Canterbury Excavations Christmas 1945 and Easter 1946

^rrfoflkp* dfantira. CANTERBURY EXCAVATIONS, CHRISTMAS 1945 AND EASTER, 1946 THE BUTCHERY LANE BUILDINGS By AUDREY WHXIAMS, F.S A., and SHEPRARD FRERE, F.S.A. THE probability, at the end of 1945, that the open cellars on the south side of Butchery Lane would shortly be filled in preliminary to building dictated immediate examination of the area. The Canterbury Excavation Committee, therefore, began work in the three cellars south-west of F. W. Finnis and Sons' premises (Fig. 1, A, B and C). The discovery of substantial remains of a Roman building in cellar C made it desirable to continue and extend excavations in the same area at Easter, 1946. In April, therefore, two deep cuttings from the surface were made behind cellar C to trace the southward extent of the pavement and trenches were cut in cellars D and E. During the excavation it became necessary, owing to the likelihood of rebuilding, to extend operations to cellars F and G also, below The Parade (Fig. 2). This work led to the recovery of parts of two Roman buildings, one of them, building 2, evidently an extensive quadrangle house, long inhabited. The main quarters of the house lie below Butchery Lane and the Parade, whither they could not be traced; the extant remains include room 2 in the east wing, 40 feet by 10 feet 6 inches, doubtless a lobby ; a corridor (room 3) 10 feet wide and traced for 40 feet 6 inches, skirting the north wing ; a small hypocaust (room 4) ; and walls of the west wing. CELLARS A AND B In cellars A and B a cutting was made parallel to Butchery Lane and at a right angle to the cellar party walls (Figs 1 and 3, AB). In cellar A the cutting was taken out only to a depth of just over 4 feet; but in cellar B excavation to 13 feet below the cellar floor (24 feet O.D.) revealed natural soil, a bright yellow loam capped by a thin streak of sandy soil. This surface, lower than any found on previously excavated sites in the city, quickly became water-logged. This condition had been remedied in antiquity by raising the surface 6 feet or more with a deposit of clean yellow loam. In this loam had been cut three pits (Rl, 2 and 3). Pit Rl contained Claudian Samian ; R2 had no datable material but was earlier than R3, from which came Samian dating up to A.D. 70. As, however, the loam to the west (in cellar C) overlay Vespasianic debris 1 4A CANTERBURY EXCAVATIONS, 1945 AND 1946 it cannot date earlier than A.D. 79 and the material from the pits must be disregarded. Over the yellow loam, and sealing the pits, lay greyish soil interlarded with streaks of gravel and charcoal, the latter brought from elsewhere and not here indicative of fires. Here again the material, including Vespasianic Samian, suggests a rather earlier date than can be sustained in the face of the evidence from cellar C where a date c. IS 20 K a — H — • • oi M3 1 V BUILDING I */A rrrr V ^ N ^ BUILDING II PERIOD i ^ M3 ' . ' t \ Qi i / / ^ MS ,' PERIOD 2 t. / /• r PERIODS 1 & 2 f2? rz m yy\h:' 7An\i A Ii i B c I ) '/A i i » i t i-jj 37-27 i « K-37 I I ! I 25 !*i i ' ' - L ^ ^ IfJ RS R I I a 37'I9 37-23l/-f R2 I PANEL! S un ^ i 3732 i^ffi •T« 1 M c l i I «* 3r-3+ KERB -v, BUTCHERY LANE ElG. 1. A.D. 90 is likely for the build-up to 7-8 feet above the undisturbed surface. BUILDING 1. Into the late first century grey soil had been sunk the foundations and lower courses of a roughly coursed flint wall. The 2 CANTERBURY EXCAVATIONS, 1945 AND 1946 3avhiVd 3HJ_ •qirmnrirfi * ^ S 3 &:•.-: •T|T-s-iai-uw . . i i j j 'iilBLCffiiTij' , f ^ ^ ~ < O JS'.SS^SSSS? kv^wwwvwwj i _.-.-.X- (J 3 CANTERBURY EXCAVATIONS, 1945 AND 1946 K&&Z iv>%iSzSopi T%mm%zz^^^?M MORTAR V W ^ ^ W W V W W W CHARCOAL GRAVf MOftTAR MORTAR ~WZ7s. \ 1 uoim> Jvmoi-iiowN io»» g 3 CUYIIH iOH j 110 1011 TI1H CHAUOtl n~) NATUKAl 1011 =• [ = 3 m imiNO rrrm MOUA* FIG. 3. SECTIONS IN CELLARS A, B and C. 4 CANTERBURY EXCAVATIONS, 1945 AND 1946 foundations of gravel with mortar were 2 feet wide and 2 feet deep ; the wall was reduced to under a foot in height. Over it rested the narrow party-wall of cellars A and B. Of the Roman wall only a length of 13 feet could be exposed before it disappeared beneath Butchery Lane. It was taken to be the south-west external wall of a building lying partly under the modern street. A patch of mortar, on the grey soil and against the wall-face, presumably a mason's patch contemporary with the building, produced part of a mortarium of late first century type (Fig. 9, 2). East of the wall a continuous layer of black soil with household debris, on top of the grey clay, suggested a floor. This could be dated to the early second century. Over the wall and on both sides of it ran a layer of yellowish loam with, above, a number of deposits of various materials—black soil, more yellow loam, mortar and red soil with charcoal—evidently materials dumped indiscriminately to level over a disused building. The pottery (Fig. 9, 3-5) points to the operation having been carried out in Antonine times or soon after. A pit (R4), of which only the lower part had survived the digging of the cellar, disturbed the Antonine layer in the interior of the building. Coins dated the pit to the late third century. Building 1 appears to have been in use early in the