut had been replaced by a timber-framed building in the second century. Pottery and two brooches from the site are discussed below. 4. N.ASH])NDEN FARM (TQ 732659) This place was a considerable manor in earlier times and is mentioned in Domesday as belonging to Bishop Odo, half-brother of William I, when it appears as Essendenes (the first of a wide variety of spellings, viz. Eiscendene4 in Domesd,ay Monachorum, Hescindena in 1115,6 Nessyndone in 1253,6 Nessindon in 1260,7 Nas􀄉wn in 13208 and Nesslvyndoune in 1347).9 In the Textus Roffensis of c. 1115,5 it is listed 3 V.O.H., Kent, iii, 225. It has been suggested (J, K. Wallenberg, Place-names of Kent, Uppsala, 1954, 126) that the name may be compound of OE Hyscen = a small house and either Denu = valley or Denn = pasture. ' Op. cu., 267. Arch,, Cant., xiv (1882), 46. 6 Ibid., xii (1878), 224. 1 Ibid., iii (1860), 250, 8 Ibid., xii (1878), 304. 9 Ibid., X (1876), 159, 98 l'/1Q/o: R. G. Foord A. The ('0111111011: Inner Fn<'P of Homnn City Wnll. PlioU>: R. 0. Foord B. ThP Co1111non: Outer Foco of Roman Cit,v \Vall. PLATE IC J•h,,to: U. f:. f'o,r>rfl .\. Th,, Co1111non: lntl,•nlntion􀁘 in Conerete Huft of Homnn City \\'nil. Crmrn Copvrighl Rt􀀔rrtd B. Polychro,ne Jug. EXOAV ATIONS IN ROOHESTER as possessing a Chapel, subordinate to the Church of St. Margaret which also received the tithes of the manor. In 1403, it was given to the Rochester Bridgewa.rdens in whose possession it remains .. The supposed site of the Chapel is a building lying near the west boundary of the farm and marked upon George Russell's map of the manor dated 1706, and belonging to the Bridgewardens. It was excavated in 1876 by 'a gentleman staying at the farm' and a sketch-plan derived from this investigation appears in Coles-Finch's book, In Kentish, Pilgrim Lanil,.10 In 1969 the Archreological Society of Sir Joseph Williamson's Mathematical School re-excavated the site, by kind permission of the owners and of Mr. Auger, the tenant, whose co-operation is gratefully acknowledged. Tlie BuilJJ,ing. It was soon apparent that the Victorian excavation had destroyed all stratification and that approximately half of the building had been destroyed when the nearby oast-house was built. However, it was possible to reconstruct the ground-plan with some certainty and to trace four periods of construction (see Fig. 3). The first building seems to have been a rectangle, 28 ft. 6 in. by, perhaps, 18 ft. internally, well-built in flint and ragstone with a bottom course of large ragstone blocks. The walls were 2 ft. 4 in. thick, rendered with mortar externally and resting upon a foundation-trench, 1 ft. deep and filled with flints. The mortar was light brown in colour and contained a large number of small shells. To this building was added an extension, 16 ft. long internally. Its walls were of similar construction, with identical mortar, but were only 2 ft. 3 in. thick and had no flint foundation. A vertical joint was found in the south-west wall at this point of junction. A doorway, 2 ft. 8 in. wide, in the south-east wall of the original building provided access to the extension. A dividing wall, 1 ft. 10 in. wide, was subsequently inserted into this extension. If, as seems likely, this was placed centrally, then the internal ,vidth of the whole building will have been 18 ft., as suggested above. This wall was of flint and ragstone set in lime-mortar which was pink in colour. Finally, the south-east wall of the second-period building was reconstructed in flint with some re-used tufa set in a white mortar of lime and crushed flint. Discussion . .Although this building is traditionally the Chapel, there is nothing ecclesiastical about its plan or its orientation and, in its final form of one large and two small rooms, it must have been in domestic use. It seems probable that this was the case from the first and that what we have here is the ground-plan of the basement of a first-floor hall, first built as a single cell and subsequently enlarged. (From the similarity of the building technique, it is not thought that there was any long interval between the first two phases of construction.) A close 10 William Coles Finch, London, 1926, 240-60, 99 .... 