Canterbury Excavations, Summer, 1946

CANTERBURY EXCAVATIONS, SUMMER, 1946 THE ROSE LANE SITES By SHEPPARD FRERE, F.S.A. IN the Summer of 1946 the Canterbury Excavation Committee, which had been examining the basements in Butchery Lane and on the north side of The Parade (Arch. Cant., LXI, pp. 1-45) extended its operations to the opposite border of the latter street on each side of the mouth of Rose Lane. Here a frontage of about 115 feet had been bombed between the premises of Messrs. J. Lyons and Messrs. Marks and Spencer. On the north-west side of Rose Lane ceUars H and I occupied the site of the Rose Hotel: across the lane were ceUars L, K and J, whUe a small area at surface level, undisturbed by cellars, occupied the street corner (K 2, Fig. 1). The first purpose of the excavations was to discover whether the Butchery Lane Roman buUding extended as far as the south side of the Parade ; but there was also the general policy of testing all available areas. In the event it was found that the cellars had usuaUy been dug too deep, and little remained below their floors except in the shallower rear part of cellar H. It was clear, however, that the Butchery Lane buUding did not extend as far as trenches H 1 or K 2. The principal discoveries were : (1) a pre-Roman Belgic ditch running N.E.-S.W. in cellar L with a contemporary occupation to the east; (2) a Roman drain running obliquely under Rose Lane. This was first found in cellar H and was picked up east of Rose Lane in a trench cut from the surface (K 2). Here the structure was better preserved (Fig. 10, PI. II), and the seating for its vault could be seen. In both sections the drain overlay an early ditch which could not be completely explored. It was probably Belgic in origin, of larger size than the ditch found in ceUar L ; its filling contained Belgic pottery and a few Claudian sherds, implying that it had been closed early in the Claudian period, doubtless as part of the tidying up of the site after the Conquest. The drain itself appears to date from the mid-second century. In Section K 2 the latest material beneath it was dated c. A.D. 100-12Cl#T|J$)&lskyer 5 above this contained no finds. In H 1 the latesfypottery beloW the ioi f l]ma CANTERBURY EXCAVATIONS, SUMMER, 1946 drain ran down rather later than this, and a date c. A.D. 150/60 seems hkely for its construction ; (3) a number of Roman and medieval pits, a group in cellar I being of especial interest; CANTERBURY MARKS S SPENCER MEDV WELL ©• 946 SRI! BELCIC OCCUPATION AREA ZJ I I / Mt -i 3: r~ STREET WIDENED rn K. 1952 TD V \ > ©*•* \ Xl i s t n %\ J Dm 11 M8 i I • MI2 =*, I Ml \ H2 lM9 HI \ -•» J % Mr H » MEDIAEVAL FOOTINGS L y SCALE : FEET LYONS paxjpGHBKi 0 20 30 40 £0 60 10 ao t METRE FIG. 1. General Site Plan (4) a large medieval foundation under the rear west wall of cellar H. This seems to have been part of a thirteenth-century cellar. CELLAR L The front part of cellar L was comparatively free of later disturbances, but had been excavated almost to the level of the natural yellow loam or brickearth. 102 CANTERBURY EXCAVATIONS, SUMMER, 1946 Immediately north-west of Marks and Spencer's appeared a Belgic occupation area, bounded by a smaU drainage ditch. The occupation had disturbed the top of the natural. Near the edge of the ditch natural soil was seen at 35 ft. O.D. capped by still clean yellow loam. But further east the mixed dirty yellow loam of the occupation went down 6 in. deeper than this, and below it was at least another 6 in. of clean yellow loam with flecks of charcoal. This may be partly accounted for as a drop of natural to the east, and partly doubtless as the result of farm-yard conditions. We ourselves found it impossible to walk in this loam after heavy rain. The occupation soil contained a large number of decayed bone fragments (horse, dog, cow, pig, sheep1) and small pieces of pottery. There was a ragged line of post-holes, surviving only 4-6 in. deep, two of which seemed to be in larger holes packed with clay. There was also a row of smaller stake-holes about l£-2 in. in diameter, though it was not always easy to distinguish these from the marks left by the decay of bones. (PI. I, 1.) THE BELGIC DITCH (PI. II, 2 and Fig. 2) Bounding this occupation to the north-west was a small U-shaped gully cut about 24 in. into the natural soil and with a ledge along its east edge. Sixteen feet of it were traced in the front part of the cellar, but towards the south it was lost in some deep medieval disturbances. This gully is of a type commonly met with on Belgic sites : its necessity for draining the sticky yellow brickearth was abundantly demonstrated by heavy rain during the excavation. It contained occupation material of great archseological interest. The filling (Fig. 2) consisted of (1) primary sUt of light greyish yellow loam containing some large bone fragments, including the jaws of a horse and a large portion of sheep's skeleton.2 (2) dirty yellow loam with much small charcoal and many pieces of pottery and bone, evidently flung in from the occupation to the east.3 (3) over this in places at the south end was a dark shelly Roman occupation layer about 3 in. thick (not seen in Fig. 2) dated c. A.D. 90-120. (4) Light yellowish grey loam, sealing 2 and 3, containing material to mid II A.D. 1 The majority of bones were of ox, a small type and often adult or old. 2 Also part of an ox jaw, and leg bones of a small horse. 3 The following charcoal has kindly been identified by Mrs. F. L. Balfour- Browne : Layer 1, Ash and Oak ; Layer 2, Ash, Oak and Willow or Poplar. 103 CANTERBURY EXCAVATIONS, SUMMER, 1946 THE BELGIC POTTERY All the large quantity of broken pottery found in layers 1 and 2 of the ditch and in the occupation layer beside it was of native Belgic type except for three sherds showing possible Roman influence, together with a handful of imported amphora fragments, one sherd of rouletted white butt-beaker, and three sherds of terra nigra. The amphorse probably came from Italy, S. Gaul or Spain : the terra nigra from Galha Belgica. Such imported fragments are common in pre-conquest Belgic finds. No Samian ware at all was found, and T V NORTH-WEST CELLAR. FLOOR. SOUTH - E A 5T MEDIEVAL CHALK o o op*" FOOTING" DISTURBANCE SCALE-FEET SECTION taaoA&a BELGIC DITCH FIG. 2 in a coUection of pottery as large as this such absence is significant. The excavations at Colchester (Camulodunum, p. 31) have proved that the importation of Samian ware did not begin before the Conquest : taken with the typologically early character of the vessels, the total absence of Samian from so large a group, though Samian is abundant in Claudian levels at Canterbury, gives good grounds for attributing this settlement, and thus the beginnings of Canterbury itself, to a pre- Claudian date. The inclusion of two or perhaps three post-conquest sherds of coarse ware implies that the ditch was filled in and the site tidied up as one of the first acts of the new authorities. Native Belgic pottery (especially combed ware) continued to be made in Canterbury certainly untU late Flavian times (Arch. Cant., LXIII, p. 101, Fig. 10, 23-25), but this later material can often be 104 CANTERBURY EXCAVATIONS, SUMMER, 1946 distinguished from the pre-Roman by a better firing and a redder colour: the present group is grey in colour and shghtly cruder in its manufacture. The finer bowls and jars, however, quickly assumed a Roman technique after the Conquest. The pottery is discussed in detaU below ; here it may be noted that the Belgic quoit-shaped pedestal bases are rare in Canterbury, and only •-. \ ms V ) m 4 S~\ / 5 X \ W \ 7 V8 t FIG. 3. Belgic pottery, from primary silt of ditch, Cellar L. (\) occur in the earliest levels : they seem always to be of pre-Roman manufacture, though occasionally surviving later as rubbish. Layer 1, primary silt of ditch (Fig. 3) 1. Bead-rim, very heavUy brush-striated grey porridgy ware. 2. Simple beaded rim, grey paste with brownish-red surfaces ; arched brush-striations. 3. Dark grey ware ; brush-striations. 105 CANTERBURY EXCAVATIONS, SUMMER, 1946 4. Smoothed bowl of grey ware ; weU-made angular cordon at base of upright neck. 5. Burnished grey-buff surface ; grey paste. 6. Spindle-whorl from cordoned shoulder of burnished grey ware. 7. Corrugated shoulder above brush-striated (?) decorations : heavy buff gritty ware. 8. Corrugated neck of coarse grey porridgy ware, surface smoothed. 