second century and abandoned in the second half of that century. This short history contrasts very markedly with the long occupation of building 2, but it could not be accounted for in the two cellars examined. However, a possible explanation can now be suggested as a result of excavations by Messrs. F. Jenkins and J. Boyle in 1946-7, in the premises of No. 47 Burgate Street (Fig. 4). A report of this work will be published later, but the plan has been prepared from information supplied by them. In the cellar at the north end of the site were found levels suggesting the demolition of a building at the end of the second, or in the early third century ; while in trenches at the south end of the area a large gravel foundation of late Roman date, mutilated by medieval pits, was found. Its full extent was not revealed, but it clearly supported some massive structure, to make room for which the earlier building had been cleared. CELLAR C. In cellar C in a cutting alongside the north-east party wall natural soil occurred at 12 feet below the cellar floor. The stratification is best illustrated by two sections across this cutting (Figs. 1 and 3, CD, EF). At the south-east end of the treneh black soil with oyster-shells covered the old surface. This layer, 12 inches deep, contained a coin of Vespasian and cannot therefore be earlier than A.D. 69-79. Above came the build-up of clean yellow loam already noted in 5 GRAVEL. FOUNDS BUILDING DEBRIS HERE f " <=Z= ^T^Z^Z y77777777777\ V, V/V////J////JV/////////;;// EXISTING SHOPS y////////\////////)y/l'////////A .///^a^^^A '/£//////< BUTCHERY LANE BUILDING I FEET f/S'W Sjy/X BUILPING II PIG. 4. Rlan to illustrate relation of Building I to excavations at 47 Burgate Street. CHAS: I FARTHING BELLABMINE o ^ - C H A L K cJ- -;CLAY.v CHARLES II FARTHING EARTH 6. RUBBL *». '£> • DIRTY CLAY & RUBBLE -^ BLACK STONY SOIL WITH SHELL . OYSTERS nORT. HARCOAL BLACK MEDIAEVAL". .CHALK' RUBBISH TESSELLATED PAVEMENl 'nTnYl1Yfffrrnt^:V:;::;;: fTUMBLED ROMAN lUM MASONRY il ^ J l l l l i i l i i i l l ^•b^•:'•:••^::•:•\•\:::::::::^^ UNEXCAVATED BUTCHERY LANE SECT/ON N-0 \MEDIAEVAL PIT (?) F I L L I N G ^ FEET 10 51 FIG. 5. CANTERBURY EXCAVATIONS, 1945 AND 1946 cutting AB. This was so firm and clean as almost to justify its acceptance as natural soil. It was, however, removed—with the reward of the Vespasianic level below. Over the sterile loam lay more black soil with oyster shells, a deposit 3 feet thick and dated by a practically complete Samian bowl of form 30 (Fig. 11) to c. A.D. 90. With this late first century date the coarse ware accords (Fig. 9, 9-12). Into the black soil had been sunk the foundations of two walls roughly aligned and running approximately parallel with the wall of the cellar. Various patches of mortar, presumably contemporary with the building, were found on top of the black soil just above the base course of the wall. Mortarium rims from these patches were of the late first-early second century type (Fig. 9, 13-14). Finds from the post-wall deposits set an upper limit to the probable period of building. Throughout the cutting 18 inches of brownishyellow loam overlay the black soil. A Trajanic semis in good condition indicates a date soon after A.D. 100 for this first post-wall layer. Above, at the south-east end of the cuttmg, a considerable fire had baked, cracked and discoloured the yellow soil and left a quantity of charcoal (cornus species (dogwood), pyrus sp. (apple, pear, rowan etc.) and quercus sp. (common oak)). From the top of the burned material came a coin of Hadrian, a rare type of sestertius, dated A.D. 134-8. Several patches of mortar against the wall-face at this height suggest that repairs were carried out soon after that date. BUILDING 2. The two walls already mentioned represent work of two periods. The first wall, in the north-west corner of the cutting (PI. I, 1, centre background) was of coursed flints with brick lacing courses over a gravel foundation. By undercutting the Butchery Lane revetting wall it was possible to uncover just the south-east angle of a room (Fig. 2, room 1) lying beneath the modern road. It was clear that the south wall of that room, in a reduced state, had been incorporated in the building represented by the second wall found in the cutting- At their junction there was a straight joint between the two walls to a height of 18 inches above the foundations. From that point the flints of the second wall rested on the reduced earlier masonry. They were much robbed at the angle but definite enough on the return wall to the west. It can be inferred that the brick laced wall was erected soon after A.D. 90 and had a short life. The second wall was of flints, random or roughly coursed, with mortar. Its foundations (3 feet deep) were of gravel, loose below but more compact, with some mortar admixed, towards the top. A base course of larger flints projected slightly over the foundations. This proved to be the north-east wall of a room (Fig. 2, room 2), of which the north wall, overlying the south wall of the earlier buildings, ran obliquely under Butchery Lane. 8 PLATE 1(1). BUILDING II. Room 2 north corner, with angle of Room I centre background. PLATE I (2). BUILDING II. Room 2, from the north. CANTERBURY EXCAVATIONS, 1945 AND 