8 r·-·-·-·-·-7 i i i i i. ! I I i ! i i i i i i i i 1 L._ ,go i ! NASHENDEN FIG. 3. r·-·-----·-•7 i i r·-·-·-·-·--·-·-·-·-, I i i F'HI /0 I EXCAVATIONS IN ROCHESTER parallel is afforded by the plan of the first-floor hall at Boothby Pagnallll (see Fig. 3). This is dated to c. A.D. 1200, and the style of construction and the shelly mortar12 at Nashenden would be consistent with a similar date. On the whole, therefore, it would seem likely that this building formed the late twelfth- or thirteenth-century manor house and that the site of the Chapel, if a separate building, must be sought elsewhere. THE FINDS ROMAN POTTERY (The Common) By A. P. DETSIOAS, M.A., F.S.A. (i) COARSE WARE The largest proportion of the coarse wares submitted for study consisted of the usual types and fabrics of bowls, dishes, cooking-pots, flagons, storage vessels, etc., characteristic of the second century A.D., with a clear predominance of vessels usually current during the second half of the century and later; there were also present several sherds of earlier material but not in significant numbers and in association with much later pottery, which must be considered as rubbish survivals. It has proved impossible to separate chronologically any of these layers on the basis of the coarse pottery as virtually all layers contain strictly contemporary sherds; this was further borne out by the evidence of the samian ware. Layer 18 contained several sherds from cooking-pots with cavetto rims, bead-rim bowls, storage vessels in Patch Grove and shell-filled wares, dishes in black-burnished fabric with straight sides and rolled rims as well as colour-coated wares in rough-cast fabric, applied decoration en barbotine and hunt cups; it cannot be dated earlier than A.D. 150, in spite of the rubbish survivals of earlier date it contained. Layer 20 contained the same mixture of second-century coarse wares, with bowls and cooking-pots having cavetto and bead rims as well as colour-coated sherds in rough-cast fabric; it must date later than A.D. 150. Layer 26 again must be dated to the second half of the second century A.D. owing to the presence in its coarse wares of colour-coated sherds. Layer 30 contained sherds of 'Poppy-head' fabric, cavetto rims and u Margaret Wood, The English Medieval House, London, 1966, 19, fig. 6. 12 Of. Arch. Oant., xxi (1895), 26. 101 llA. EXCAVATIONS IN ROCHESTER several colour-coated pieces in rough-cast fabric. Date: .A.D. 150-200. Layer 33 was the only deposit where the absence of colour-coated sherds might have indicated a dating earlier than .A.D. 150; on the other hand, the presence in this assemblage of sherds from at least one straight-sided dish in black-burnished fabric urges caution. (ii) S.AMIAN WARE Abbreviations and References OGP J. A. Stanfield and Grace Simpson, Central Gaulish Pott,ers, London, 1958. D. J. Dechelette, Les Vases dramiques ornes de la Gaule romaine, ii, Paris, 1904. Knorr 1919 R. Knorr, Tiipfer und Fabriken verzierter TerraBigill,at,a, des ersten Jalwhunderts, Stuttgart, 1919. Knorr 1952 R. Knorr, Te1·ra-Bigill,at,a,-Gefasse des ersten Jahrhunderts mit Tiipfernamen, Stuttgart, 1952. ·LauriMwm P. Karnitsch, Die verzierte Sigillata von Lauriacum (Lorch-Enns), Linz, 1955. Lu. V. W. Ludowici, Stempel, Namen und Bilder riimisclier Tiipfer aus meinen AU8grabungen in Rheinzahern 1901-14, Kata.log V, Jockgrim, 1927. 0. F. Oswald, Index of Figure-Types on Terra Sigillat,a, ('Samian Ware'), i-iv, Liverpool, 1936-7. 