Canterbury Excavations in 1945

( 68 ) CANTERBURY EXCAVATIONS LN 1945. BY AUDBEY WILLIAMS, E.S.A. I. A SITE IN BUBGATE STBEBT. IN 1868 a mosaic (now in the Royal Museum, Canterbury) was found at a depth of eight feet below the carriage-way of Burgate Street, about eighteen yards west of Iron Bar Lane. In 1871 more of the ^ _ B.M.47-2 *o APPROXIMATE N SITES OF EARLIER *r EXCAVATIONS S J> *e * *r s EXCAVATED AREAS BLACK GROUND FLOOR STIPPLED FEET 100 200 - I 1 1 PIG. 1. BUEGATE STBEET SITE. buUding to which this pavement belonged was discovered beneath the adjacent south pavement, and in a yard adjoining, where there appeared " a waU of sohd concrete " and a second mosaic. Other walls and floors were said to he nearby.1 With these facts in mind the Canterbury Excavation Committee selected as its second site for excavation three ceUars on the south side of Burgate Street with a frontage twenty to forty-five yards west of Iron Bar Lane (Fig. 1, 1-3)- 1 V.O.H., Kent, p. 72, tabulating Phlbrow's notes. CANTERBURY EXCAVATIONS JANUARY AND APRIL 1945 MODERN WALL BURGATE STREET .37-30 PARTLY EXCAVATED I FLOOR, (PERIOD II) PH5 8,6 y-2'10") X'38-42 ~ PH7(+2'5") CELLAR NATURAL) CELLAR MEDIEVAL PIT CELLAR 44-95 ^ S ^ 44-76 3 FEET ROMAN W> FOUNDATIONS m WALL 1 '//, FOUNDATIONS] Hill FOUNDATIONS PERIOD H P i - 4 PITS PH1-7 POSTHOLES PERIOD I •PERIOD I (+11") HEIGHT OF THE TOP OF ETC A FEATURE ABOVE NATURAL SOIL. OTHER MEDIEVAL FEATURES RUN UP TO SOS WALL THE CELLAR FLOORS _ 38.96 LEVELS IN FEET ABOVE ETC O.D. Eio. 2. BURGATE STREET SITE : PLAN. CANTEBBUBY EXCAVATIONS IN 1945 69 Here part of a buUding of some size was uncovered and dating evidence, tantahziagly absent from earher records of Roman Canterbury, secured from weh-stratified deposits. Although the results are here coordinated the actual work was carried out in two stages. In January 1945 two exploratory trenches were dug, one roughly north-south in ceUar 1 (Fig. 2, AA') and the second east-west in ceUar 2 (Fig. 2, DD'). Concrete foundations of Roman date having appeared at the north and south ends of the first cutting, April 1945 was devoted to foUowing up this clue. The inner face of the concrete, indicating the north, west, and south walls of a buUding, was traced in ceUar 1. The north-west outer face was found in ceUar 2 and the north-east external angle in ceUar 3 (Fig. 2, CC). The total width of the foundation was accessible in an entrance through the exceptionaUy wide party-waU between ceUars 1 and 3 and the outer face of the concrete could be foUowed from this point almost to its return to the south. A cutting was taken at the south-west corner of ceUar 1 through the ground floor area to the west (Fig. 2, EE'). The later Roman levels, the primary objective of this section, had been destroyed and the whole of the make-up was of medieval date. It was, however, gratifying to find that part of the waU buUt on the concrete foundation had here survived, incorporated in a later party waU! Further search for Roman levels later than those encountered in the ceUar was made south of ceUar 1. It was hoped to get a more or less continuous section across this open ground floor level but the two trenches dug showed intensive use of the area in comparatively recent times. In each case the cutting was taken out to slightly below the level of the ceUar floor but vaults, later in-fiUed and bricked up, cess-pits and rubbish pits had destroyed the deshed evidence. A smaU trial trench in a ceUar on the north side of Burgate Street (Fig. 1, 4) showed stratification simUar to that of cutting A in ceUar 1 but was featureless. SECTION A (FIG. 3). Virgin soU, a yeUow clayey loam, occurred at 6-8 feet below the modern floor, at about 30 feet above O.D. The fact that natural ground here was low compared with other parts of the Roman town1 explains the necessity for the succession of build-up deposits found throughout the site. They were evidently intended to combat marshy conditions. The various layers are referred to below by the numbers given them in the drawn section, Fig. 3, AA'. The earhest feature was a small guUy dug in natural soil (PI. I, 1). I t had short sloping sides and a wide flat bottom. Its width was 1 44 feet O.D. in tbe Burgate Lane area. Arch. Oant., Vol. LIX, p. 69 and Boman Canterbury, No. 2, p. 7. 70 CANTEEBUEY EXCAVATIONS IN 1945 2 feet 6 inches and its normal depth 7 inches, though to the west (in cutting C) it deepened to 10 inches to give a shght faU to a sump 20 inches deep. Its further course, if any, lay outside the excavated area. The guUy probably served as a drain for surface water. Evidently it had been in use when the first deposit of heavy grey loam (1) was laid down. In section it showed no sUting ; its filling consisted entirely of the grey loam. The two sherds found in it, a scrap of Belgic combed ware and a fragment of wheel-made fine black ware, belong to the grey deposit and not to what must have been part of the earhest occupation of this area. The grey loam, a foot (more or less) thick, contained Claudian Samian and early types of coarse ware. Above came an occupation level (2). Black soil, rich in potsherds, broken animal bones, oyster sheUs, crushed charcoal and scraps of daub, ran the whole length of the cutting and was found again to the west in cutting C. Associated with this occupation in cutting A were three