1946 Within the limits of cellar C two rooms could be identified. In the north-west corner of the cellar appeared the east end of room 3, a corridor 10 feet wide with a tessellated floor. Its south wall, of flints with brick lacing and, as far as could be seen in so small a space, of two periods, was found, but the presence in room 2 of a pavement which had to be preserved precluded search being made for the east wall of room 3. Its probable line was covered by a wall inserted at a later date, the north part of the north-west wall of room 2. This overlapped the edge of the pavement of room 3. The south part of the north-west wall of room 2 had a straight joint with the south wall of room 3 and ran parallel with the external wall of the house giving room 2 a width of 10 feet 6 inches. The pavement of room 2 had been laid rather more than 3 feet above what might be expected to be the original floor level of the room and was 6 inches higher than the pavement of room 3. Its preservation prevented investigation of its antecedents and no definite date can safely be suggested for this rehabilitation of the room. At this stage it had pink-plastered walls, vestiges of which remained standing (PL I, 1, right) and a quarter-round moulding of pink mortar. The tesserse were bedded in similar mortar. The pavement, in all 40 feet long,1 consisted of three mosaic panels at intervals along the long axis of the room and a margin of coarse grey tesserse. The panels, 5 feet by 4 feet, were executed in red, yellow, white and black. The north and south panels had centrally a conventionalized flower and leaf design, with a " wave " border above and below. The whole was framed with a three-strand guilloohe. The centre panel differed in that the flower motif was bolder, the leaves absent; a diamond border replaced the " waves " and the guilloche had two strands. Microscopic examination showed that the tesserse used in room 3 and for the margin of room 2 were of a sandstone derived from the Upper Greensand, probably from the Dorking—Godstone area. The red and yellow tesserse were cut from bricks or tiles. The fine-grained dolomitic limestone of the black tesserse belongs to the Carboniferous Limestone series; it may have been brought from the Mendips or possibly from Leicestershire. The pavement was in use long enough to warrant patching. A worn place on the north-west margin had been made good with part of a brick and a slab which may have been a re-used sillstone (PI. I, 2). Superficially the slab appeared to be of marble ; microscopically it proved to be a shelly limestone of which the source could not be determined. Two major disturbances of the pavement had occurred. A pit (PI. I, 2, foreground) had destroyed a corner of the north mosaic. Later 1 This measurement includes the area exposed later, south-east of cellar C. CANTERBURY EXCAVATIONS, 1945 AND 1946 than the debris covering the floor, the pit cannot date earlier than the fifth century. The second blemish was a cavity against the face of the south-west wall. This hole was stone-Hned except on the east side which was probably left as a soak-away. It contained sticky black soil with streaks of lime and was obviously the lowest part of an earthcloset dated by an eighteenth century token. The evidence for the end of this part of the house was clear. Over the pavements and the internal wall lay a mass of flints, pink plaster and mortar, evidently the ruins of the walls. Only the first foot of this debris remained, immediately under the cellar floor. It contained coins up to the late fourth century, including one of Valens (A.D. 364-78) in good condition and late pottery (Fig. 9, 16-17). Rather later the external wall was robbed to, or just below, pavement level. The relationship between the mortar debris and the fine black soil covering the denuded wall was obvious in the section cutting. None of the soil had drifted into the room ; the edge of the debris remained clean and regular. Worn late fourth century coins point to a fifth century (or later) date for the depredations. ABEA SOUTH-EAST OF CBLLAB 0. The evidence obtained in cellar C was confirmed and amplified in the trenches cut from modern surface levels south-east of it (see Fig. 5, section N-O). Above a 4-6 inch accumulation of dirt on the tessellated floor came the collapsed masonry of the walls, a thick layer of flints and mortar, containing pieces of window arched in tile. The first of these layers produced 2 antoniniani of Radiate Crown type and 2 of Tetricus I together with a barbarous radiate ; the rubble contained an antoninianus of Tetricus I. All these coins were circulating in the fourth century, but a clearer date for the collapse of the wall is given by the coin of Valens mentioned earlier. The pottery found on the floor is shown in Fig. 12, 20-3. The south-east wall of room 2 was only found in November 1947 during excavations connected with the basement which to-day houses the mosaics. It was of very doubtful and fragmentary character, and all that can be certainly stated is that the north-east wall continued beyond it towards the south-east. A second wall survived in better state running from this to the north-east; but their point of junction had been destroyed by a pit. After the collapse of the