9. Striated simple beaded jar, very sooty outside. In addition, this layer contained the following unfigured pieces : (a) pedestal base, as 30 ; (b) rippled shoulder as 19 ; (c) lower part of bowl with sharply curved shoulder (cf. Swarling 22, 24) lightly treUised : dark porridgy ware with black glossy surface ; (d) base of thick clumsUy made dish like 40 : also one in thinner better paste ; (e) two fragments of reddish gritty amphora and one in hard cream coloured ware with whitish surface ; ( / ) several fragments of burnt daub. Layer 2, secondary filling of ditch (Figs. 4, 5, 6) In addition to the bronze fibula (p. 140), this layer contained : (a) two pieces of terra nigra platter of the best period, in whitishgrey ware with silvery surface ; (b) three pieces of spindle-shaped amphora (Camulodunum 181-4), two in gritty red ware, one in hard whitish ware ; (c) one small fragment of native jug copying an imported Gallo-Belgic model in fine reddish sandy paste with external cream slip. The sherd is very thin (only •£ in.). Camulodunum 165 ; (d) one chip of rouletted pipe-clay butt-beaker. 10. Large jar with corrugated neck and shoulder, decorated with brush-striations surmounted by a row of jabs, probably done with a twig-end: grey porridgy paste, hght buff-grey surface, smoothed on neck and rim. 11. Similar vessel in grey ware with black smoothed outside. 12. As 10, 11 : patchy buff red to black surface : corrugations smoothed. Fragments of at least three others were present; cf. also Butchery Lane, Arch. Cant., LXI, p. 34, No. 58. 13. Large thick jar, dark grey ware ; rounded Up and corrugations smoothed, below which a zone of jabbing apparently with a 3-pronged fork done on a slow wheel. 106 CANTERBURY EXCAVATIONS, SUMMER, 1946 14. Smoothed biconical jar in grey ware with leather-coloured outside surface : neck shghtly rippled with poor low cordons, shoulder grooved. FIG. 4. Belgic pottery from ditch, layer 2, in Cellar L (£) 15. Spindle-whorl from a shoulder sherd of striated grey ware with reddish-brown surface. 16. Wide-mouthed heavy jar in porridgy grey ware with smoothed black surface. 107 CANTERBURY EXCAVATIONS, SUMMER, 1946 T \ k FIG. 5. Belgic pottery from ditch, layer 2, in Cellar L. (£) 108 CANTERBURY EXCAVATIONS, SUMMER, 1946 ) bifesii-satj: / 53 / 55 57 FIG. 6. Belgic pottery from ditch layer 2 (41-49), and occupation layer (50-57), in Cellar L. (£) 109 CANTERBURY EXCAVATIONS, SUMMER, 1946 17. Corrugated bowl in hard sandy grey ware : below the shoulder a matt area with tooled pattern. 18. Very large heavy storage jar : coarse porridgy ware, black surface. Fragments of three others present. 19. Bowl with weakly corrugated shoulder : hard porridgy grey ware, reddish-brown surface. 20-22. Bead-rim bowls in coarse dark grey or black ware, striated surfaces. 23. SimUar bowl in light grey shghtly porridgy ware, rim flattened to a bevel. 24. Coarse grey ware, buff-brown surface, cf. 2. 25. Coarse dark grey to black ware. 26. Funnel-shaped beaker in smoothed somewhat sandy dark grey ware. This is the best local Belgic ware. 27. Biconical jar with rippled shoulder, smoothed dark grey ware. 28. SimUar jar in leathery brown ware with grey surface. 29. Moulded pedestal base in simUar ware to 26, possibly belonging to it. 30. Broken fragment of pedestal base with rough cordon : ware as 26. 31. Quoit-shaped pedestal, porridgy red ware with leathercoloured surface. 32. Bowl in porridgy grey ware with smoothed grey-buff surface. Recurved rims of this type are rare. Cf. 45. 33. Grooved rim in rough black ware, surface pohshed, furrowed shoulder. 34. Highly pohshed dark grey to black cup or bowl in fine sandy ware. 35. Moulded base in Roman-type fine hard light grey ware with darker surface, probably from a carmated beaker Camulodunum 120 A. A second simUar base is present in shghtly more native-type ware: these together with a third smaU sherd of grey ware are the only suggestions of local Roman influence. 36. Domed base, grey ware. 38. Carmated bowl of smoothed shghtly sandy grey ware, perhaps copying a metal cup. 39. Cordoned jar. Cf. 16. 40. Smoothed shghtly sandy grey ware : a somewhat crude native copy of a terra nigra dish, Camulodunum 22. 41-43. Grey dishes simUar to 38. 44. Grey ware with dark burnished surface and groove at base of neck. 45. Recurved rim, hght grey ware, smoothed surface. Cf. 32. 110 CANTERBURY EXCAVATIONS, SUMMER, 1946 46. Base of grey ware with brown surface, striated. A very large proportion of otherwise featureless sherds were of this type. 47. Shoulder in hard grey ware, with grooves. 48. Rough grey ware with brownish surface, inside much decayed : flat shoulder, smoothed above, matt below grooves. 49. Deeply corrugated shoulder of rough leathery ware : exact slope uncertain (restored after Wheathampstead 11-12, but cf. Prae Wood, Verulamium, Fig. 9, 3). Occupation layer S.E. of ditch (Fig. 6) This contained (a) one base fragment of terra nigra dish of the best period in hght grey paste and fine sUvery surface, from a platter over 1 ft. in diameter (cf. Camulodunum 2-5) ; (b) one piece of spindle-shaped amphora in hard reddish paste with creamy shp. Camulodunum 181-184 ; (c) the foUowing significant pieces which extend the type series in some degree. 50. Grey ware dish, as 38. 51. Brown porridgy ware with dark grey-brown smoothed surface. 52. Coarse brown porridgy ware with black smoothed surface ; cf. 16. 53. SimUar rim in very porridgy dark grey ware with hght reddishbrown surface, badly crumbled inside. 54. Hard well-made buff-grey ware with very feintly tooled trellis. Cf. Aylesford 16 (Arch. 52, PL ix, 7). 55. Native butt-beaker in brownish leathery ware. This and 56 are the only two pieces of native butt-beaker present. Cf. Prae Wood, Fig. 14 (but see also 85, 86 below, from ceUar H I ) . 56. Smoothed buff-grey ware : apparently lower portion of buttbeaker. 57. Porridgy grey ware with smoothed cordons ; crumbled inside. Cf. 10-12. We have here the domestic pottery of the later Kentish Belgic culture, for the first time in a large group. There are interesting distinctions or changes of emphasis to be seen when comparison is made with the funerary groups in which the culture was first distinguished, and by which it is best known. In particular the pedestal urn with its high pear-shaped body and recurved lip is rare or virtuaUy absent. The butt-beaker and its local copies is rare1 (55, 56) : only the imported 1 Two others were found in HI, Nos. 85, 86 below. It should be stated, however, that in the contemporary Belgic site excavated in 1953 at Whitehall Road butt-beakers were plentiful. I l l CANTERBURY EXCAVATIONS, SUMMER, 1946 platters seem to have caused a demand which had to be catered for by the local potters : these were on the whole conservative. Elsewhere in the Belgic region such domestic groups are now known, as at Wheathampstead and Prae Wood1 near St. Albans and at Camulodunum, the Belgic Colchester : but this is the first large Kentish group. It is interestmg also, therefore, to identify the slight variations and local characteristics which must indicate tribal individuality. The group is late : it is tied down to the years immediately preceding the Conquest by (a) stratigraphy ; (b) the brooch, and (c) the presence of terra nigra and its copies, of imported amphorse, and two or three sherds which may be post-Conquest. The pedestal bases are of the late, flat, type : 29 can show some attempt at modelling and lightness, but it does not approach the true early type (Swarling 1-3) ; it is closer to Swarling 8, dated to this same second quarter of I A.D.2 The group being late, an interestmg problem is presented by the corrugated and striated jars 10-12, 57. This type was also found in a just pre-Roman level below the Butchery Lane house.3 At Swarling a similar jar, 31, was found dated early in the series (c. 50 B.C.). This accords with the Wheathampstead evidence. It wUl be remembered that of the two Belgic sites published in Verulamium, Wheathampstead was shown to date before 10 B.C. and is generally regarded as the capital of Cassivellaunus which Caesar captured in 54 B.C. The finds from Wheathampstead therefore give us a picture of the first phase of Belgic culture. Prae Wood, on the other hand, began when Wheathampstead ended, and may fairly be regarded as its successor. At Prae Wood we have material dating 10 B.C.