stake-holes, aU dug in the grey loam. FUled with black soil they were uniformly chcular, 5 inches across, 7 inches deep and bluntlypointed in section. One of them, as PI. I, 1 shows, had flints ringed round its mouth. The holes must have held the timber supports of a wattle-and-daub structure, either the waU of a hut or of a fence. The level is dated to the middle of the first century by Claudian Samian and a coin of Claudius I of A.D. 41 lost whUe in good condition. Part of this occupation layer was sealed by a deposit of yeUow loam (3), clean enough to have been taken from virgin soil elsewhere, and, in fact, at first regarded as undisturbed. It did not, however, run the whole length of the cutting and appeared to have been laid down to level up the subsidence in the preceding deposit over the guUy area. It produced only three fragments of pottery, aU Samian of Claudian date. The next deposit, of gravel (4), was thinly spread over the yeUow loam, more thickly (but stiU only 8 inches thick) over the unsealed occupation layer to the south. It produced no datable material. Above came more yeUow loam (5). This contained lumps of flint and chalk and, for about half the length of the cutting, a thin, grey, somewhat sticky band taken to represent decayed turves.1 The pottery from this second yeUow loam included Flavian Samian. Succeeding the late first century deposit was another thin layer of clean gravel (6), which could not be dated from a solitary smaU pot. In cutting 0, however, the same gravel produced a sherd of Trajanic Samian as did black soil at the same level at the west end of that trench. It foUows that the next deposit (7) in cutting A, though it contained Flavian Samian and no closely datable coarse ware, cannot be earher 1 Similar bands frequently occur where turves have been used in barrowbuilding. HEARTH MODERN BUILD-UP / 7 / S ^ E ^ S ; TTTZMZZZZZZZZte.- MID -2ND CENTURY v x n CULLY ^(PROJECTED) PH3 N I X v ^ - • - . s x ^ ^ MODERN BUILD-UP TTTHXI BRICKS ft PIT 2 ^ CENTURY-J MODERN BUILD-UP LATE 2ND CENTURY. X-WALL MEDIEVAL PIT MODERN DIEVAL FACING EPOS1TS MODERN TRENCH MODERN BUILD-UPj, M0DERN TRENCH '7S^U^4£ZL: K ^ ^ ^ S ^ N S (PROJECTED) PH 4 W I I I I I MODERN BUILD-UP MODERN TRENCHES I FOUNDATIONS = ^PERIOD I '//, PERIOD H \\GREY LOAM ^ GREY LOAM WTH FLINTS :•:•:•:•: BLACK SOIL «K BLACK SOIL WITH SHELLS <% RED SOIL //YELLOW LOAM :••« GRAVEL 1.1 ! MIXED SOILS T~TNATURAL LOAM «o H FIG. 3. BURGATE STREET SITE : SECTIONS. CANTEEBTJBY EXCAVATIONS IN 1945 71 than the early 2nd century. This black sheUy soU ran for most part thinly over the gravel, but at the north end of the cutting it had patches of yeUow loam either beneath or mixed with it. In the deposit (8) immediately beneath the make-up for the ceUar floor various materials mingled; yeUow and grey loam, reddened soil, black soU and, to the south, a few bricks formed a confused mass. Towards the north end of the trench a hoUowed area had been used as a hearth, now visible as red soU, cracked stones and a quantity of black ash. Nothing datable came from the hearth, but it was obviously later, however shghtly, than layer 8 which was dated by pottery and, in cutting C, by a coin to the second half of the 2nd century. From layer 8 came part of a mortarium stamped by Juvenahs. The only previous record of this stamp is from Boulogne. The foundations encountered in this and other sections are discussed later. SECTION B (FIG. 3). The south half of the west face of cutting A gave a section through two later pits, one of whioh straddled the trench. On plan Pit 1 was roughly chcular, just over 8 feet across. In profile it was bag-shaped and 13 feet deep, running from the top of layer 8 down into natural soil. It had been fiUed with heavy grey soU and large nodules of flint. From it came a coin of A.D. 141 and Hadrian-Antonine pottery. The pit must have been cut and fiUed in quite soon after layer 8 had been deposited in the middle-late 2nd century. In the 4th century a curious funnel-shaped pit (Pit 2) was dug in the filling of Pit 1. Wide at the mouth (4 feet 6 inches by 5 feet) it tapered to become a square shaft, 2 feet 2 inches across by 6 feet 3 inches deep out of the pit's total depth of 8 feet 6 inches. Its sticky black filling suggested a predominance of decayed vegetable matter. SECTION C (FIG. 3). The stratification in cutting C (ceUar 2) equated fahly closely with that of cutting A, as the numbering of the deposits in Fig. 3 shows. As might be expected materials varied from cutting to cutting; here less gravel and more occupation soil were encountered. Within the narrow hmits of the excavation it was difficult to differentiate between occupation soil in situ and that brought from elsewhere as make-up. Possibly some of the black soil layers may be associated with postholes just outside the cutting, or form part of floors of which only a narrow strip has been seen. At a shghtly lower level than the stake-holes of cutting A, but belonging to the same Claudian occupation, cutting C had a large posthole surrounded by a patch of black soil (2A). The hole penetrated the first grey loam deposit (1) to the underlying natural soil. It was 72 CANTEBBTJEY EXCAVATIONS IN 1945 16 inches in diameter and 18 inches deep. Its post, a substantial one, seems to have been withdrawn before the floor was covered with further grey loam, for in the