walls, the site seems to have lain desolate for almost a thousand years. The next deposit was a thick black layer of medieval rubbish containing many bones, oyster shells, and sherds, about 3 feet 6 inches thick. The date of this layer, which contained a halfpenny of Henry II, is indicated by the pottery, which is of early 10 CANTERBURY EXCAVATIONS, 1945 AND 1946 thirteenth century type (Fig. 17). The thick layer of domestic refuse was doubtless put down to level off and bury the ruins ; the exterior north wall was still standing well above the pavement at this time, as was shown by its robbing trench, traceable certainly to within 16 inches of the top, with indications that it had not in fact been dug until after the whole layer had been deposited. Almost at once rubbish pits were dug in this layer ; the section shows pit M4 as earlier and pit M3 as later than an occupation layer on top of the rubbish ; while the whole southern end including the pavement had become distorted through collapse into a very deep excavation below M3, which we were unable to excavate completely. The pottery from all these layers and pits is closely similar (Figs. 17, 18) and shows that the whole sequence occupies but a short time in the thirteenth and early fourteenth centuries. Indeed, the filling of many pits is very difficult to distinguish from the surrounding soil; sometimes its consistency differs ; sometimes as in M4 the decay of vegetable matter has resulted in subsidence and a hollow. The next period of activity is the second half of the seventeenth century, to which belong two unrelated footings seen in the section, and also the angle of a brick-built room west of the section, associated with a chalk floor and a brick platform for a fireplace along its southern side. There is a scatter of Charles I and Charles I I farthings—9 in number1 in these layers, but nothing later. A seventeenth century Bellarmine jug was found almost intact beside a brick and stone wall running at right angles from the brick-built room into the trench, and then turning south. At the angle occurred a large number of ox-bones, perhaps explaining the origin of the name Butchery Lane.2 There had been very little accumulation in modern times after the removal of the seventeenth century buildings. CELLARS D AND E. (a) Roman. Section GH (Figs. 6 and 7) gives an indication of the sequence of 1 Together with one of James I and 6 counters. 2 Mr. W. G. Urry tells us that the lane has had various names in times past. In the thirteenth century it was Sunwineslane (Charta Anliqua, Ch. Ch. Cant. no. C. 697, before 1207, and Rental 33, 1, c. 1232). About 1300 it was called Salcockeslane or Clementeslane (Register J., 231 ff). About 1640 William Somner in the Antiquities, p. 347 calls it Angell-lane. This name arose from the angel standing on the apex of the gable of the SW transept of the Cathedral, dominating the lane. Richard Culmer, the iconoclast, speaks in his Cathedral News from Canterbury (1644), 23, of a " statue of Michael the Archangel looking straight to a lane right over against it, in Canterburie, called Angell-lane." By 1768 the modern name had come for it is marked Butchery Lane in the Plan of Canterbury of that date by Andrews and Wren. The City Shambles lay in the road at the south end until 1740 (Gostling, Cant., ed 1, 1774, 31). 11 CANTERBURY EXCAVATIONS, 1945 AND 1946 layers found in cellar D below the Roman building. The most remarkable fact was that the natural soil was found at only 33 feet O.D., whereas below the outer wall in cellar C, only 18 feet farther east, it was 8 feet lower at 25 feet O.D. Somewhere below the pavement occurs a very rapid rise. This was verified in pit M7, which gave a section ; natural loam occurred at 34.17 ; at 30.26 occurred a change, greenish to ir z« b FEET ROMAN WALL,PERIODS m' PERIOD 2 R 1 .M1 &c ROMAN AND 1 ». 2. OVERLAP. MEDIAEVAL PITS £ \ Rf3 7 Dl r / >l VII / flft.C* M» J ££* ** L_ Hi / \ y "\ FIX l l \ \ \ -\ s- \HHl CLAUDIAN PMT-HOLES '//' '/, Cb c l . - l , 1 1 t i \ FIG. 15. SMALL OBJECTS (f). Nos. 1, 2, p. 35 ; 3, p. 31; 4, 5, p. 26 ; 6, p. 32 ; 7, Roman Bronze Pin from Trench V; 8, p. 40; 9, p. 41. Other Finds. 1. Bronze Brooch (Fig. 15,3). Collingwood, group P. Gamulodunum type XVIII B. Incomplete specimen: the bow is fluted, the central ridges being knurled, the outer plain; the foot flat and diamondshaped, and decorated with punch marks. Period : Claudius-Nero. 2. Glass handle, ribbed: deep amber ; from a jug with angular handle. M.J. type 58. Later first century A.D. 31 CANTERBURY EXCAVATIONS, 1945 AND 1946 (f) Layers sealed by the mortar spread of the courtyard. Samian. 1. Form 27, small fragment, trench 5b. 2. Form 29, small fragment with godroons ; possibly Claudian, trench 5a. 3. Form 37, ovolo, indeterminate, trench 8. Coarse Ware (Fig. 14). 45. Bowl with reeded rim ; granulated light grey ware. Also fragment of a second. These bowls are common at Canterbury ; they appear first in Claudian times at Camulodunum (form 246) and are common in Flavian times, Richborough I, 79, Collingwood 18-20. 46. Rim of brown ware with grey surface, of poppy-head beaker type. Early second century. 47. Rim, groved for lid, grey ware cf. Richborough IH, 245, dated A.D. 80-120. 