-A.D. 43. At Wheathampstead, then, simUar corrugated (2) and striated (17-20) urns with notched ornament (18, 19) were found ; this group dates before 10 B.C. and probably goes back to Caesarian times and before. There is no parallel to this corrugated type at the later site of Prae Wood. Nor is there a paraUel in Essex, which became Belgic at the end of the century. North of the Thames, therefore, it belongs to the early Belgic period, before Christ. Our group offers other parallels to that at Wheathampstead : e.g., Nos. 14,19, 27 to ibid. 8 which is not found at Prae Wood ; Swarling 19, a comparable form, however, shows that in Kent unlike Hertfordshire this type, too, continued late. No. 4 resembles Wheathampstead 10, and 38 is like ibid. 9 ; these also do not appear at Prae Wood. Essex was inhabited by the non-Belgic Trinovantes, the story of whose resistance to the house of Cassivellaunus and his successors is 1 Both these sites are published in Verulamium, to whioh the references relate. * Cf. also Prae Wood 49 a, b and o. » Arch. Cant., LXI, 34, Fig. 16, 58. 112 PLATB I to I. < v|l«r U Belgie post-bolei 2. Cellar L, Belgic ditch looking north timet p. lit PLATE II # * . .""S •$ 1. Cellar H, Roman drain looking west 2. Section K2, Roman drain looking north-east CANTERBURY EXCAVATIONS, SUMMER, 1946 weU known.1 When Addedomarus, the last native Trinovantian king, ceased to reign about A.D. 1, his coins were superseded, however, by those of DubnoveUaunus of Kent. It seems that there was an invasion across the estuary ; and Hawkes and HuU, noting that the distribution of Belgic pottery-types suggests infiltration up the rivers from the coast, have not hesitated to suppose that it was DubnoveUaunus " who was responsible for the decisive drenching of the Trinovantes with Belgic culture."2 It is odd if this were so, that the comparatively common Canterbury form of large jar with corrugated shoulder and striations should be unknown in Essex. The explanation can hardly be that the Canterbury group is earlier than here suggested, nor that it did not become current till after A.D. 1 : perhaps Kentish settlement did not play so large a part in the Belgic culture of Essex, despite the distribution, as did influence from Catuvellaunian sources in Hertfordshire. Other individual Kentish types not found or rarely found in Essex include 26 and 38. To the former a vague cousinly parallel is known from Billericay not far from the estuary, but it is not closely simUar. No. 38 is akin to Camulodunum 214 B, but the metallic projecting carination of the Wheathampstead prototype, present at Canterbury, is lost at Colchester, and again there is only a general resemblance. No. 54 is an example of a generalized Belgic type of wide-mouthed cordoned bowl, but its particular feature of upright neck, double cordon, and treUised shoulder are not found together at either Prae Wood or Camulodunum. There is, however, an Aylesford parallel (no. 16) : so this, too, may be Kentish Belgic. SimUarly the neckless in-sloping rims 7, 13 and 44 are unknown at Colchester as, too, at Prae Wood, though a prototype for 7 and 13 can be seen in Wheathampstead 8. In general, too, the Camulodunum series is remarkable for the range of recurved rims on bowls or jars with simple bulging shoulders demarcated by cordons : Prae Wood, too, is mainly characterized by necked jars. Such rims and shoulders are markedly rare at Canterbury (32 and 45 are the only such rims actuaUy found) ; instead the bowls seem to be mainly bead-rimmed. The bead-rim, and in particular the furrowed bead-rim bowl, is an integral feature of eastern Belgic culture as has been recognized from the beginning,3 though this fact has tended to be overlooked by the popularity of the type among the western Belgae. The type with incurving shoulder and furrowed decoration, in particular, has been recognized as a Kentish type at Richborough 1 Caesar, B. G., V. 20, 23. Allen, Arch. XL, 16, Hawkes and Hull, Camulodunum 6. 2 Camulodunum 6. 3 Hawkes and Dunning, " The Belgae of Gaul and Britain," Arch. Journ., LXXXVH, 278, 288-90. 113 11 CANTERBURY EXCAVATIONS, SUMMER, 1946 where it occurs plentifuUy in Claudian deposits.1 The type has been studied by Bushe-Fox who points out that the Essex forms usually have not a beaded but a recurved rim.2 It is thus possible to suggest that this Canterbury group displays a tribal individuality, somewhat archaic in form, distinct from the Belgae not only of Hertfordshire but also of Essex ; and that this individuality survived the political vicissitudes of the times3 until the Claudian conquest itself. THE ROMAN AND LATER LEVELS, CELLAR L Layer 3, A.D. 90-120 (Fig. 7) Over the Belgic filling of the ditch was a thin dark sheUy Roman occupation soil; this only occurred in the vicinity of Pit M I (which was 1 58 y f 59 62 *5> \ Kwswl J 64 % 60 I LT ~ ~ > #61 F 65 6 6^**''M®#S?2K5i 7 FIG. 7. Pottery from Roman layers, Cellar L. (£) in fact of the eighteenth century). This Roman layer contained : (i) a small fragment of mosaic, six white chalk cubes c. 0-5 by 0 • 7 in. cemented together ; (U) a flue tUe fragment with roller-die pattern (see p. 115); (in) a flagon neck of granulated orange-buff ware; the rim is missing but appears to be of late I A.D. ; (iv) a flagon rim fragment apparently of Camulodunum form 136 A (cf. ibid. Fig. 51, 7) in similar ware to (U). This is Claudian or earlier, and is clearly a rubbishsurvival. Also the foUowing : 58, 69. Two deep coarse grey-ware bowls with reeded lip. 60, 61. Coarse grey-ware bowls of types 71, 72 below (Fig. 8). 1 Richborough, II, 135, 136 (pp. 97-99). 2 Cf. Oamulodunum 267, a rare type there, and mistakenly, called Roman. » Arch., 90, 29-36. 114 CANTERBURY EXCAVATIONS, SUMMER, 1946 62. Girth-beaker in terra rubra 3.1 63. Rather larger girth beaker in terra rubra 3. 64. Cup of terra nigra, whitish paste, fumed dark grey surface. 65. Rough-cast cup, reddish hard ware with dark slate-to-brown coloured shp : also, unfigured, a sherd in white paste with purplishchocolate shp. Nos. 60, 61 are a common Canterbury Flavian type : 58 and 59 are descended from Camulodunum 246, but such thick large and heavy bowls are not common before Flavian times and can be as late as A.D. 120. No. 65 is contemporary. The occupation layer therefore dates about A.D. 90-120, with rubbish survivals from the underlying Belgic levels. Layer 4. A.D. 120-150 (Fig. 7) Layer 3 was sealed by layer 4, a fighter loamy earth which contained (i) Samian form 18/31 stamped BIGA[ -FEC (Fig. 23, 3). Biga of La Graufesenque, c. A.D. 90-100. (ii) Samian form 37, perhaps by CINNAMVS of Lezoux,2 A.D. 140-150. (iii) Two other unidentifiable fragments of form 37. 66. Small piece of Gallo-Belgic platter. Cf. Camulodunum form 3 in terra rubra 2. 67. Fragment of wall of terra nigra platter, Camulodunum form 13 ; white paste, sUvery grey surface. Both of these are rubbish survivals. Further south a trench was cut to trace the Belgic occupation below a stUl-standing cellar-vault; but the area was found to be deeply disturbed by medieval pits which would have lain behind the houses fronting the Parade ; they contained nothing of consequence. TRENCH K 3 (Fig. 1) This trench was a small test hole dug to explore what turned out to be a Roman rubbish pit, R 2, having useful contents. The stratification was : (a) blackish sheUy occupation material sealed by (b) mixed clay. Both layers contained pottery, and the former also a flue-tile fragment with a roUer-die diamond pattern.3 This is die 16 : a second piece from this die was found in ceUar L, layer 3 (see p. 114) closely adjacent. A piece from die 41 (plain chevron) was found in Pit M 14, ceUar I. It should also be recorded that two pieces from die 29 (plain chevron) came from the Butohery Lane site, one from the medieval " black stony soil with shells " in Fig. 5 (Arch. Cant., LXI) and one from Pit M 6 1 For description see Camulodunum, p. 204. 2 Kindly identified by Dr. F. Oswald, F.S.A., together with all Samian mentioned in this report. 3 A. W. G. Lowther, A study of the patterns on Roman Flue-Tiles and their Distribution, Surrey Arch. Soc. Research Paper No. 1. 115 CANTERBURY EXCAVATIONS, SUMMER, 1946 (ibid. Fig. 6). Two pieces from die 43 (plain chevron) came from ceUar F in the same report, in the buUders' debris associated with the first phase of the west wing (ibid. Fig. 7, section LM). In so far as these tUes are stratified their date appears to be late I-early I I A.D., which is in accordance with Mr. Lowther's conclusions. The pottery from this Pit is here published, as it forms a useful group ; with it were associated the foUowing Samian sherds : (a) lower level, form 36 (Fig. 24, 14). Trajan-Hadrian. (6) upper level form 36, ? Vespasianic, A.D. 70-80. form 27 ? Flavian. (a) Lower Level. Hadrianic (Fig. 8, 68-71) 68. Bowl with inturned rim shghtly beaded. This shape is found in Flavian times (Caerleon Amphitheatre, Arch. 78, Fig. 20, 22) but the heavy fumed grey fabric, pohshed surface, and tooled trellis suggest an early II A.D. date. Cf. Needham (Norfolk Arch., XXVIII, 187), No. 9, A.D. 100-150. 