filling loam occurred mixed with and beneath black soU. Over the grey layer, at the east end of the cutting, lay black soU which was evidently the occupation level (2) of cutting A petering out. In deposits 4 and 5 Flavian material began to appear. An interesting item from the yeUow loam (5) was a mortarium, practicaUy complete, stamped by Orbissa, a Flavian potter whose work was hitherto known in this country only from London. It may be that the deposits numbered 4-7 were laid down in quick succession in that three joining pieces of a Samian bowl by the Vespasianic potter Vitahs came from the black bands 4, 6A and 7. On the other hand, since nothing can be known of the history of these fragments theh presence in these layers may be mere coincidence. More important is the fact, aheady mentioned, that layer 6 contained an early 2nd century element; Trajanic Samian occurred in both the gravel (6) and the black soU (6A) to the west. Associated with the black soU of layer 7 was a mass of red soil, indicative of fires; it extended across the cutting, was 6 feet long and 6 to 8 inches deep at its maximum. The top layer (8) produced Hadrian-Antonine pottery, and a sestertius of Antoninus Pius, A.D. 145-61, in good condition. At the east end of the cutting were two rubbish pits, both obviously later than the mid-2nd century layer. It was evident that pit 3 also post-dated the concrete foundation. On plan it remained a half-circle, an obviously incomplete form, against the outer face of the concrete. In section the chronological relationship between the features was clear. The pit ran deeper (4 feet 4 inches below the modern make-up) than the concrete (2 feet 6 inches deep) and had it been the earher its black tilling must have appeared beneath the concrete. In fact, the concrete rested on normal loam, and the waU of the pit continued the vertical line of the face of the foundation. The material from the pit is of 3rd to mid-4th century date, by which time the waU of the buUding was presumably obhterated at this point and the pit-diggers started work at a higher level, ignorant of the foundation below. Pit 4 dated to the late 2nd-3rd century and was probably, as wUl be seen, contemporary with the buUding. Bag-shaped in profile it had the same black filling as pit 3. It was just over 9 feet deep, running down through the grey loam and destroying the sump of the guUy dug in the undisturbed soil. SECTION D (FIG. 3). In ceUar 3 the stratification was rather simpler than elsewhere but the dating material corroborated the evidence from the previous cuttings. Undisturbed soU lay 6 feet (more or less) below the ceUar \ .. -^ PLATE I (1). BURGATE STREET SITE A gully in natural soil (marked by fork) Bad a century post-hole (left centre). raid- 1st PLATE I (2). BURGATE STREET SITE. Clasping buttress at north-west corner of building. PLATE II (1). BURGATE STREET SITE. Late 2nd century wall in cutting E, north face. PLATE II (2). BURGATE STREET SITE. Late 2nd century wall in cutting E, south face. CANTEBBUBY EXCAVATIONS IN 1945 73 floor. Over it came a thick band of yeUow-brown loam (1). Above this the west end of the cutting had been disturbed by pipe-laying trenches, one along the cutting, the other across it. To the east the sequence was black sheUy soU (2), grey loam (3), more occupation soil (4) and a top layer of mixed soils (5) familiar as layer 8 of cuttings A and C. The earliest three layers, to a height of just under 3 feet, can be taken as Claudian or rather later. Layer 4 had Flavian material with Claudian survivals and the final " mixed " deposit was Antonine. In an extension south from the west end of the cutting was found a group of three post-holes, aU close together but one at a shghtly lower level than the other two. AU were dug in the pre-Flavian loam and seemed to be associated with the Flavian occupation soU (4) but it was not possible to decide whether they belonged to a definite timber structure. One post-hole was rectangular, 6 inches by 4 inches by 12 inches deep. The other two were round, 5 to 6 inches across and 10 inches deep. They contained scraps of carbonized wood and a couple of Flavian potsherds. THE BITILDING. The concrete foundations (Fig. 2, hatched) enclosed an area 35 feet north-south by 22 feet east-west. The top of the concrete lay normahy about 12 inches below the ceUar floor; in some places, for instance at the south end of cutting A, it had been cut away and a greater depth of modern make-up covered it. The foundations were just over 3 feet wide and had a depth of 3 feet or more. At the north corners, where clasping buttresses (PI. I, 2) occurred, they ran down to 6 feet. Of the south corners only that on the west could be investigated. There, by undercutting the modern wall and also the ground floor area south of cutting B, it was found that the foundation of a waU running away to the south joined the outer face of the south foundation of the excavated buUding. This second foundation was of concrete, 2 feet 6 inches wide, its outer face projecting shghtly beyond that of the waU it joined. A gap in the main foundations suggested a doorway in the south waU. This may have given into the room to the south. It may, however, have been an external opening and the south room may he west of its one known waU. In either case this second foundation showed that the excavated