0 & P F. Oswald and T. Davies Pryce, An I ntrowucti,on to the Stuwy of Terra Sigillat,a,, London, 1920. Ricken-Ludowici H. Ricken-W. Ludowici, Die Bilderschusseln der riJmischen Topfer von Rheinzabern, Katalog VI, Speyer, 1948. Ricliborough II J.P. Bushe-Fox, Excavation of the Roman Fort at Ricliborough, Kent, II, Oxford, 1928. Wela P. Karnitsch, Die Reliefsigillat,a, von Ovil,ava (Wela, Oberiisterreich), Linz, 1959. CG = Central Gaulish EG = East Gaulish SG = South Gaulish A considerable amount of samian sherds was submitted for examination and reporting and is dealt with below in two sections of plain forms and decorated bowls, respectively. (a) PT,ain Forms. Layers 9-10 contained sherds of Forms 27 (CG), 18 (both SG and CG), 18 or 18/31 (CG); 36 (CG) and fragments of a shallow platter, also occurring in Layer 26, in EG fabric (of. 0 &, P, pl. lxvi, 3). 102 EXOAVATIONS IN ROCHESTER Layer 18 contained sherds of Forms 27 (1 SG vessel and 3 CG), 33 (1 CG and 2 EG vessels), 18 (SG), 18R (SG, with a partial graffito ]MO, 18/31 (CG), 31 (Tq)(EG), 31R (OG), 45 (OG), 38 (CG, 2 vessels) and 36 (CG, 3 vessels); also, decorated nos. 6, 8, 13 and 22 (Figs. 4-5, below). Layer 20 had sherds of Forms 27 (CG, 3 cups), 33 (1 CG and 3 EG cups), 18 (SG), 18/31 (CG, stamped PATERATIOF; see Fig. 4, below), 31 (EG), 31R (CG); also, decorated nos. 4-5, 7, 9, 14, 16 and 17-19 (Figs. 4-5, below). Layer 26 contained sherds of Forms 18 (SG and CG), 27 (CG), 31 (CG) and fragments of a shallow platter as under Layers 9-10, above. Layer 30 had sherds of Forms 27 (SG and CG) and 18 (SG, stamped [C]ALVI; see Fig. 4, below, and CG); also, decorated nos. 10-11 (Fig. 4, below). Layer 32 contained one sherd only of Form 36 (CG), and Layer 33 a sherd of Form 18/31 (SG, with the partial impression of the stamp of VITALIS SERVVS; see Fig. 4, below). (b) Decorated Forms (Figs. 4-5) 4-5. Form 37. CG, good condition and glaze. Seven sherds, six conjoining, from a bowl in the style of the QVINTILIANVS group of potters, with a decoration including their ovolo no. 4 (OGP, fig. 17, p. 145) over a wavy-line border and such borders divide the rest of the decorative scheme into panels. The surviving panels contain beaded rings (detail 5), astragali (detail 13), also used as terminals to the borders, acanthus leaves (data.ii 17; cf. PATERCLVS, OGP, pl. 72/34), and two figure-types: the hindquarters of a feline which cannot be reconstructed with certainty and Woman (D.193=O.339), 􀃵 on a sherd from Leicester (OGP, pl. 71/26). Date: c. A.D. 125-150. (Layer 20.) 6. Form 37. CG, good glaze and condition. A small sherd in the style of CETTVS, with his ovolo no. 2 and detail no. 4 (OGP, fig. 42, p. 245). Date: o • .A.D. 160-195. (Layer 18.) 7. Form 37. CG, good glaze and condition. Two conjoining sherds (a third one with a remnant of the ovolo only is not illustrated) in the style of the PAVLLVS group of potters (PSAS, xoiv (1960-61), fig. 6, no. 8), with CINNAMVS's ovolo no. 3 (OGP, fig. 47, p. 267) and a free-style decoration of animals and decorative details simulating grass. The figure-types, which have not impressed well owing to the shallowness of the bowl, are: Bear to left (D.820 = 0.1627), as on a signed sherd from York (OGP,pl. 165/3) and Deer to left (D.867 = 0.1772), as on the Mumrills sherd referred to above. Date: o. A.D. 150-190. (Layer 20.) 8. Form 37. CG, good glaze and detail. A small scrap from the ovolo band which could belong either to DRVSVS I (of. A. P. Detsicas, The Anonymoua Central Gaulish Potter known as X-3 and his Connections, Collection Latomus, vol. lxiv, Bruxelles, 1963, and Rei Oretariae Romanae Fautorum, Acta VII (1965), 67-9) and his contemporary 103 1111 14􀀃 FIG. 4. Sa .a n Ware (½ ) (Stamps 1 : l) 104 90I '(f) 8.1"8.M U11ftl1fl$ '!/ 'DJ.i{ 'Hfil:J.S[[HOQ'H NI SNOIJ.VAVOXfil: EXCAVATIONS IN ROCHESTER Potter X-2 (OGP, fig. 3, p. 7) or to the QVINTILIANVS group of potters (GGP, fig. 17, p. 145). Date: c. A.D. 100-150. (Layer 18.) 9. Form 37. CG, fair glaze, blurred relief. From a bowl in the style of CINN.AMVS, with his ovolo no. 3, large leaf (detail 38) and smaller leaf at the end of a tendril (OGP, fig. 47, p. 267), in a style similar to a signed bowl from London (OGP, pl. 162/61). Date: c. A.D. 145-190. (Layer 20.) 