48. Reeded rim in grey ware, cf. Camulodunum 250. Other Finds (Fig. 15). 6. Bone needle. (g) Wall trench of hypocaust room 3. Samian. 1. Form 30 or 37, small rim fragment. 2. Form 18, indeterminate. 3. Form 18/31, indeterminate. 4. Form 27, indeterminate. This piece was in the clay bedding for the gutter, the other three below, in the wall-trench. Coarse Ware (Fig. 14). 50. Rough-cast beaker, brown paste. This, with its small rim and low belly is not earher than early second century. Verulamium Fig. 27,9. 51. Dish-, light grey " porridgy " ware. Also body sherds of grey ware beaker with barbotine dots. Other Finds. Pieces of polished stone. See Appendix I, Nos. 2-7. B. DEPOSITS CONTEMPORARY WITH BUILDING. (a) Mortar spread (Courtyard floor). 49. Bowl, probably carinated, in granulated grey ware, of. Richborough I, 22, III, 215 : second half of first century. 32 CANTERBURY EXCAVATIONS, 1945 AND 1946 (b) Gravel layer (section GH) above grey loam. Samian (Fig. 13). 2-3. Form 67, two pieces; period, probably Nero, c. A.D. 60-70. (c) Pebbly clay above (b). Samian. 1. Ritterling type 1 ; probably Claudian (A.D. 50-60). 2. Form 18, indeterminate. Coarse Ware (Fig. 14). 52. Bowl with reeded rim in grey ware ; three different vessels. 53. Base of olla with lattice pattern. 54. Lid in grey ware. Glass. 1. Fragment of rim of large flask, diameter of rim 2\ inches, greenish, rim folded upwards and inwards. Form not in M.J.; second-third centuries. (d) Pit R\. Fourth century. Coarse Ware (Fig. 14). 55. Dense smooth whitish mortar. Late third or early fourth century. Cf. Richborough I, 97-102 for form only (early to mid fourth century) ; Oswald Margidunum in Ant. Journ., XXIV, Fig. 7, 61 (late third century). 56. Soapy black rim ; perhaps a late wide-necked flagon like Richborough I, 118 (fourth century), but perhaps a native version of a Claudian butt beaker. 57. Neck and rim of Castor beaker ; pink paste, red interior, dark slate-coloured glaze outside. The rim form is of late third century type cf. Verulamium Theatre, Arch., LXXXIV, 255, Fig. 10, 14. C. POST-DEMOLITION DEPOSIT. (a) Rubble filling of hypocaust, room 3. The layers were as follows (a) 3 inches of dark silt on the concrete floor, probably contemporary with its use ; (6) fallen building debris above (a) ; (c) thin deposit of earthy silt above (b), perhaps washed into the hollow by weathering. Coins. 1. Claudius H (posthumous) A.D. 270. Antoninianus: condition fair. M. and S. 261/2 (b). 2. ? Tetricus I A.D. 270-3. Antoninianus (a). 33 « CANTERBURY EXCAVATIONS, 1945 AND 1946 3. Tetricus H (A.D. 270-3). Antoninianus: condition good. M. and S., 270 (b). 4. Radiate Crown c. A.D. 270. Antoninianus. 5. Radiate Crown c. A.D. 270 or later. Antoninianus (small) (c). 6. 1 Barbarous Radiate. Late third century (b). 7. Barbarous Radiate. Late third century (b). 8. Carausius (A.D. 287-93). Antoninianus: condition good. M. and S. 101 (b). 9. Constantine I (A.D. 307-37). SM : mint condition. Cohen 487 A.D. 320-4 (b). 10. Constantine I, 3 M A.D. 330-5 : mint condition (c). J 5 9 \ 58 59 y 60 62 ( 6\ I 6 5 64 ; 67 68 7\ < . v? i 70 69 72 ) FIG. 16. COARSE AND OTHER POTTERY (i). 34 CANTERBURY EXCAVATIONS, 1945 AND 1946 11. Constans (A.D. 337-50). 3M A.D. 345-50 : condition good (b). 12. Constantius I I or Constans. 4 M size, A.D. 345-61 (b). (b) Also the following. 1. Postumus (A.D. 259-68). Antoninianus c. A.D. 260 : on tile gutter. 2. Tetricus I (A.D. 270-3). Antoninianus, condition fair. M. and S., 79/80. On pavement of corridor. 3. Tetricus I. Antoninianus, condition fair. M. and S. 100. On pavement of corridor. HI. WEST WING, CELLARS F and G (Section LM). A. PRE-BUILDING DEPOSITS. (a) Dirty loam on natural soil. Coin. 1. British tin coin (Fig. 15, 1). Derek Allen's class 2 1 : diameters 0 • 56 and 0 • 48 inches. Coarse Ware (Fig. 16). 61. Bead-rim bowl with burnished neck and brush-striated sides, in fight grey " porridgy" ware, black surface. Pre-Roman type, paralleled in the Belgic ditch (cellar L) on the south side of St. George's Street (to be published) cf. Gamulodunum 258, which began before the conquest, but reached its greatest popularity in period IV (A.D. 50-61). Other Objects. 1. Brooch—some fragments which Mr. Hull thinks may have been the shape of Collingwood's group A (Gollingwood, Fig. 60, 2). Possibly pre-Claudian. 2. Bronze finger-ring, decorated with three grooves on face (Fig. 15, 2). (b) Pebbly-ham above (a). Terra sigillata. 1. Arretine platter (form 18?) with stamp C. SENTI as at Haltern (Loeschcke, Haltern H, PI. XXIX, 203, 204, and p. 182). His stamps even occur on the Augustan site of Mount Beuvray : this piece may be only Tiberian (Fig. 13, 6). ' 2. Form 29 : too small for exact identification: probably Vespasianic (A.D. 70-80). Also 2 unidentifiable fragments. Trans, of International Numismatic Congress, 1936, 351-7. 35 CANTERBURY EXCAVATIONS, 1945 AND 1946 Coarse Ware (Fig. 16). 58. Large jar with corrugated neck and brush-striated shoulder. Typologically early Belgic form, cf. Wheathampstead No. 2. (Verulamium PI. XLIX, 2), Swarling No. 31 ; it does not occur at Prae Wood or at Colchester. 59. Bead-rim bowl; fight brown " soapy " ware with dark grey core. 60. Pedestal base, in " porridgy " grey ware. 62. Biconical bowl in buff-grey ware ; shoulder bears lightly tooled grooves to give corrugated effect. (c) Pit R8. Samian. 