building was but part of a larger structure. On the south-west (in cutting E, described below) the outer face of the waU itself survived above the concrete (PI. II, 1 and 2). Its inner face had been cut back and faced with cement to form the existing cellar wall. The outer face was uncovered for a few feet either side of the section cutting. It may be assumed that a length of it was preserved beneath the ground floor area south of ceUar 2. The waU and the foundations were examined with great care for signs of later, possibly medieval, waU-building on the earher foundation but there 74 CANTEBBDBY EXCAVATIONS IN 1945 was not the slightest line of demarcation ; the base course was firmly bedded in the concrete and the whole structure was obviously of one buUd. The waU face, of roughly coursed dressed flints, stood almost 5 feet high above a shghtly projecting base course. Patches of adhering plaster indicated that the building was originaUy plastered externaUy. There is sufficient evidence for assigning the buUding to the late 2nd-3rd century. It must be later than the mid-2nd century deposit into which its foundations were sunk. Also its internal partition (see below) crossed the filling of a mid-2nd century pit (Pit 1). On the other hand it antedated, as has been seen, a 4th century pit (Pit 3) against the north-west buttress. In view of the general insecurity of the 3rd century, the late 2nd century would perhaps be the more hkely date for an urban buUding of some size. In the absence of later Roman levels no estimate as to the length of the buUding's life can be offered. Turning to the interior of the buUding—clearance of the area immediately inside the west foundation, revealed the foundations of three cross waUs (PI. Ill, 1). (i) The most northerly foundation was at the same level and simUar to a slight (1 foot 6-9 inches wide and 9 inches high) foundation found crossing cutting A. Possibly they were originaUy linked by a return waU lying beneath the unexcavated balk. The concrete of this cross foundation was more pebbly and darker in colour than that found elsewhere. Although now vestigial against the main foundation the cross waU was certainly the earher feature. Its foundation was sunk into a late lst-early 2nd century deposit (7) and the waU itself had been razed before the overlying mid-2nd century material (8) was laid down. Whatever buUding it represented had been obhterated before the main buUding was erected. (ii) The south foundation, flints set in clay, was of later date than the main foundation which had been cut into to accommodate its end. It was 1 foot 10 inches wide and 2 feet deep, running right up to the modern make-up. Little of it survived, but it may have been a shghtly later substitute for foundation ih. (iii) This was contemporary with the main foundation into which it was bonded. Its waU apparently served as an internal partition; it was just over 2 feet 6 inches wide and ran down from the modern make-up to a depth of 2 feet. In the north-west corner lay a rectangular patch of concrete (PI. Ill, 2) similar to, and on the north tight against, that of the waU foundations. It was 5 feet long (north-south) by 3 feet 6 inches wide. Its upper surface lay 6 inches below the ceUar floor and 3 inches below ftysvjy.'j PLATE III (1). BURGATE STREET SITE. The inner face of the foundations of the west wall of the building, with foundations of three cross walls. .irJwm*aTi *-*"'' / h M MHHHHBHIHHMHHHIHHHnnHSBBBHHK:-'': V -\* T: PLATE III (2). BURGATE STREET SITE. Concrete floor, with medieval wall over, in north-west corner of building. PLATE IV. BURGATE STREET SITE. Decorated Samian ware. CANTEBBUBY EXCAVATIONS IN 1945 75 the top of the waU foundations. It appeared to be the foundation of the floor of the buUding. It was mutilated on the east by a trench for modern pipe-laying and had also been cut back on the south. Over it ran the base course of a narrow waU, of medieval date in that two scraps of glazed ware were incorporated in its mortar. SECTION E (FIG. 3). That part of cutting E which lay outside the face of the Roman waU (Fig. 2) faUed to reward the search for later Roman levels than could be expected to survive in the ceUars. The make-up from the base of the outer face of the Roman waU to within 2 feet of the existing ground floor was of medieval date. There were six distinct layers ; yeUow clay occurred three times, black sheUy soU twice, with grey clay as the final deposit. These layers seemed to be banked up against the Roman waU to counteract its tendency to coUapse outwards. Its threatening bulge was more pronounced south of cutting E. A smaU quantity of pottery from these tips indicates a late 13th century date for the shoring-up. It was noted that the lines of the lower tips were stiU sagging fahly strongly at the bottom of the excavated trench, over a foot below the base of the waU, and an underlying earher pit may be suspected. Further west there was a later medieval pit starting immediately beneath the modern make-up and fiUed almost enthely with flints and large lumps of chalk. This could not be dated ; its most determinate sherd was of 3rd century Samian ware. SUMMABY. To sum up : the site bespeaks continuous activity from the middle of the 1st century to the end of the 2nd ; with sporadic use of the area as a rubbish dump in the 3rd-4th centuries. The