10-11. Form 30. SG, fair glaze, squashed relief. A total of six pieces, four of which conjoin in pairs; a seventh shard occurred in Layer 30. The ovolo, with a tongue projecting below it and ending in a large star, is one used by ORESTIO (Knorr 1952, Taf. 17B), CADMVS (Knorr 1952, Taf. 14M) and CALVS (Knorr 1952, Taf. 70F); the pointed leaf was used by MCRESTIO (Knorr 1919, Textbild 9 and Taf. 28/14); the large rosette within the double half-medallion was used by OADMVS (Knorr 1952, Taf. 13D-E) and MASOLVS (Knorr, 1919, Textbild 11 and Taf. 52/17); the terminal of the wavy-line border was used by MCRESTIO (Knorr 1919, Textbild 10 and Taf. 28/15). The one surviving figure-type is not clear enough but looks like Goose to left, looking back (0.2294). Though definite attribution to the style of any one of these South Gaulish potters is unwise, it seems likely that this vessel dates from the last quarter of the first century A.D. rather than earlier. (Layer 33.) 12. Form 37. CG, good glaze and relief. A scrap with AVSTRVS's ovolo no. 2 (OGP, fig. 25, p. 180). Date: c . .A..D. 125-150. (Layer 20.) 13. Form 37. CG, good glaze and relief. A small sherd preserving a remnant of the ovolo and a vine-leaf on a tendril, used by many Central Gaulish potters, though attribution to the style of any one of them is not possible on such scanty evidence. Date: second half of the second century A.D. (Layer 18.) 14. Form 37. CG, worn glaze and relief. A scrap from an indeterminable ovolo band. Date: A.D. 150+. (Layer 20.) 15. Form 37. OG, good glaze and condition. The ovolo is very similar to IVSTVS's no. 1 (OGP, fig. 31, p. 201), though this potter is not recorded as using beaded borders; the figure-type is recorded by Oswald (p. 93) on a Lezoux sherd in the style of ADVOCISVS, but the general appearance of the decoration on this sherd does not appear to indicate the style of either of these two potters. Date: A,D. 150+. (Unstratified.) 16. Form 37. CG, good condition, squashed relief. Two fragments conjoining from a shallow bowl in the style of CETTVS; his ovolo no. 2 (OGP, fig. 42, p. 245), without an enclosing bead-row border. The decoration, in panels divided by fairly square bead-rows, consists of double-ringed medallions, narrow spaces filled with double rings containing astragali and one :figure-type, Man (D.382 = 0.660), as on a 106 EXCAVATIONS IN ROCHESTER sherd in the style of CETTVS from London (OGP, p. 142/17). Date: o. A.D. 160-195. (Layer 20.) 17-19. Form 37. CG, good glaze but extremely poor relief. Eight fragments in all, some of which conjoin, from a fairly large bowl in the style of CETTVS, with his ovolo no. 3 (OGP, fig. 42, p. 245) over a panel decoration of animals and human figures. The figure-types are as follows: (no. 17) Lion to left (D.766 = 0.1450), as on sherds from Corbridge and Leicester (OGP, pls. 143/42 and 142/33), Boar to right (D.826 = 0.1641), as on the above Leicester sherd, Man (D.382 = 0.660), as on a sherd from London (OGP, pl. 142/17) and no. 16 above, and an unidentifiable feline to right; (no. 18) Lion to right (0.1404), as on several sherds in CETTVS's style (e.g. OGP, pl. 141/10, from Silchester), and another badly squashed feline to right; (no. 19) Lion to left (D.769 = 0.1457), as on sherds from Corbridge and London (OG P, pls. 144/50 and 144/49), other figure-types present on nos. 17 and 18, and Man (0.637), as on sherds from Corbridge and Leicester (OGP, pls. 143/37 and 142/33). Date: c. A.D. 160-195. (Layer 20.) 