69. Heavy bowl or possibly lid : ware as 68 but no trellis. The angle is not certam, nor easy to parallel. Cf. Verulamium (Arch. 90, Fig. 17, 9), A.D. 200-250 ; but this is too late for our group. See 82. 70. Rough-cast cup, cf. Richborough III, Nos. 300-302. 300-301 are the bulbous type, A.D. 80-120 ; 302 straighter, as here A.D. 90-140. Also Caistor (Norfolk Arch., XXXVI, 197), T 2, A.D. 110-160. Ours not so degenerate as Verulamium, Fig. 27, 9, A.D. 160-90. Hadrianic. 71. Cooking pot, soot-encrusted, with rim grooved for lid ; coarse light grey ware. See 108 there dated A.D. 100-20 ; also 60-61, 72. (b) Upper Level (Fig 8, Nos. 72-84, A.D. 100-150) 72. Ware as 71. Cf. bulbous jar, Richborough IV, 405, A.D. 75-100 ; for simUar grooved rim, but on necked jar, cf. Leicester, Fig. 25, 9 ; 37, 27 (Flavian and down to A.D. 130) ; cf. Verulamium, Fig. 31, 42 (A.D. 120-60) but in different ware. 73. TreUised pie-dish, ware as 68 ; cf. Verulamium, Fig. 27, 6, A.D. 160-90, but there chamfered ; for an earher example of Richborough III, 339 (A.D. 80-120). This type is uncommon before c. A.D. 150, and is normal in Antonine times. 74. Cooking pot with girth-groove ; coarse brown-grey granulated ware. 75. Flagon, reddish paste, yellowish-orange slip; cf. 102, apparently a first century type at Canterbury. Flavian. 76. Carinated bowl with reeded rim ; coarse hard grey ware. This type is Flavian and continues to c. A.D. 150 ; ours with its pronounced carination should be of I A.D. 116 CANTERBURY EXCAVATIONS, SUMMER, 1946 77. Carinated jar, fine hard thin grey ware with thin grey-brown shp ; cf. Richborough I, 76, 77 (first half of I I A.D. or earher). 78. Pie-dish with everted grooved lip (two examples); cf. Angmering (Sussex A.C., 86, 21), Fig. 10, 39, late I A.D. 79. Pohshed grey pie dish with internal bead, a Belgic derivative, cf. Leicester, Fig. 41, 23 (A.D. 125-30) but in different ware. 68 >69 r7 0 S 7 i 7 7\ A 72 \ \ 73 7 4 I / 75 *> 76 :> 77 78 7 79 ? BO \ E=^ \ 82 mwmmmmm®m%% mM M * 81 *f 85 | V ^ 841 7 FIG. 8. Roman pottery from K 3 ( | ) 80. Jar in Belgic porridgy fabric which survives at Canterbury into Flavian times. 81. Poppy-head beaker in fine grey paste with thin white wash over outside and down inside rim; pohshed to lip. An early style of rim ; cf. Leicester, Fig. 26, 33 (Trajan, Hadrian), 42, 1 (A.D. 125-30); Verulamium, Fig. 31, 39 ; Richborough IV, 418 (A.D. 90-125). 117 CANTERBURY EXCAVATIONS, SUMMER, 1946 82. Pie dish or lid with light trelhs pattern, ware as 68, 73 ; vide 69 ; cf. very simUar vessel from Pit R I at 47 Burgate Street, Arch. Cant., LXIII, Fig. 11, 28 (Flavian). 83. Jar with reeded everted rim, coarse light grey ware. This type has a Claudian ancestry, Richborough I, 11 : a Hadrianic parallel exists at Verulamium (Arch., 90, Fig. 15, 17); cf. Leicester, Fig. 27, 2 (late first century and down to A.D. 160) ; also 42, 1 (A.D. 125-30) if ours is a bowl rather than a jar. 84. Small jar or cooking pot; cf. Leicester, Fig. 26, 3 (A.D. 125-30), rare by middle of II A.D. ; Richborough III, 320, a jar of larger size, A.D. 90-140. The majority of the pottery is therefore to be dated to the reign of Hadrian or earher, but 73 and perhaps 69, the latest pieces of the group, may bring the date down to c. A.D. 140/50. THE ROMAN DRAIN : CELLAR H TRENCH I, AND SECTION K 2 SECTION H 1 (Figs. 1 and 9) West of Rose Lane trenches were cut in cellar H, which had been part of the Rose Hotel. The north part of this cellar had been deepened w m PAR EART 1 PIT M9 SSs PAuB I ROMAN LICIT. \ P I T AM LI t-t.-i S ' . ' ^ l i RRTH ±fr\-t XIII CENT TIP £*\6 BLACK SHELLY OCCUPATION PO HOL 6 FLOOF mini ASH RN LOAM DARK GREY EARTH \ .;.. I YELLOW LOAM I ' I WITH MORTAR 34 65 0D * BRICK. 2, TILE KOSE LANE SECTION K 2 BUILD-UP ®.«% SCALE, FEET BSBKSU 4- a LOAM FIG. 10. Section from surface, east of Rose Lane SECTION K 2 (Fig. 10 and Plate II, 2) Here again the early ditch was picked up, but the depth of the section prevented fuU excavation. (This ditch does not appear in cellar L). Its top here had been consolidated with a layer of buhders' debris, over which was one of dark earth corresponding to layer 4 in H 1. The drain itself was built in a trench cut into layers 5 and 4 ; layer 5 1 The Roman Pit R 3 contained a British Tin coin, Allen class 2, which unfortunately dissolved in cleaning. 8 Kindly identified by Dr. K. C. Dunham as a greyish-white siliceous glauconitic limestone, with a few quartz grains and glauconite pellets up to 0 • 1 mm. diameter in a matrix of calcite carrying abundant chaloedonic silica, in places showing perfect spherulitic structures. Large .shell-fragments, partly silicified, are present. Cf. Nos. 5, 6, Arch. Cant., LXI, pp. 42, 43. 119 CANTERBURY EXCAVATIONS, SUMMER, 1946 contained no closely datable objects ; layer 4 was dated c. A.D. 100-120. The structure was better preserved here and the seating of a tUe voussoir could be seen in the east face. On the floor lay a coin of Constans (3AE, A.D. 337-342). The fall was towards the west. In K 2 the floor was at 34-65 O.D.; two consecutive levels in H read 34-53 and 34-47 going west; the third, however, taken at the end of the surviving portion, read 35-12, but this rise is probably due to later distortion. At any rate a fall to the west is in accordance with the natural faU of the vaUey-floor towards the Stour. The drain was robbed in Saxon times ; layer 6 contained a sherd (112) of coarse Saxon pottery of about the eighth century, and the dark grey compact fine soU of layer 7 also appeared to be of this date. It contained very little apart from a coin of Constantine1 and a few scraps of Roman pottery, but one sherd (113) is probably from a seventh or eighth century vessel. Through this layer were cut two pits, Pit S I containing nothing but a few late Roman scraps, and Pit S 2 (not shown on section) which contained two Saxon sherds simUar to 112 and probably eighth century. Next came a late eleventh century occupation consisting of a yellow loam floor associated with post-holes, and burnt red round a central hearth. Above this were layers of thirteenth century buUd up. THE FINDS, SECTION H 1 (Fig. 11) Layer 1 85. Butt-beaker in smoothed soft grey ware, identical with the best native Belgic fabric in cellar L. Layer 2 86. Butt-beaker with neck cordon, as 85. 87. Gallo Belgic terra nigra platter in the best whitish fabric with slaty grey surface, silvery outside below collar. Also, unfigured, ribbed handle of white pipe-clay Gallo-Belgic jug (Camulodunum 161) and a rough flint cleaver. Layer 3 88. Narrow necked butt-beaker : leathery grey ware. 89. Soft grey-ware Belgic platter-base with almost rudimentary foot-ring, imitating Gallo-Belgic form. 90. Terra nigra platter in same fabric as 87. 91. Roughly-made bowl in grey ware with leathery surface ; the ware is hardly yet Roman, but this form with constricted neck seems 1 A follis (Prinoipi Juventutis type). 120 CANTERBURY EXCAVATIONS, SUMMER, 1946 to be post-Conquest. The lower part inside and out has a black coating which might be soot but looks like a bituminous shp. 92. Pedestal base of hard grey ware bearing same black coating on both surfaces of pedestal; this links it with 91, though the ware is not quite identical. 93. Romanized bowl in simUar native fabric ; shoulder rUled as 94 which is in a harder dark grey more Roman ware. ; l 96 97 T 98 D % 100 1 } 10] EIG. 11. Belgic and Roman pottery, Section H 1 (£) 95. Rounded bead-rim jar in native smoothed buff-grey ware. Also, unfigured, rim and corrugated shoulder of large jar as 10-12 ; and a smaU fragment of Samian, unidentifiable. Layers below drain 96. Bowl with reeded flange and girth groove ; brick coloured ware with small granules, somewhat fire-warped : possibly a product of the St. Stephen's kilns ; Pit R 3. 97. Heavy grey ware bowl: layer 7. 121 CANTERBURY EXCAVATIONS, SUMMER, 1946 98. Dark grey pie-dish : layer 10. 99. Dirty whitish mortarium with yeUow surface : ibid. 100. Light grey pie-dish : Pit R 3. 