significance of the guUy dug in the undisturbed surface cannot be stressed since that level could not be sufficiently explored for post-holes and habitation soil. It seems however, to hint at pre-Roman occupation. There were signs of more than one timber structure in the 1st century. The first stone building, of which only a sbght trace survived, was put up in the early-middle 2nd century and was apparently fahly quickly demohshed. The late 2nd century saw the erection of what may have been some kind of public building of which part only has been excavated. It was however, clear that its northern hmit lay beneath what is now the south edge of Burgate Street, that its one room, with internal partition and south doorway, occupied roughly the area now enclosed in ceUar 1, and that a second room lay to the south. The building evidently stood free of its neighbours ; it was unconnected" with PiUbrow's buUding on the east and had an open space to the west. 76 CANTEBBUBY EXCAVATIONS IN 1945 THE FINDS. A. ROMAN. As has been shown, the deposits, despite shght variations from section to section, faU into three successive chronological groups. It seems advisable to catalogue the finds according to this sequence rather than typologicaUy. They are, therefore, grouped as mid-lst century, late lst-2nd century and mid-2nd century. The find-spots are indicated by the letters and numbers given to the cuttings and deposits in the section drawings (Fig. 3). The material from the pits is hsted separately. The foUowing abbreviations are used : Behn Romische Keramik mit Einschluss der hellenistichen Vorstufen 1910. Birdoswald Cumberland and Westmoreland Antiquarian and Archceological Society. Transactions XXX—New Series. C.I.L. Corpus Inscriptionum Latinarum. Haltern Loeschcke, Mitteilungen der Altertums-Kommission fur WestfaUn V. Hofheim Ritterling, Nassauische Annalen XL. Knorr 1919 Knorr, Tbpfer und Frabrilcen verzierter Terra-Sigillata des ersten Jahrhunderts, 1919. M. 'and S. Mattingley and Sydenham, Roman Imperial Coinage. 0. Oswald, Index of Figure Types on Terra Sigillata. 0. and P. Oswald and Pryce, An Introduction to the Study of Terra Sigillata. Richborough Bushe-Fox, Society of Antiquaries Research Reports I, II, HI VI, VII and X. Verukmium Wheeler, Society of Antiquaries Research Report XI. Wroxeter Bushe-Fox, Society of Antiquaries Research Report I. 1. FROM THE MID-1ST CENTUBY DEPOSITS. (a) Coin. Claudius I. As of A.D. 41; worn condition. M. and S.66. A2. (b) Samian. (i) Decorated. 1. Form 29 ; lower frieze. The leaf-wreath has seven bifurcated stalks as used by COTOVS. Cf. 29, OF COTOI, Worms (Knorr 1919, 85C).Claudian. Al. 2. Form 20; lower frieze. Scale ornament and leaveB as used by AQVITANVS. Cf. 29, AQVITANVS, Nymegen (Knorr 1919, '8A). Claudian. Al. 3. Form 29. By SENECIO of La Graufesenque. For his vineleaf, grapes and bird see 0. 2260A. Claudian ; A.D. 40-50. DI. CANTEEBUBY EXCAVATIONS IN 1945 77 4. Form 29, with high glaze and large beads. The double ring under the spirals of the medallion is characteristic of MASCLVS. Cf. 29, OF MASCLI, Vienne (St. Germain Museum). Claudian. A2. 5. Form 30, fragments from rim and body. Style of MASCLVS. Claudian. (ii) Plain. 1. Form 24. cf 0. and P. XL, 10, COTTV, Aishngen. Claudian. A2. (Fig. 4, 1.) 2. Form 24, base. Stamped OF PRIMI. Cf. 0. and P. XL, 3, Hofheim. Claudian. A2. (Fig. 4, 2.) n—pg3W GMMJSEDi C£MAs£U>4 IVtRECV E5HD8 "irnBEHUfaiWr" 0S5KE6 FIG. 4. BURGATE STREET SITE. Samian ware. Nos. 1, 5 and 13 (J) ; potters' stamps (f). 3. Form 18, with high glaze. Stamped TERTIVS-F as at Hofheim. Claudian. D2. (Fig. 4, 3.) 4. Form 29, base. Stamped OF MASCLI. By Masclus of La Graufesenque. Claudian; A.D. 45-50. D3. (Fig. 4, 4.) Among 19 additional fragments the foUowing forms occur : 29, 24 (larger than Fig. 4, 1), 15, 15/17 (smaU), 17 (Tiberius- Claudian), 18, and Ritt. 12. (o) Coarse Ware (Fig. 5, 1-17). Belgic types with cordoned necks and combed decoration were strong in this series and persisted, but more weakly, in the later deposits. Indeterminate fragments of this ware occurred in quantity. 1. Storage jar, recurved rim, shght offset at base of neck ; coarse buff ware, combed decoration. Belgic. Cf. Richborough I, 29 ; midlst century. A2. 78 OANTEBBUBY EXCAVATIONS IN 1945 s 7 EIG. 6. BURGATE STREET SITE. Coarse ware from the mid-lst century deposits (J). OANTEBBUBY EXCAVATIONS IN 1945 79 2. Rim of smaU jar; high round shoulder, short neck, simple rim; coarse black ware, combed decoration. Belgic. Gi.Richboroughl, 21; mid-lst century. A2. 3. Rim of storage jar, shghtly concave body, simple flat-topped rim thickened internaUy; grey ware, deep combed decoration. Belgic. 01. 4. Storage jar, everted rim with squarish section, shaUow cordon on neck ; coarse buff ware, decorated obhque combing. Belgic. C3. 5. Storage jar, everted rim, cordoned neck ; pohshed black ware, combed on body. Belgic. C3. 6. Jar, outbent rim, cordoned at base of neck ; burnished black ware. Belgic. A2. 7. Jar, outbent rim, cordoned neck; coarse grey-black ware. Cf. Richborough III, 252-3, with pre-Flavian and Flavian pottery. A2. 8. Jug-neck, flanged rim, neck tapering shghtly downwards ; strap handle; pohshed reddish ware. There is no close paraUel for the smooth sweep of the coUar to the neck but for the general type cf. Haltern, Taf. XII, 47. Early 1st century. A2. 