1. Form 30, fragment of rim, high glaze ; probably Claudian, A.D. 40-50. Other Pottery (Fig. 16). 63. Gallo-Belgic terra nigra platter. Camulodunum form 7b. 64. Platter in fine pohshed grey-black ware imitating terra nigra. Cf. Camulodunum 26-7. 65. Dish in coarse granulated buff-grey ware. 66. Lower part of girth beaker in pink ware with darker surface ; vertical fine combing. 67. Jar of more Romanized appearance ; dark grey smoothed surface ; paste light grey and " porridgy." (d) Post-Hole L. 68. Lid (?) in reddish sandy paste, surfaces pohshed dark grey; band of incised comb-rilling. (e) Clean loam above pit 228 below builders' debris east of wall. 71. Incense-cup : two fragments in red granulated ware ; burnt dark on inside. Not closely datable ; cf. late first century example Richborough I, 30. B. DEPOSITS CONTEMPORARY WITH BUILDING. (a) Glean clay build-up below mortar floor. Samian (Fig. 13). 4. Form 37 style of MEMOR of La Graufesenque. His ovolo as on 37 at Pompeii inscribed MIIMORIS (retro), with his Nile goose (Oswald 2244). Period : Vespasianic, A.D. 70-80. - 36 CANTERBURY EXCAVATIONS, 1945 AND 1946 Coarse ware (Fig. 16). 69. Flagon neck in soapy grey paste with orange surface. This seems to be not later than Claudian (cf. Gamulodunum Fig. 51, 3) and was doubtless incorporated when the clay was brought from elsewhere, together with a biconical bowl of Belgic make resembling Camulodunum form 229 B. ; ) s 2-SSL n 4 y> 7 J COOT reenter ewro 8 Moroo»(»«o«i>r m 9 10 I mii PIG. 17. COARSE MEDIEVAL POTTERY, EARLY 13TH CENT. (J) (b) Builders' debris east of wall. 70. Dish in grey ware ; early second century cf. Richborough I, 46, IH, 339 (both late first or early second centuries) and Newstead p. 258. (c) Black filling of gutter east of wall (late third century). Coins. 1. Salonina ? (A.D. 260-8). Antoninianus. 37 CANTERBURY EXCAVATIONS, 1945 AND 1946 2. Victorinus (A.D. 268-70). Antoninianus: condition good. M. and S. 118. 3. Claudius H (A.D. 268-70). Antoninianus: condition fair. M. and S. 81. 4. Tetricus I (A.D. 270-3). Antoninianus : condition good. M. and S. 100/2. 5. Tetricus I. Antoninianus : condition good. M. and S. 147. 6. Radiate crown A.D. 270. Antoninianus (small). 7. Radiate crown c. A.D. 270. Antoninianus (small): condition fair. 8. ? Radiate Crown c. A.D. 270. 9. Radiate Crown c. A.D. 270. Antoninianus (small), semibarbarous. 10. Barbarous Radiate Crown c. A.D. 270-90. Antoninianus: condition fair. 11. Illegible 4 M size. Third or fourth century. Probably diademed head. Coarse Ware (Fig. 16). 72. Jar rim, coarse grey ware. (d) Occupation of rooms. Coins. 1. Probably Radiate Crown (mid to late third century type) ; on tessellated pavement in cellar F. 2. Constantine I (A.D. 307-37). ConstantinopoUs type (A.D. 330-7), 3 M ; on upper tessellated pavement in cellar G. MEDIEVAL AND RECENT. A. LEVELS STRATIFIED ABOVE ROOM 1 (Section NO). (a) Black medieval rubbish above Roman debris. Early thirteenth century. Coin. Henry II. Cut Halfpenny (Canterbury ?). Mule : Tealby C/Short Cross la. c. A.D.1180. Pottery (Fig. 17). 1. Bowl, reddish brown hard granulated paste. 2. Bowl, dark grey ware ; simple almost Roman shape. 3. Coarse red ware, very shelly; dimpled pie-crust rim with pronounced Up. 4-6. Hard brick-red to brown granulated paste. 38 J ) FIG. 18. COARSE MEDIEVAL POTTERY FROM PITS (1). 39 CANTERBURY EXCAVATIONS, 1945 AND 1946 7. Narrow-necked jar with collar-rim, grey ware, decorated with bands of coarse rouletting. 8. Rim, grey ware, with applied band in hollow of neck. 9. Grey-black granulated ware. 10. Coarse light grey granulated cooking pot. 11. Reddish jug, with a little shell grit, unglazed. Other Objects (Fig. 15). 8. Bronze key with circular notched bow in axis at right angles to that of the bit, which is welded. A terminus post quern for this group is provided by the coin of Henry H, which shows appreciable wear, yet would have been pushed out of circulation by succeeding issues fairly soon. The deposit will not then be earher than c. 1200 ; on the other hand the pottery it contains is on the whole earher in type than the late thirteenth century wares represented in the pits (Fig. 18). In general these vessels are more curved, less angular or flattened in the neck and rim than their successors, a feature also seen in the Norman pottery from Caesar's Camp, Folkestone (Arch., XLVH, PL XX) and in the twelfth and early thirteenth century pottery from Rayleigh Castle, Essex (Trans. Essex Arch. Soc, XII, 182). The slight shoulder and tall straight rim of 10, again, is a type found stratigraphically earher than a late thirteenth century pit below the Rose Hotel (to be published) ; and the internal swelling of 8 and 9 is parallelled in twelfth century pottery from Alstoe Mount, Rutland (Antiq. Journ., XVI, 403, Nos. 11, 13). Rouletting, too, as on 7, with square or diamond-shaped notches, is a feature inherited from late Saxon times (Oxoniensia, V, 46), and it is found on Norman wares of the twelfth century (Old Sarum, Antiq. Journ., XV, 185). For an example more nearly contemporary with ours cf. White Castle, Monmouthshire (Ibid. XV, 330-4 and Fig. 3, 14). (b) Medieval rubbish pits later than (a) (Fig. 18): late thirteenth century. Pit i¥4. 