20. Form 37. EG, good glaze and relief. In the style of either COMITIALIS V or LATINNVS, with an ovolo illustrated in RiokenLudowioi (Taf. 262/17). The decoration consists of leaves in the field (Lu. V. P.49) and two figure-types, Panther to right (0.1518 = Lu. V. T41), as on a sherd from Lorch-Enns (Lauriaoum, Taf. 31/10), and Lioness to right {0.1505 = Lu. V. T55, 57), as on another Lorch-Enns sherd (Lauriaoum, Taf. 32/7). Date: c. A.D. 140-170. (Martin's Bank.) 21. Form 37. CG, good glaze and relief. From a bowl in the style of SACER, with his ovolo no. 1 and detail no. 6 (OGP, fig. 22, p. 163). Date: c. A.D. 125-150. (Martin's Bank.) 22. Form 37. SG, good glaze and relief. Three sherds, of which only one is illustrated, from a bowl in the style of BIRAGILVS, with his ovolo and characteristic trifid tongue bent to right (Knorr 1919, Textbild 5), his serrated leaf (Knorr 1919, Taf. 16/20), used also by several other contemporary South Gaulish potters, and remains of a winding-scroll sub-divided into two compartments by a wavy-line border ending in a neat rosette; similar wavy lines, placed diagonally, occupy part of the lower compartment with the remnant of an arrowhead to their left. An animal figure-type, which cannot be restored with certainty, filled the upper compartment. Date: c. A.D. 75-95. {Layer 18.) 23. Form 30. CG, good glaze and relief. A very small sherd with the remnant of a neat double-bordered ovolo, not unlike ovolos used by DOECCVS and MERCATOR II; attribution is impossible. Date: o. A.D. 150+. {Martin's Bank.) 24. Form 37. SG, good glaze and condition. A scrap with the remnant of a figure-type, Dog to lefi (0.2004), over leaf-tips, in the 107 EXCAVATIONS IN ROCHESTER style of MASCVVS (Wels, Taf. 20/8). Date: o. A.D. 80-100. (Martin's Bank.) 25. Form 37. SG, good glaze and relief. From the lowest part of the bowl, with a basal wreath similar to one in use by NATALIS (Wels, Taf. 26/4) and obtaining at Riohborough, stamped MERCATO (retro.) (Richborough II, p. xxvii, no. 11). The remnant of the decoration, which is rather blU1Ted, above the basal wreath consists of a Deer to left (D.862 = 0.1746), facing a tree-a motif used by several South Gaulish potters. Date: o. A.D. 80-100. (Martin's Bank.) MEDmVAL POTTERY (Fig. 6) (Old Corn Exchange) By P. J. TESTER, F.S.A. A and B. Cooking pots with simple straight everted rims, a form which can be compared with examples from Canterbury found in a probable twelfth-century context.13 The outer face of A varies from dark grey to reddish brown, while internally and in the fracture the ware is lighter grey.B is similar but is reddish internally. Clay liberally mixed with shell in both cases, whereas the Canterbury examples cited were in very coarse granulated ware. Shell :filling persisted in Kent throughout the thirteenth century and in the present absence of a well-dated series of medieval rim-forms from the Rochester area the dating of these pots can be only tentative. Clearly they are earlier than the common flat-topped flange type which hitherto has been assigned to the late thirteenth century on the Canterbury evidence, 14 but which Messrs. Rigold and Mynard have now reported in a context at Dover appMently as early as 1248.15 A late-twelfth or early-thirteenth-century date for these two Rochester pots is therefore suggested, though admittedly this may call for revision in the light of future evidence. Pit 2. C. Unglazed jug, dark grey on outer face, reddish inside. Handle has deep cuts or slashes. Thirteenth century. From cellar steps. D. Jug of hard grey wMe, with uneven green glaze. Handle has a few widely-spaced stabbings. Thirteenth century. Privy pit. 1s Ibid., Ixviii (1954), 132, fig. 17. 1' Ibid., 132-4. 15 Ibid., lxxxiii (1968), 20-1. 