9. Jug-neck with disc rim ; cream ware. Cf. Richborough III, 191 ; A.D. 50-70. A2. 10. Jug-neck with double coUar, underside of lower coUar stepped ; cream ware. Cf. Richborough I, 66 ; early 1st century type. D2. 11. ShaUow dish with moulding at base of waU internaUy; pohshed black ware. In imitation of Samian form 15, resembling a weU-known Belgic platter of Arretine origin. Cf. Richborough I, 58 ; probably mid-lst century. A2. 12. Dish with simple rim and footring ; smooth buff-black ware. D3. 13. Bowl, straight coUar; smooth black ware. Probably a simplified version of an Arretine form cf. Haltern, Taf. X, 8, A.D. 11-16 and Verulamium, Fig. 11, 1-2, first half of 1st century A.D. A2. 14. SmaU bowl, reeded flat-topped rim; coarse grey-black ware. Al. 15. Cooking-pot, flanged rim, girth grooves ; sandy grey-black ware. DI. 16. 011a, undifferentiated form; buff-black ware. Al. 17. Mortarium, bead and roU rim; cream ware. DI. II. FROM THE LATE IST-EARLY 2ND CENTURY DEPOSITS. (a) Samian. (i) Decorated. 1. Form 29. Style of VITALIS of La Graufesenque. Similar fan-taUed plant flanked by diverging serrated pointed leaves with pinnate leaflets occurs on 29, OF VITAL, Vindonissa (Knorr 1919, 80 OANTEBBUBY EXCAVATIONS IN 1945 83 E and 84 K) and on 29 RottweU (ibid. 84M). Vespasianio. C4 with joining fragments from C6 and C7. (PI. IV, 1.) 2. Form 37, two fragments. By BIRACILLVS of Banassac. Of. O. 2050 and 2096. Domitianic; A.D. 80-90. D4. (PL IV, 2.) (ii) Plain. 1. Form 18. Stamped NI[. By NIGER of La Graufesenque. Claudian; A.D. 45-50. D4. (Fig. 4, 5.) 2. Form 27. Stamped OFYAL. by VITALIS of La Graufesenque. Vespasianio. C7. (Fig. 4, 6.) 3. Form 27. Stamped OF VITA. By VITALIS of La Graufesenque. Although the stamp occurs at Claudian Hofheim the absence of a groove on the footstand indicates a Vespasianio date here. D4. (Fig. 4, 7.) 4. Form 33. Stamped BIGA FEC. By Biga of La Graufesenque. Domitianic ; A.D. 80-90. D4. (Fig. 4, 8.) The remaining Samian sherds, 39 in number, are mainly Flavian. There are a few survivals, five Claudian and two probably Neronic. There is also a very slight Trajanic element limited to two fragments, from a smaU 33 and a 30. The forms represented among the Flavian sherds are 18, 18/31, 27, 33, 35, 36 and 37. (b) Coarse Ware. (Fig. 6, 1-14.) 1. Cooking-pot with simple thickened rim ; slight offset at base of rim ; smooth black ware with combed decoration. Belgic. Cf. Richborough I, 21 ; mid-lst century. A5. 2. Storage jar with heavy roUed rim; groove at base of neck ; buff ware with combed decoration. Belgic. C5. 3. Jar with cordoned neck and shoulder ; fine grey ware. Belgic. Cf. Richborough II, 144; at Richborough probably 1st century and compared with a London.example dated A.D. 70-100. A5. 4. Jar with cordoned neck; fine burnished black ware. Belgic type related to No. 3. A5. 5. Jug-neck with smaU disc rim; fine red-buff ware. D4. 6. Bowl with down-turned rim, shghtly convex rim ; burnished black ware with tooled trellis pattern. Cf. Richborough III, 339 ; A.D. 80-120. C4. 7. Carinated flanged bowl with ledge for hd ; sandy blaok ware. C£ Richborough I, 22 and HI, 215 ; mid lst-early 2nd century. Here the neck is shorter. C6. 8. Flanged bowl with girth grooves on body, sandy grey ware. C6. 9. Bowl in imitation of Samian form 18/31 ; smooth reddish ware. Early 2nd century. A7. 10. SmaU ovoid jar with short concave neck and beaded rim.; rough-surfaced reddish ware. A6. OANTEBBUBY EXCAVATIONS IN 1945 81 pwwsssws ? 7 <> 1 EIG. 6. BURGATE STREET SITE. Coarse ware from the late lst-early 2nd century deposits (J). 82 OANTEBBUBY EXCAVATIONS IN 1945 11. Jar with reeded thickened rim ; rough grey ware. Somewhat akin to Verulamium, Fig. 31, 42, a type there common in the first half of 2nd century. A7. 12. Beaker with short sharply everted rim and bulbous body; thin smooth grey black ware decorated rouletting. Cf. Richborough I, 62. 04. 13. Carinated jar with cordon at base of neck ; thin smooth black ware. Cf. Richborough III, 289/90; A.D. 80-120. A7. 14. Mortarium with wide flanged and beaded rim ; cream ware. Stamped ORBISSA FE cf. C.I.L. XIII, 10006, 64 from Beauvais and Boulogne. A London example in GuildhaU Museum. Flavian. C5. I I I . FEOM THE MID-2ND CENTURY DEPOSITS. (a) Coin. Antoninus Pius. A.D. 145-61. Sestertius; good condition. M. and S. 772. C8. (b) Samian. (i) Decorated. 1. Form 37. By CENSORINVS of Lezoux. Cf. the Bacchus, acanthus, bud rosette, wavy line on a London specimen (Oswald coU.). Trajan-Hadrianic ; A.D. 115-125. D5. (PI. IV, 3.) (ii) Plain. 1. Form 27, base, rather large. Stamped VERECV. By VERECVNDVS of La Graufesenque. Vespasianio. C8. (Fig. 4, 9.) 2. Form 18. Stamped PATR/. By PATRICD7S of Lezoux. Trajanic. A8. (Fig. 4, 10.) 3. Form 27, base. Stamped /CVRVS F. By COCVRVS of Lezoux. Cf. 27, COCVRVS F, London (B.M.). Probably Trajanic. A8. (Fig. 4, 11.) 4. Form 18/31, smaU. Stamped Z>VCI. By ALBVCIVS of Lezoux. Trajan-Antonine; probably Hadrianic. A8. (Fig. 4, 12.) 5. Form 38, with high glaze. By ALBVCIVS of Lezoux. Cf. 0.andP.LXXII,Z,Silchester. Hadrian-Antonine. C8. (Fig. 4,13.) The unUlustrated sherds (16) represent forms 15, 27, 18/31, 31, 33, 37, 38 and 81. They include survivals as weU as Hadrian-Antonine material. (c) Coarse Ware. (Fig. 7, 1-9.) 1. Cooking-pot with short everted rim, cordoned neck; smooth black ware, rough below neck, decorated obhque slashes. Belgic. Cf. Richborough TH, 244. C8. 