12. Spouted flagon, in hard grey ware with light golden brown surface and glaze, traces of white slip in places below it. Surface combed horizontally; below the spout (missing) an arcading of low raised ribs. 13. Reddish hard granulated ware. 14. Coarse shell-gritted ware. 15. Cooking pot, ware similar to 2. 16. Similar. 17. Jug, hard grey ware with patches of olive-green glaze. 40 CANTERBURY EXCAVATIONS, 1945 AND 1946 18. Reddish jug with brown glaze. 19. Similar. Pit Ml. 20. Conical cresset-lamp for insertion in a bracket. Pit M6. 21. Cooking pot, roughly shaped, hard brown sandy ware. 22. Cooking pot, grey ware. Pit jura. 23. Hard grey granulated jar, hardly any shell grit; internal bevel. Two examples. 24. Large dish, diameter 14 inches, grey ware with copious shell grit, reddish brown surfaces. 25. Bowl, hard sandy ware, buff brown, with a little shell grit. 26. Very large dish, diameter 2 feet, ware similar to 13 but containing much more shell grit. 13 and 14 both have a row of small pinincisions round the flat top of the rim, c. -| inch apart. 27. Oblique sided dish, brown sandy ware. 28. Dish or lid (?) similar ware. Also present, a number of pieces of flat heavily-fired floor tile, c. -| inch thick, sometimes thickly greenglazed. In this group, while there are obvious similarities with the earliest group e.g. Nos. 21 and 22, on the whole the rims are flatter and the shoulders more pronounced or angular. Glaze, too, is more common. This group bears comparison with the late thirteenth century pottery from Watling Street (Arch. Cant., LX, 99). NOTE.—There was also found, unstratified, a glazed stamped tile of the pattern figured as No. 45 of the London Museum Medieval Catalogue list (p. 245). This type is already recorded from Canterbury Cathedral. (c) Seventeenth century layer. Fig. 15, 9. Bronze face of small portable sun-dial. This was found in the layer at the north end of section NO, 16-24 inches below the surface. The gnomon, which is missing, was apparently held in a small socket rivetted to the central arm. APPENDIX I PETROGRAPHIOAL REPORT By K. C. Dunham, D.Sc, S.D., of the Geological Survey and Museum 1. Pohshed slab, 0 • 55 cm. thick, with one obtuse angle of the original outting (102°) remaining. Cellar D, Pit Ml. A red hornblende porphyrite composed of white phenocrysts of oligoclase-andesine up to 3 mm. long, highly sericitized, and phenocrysts of oxyhornblende up to 41 CANTERBURY EXCAVATIONS, 1945 AND 1946 1-25 mm. long in a fine-grained groundmass containing quartz and feldspar with abundant disseminated cryptocrystalline ?hsematite, imparting a red colour to the rock. The oxyhornblende is pleochroic pale yellowish-brown to deep brown, and many crystals show dark margins ; it is optically negative with high optic axial angle and low extinction angle. A little calcite is present in the groundmass. It is possible that some of the phenocrysts may have been orthoclase rather than plagioclase, but the extensive sericitization makes positive determination impossible. This may be identified with confidence as the rock known as Porfido Rosso Antico, obtained from quarries at Jebel Dokhan in Egypt. The specimen submitted is identical with our specimen from that locality (F.3136) and the rock-type is an unusual one. 2. Floor-slab, one surface only ground, 2 • 2 cm. thick, other dimensions at least 15 x 14 cm. Coarsely crystalline shelly limestone, containing abundant remains of the gasteropod genus Viviparus (Pdludina). The shells are recrystallized, and partly filled with calcite mud. They are cemented by a matrix of coarsely crystalline clear calcite, with crystals up to 0*5 mm. 3. Floor-slab, similar to 2, shelly limestone rich in remains of Viviparus. 4. Slab ground both sides, 2 • 3 cm. thick, other dimensions at least 7x5*5 cm. A coarse shelly limestone rich in Viviparus remains and containing patches of yellowish calcite mudstone, possibly with some dolomite, enclosed in a matrix of very coarsely crystalline calcite. Some of the sheUs are partly replaced by hmonite, possibly after pyrite. Fragments of crystalline phosphate and rare quartz grains are present. These shelly limestones rich in remains of Viviparus (Nos. 2, 3, 4) are characteristic of certain beds in the Purbeck (upper Jurassic) and Wealden (lower Cretaceous) series. The former have been extensively worked for marble (" Purbeck Marble ") in the neighbourhood of Swanage, Dorset, while the latter has been worked under the name " Sussex Marble " at Gorlinger, Petworth, Bethersden and Coolham, Sussex. The species of Viviparus present in the Wealden is somewhat larger than in the Purbeck, and my coUeague Mr. Edmunds is of the opinion that the specimens submitted are more likely to have come from the Purbeck than the Wealden. In this case, they may have been derived from thin beds which occur in the vicinity of Battle in association with iron ores probably known to the Romans, and need not necessarily have been brought from Swanage. 5. Small fragment, 3 • 3 cm. thick, ground on one side and partly on the other. A glauconitic sandy limestone : composed of quartz grains averaging 0-2 mm. but reaching a maximum length of 1-5 mm., and abundant green glauconite pellets averaging 0-15 mm. diameter in a 42 CANTERBURY EXCAVATIONS, 1945 AND 1946 calcite mosaic matrix. A little pyrite, partly altered to hmonite, is present. 