108 EXCAVATIONS IN ROCHESTER •· ... A FIG. 6. Medieval Pottery and Adze (¼), 109 EXCAVATIONS IN ROCHESTER PoLYOHROME JuG (Fig. 7 and Plate II, B) By G. C. DUNNING, B.Sc., D.L1T., F.S.A. Small jug of polychrome ware, 7} in. high. The ware is pink in the core, whitish on the surfaces. Thin colourless glaze covers the entire outside and extends partly down the neck inside. The jug has a cylindrical neck and an ovoid body, and its structural features, such as the moulded rim, the large bridge-spout and the straphandle, are normal on this class of ware. The part of the rim behind the bridge-spout was, however, cut away before firing, and this feature has been noticed on a number of polychrome jugs found in both England and France. r Fxo. 7. Polychrome Jug(¼), The painted decoration is repeated on both sides of the jug. It comprises two separate parts; a yellow four-pointed leaf and a green stem on the body, and a yellow bud placed horizontally on the neck, all outlined by dark brown lines. The two parts are really the components of a scroll pattern with leaves and an upright bud, such as appear on jugs from London, Carisbrooke Castle and Kidwelly Castle.16 On the Rochester jug the design has undergone stylistic disintegration, whereby the scroll is reduced to a loop enclosing a single leaf and the detached bud is placed above it on the neck. In this position on its own the bud has lost its original meaning, and in fact looks rather like a fish. On another polychrome jug, found in London,17 this process has gone even further; the stem and leaf are enclosed in a triangular border resembling a shield, and the bud has vanished altogether. 1a Archr.eologia, lxxxiii (1933), 129, fig. 14, a-b, and 11, fig. 6. 11 Ibid., 128, fig. 13, a. 110 EXCAVATIONS IN ROCHESTER It should be noticed that although the designs on the Rochester and London jugs may be regarded aa degeneraoo, the technique of painting and the details of the various parts maintain the delicacy and sureness of the complete scroll patterns on the other jugs. There is therefore no need to suppose that these jugs are any later in date than the rest of the series. The Rochester jug is thus an interesting addition to the polychrome wares made in SaintongeJ.8 and imported into Britain; it may be dated to c. 1300. BROOOHES By A. P. DE'l'SIOAS, M.A., F.S.A. Two bronze brooches were found and are described below: 1. This brooch measures 2¼ in. (55 mm.) and has a long, curved bow, with a D section, an iron spring and a perforated catch-plate which is partly broken though it preserves two key-shaped perforations. It belongs to the well-known La Tene III brooch ofCollingwood's Group F, usually referred to as the 'Colchester brooch' from its prevalence at that site; it is quite common in Kent and several such brooches have been found in recent years at the Eccles Roman villa site. This brooch is closely similar to Colchester Type III (C. F. C. Hawkes and M. R. Hull, Oamu!,orl!un1im, Oxford, 1947, pl. lxxxix, nos. 12-13), dating to A.D. 49-61; to one found at Richborough (J.P. Bushe-Fox, Excavation of the Roman Fort at Richborough, Kent, i, Oxford, 1926, pl. xii, no. I), though this brooch has rounded perforations; and is paralleled by others found at Leicester (K. M. Kenyon, Excavations at the Jewry Wall Site, Lei,cester, Oxford, 1948, fig. 80, no. 4) and Allington (J.P. Bushe-Fo:x:, Excavation of the Late-OeUic Urn-field at Swarli1'1{1, Kent, Oxford, 1925, pl. xv, no. 14). All these brooches are datable to about A.D. 60, though the type seems to have survived in use 'into the early second century' (R. G. Collingwood and Ian Richmond, The Archreology of Roman Britain, London, 1969, 292), which may be closer to the context of the Rochester brooch. 2. The second brooch belongs to Collingwood's Group W and is very close to the brooches in Colchester Type X; its catch-plate is perforated like the brooch illustrated in Oamulorl!unum, pl. xciii, no. 73, but its bow, which is attached to the plate, is much nearer to no. 76. Dating range in Colchester is A.D. 43/44 to c. A.D. 65. 18 K. J. Barton, 'The Medieval Pottery of the Saintonge', Arcli. Joum., oxx (1963), 201-14. Ill EXCAVATIONS IN ROCHESTER IRON ADZE (Fig. 6) This adze was found in close association with the polychrome jug described above and is therefore closely datable to the last quarter of the thirteenth century. No exact parallel has been found, but it is similar, though broader, to the one from Hurbuck, Co. Durham, illustrated by D. M. Wilson, in 'Anglo-Saxon Carpenters' Tools', StU