2. Dish, simple rim and obhque sides; burnished black ware decorated tooled lines. C8. OANTEBBUBY EXCAVATIONS IN 1945 83 3. Poppy-head beaker ; fine black ware decorated panels of studs. Cf. Richborough I, 54 ; Antonine. A8. 4. SmaU jar, wide-mouthed, recurved rim, girth-groove round body ; sandy grey ware. Cf. Richborough II, 144 ; 1st century. A8. 5. SmaU jar ; black burnished ware, roughened on neck and body, tooled decoration. C8. 6. Cooking-pot, everted rim ; smooth black ware. C8. ?i j % 3WJ MS X t ^#¥ , 7 S & \ 7 } FIG. 7. BURGATE STREET SITE. Coarse ware from the mid-2nd century deposits (J). 7. Mortarium, bead and flange rim; buff ware. Stamped IWENALIS. The stamp of this Flavian potter is previously recorded only from Boulogne (C.I.L. XHI, 10006, 42). A8. 8. Mortarium, roUed rim and bead ; red ware. Cf. Wroxeter 1912, type 98; the latter part of 2nd century at Poltross Burn. A8. 9. Mortarium, bead and flange rim; red ware. A8. (d) Glass. 1. Fragment from a large rectangular green bottle, with monte pattern of concentric circles on. base. lst-4th century. A8. 2. Fragments (2) of green window glass. C8. 9 84 OANTEBBUBY EXCAVATIONS IN 1945 (e) Other Objects. 1. Bronze disc brooch, decorated with triangles of enamel, an inner chcle of red and blue alternately, an outer chcle of blue only. Cf. Richborough I, 7. A 2nd century type which persists later. A8. (Fig. 9, 1.) 2. Bronze beU, rectangular. Cf. Richborough I, 15. A8. (Fig. 9, 2.) 3. Bronze beU, campaniform. Cf. London Mus. Cat., Roman London, PI. XLVHI, 4, Bank of England. A8. (Fig. 9, 3.) IV. FROM THE PITS. Pit. 1. Mid-2nd Century. (a) Coin. Faustina. Sestertius, A.D. 141. (b) Samian. (i) Decorated. 1. Form 29. Probably by PRIMVS. Cf. 29 (two examples), PRIMI, Mainz (Knorr 1919, 66B and j). Probably Claudian. (PI. IV, 4.) 2. Form 37. Style of DOECCVS of Lezoux. Cf. 0.344. Hadrian- Antonine, A.D. 130-40. (h) Plain. 1. Form 27. Stamped OFI, a much abbreviated stamp of FIRMO, as at Trion. Claudian. (Fig. 8, 1.) (c) Coarse Ware. 1. Flagon, saucer-necked; pink-buff ware. Cf. Richborough I, 148, there A.D. 150-200. (Fig. 8, 2.) 2. Mortarium, bead-and-fiange rim ; pink ware with grey core. Probably first half of 2nd century. (Fig. 8, 3.) (d) Other Objects. 1. Bracelet, Kimmeridge shale. (Fig. 9, 4.) Pit 2. 4th Century. (a) Coins. 1. Septimius Severus. Denarius, A.D. 202-10. 2. Tetricus I (A.D. 270-3). Antoninianus. M. and S. 87. 3 and 4. IUegible. Probably 4th century. (b) Coarse Ware. 1. Bowl, oblique-sided, bead-and-flange rim; black burnished ware with tooled interlacing arcs. Cf. Richborough I, 121. 4th century. (Fig. 8, 4.) 2. SimUar bowl with blunter flange and undecorated. (Fig. 8, 5.) 3. Jar with reeded thickened rim ; fumed black ware. (Fig. 8, 6.) OANTEBBUBY EXCAVATIONS IN 1945 85 \M > % x s \ 7 5> 7 J x A r^\ /M C - * < _} ^ « 14 7" FIG. 8. BURGATE STREET SITE. 1-3, Pit 1; 4-6, Pit 2; 7-12, Pit 3 ; 13-14, Pit 4. (J) 86 CANTERBURY EXCAVATIONS IN 1945 (c) Glass. 1. Fragments from body of unguentarium ; heavy pale-green glass with sphal threads of opaque white mavered in. 3rd-4th century. 2. Rim (diam. 3 inches) and body fragments of oUa; greenish glass. 3rd-4th century. O &*.:& i j -A * £&& IF* 3=s "•t* / s I II \ ,1 I I \ , v r ^ / 5 FIG. 9. BURGATE STREET SITE. Objects of bronze and shale (4) (J). 3. Fragment from base of bowl; colourless glass with greenish tinge. Late 3rd-4th century. Pit 3. 4th Century. (a) Coin. Tetricus I (A.D. 270-3). Antoninianus. M. and S. 100. (b) Coarse Ware. 1. Rim of beaker, smooth buff ware. 3rd-4th century. (Fig. 8, 7.) 2. Dish, bead rim, beveUed base ; grey-buff ware, treUis decoration. 3rd century. (Fig. 8, 8.) 3. Bowl, bead-and-flange rim; burnished grey ware. Similar to Pit 2 bowls, Fig. 8, 4 and 5. 4th century. (Fig. 8, 9.) 4. Imitation of Samian form 38; pohshed red ware. Cf. Richborough II, 170. 3rd century. (Fig. 8, 10.) 5. Mortarium, bead-and-flange rim; cream-pink ware. Cf. Wroxeter 126. Probably late 3rd century. (Fig. 8, 11.) OANTEBBUBY EXCAVATIONS IN 1945 87 6. Mortarium, variant of hammer-head rim; cream-pink ware. (Fig. 8, 12.) (c) Glass. 1. Rim of bowl (diam. c. 7 inches); dark amber glass. 3rd-4th century. (d) Other Objects. 1. Bronze ring, with sharp inner edge. (Fig. 9, 5.) 2. " Melon " bead of blue frit. Pit 4. 2nd-3rd Century. (a) Samian. 1. Form 45. Late 2nd century. (Fig. 8, 13.) (b) Coarse Ware. 1. Mortarium, bead rising above down-bent flange; cream ware, herring-bone pattern stamp. Cf. Wroxeter 106. Late 2nd-3rd century. (Fig. 8, 14.) (c) Glass. 