101. Poppy-head beaker, grey ware : layer 10. The significant type is undoubtedly the pie-dish. The various levels earher than the drain, deficient as they were in datable pottery, produced in aU ten examples. This type of dish (Leicester, type A) is most popular in the reigns of Hadrian and Pius. The reeded rim jar 96 is of a type common from late I to mid I I A.D., and 101 is of the early type c. A.D. 80-120 ; 82, also, is paralleled among mortars of the first half of I I A.D., cf. Arch. Cant., LXI, p. 24, Fig. 9, 13 (Butchery Lane). A terminal date, therefore, for these layers about A.D. 150/60 seems likely, emphasized by the predominance of pie-dishes ; and this is borne out by the few Samian sherds found. Layer 4 produced a scrap of form 27 (c. A.D. 60-75) and form 29 (c. A.D.70-80) : the Roman pit, a scrap of form 33 ; layer 7, a form 38 as Oswald and Pryce, LXXII, 2 (c. A.D. 115-25) and layer 10 a piece of form 31 described by Dr. Oswald as ? Antonine. The medieval pottery from pits M 8 and M 9 is discussed on p. 128. THE FINDS, SECTION K 2 (Figs. 12, 13) Layer 2, c. A.D. 60-80 102. Flagon, red granulated paste, dirty buff surface ; flanged lip not easy to parallel : for shape cf. Richborough I, 39 (? mid. I A.D.) and ibid. I l l , 200 (A.D. 70-100) ; see 75 above. 103. Cordoned jar in porridgy grey clay ; laiife-smoothed surface ; a first century type derived at no great remove from Belgic. Also present, similar rim in hard fine grey paste with brownish-grey smoothed surface. 104. Jar in reddish paste, granulated grey surface ; late first century type, Richborough III, 267 (A.D. 70-100), May, Colchester 238 (A.D. 50-100). Layer 3, A.D. 80-100 This contained three fragments of Samian form 18, one of them identifiable as ? Flavian, as well as 4 lids, and one chamfered dish base with traces of gold mica dust ; also one graffito, see p. 142, No. 13. 105. Dark polished cordoned jar, mid first century ware. 106. Jar in coarse grey ware. Cf. Richborough III, 263 (A.D. 70- 100). 107. Everted rim, coarse grey ware; a late lst-early 2nd century type not easily paralleled, but cf. Richborough I, 28 (late I A.D.), Leicester, Fig. 42, 38 (A.D. 125-30). 122 CANTERBURY EXCAVATIONS, SUMMER, 1946 103 104- T 109 i ) 111 106 \07 108 T 10 7 T 119 W 7 115 SZ7 116 T 18 ^ ; 120 Fio. 12. Roman, Saxon and medieval pottery from Section K 2 (£) 123 CANTERBURY EXCAVATIONS, SUMMER, 1946 Layer 4, A.D. 100-120 This contained one orange sherd with gold mica dust, one grey reeded handle, and the foUowing Samian : form 18 and form 27, both not closely datable, and a form 30 perhaps Trajanic (A.D. 110-20). 108. Bowl of granulated grey ware with rim grooved for hd ; cf. Richborough I, 22 ; III, 215 (A.D. 50-75), but a simUar form is found at Verulam in the first half of I I A.D., Verulamium, Fig. 30, 31. 109. SimUar rim, coarse grey ware ; cf. Richborough IV, 405 (A.D. 75-100); but this type can go down to A.D. 120 (Leicester, p. 145, Nos. 7, 8). 122 T 124 J 127 FIG. 13. Medieval pottery, section K 2, layer 9 (J) 110. Thin walled bowl, probably carinated ; coarse granulated grey ware ; a late first-early second century type, cf. Richborough I, 22 ; III, 216 (A.D. 50-75); also Verulamium, Arch., 90, 110, No. 9 (A.D. 55-65). 111. Bowl in hard pohshed grey ware with wide grooved cordon ; cf. Richborough II, 144 ; III, 271, 273 (A.D. 80-120). Layer 5 contained no datable finds. Layer 6. Saxon, Eighth or Ninth Century 112. Coarse hand-made rim of dark grey granulated paste roughly burnished below neck. Saxon, eighth or ninth century. It approximates in paste to the late ninth century pottery from Canterbury Lane, but the rounded rim and burnished surface have an earlier appearance. It is not yet certain when the hard granulated paste of medieval fabric appeared in Canterbury. It seems Hkely that in VII A.D. they were stUl using straw-fUled ware. 124 CANTERBURY EXCAVATIONS, SUMMER, 1946 Layer 7. Saxon 113. CoUar-rim of soft fine grey ware apparently wheel made. This has been restored after a sixth century vessel from Gresham Street1 but Mr. G. C. Dunning tells me its paste resembles the eighth century pitcher from Richborough (Richborough, III, 362). Pit S 2. Saxon 114-115. Two rough coarse Saxon pots, 114 with a httle sheU grit, 115 in sandy granulated ware, surface shghtly burnished. ? VIIVIII A.D. Layer 8. XI A.D.2 116. Small crucible, fine grey ware slightly fused ; on floor. 117. Cooking pot with sparse shell-grit; flattened everted lip ; on floor. 118-120. Cooking pots, coarse sandy grey to black ware. 121. Bowl. Nos. 117-121 are all of the simplified flaring type of profile which is certainly pre-thirteenth century. A certain amount of flattening or bevelling of the rim is perceptible, and evidence from Pits M 14, M 8 and M 9 points to this being an introduction of late XI A.D., the vessels of the middle of XI A.D. being plainly beaded. A late twelfth-early thirteenth century group from Butchery Lane may be contrasted (Arch. Cant., LXI, p. 37, Fig. 17). Layer 9. XIII A.D. 122-123. Coarse grey-brown jars, carefully knife-trimmed inside and out; short everted rims. 124. Flagon rim, with traces of green-brown glaze below carination. 125. Bowl or large lamp, crudely made grey ware. 126. Cooking pot, cf. 118 ; twelfth century survival. 127. Club-headed bowl, red ware. 128. Flat-rimmed bowl, red ware. 124,127-128 are recognizable thirteenth century vessels, 127 amongst the early thirteenth century group (Butchery Lane, I.e.), 124 and 128 later in the century (ibid., Fig. 18). Also of late XIII A.D. were four sherds with scratched decoration below greenish-brown glaze, and one sherd of cream-coloured shghtly granulated ware, with a good 1 Wheeler, London and the Saxons, 166, Fig. 32, 1. 2 This layer contained a fragmentary silver coin which Mr. R. H. Dolley says cannot be earlier than Edward the Elder nor later than Aethehed I I : he is reasonably sure that it can be placed between A.D. 916 and 970. This coin must be a survival : the pottery cannot be put back a century. Vide also Fig. 23. 6, a bone tool with Saxon affinities, from this layer. 125 CANTERBURY EXCAVATIONS, SUMMER, 1946 thick yeUow glaze over a shght horizontal ribbing, the ribs being 7-8 mm. apart. This is an Enghsh but non-local vessel. There was also a piece of Niedermendig lava quern. CELLAR K. TKENOH 1 Pit M 3, a straight, vertical-sided cut fuU of animal bones contained only three sherds, of which one might be Saxon. Pit M 2 contained two medieval sherds, and Pit M 4 a twelfth century cooking pot. Around these pits was about 16 in. of dirty loam becoming cleaner towards the base ; natural soil was 29 in. below the cellar floor. CELLAR J Two trenches were cut parallel with Rose Lane; they produced little. Natural soil was about 30 in. below the cellar floor ; above this was about 8 in. of dirty loam with little pottery, and above this a thin layer of dark later Roman earth.1 Around pit M 5 this contained lumps of disintegrated opus signinum. Pit R 4 was cut through the dirty loam ; around it the surface of the loam was very hard and beaten, yet clean. This could not be explained at the time of excavation, but later knowledge suggests that one of the east-west streets of the Roman city crosses the south end of this ceUar (removed by the ceUar excavation) and this may have caused compression of the loam ; yet there was no trace of metal immediately above, nor was there enough occupation material for it to have been a hut floor. THE MEDIEVAL FINDS, CELLARS H AND I TRENCH H 4 (Fig. 1) Along the west and south sides of cellar H ran a very deep medieval footing of loose gravel flints and mortar-dust. The area was very deeply cut by intersecting medieval pits ; most were earlier than the foundation but one pit, which was later, contained spUls of mortar from the footings. At one point these had been dug down 9 ft. from the ceUar floor in an endeavour to reach natural soU through the pits. It seems that the foundation was for the medieval predecessor of the ceUar, and that the modern cellar had been enlarged by cutting back the thick medieval wall. The pottery from the pits through which the foundation has been cut is of late XII A.D. Rims of the flattened or beveUed type (cf. Pits M 8, M 9, Nos. 14-16, 19-22 below) are common, and there is one 1 Containing three coins, (a) 1 Barbarous Fel. Temp. Reparatio, ? diademed head ; (6) ? Radiate crown of c. A.D, 270 ; (o) fragmentary, third or fourth century. 