6. Paving slab, at least 14 x 12 • 5 cms. A fine-grained limey glauconitic sandstone : containing highly angular quartz grains averaging 0-05 mm., with some brownish-green glauconite, cemented by calcite. Minor constituents include brown staurolite, zircon, oligoclase, ilmenite, magnetite, leucoxene. The glauconitic hmestone and limey sandstone (Nos. 5, 6) may be assigned with confidence to the " Kentish Rag " facies of the Lower Greensand. This facies extends from the neighbourhood of Maidstone to beyond Ashford, Kent, and is extensively quarried at Maidstone, East Mailing, Otham, Larkfield and Chilmington. NOTE.—Nos. 2-6 were found in the foundation trench of Room 4. With them was (7) another piece of shelly hmestone in the shape of an irregular rhomboidal quadrilateral, 8-8x5-6 cm., and 1-3 cm. thick. Its edges were slightly bevelled to give a good joint; its face is pohshed, the other sides roughly ground. This is clearly opus sectile, as also is probably No. 1. Similar fragments of Wealden Marble, cut for an opus sectile floor, were found at Angmering, and a discussion will be found in Sussex Arch. Colls., LXXIX, 15-19. APPENDIX II. SKELETON OP INFANT FROM HYPOOAUST-FILLING OF ROOM 4. By I. W. Cornwall, B.A. The remains were extraordinarily complete, for the most part, even paper-thin bones, such as the inter-alveolar septa of the jaws being preserved. A few parts were missing altogether : the nasal, palatine, ethmoid and lachrymal bones of the skuU, the left tibia, both fibulae, all the bones of hands and feet and several vertebrae. The skull was, with some difficulty, approximately reconstructed. Though the still-separate bones of the base, and some of the face, showed good contacts, those of the vault were evidently still widely separated during life and considerable fragments of them were missing. As they were, moreover, mostly no more than 1 mm. thick, large errors were inevitable in making edge-to-edge joins during reconstruction. The final result, therefore, was not exactly symmetrical, though the errors were distributed as widely as possible to enable the general appearance of the skull to be reproduced fairly accurately. Even the posterior and antero-lateral fontanelles of the braincase 43 CANTERBURY EXCAVATIONS, 1946 AND 1946 were wide open. These are usuaUy filled in by growth of the surrounding bones within 2-3 months after birth. Only one tooth was found, the germ of a lower medial incisor, the first tooth to cut the gum, generaUy between 6 and 9 months after birth. The root was quite unformed and the tooth was evidently in life, stiU buried in its alveolus in the mandible. This indicates an age probably less than six weeks. On one side the tympanic ring of the temporal bone was still separate from the squamous part and very weU preserved. On the other side the corresponding part was already fused with the rest of the bone. This union generaUy takes place shortly before birth, so that its incompleteness in this instance indicates at most a newlyborn infant and possibly even a fuU-term foetus. BIBLIOGRAPHY. Arch. Archceologia (Society of Antiquaries). Camulodunum Hawkes and HuU, Society of Antiquaries Research Report XIV. Cohen Cohen, Description historique des monaies frappdes sous Vempire romain. CoUingwood CoUingwood, The Archceohgy of Roman Britain, 1930. J.R.S. The Journal of Roman Studies (Society for the Promotion of Roman Studies). Knorr, T.S. Knorr, Topfer und Frabiken verzierter Terra-Sigillata des 1919 ersten Jahrhunderts, 1919. M. and S. Mattingly and Sydenham, Roman Imperial Coinage. M.J. Morin-Jean, La Verrerie en Gaul sous Vempire romain, 1913. Newstead Curie, A Roman Frontier Post and its People, 1911. Ospringe Whiting, Hawley and May, Society of Antiquaries Research Report VHI. Oswald Oswald, Index . of Figure Types on Terra Sigillata, (Supplement to Liverpool Annals of Archaeology and Anthropology). O. and P. Oswald and Pryce, An Introduction to the Study of Terra Sigillata. Richborough Bushe-Fox, Society of Antiquaries Research Reports / , II, III VI, VII, X. Swarling Bushe-Fox, Society of Antiquaries Research Report V. Verulamium Wheeler, Society of Antiquaries Research Report XI. Wroxeter Bushe-Fox, Society of Antiquaries Research Report I. 44 CANTERBURY EXCAVATIONS, 1945 AND 1946 ACKNOWLEDGEMENTS. The writers desire to thank their many friends who have assisted the work. Mr. A. Martyr Smith made the prehminary surveys of the sites, and Mr. J. Mann took the levels. The Samian ware has been reported upon by Dr. Felix Oswald, the glass by Mr. D. B. Harden, and the coins by Mr. B. H. St. J . O'Neil. Mr. Edmunds and Dr. Dunham examined geological material, the Rev. Dr. S. G. Brade-Birks a soilsample. Mr. W. F. Grimes drew Fig. 11, and Mr. Cornwall reported upon the human remains. The thanks of the Canterbury Excavation Committee are due to the foUowing for permission to dig on their property : the Directors of Court Brothers (Electrical Contractors) Ltd., Messrs. George, Beer & Rigden and Taylor Brothers. Thanks are also due to F. W. Finnis and Sons for the use of their bakery ceUar as storing-place for tools and Roman pottery, and for many kindnesses. F.W.T. 45