126 CANTERBURY EXCAVATIONS, SUMMER, 1946 example of the clubbed type (Butchery Lane, Arch. Cant., LXI, Fig. 17, 9, there dated c. A.D. 1200 ; cf. No. 127 above). This type does not appear in any of the twelfth century groups pubhshed below, and this fact reinforces the Butchery Lane evidence. The foundation, therefore, was not constructed untU the end of XII A.D. Overlying the pits was a gravelly layer containing late thirteenth century pottery ; this was probably the medieval cellar floor. Y SOUTH - WEST NORTH CONCRETE FLOOR OF CELLAR. I I I I I I I I I I I I I I I I I I I I I I I M 1 I I I I I I I I I I I I I I I t » / i ACy 7/ ?. DIRTY t . < l MORTAR PIT ^ \ \ YELLOW LOAM I DISINTEGRATED MORTAR a-SteL£u?c? o ^ DIRTY PIT LOAM R 7 DARK BROWN OAM SECTION 1,1 SCALE : FEET ? FIG. 14. Section Cellar I CELLAR I (Figs. 1 and 14) A trench cut in cellar I produced an interestmg sequence of medieval pits at its north end. M 12 contained a plain stiff black filling with a little late Saxon pottery and two pieces of Niedermendig lava quern. It cut M i l and was cut by M 13 which was full of loose soft brown vegetable matter suggestive of a cess-pit. M 14 contained a soft black sticky filling, with very large quantities of animal bones1; the pottery dated from the middle of XI A.D. Cutting M 14 was M 15 with much stiffer black filling, many oysters and flints, and a large series of restorable pots of late XIII A.D. (Figs. 18-22) together with an ampulla- 1 Goat horn, cut; sheep or goat, jaw; sheep, skull; calf; ox, tibia and mandibles ; horse, pony size, tibia and fore cannon bone. 127 CANTERBURY EXCAVATIONS, SUMMER, 1946 mould portraying St. Thomas of Canterbury (PI. I l l , 3) and a roof slate (Fig. 24, 9). South of this series of pits was a Roman Pit R 7 containing two sherds, sealed by a layer of chalk lumps and a mass of disintegrated yeUow mortar recalling the footing on Trench H 4. Above this was a yeUow loam floor sealed by black medieval pit material. There is no real clue to the date or purpose of this mortar. It is pre-Norman, for Pit M 14 cut it on the east face of the trench ; but when the west part of this pit feU away after heavy rain, it was seen that the mortar did not continue northwards ; natural loam lay 28 in. below the cellar floor, and was capped by dirty brown loam (Fig. 14). The original north edge of the mortar thus lies where shown in the Section, and its north termination has been disturbed by the deepening of the medieval black soU. All that can be said is that it is later than Pit R 7 and the dirty loam over natural soU (neither of which are closely dated) and earlier than Pit M 14 : if it is connected with the buUding of the medieval ceUar, this ceUar is a century older than the one investigated in H 4. The Medieval Pottery The Rose Hotel site covered a valuable series of medieval pits, the pottery groups from which are Ulustrated below. To some extent, in the absence of coins, the dating is still subjective ; but as associations are increasingly noted and more homogeneous groups are published from the city, this element of uncertainty will decrease. The grounds of dating are discussed below : they are in part stratigraphical and in part typological. As far as typology is concerned, it may be said that the late llth-early 13th century pottery of Canterbury is very homogeneous in ware and shape, and distinction is rarely possible except in a group. Overlaps occur, so that it is not yet possible to place individual pieces with certainty ; but a group from a pit or layer does seem on analysis to possess a subtle distinction, unapparent though this may be at first inspection. GROUP I. PIT M 12 (Fig. 15, Nos. 1-5) In this group the ware is the usual coarsely granulated dark grey or buff ware. Nos. 2, 3, 5 have been wheel turned ; 1 and 4 lack obvious traces of this. Except 2, they all bear traces of vertical knife-tooling. The neck of 3 has been slashed as the knife has been drawn up ; the neck of 5 has been facetted, apparently by downward strokes of a knife ; and 1, 3, 5 all bear shght facetting on the body. Nos. 3 and 5 appear to have horizontal knife-tooling inside the shoulder. The vertical smoothing, indicated by the dragging of sandy particles on 128 PLATE 111 1. Thirteenth century jugs from Rouen. Museum of Antiquities. Rouen fnrhes < • '2 * i. Sherd of imported French pottery, Pit M 16 (p. 138) 3. Ampulla-mould from Cellar I, Pit M IS (p. 13(1) CANTERBURY EXCAVATIONS, SUMMER, 1946 the surface, is not found on any of the other groups and is a late Saxon characteristic at Canterbury.1 This conclusion is borne out by the simple form of the rims ; 2 is the only rim of more developed character, which if isolated might be thought to be later. A date around A.D. 1000 would seem likely for this group. i t >•' * ( ? ?- FIG. 15. Medieval pottery, Group I, from Pit M 12. ? Early XI A.D. (J) GROUT? II. PIT M 14 (Fig. 16) 6. Cooking pot of dirty brownish granulated ware ; simple everted rim from shght shoulder ; roughly beaded lip : knife-smoothed (rather than knife-trimmed) round base. 7. SimUar to 6 ; greyish-brown granulated ware. 8. SimUar ware, more carefully modelled rim with flat top and thinner walls though of same diameter. 9. Reddish granulated ware; simple beaded lip flattened and sharply angular internally. 10. Grey granulated ware; large cooking pot; J-round rim section. 11. Large cooking pot; hard brownish grey ware; roughly beaded lip, and base knife-smoothed. 12. Reddish-grey granulated ware ; simple rim concave internally perhaps for lid. 13. Reddish-brown ware ; simple beaded lip, bevelled internally. Also in the pit was a small sherd of hard white granulated paste c. 3 mm. thick decorated with a whorl of brownish red paint. Mr. G. C. Dunning tells me this is certainly imported and may be late Rhenish, from Pingsdorf or Dutch Limburg or possibly from Normandy. This group is useful both because it is stratigraphically senior to that in Pit M 15, and also because of the uniformity of its contents. 1 Eor a slightly different teohnique see the Norman " scratch-marked" Pottery discussed in Arch. Journ., CVII, 34. 129 12 CANTERBURY EXCAVATIONS, SUMMER, 1946 The knife-smoothing round the base does not amount to actual trimming noticed in late Saxon wares of East Anglia and the Oxford region1 but it is likely to belong to the same tradition. The group noticeably does not contain any of the thickened rim sections of the late 12th-early 13th century Butchery Lane group,2 still less any of ) 7 M y 10 Y r FIG. 16. Medieval pottery, group II, from Pit M 14. ? second half of XI A.D. (J) the flattened hammer-head type common hi the late 13th century (ibid., Fig. 18, and Nos. 25-36 below). On the other hand, although the lips are in some cases irregularly moulded, they do have an incipient bead and the pottery is on the whole competent and cleanly buUt up, 1 See Oxoniensia, V, 42-6 ; VII, 73, and the references given there ; also Leicester, Fig. 59. a Arch. Cant., LXI, p. 37, Fig. 17. For the type see Fig. 13, 127, above. 130 CANTERBURY EXCAVATIONS, SUMMER, 1946 and not simUar to the black rather crudely-made wares of Alfred's time found in Canterbury Lane, nor yet to Group I here published from Pit M 12 and assigned to the beginning of XI A.D. We may notice, too, that the internal bevel (13) is rare and simple : clearly it is only just beginning. The group therefore must be placed earlier than Group III from Pits M 8 and M 9, where the internal bevel is more pronounced, leading to a definite flattening on top ; in these pits this is the leading type, and it is fairly clear that the typological sequence is from Group I I to Group III rather than the reverse, because of the greater sophistication of this flattening, which leads on to the angular rims of Group IV. Group II, then, is typologically earlier than Group III, and pottery of both groups was found beneath the Castle Keep in 1953. The date of the Keep is doubtful and cannot be discussed at length here. It may go back to WUliam I, as Toy1 believes ; it was certainly in existence by 1158, when the Pipe Rolls begin, for under the years 1172-4 they record expenditure on the Keep, insufficient, however, for building it de novo. It is reasonable to believe that the Keep dates at least from the first quarter of XII A.D. Pottery of Group III was found beneath it and also in association with its earliest floor ; the group therefore dates at least from c. 1100-1150. In 1954 pottery was found stratified beneath the buUding level of the undercroft of Lanfranc's Dormitory in the Precincts ; this suggested that the transition from Group II to I I I was in fact occurring c. 1070-5. Group I I is earlier, as both typology and the Castle stratification suggest. We may provisionally date it to the middle of XI A.D. (C. 1030-70). GROUP III. PITS M 8 AND M 9 The pottery from these pits is contemporary and simUar, and together gives a conspectus of the pottery of the first half of XII A.D. (see above). Sherds of this type were found in 1953 to be contemporary with the erection of the Castle Keep. In Pit M 14 the rims were simple straight everted types rising from unobtrusive shoulders ; the lip was shghtly beaded and rounded; the flattened or bevelled type was only beginning to appear. In Group III, however, the simple rounded beaded type survives as a rarity (17), and the internally bevelled (15, 19, 21, 22, 23) or flattened (14, 16, 20) type has become the rule. This is reaUy only an advance in finish : the basic form is stUl the straight everted rim type, though the shoulders are now perceptibly more rounded; the curved clubbed form of lip dated at Butchery Lane, c. A.D. 1200 is stiU absent. The group, therefore, may be provisionally placed in the decades A.D. 1100-1150. 1 S. Toy, The Castles of Great Britain, 1953, p. 64. 131 CANTERBURY EXCAVATIONS, SUMMER, 1946 The ware of all pots is a very coarse granulated plain buff, reddish, or grey, and is irregularly modelled (though wheel made), so that the thickness varies round the perimeter at the same level. Knife-smoothing round the lower part is still present.1 T 14 7 Y YY 18 T 22 LO \ 23 \ 24 FIG. 17. Medieval pottery, Group III, from Pits M 8 and M 9. ? First half of XII A.D. (£) GROUP IV. PIT M 15. (SECOND HALE OE THIRTEENTH CENTURY) This pit contained the stone ampulla-mould of St. Thomas (PI. I l l 3), a roofing slate (Fig. 24, 9), and some sherds of a jug imported from Normandy as well as a large group of local wares (Figs. 18-22). 25-36. Cooking pots mainly in coarse buff or reddish granulated ware, often soot-encrusted outside and burnt grey. 25 is in black very shelly ware with rim rather warped. 28 is of reddish-brown fine paste with medium shell-grit, mainly dissolved out on inside. Most of the rest bear a little fine or medium shell grit round the shoulder, but this rarely appears on the fracture or inside, as if it were dusted over the finished pot. Some have low vertical ribs on the body : the rims are carefully moulded and often bear small spaced stabs on top. 37-38. Crudely made conical cresset lamps, 37 with a small pinched spout. 39. Very large dish, 21 • 8 in. in diameter with heavy hammer-head rim bearing two rings of stabs ; red granulated ware with scattered shell-grit on surface, but none in the breaks. 40. Smaller simUar dish with brownish-green glaze-patches within. 1 Pit M 9 also contained a second century mortar rim with chevron stamp (Fig. 24, 10). 132 Fro. 18. Medieval pottery, group IV, from pit M 15, second half of XIII A.D. (£) 133 58 n FIG. 19. Medieval pottery, group IV, from pit M 15, second half of XIII A.D. (J) 134 CANTERBURY EXCAVATIONS, SUMMER, 1946 J «Mll4 .ffll.l « FIG. 20. Medieval pottery, group IV, from pit M 15, second half of XIII A.D. (•}) 135 CANTERBURY EXCAVATIONS, SUMMER, 1946 41. Top part of fine pitcher, reddish granulated ware, grey in section, brownish-green glaze up to carination of neck ; wide deeply stabbed handle bordered each side by a moulded rib deeply incised to resemble cabling, glazed ; a zone of oblique incisions round neck and on shoulder. •*-*?!? 48 49 V / V. 50 u i II FIG. 21. Medieval pottery, group IV, from pit M 15, second half of XIII A.D. (£) 42. Pitcher in granulated dirty brown ware, grey in section ; body rUled by parallel grooves and covered with greenish-brown glaze up to neck and patchily to lip ; stabbed strap handle. 43. Simple squat pitcher unglazed except for small patch on shoulder ; sandy brown ware ; soot-stained round base, so has been used for cooking. 136 CANTERBURY EXCAVATIONS, SUMMER, 1946 44. Pitcher in fine bright orange sandy ware decorated down to base of ribs with white shp covered by a patchy green glaze caused by inclusion of copper filings in glaze ; over the white slip but under the glaze are vertical ribs of brown barbotine. A non-local English product, possibly from the London region. 45. Buff sandy pitcher with traces of yellowish-green glaze near shoulder ; base pierced after firing by seven smaU holes. 46. Part of shoulder of pitcher in fine sandy red ware, perhaps non-local; the surface is worked up into vertical ridges, covered with a glossy bottle-green glaze. Also present another simUarly but more •e=» 51 s fr \ (/ i/ % V/ •I >s ? as«» rtitKtiiTif-- ^Cj=^«3SSii FIG. 22. Medieval pitcher, group IV, from pit M 15, second half of XIII A.D. (\) crudely ridged shoulder, patchUy covered in greenish-brown glaze over local ware as 42 ; this is imitating the non-local 46. 47. Central part of pitcher, ware as 41 ; two horizontal grooves on shoulder which is sparsely splashed with greenish-brown glaze : below the bulge the surface has been finished with a vertical movement which has left striations. 48. Pitcher in red sandy only slightly granulated ware with grey core ; thin and rather friable with a disproportionately heavy plain handle fixed at the top by two finger impressions and facetted with a knife : the body decorated with applied ribs of reddish clay running slightly obliquely except behind handle where the small groove and cordon demarcating the neck still shows : covered with good yellowish to bright green glaze mottled with copper-filings, except patchily on 137 CANTERBURY EXCAVATIONS, SUMMER, 1946 handle and round rim and base. An ambitious yet unskilful local product. 49. Ware and patchy glaze as 41 ; spout very slightly pinched out. 50. Large pitcher in thin reddish-brown granulated ware with buff-grey core, crudely decorated with zones of obhque comb-incisions separated by grooves ; the shoulder bears a decayed patchy greenishbrown glaze. 51. Orange-brown fine sandy paste : thin patchy glaze showing orange-red on upper part of body : fine wheel-turned tilling over upper part, but hardly showing where covered by glaze. IMPORTED FRENCH POTTERY FROM P IT M 15 (PI. I l l 2, and Fig. 24, 11) Mr. G. C. Dunning, F.S.A., has kindly contributed the following note : The fragment (PI. I l l 2) is of fine whitish ware, light buff inside and showing fine wheel marks ; the ware is very hard and thin and skilfully turned, and the thickness is uniformly only 2-5 mm. The outside has a dark brownish-red shp, on which are seven smaU pellets of white clay in no obvious arrangement. On the right edge of the sherd is a white apphed strip with deeply impressed square-cut rouletting. The sherd has an overaU thin lustrous pale yellow glaze. A second smaller sherd from the same vessel (not illustrated) also has red slip and a white strip, now mostly flaked away. A third sherd (Fig. 24, 11) bears a neck cordon as PI. I l l 1, b and has a matt scar where a rib has flaked away. The fineness and lightness of the fabric and the style of decoration are characteristic of a series of highly decorated jugs found at Rouen, and preserved in the Museum of Antiquities at Rouen. Two Rouen jugs are illustrated in PI. I l l 1 chosen to show the ranges in types and decoration. The jugs may be either tall with a retracted base or squat and globular, and both shapes have a cylindrical neck with grooved surface, sometimes emphasized as broad corrugations. The decoration on the body is usually carried out in applied strips and pellets, often forming a series of medalhons or panels, as on the two jugs Ulustrated. The inside of the motifs is usually covered with dark red slip, which serves to emphasize the major elements of the pattern and contrasts with the white strips and pellets forming the subsidiary elements of the decoration. Sometimes, as on the larger jug Ulustrated, the strips are rouletted by diamond-shaped or square notches made by a roller-stamp, as on the Canterbury sherd. Pottery imported from Normandy, and almost certainly originating at Rouen, has been found at two other sites on the S.E. coast of England. At Pevensey Castle several green glazed sherds with narrow applied strips or ribs running vertically, exactly matched on a small jug at 138 CANTERBURY EXCAVATIONS, SUMMER, 1946 Rouen, were found in 1936 sealed beneath a layer of clay in the outer bailey of the Castle.1 The clay is that removed for the foundations of the curtain walls and towers enclosing the courtyard of the Castle and the surrounding moat, which belong to the middle of the thirteenth century and were buUt by Peter of Savoy, who became Lord of Pevensey in 1246. The other site is Stonar, the medieval town and port near Sandwich, where for many years Mr. W. P. D. Stebbing, F.S.A., has watched the commercial digging, rescuing a considerable amount of pottery, and other finds, and also conducted excavations.2 The foreign trading relations of Stonar are already demonstrated in the twelfth century by sherds of red-painted pottery imported from Normandy.3 In the thirteenth century there is a much greater quantity of imported pottery, showing the increasing prosperity of the place. The majority of the sherds are of the polychrome type from W. France,4 together with green glazed jugs of simUar origin, but also including a few sherds with dark red slip and small pellets, exactly comparable with the Canterbury sherd. The interest of the imported pottery from these three sites in S.E. England is therefore in showing the trade connection with Normandy, established in the Norman period, continuing into the middle and second half of the thirteenth century. At that time the intensive wine trade with Gascony was bringing along a much greater quantity of pottery from Aquitaine,5 which reached ports mainly on the W. andS. coasts of Britain. AMPULLA-MATRIX FROM PIT M 15 (PI. I l l 3) The matrix comprises half a mould for the manufacture of lead or pewter ampullae. These small flasks were sold to pilgrims to take away holy water, oil, or water from Becket's well, as a souvenir of their pilgrimage.8 The neck is missing, but two dowel-holes can be seen, and the crudely incised figure of St. Thomas Becket in mitre, chasuble, cope and pallium, holding crozier and in act of blessing. Dr. K. C. Dunham kindly reports that the rock of which the matrix is composed 1 Ministry of Works Guide to Pevensey Castle (1938), p. 7. 2 Arch. Cant., LIII, p. 62 ; LIV, p. 41 ; and LV, p. 37. 3 Cf. the red-painted pitcher found at Dover, Antiq. Journ., XXV, 153. 1 Arch. Oant., LIV, p. 56. 6 For further discussion of the trade connections with France in the thirteenth century, see " Dating Medieval Pottery," Part XV, Archceological News Letter, August, 1949. 0 See Arch., LXXIX, 33, London Museum Mediaeval Catalogue 261, and references there given ; for other moulds see British Museum Guide to Mediaeval Antiquities, 1924, 24. It is noteworthy that ours is of considerably cruder workmanship than the majority illustrated. 139 CANTERBURY EXCAVATIONS, SUMMER, 1946 is a clay ironstone.1 He suggests that it was made from a clay ironstone nodule, and that its most probable source would be Wealden ironstone. The rock occurs as nodules in the Wadhurst Clay and Ashdown Sand of the Kent and Sussex Weald. It is interestmg to note that G. S. Sweeting2 states that the ironstone was being quarried during the Middle Ages, and mentions for example that important works were opened up at Tudeley near Tonbridge during the reign of Edward III. The present example, however, must be at least half a century older than this. THE SMALL FINDS (Figs. 23, 24) 1. Bronze brooch, Belgic, cellar L, ditch layer 2. Mr. M. R. Hull, F.S.A., has kindly examined this brooch and writes that it is of the Aucissa group but thinner and narrower than the true Aucissa; the bow seems to have run into the foot directly, not through a moulding. The type is very rare; one was found at Camulodunum in a deposit dated to A.D. 43 (op. cit. no. 133) and Schulz shows another from Stossen, Kreis Weissenfels, which has the same three frond-like ornaments on the head (Germania, X, 112 abb. 3). With this brooch was found part of the pin and spring of a very large brooch of Camulodunum type 111 or Swarling, PI. XIII, 9. 2. Very decayed bronze fitting from H I layer 3. It is thus of Claudian date. It appears to be a hinge, now bent double, but illustrated in a restored form in the left hand drawing ; each plate is double, the top bearing two rivets and the lower one. This closely resembles the hinges from the back of the legionary cuirass (Camulodunum, pp. 337-8, PL CH, 15 ; Richborough, I I I , PI. XII, 39, g-j). If so, this is the first piece of military equipment identified in the excavations, and it indicates that the ditch was filled in during the earliest phase of the occupation. 3. Samian potter's stamp BIGA[-FEC from CeUar L, layer 4 (see p. 115). 4. Small solid bronze figure crudely modeUed and with divided base to receive a tang or flange ; from Pit M 1, eighteenth century or earher. 5. Small fluted metal button from the gravel floor of medieval cellar, trench H 4 ; thirteenth century. 1 Consisting of a mixture of ohalybite in rhombs averaging 0-005 mm., and lhnonite. Angular grains of quartz of about 0-02 mm. average size and tiny flakes of detrital mica are scattered through the rock. Fossil shell fragments, preserved in calcite, are present, and there are a few small grains of the phosphate collophanite, and a few streaks of pyrite. 2 " Wealden Iron Ore and the History of its industry," Proc. Geological Assoc, LV, 1944, 1-20 ; reference from Dr. Dunham. 140 CANTERBURY EXCAVATIONS, SUMMER, 1946 6. Very highly polished bone weaving tool pointed at both ends, but seemingly recut at one. Perhaps a thread-picker, as suggested for the simUar Saxon object at Harston (G. C. Dunning, Trans. Leics. Arch. Soc, XXVIII, 49-50). Section K 2, layer 8. SE=> V\ ^Yll -» mB. @ \m •*-. c 2 2 4- 3 5 6 • r mimiUWHUi/ ifihi&tout,^ 8 FIG. 23. The small finds (£). 7. Bronze signet ring with incised W surmounted by coronet. K 2, layer 10, a layer dated by half-pennies of WUliam and Mary and George II. 8. Roman blue glass melon bead, a stray in the thirteenth century layer 9 in section K 2. This is a first century type, cf. Leicester, Fig. 93, 8. 141 CANTERBURY EXCAVATIONS, SUMMER, 1946 9. Roof-slate from Pit M 15 ; late thirteenth century. It has been compared with a number of Scottish, Enghsh and Welsh slates at the Geological Survey, but cannot be assigned to a particular locality of origin, nor has Mr. J. Setchell been able to identify it. 10. Pale yellow mortarium with maker's mark ; late first century type and probably a local potter. Pit M 9. 11. Neck sherd of imported Normandy ware, Pit M 15; see p. 138. 12. Graffito on Samian form 18/31 ; from medieval pit in trench H 4. 13. Graffito scratched on tile before firing : K 2, layer 3. Mr. R. P. Wright, F.S.A., reads the upper line as ]GE[ and the lower line as ]AVA[ a poor V between two A's. (J.R.S., XXXVII, 1947, 182, no. 16.) 14. Samian form 36, half the vessel present; diameter c. 186 mm.; groove inside below flange. Cf. Oswald & Pryce, PI. LIII, 10 (Trajan- Hadrian). Trench K 3, layer A, p. 116 above. BIBLIOGRAPHICAL ABBREVIATIONS The following are additional to those already cited in previous reports in this series. Oxoniensia Journal of the Oxford Architectural and Historical Society. Norfolk Arch. Norfolk Archseology, journal of the Norfolk and Norwich Archseological Society. Arch. Journ. The Archseological Journal, of the Royal Archseological Institute of Great Britain and Ireland. Leicester K. M. Kenyon, The Jewry Wall, Report of the Research Committee of the Society of Antiquaries, No. XV. May, Colchester T.May, Catalogue of Roman Pottery in the Colchester and Essex Museum. Sussex A.C. Sussex Archseological Collections, journal of the Sussex Archseological Society. ACKNOWLEDGMENTS The writer records his grateful thanks to all who helped with the work. To Major F. W. Tomlinson, F.S.A., the first Hon. Secretary of the Excavation Committee, especial gratitude is due for all he did in arrangement of preliminaries, smoothing out of difficulties, and encouragement whUe excavation was in progress ; to the owners of the sites for permission to dig; to Dr. F. Oswald, F.S.A., for examining and 142 zzzzz^SZZZ2ZZZ^2ZZ2^ZZffiZZZZZMZ YELLOW GLAZE MATT FIG. 24. The small finds (§). [face p. 142 CANTERBURY EXCAVATIONS, SUMMER, 1946 reporting on the Samian ; to Mr. M. R. Hull, F.S.A., for helping with the Belgic material, and for drawing the brooch : and to the various experts to whom acknowledgment is made in the text. Mr. B. H. St. J. O'NeU, F.S.A. as usual reported on the coins ; Mr. A. Martyr Smith made the preliminary surveys ; and Mr. J. Mann took the levels. Without the help of these, and of all the volunteers in the field, especially Mr. V. J . Newbury and Mr. J. S. Wacher, no progress could have been made. 143