1. Fragment from base of bowl; colourless glass with cut circular and oval facets; finest ware. 2nd-early 3rd century. B. MEDIEVAL. None of the medieval sherds from cutting EE' warrants Ulustration. Most of them came from layers 4 and 5. AU were of late 13th century date. The long-necked cooking pots of local type aheady famihar from the Burgate Lane site1 were present, as were fragmentary flagons. The ware was uniformly hard and sandy, buff or grey in colour. A few sherds were glazed, a few decorated with apphed finger-printed strips. II. A SITE IN WATLING STREET. Under the auspices of the Canterbury Excavation Committee excavation was carried out in August-September, 1945 on the site of the Countess of Huntingdon's chapel on the north side of Watling Street, 80 yards from the Riding Gate (Fig. 10, 1, A). The open ceUar offered for examination an area 95 feet north-south2 1 Arch. Oant., Vol. LIX, p. 70 and Roman Canterbury, No. 2, p. 9. a For convenience the long axis of the site is said to lie N-S. It is in fact NE-SW. 88 OANTEBBUBY EXCAVATIONS IN 1945 by 40 feet east-west. The obstacles which conditioned the work must be mentioned. Between the south waU of the ceUar and the north pavement of the street the ground was almost at street level but much disturbed by recent structures. The ceUar itself was divided into two parts by a raised platform of sohd masonry 4fe*> X Zl *•<£*•>. Ofi er-J RIDING GATE / BDxrn—j/ 1 FEET / 0 0 J \ WEST B GATE ui 40 Ui ? LINE OF IU t to ROMAN V-<'***£: \n1 1 WATLING ST RIDINC GATE X (/> I V ZT MILE FIG. 10. WATLING STREET SITE. towards the north end. Also, unsuspected substantial foundations for piers carrying the ground floor of the chapel appeared at intervals down the centre of the ceUar, where the initial cutting was made, and considerably reduced the stratified deposits and the quantity of finds. CANTERBURY EXCAVATIONS AUGUST WATLING MODWSJSrr 43.30 /////////, EAST FACE OF ROMAN FOUNDATIONS NOT EXCAVATED 24Dl26 W m MODERN FOUNDATIONS W V, RAISED I PLATFORM 4 0 . 06 ROMAN_ ^FOUNDATIONS fSS ROAD METALLING MEDIEVAL., PITS LEVELS IN FEET .39.87 ABOVE O.D. FIG. 11. WATLING STREET SITE: PLAN. OANTEBBUBY EXCAVATIONS IN 1945 89 A. ROMAN. SECTION A (FIGS. 11 AND 12). This cutting proved intensive use of the area in medieval times but enough remained of the Roman deposits to give some idea of theh sequence. No trace of buUding either in timber or stone appeared but two bands of compacted gravel were encountered. Natural gravel lay 8 feet 6 inches below the existing floor, at roughly 31 feet O.D. On this old surface were found a few fragments of Belgic combed ware. Above came 3 feet, more or less, of clean yeUow loam, familiarly described as brick-earth. This seemed to have been brought from vhgin soU elsewhere. It contained no debris save a single scrap of bone and three diminutive sherds pf Belgic ware. On this loam rested the two bands of gravel mentioned above. The first running across the south end of the cutting (PL V, 1), was 11 feet below present street level. Where it disappeared beneath Watling Street it was 6 inches thick but taUed to 2 inches at its north edge. Its fuU width remains uncertain, only 7 feet being exposed in the cutting. Obviously its south margin hes beneath the modern road ; there was no sign of it in the ceUar opposite on the south side of the road (Fig. 10, 1, B) where a trial trench was dug. The gravel was tightly packed and the soU below unusuaUy stiff and intractable as though dehberately rammed. Both this and the gravel band next described were taken to represent road metalling. The second gravel band, 55 feet north of the first, ran obhquely across the cutting. Its south edge was seen just south of the raised platform. Its north edge could be traced for a length of 9 feet despite complications introduced by medieval pits. Its further course could not be foUowed ; east, the area was known to have been disturbed by modern weU-digging ; west, pits and the platform made search impracticable. Here again the gravel was 6 inches thick, diminishing towards the edges. Over the south part of the cutting the succeeding layer was a thin band of yeUow loam, not more than 12 inches thick. It rested immediately on the gravel but elsewhere was differentiated from the first deep deposit of yeUow loam only by a thin streak of grey soU or a scatter of gravel. This second loam layer, post-dating the gravel, contained Nero-Vespasianic Samian ware and must therefore be assigned to the late 1st century A.D. or later. Over the north road it was absent. The final deposit throughout the cutting consisted of grey-brown soU to a height of 6 feet above natural ground in the open ceUar and to 10 feet behind the south revetting waU. For this layer pottery indicated an Antonine date. 90 OANTEBBUBY EXCAVATIONS IN 1945 SECTION B (FIGS. 11 AND 12). This was the first of a series of trenches taken transversely from cutting AA' to the east and west walls of the ceUar, testing the remaining area, except where it was known to have been disturbed comparatively recently, and also pursuing the possibility that the modern walls might foUow the lines of Roman foundations as at the previous site in Burgate Street. The west cuttings proved devoid of interest but those on the east showed that a foundation of flintswith- gravel and patches of mortar underlay the modern waU-footings (PI. V, 2) and extended into the ceUar. The eastern face of this foundation was not uncovered, lying as it did beneath the waU or in the adjoining ceUar, but the width exposed (6 feet) was sufficient to suggest a massive waU. The depth of the foundation was 18 inches here, and nowhere exceeded 2 feet. As the plan and drawn sections show it had been sadly mutUated by medieval diggings. In this cutting about 12 inches of yellow loam covered the undisturbed gravel. Evidently the first build-up of yeUow loam varied in depth over the area. Here a deposit of pale greyish heavy loam foUowed, the two layers together equating with the 3 feet of yeUow loam in cutting AA'. Into this grey soil the foundation had been sunk. Above came grey-brown soil variegated with scattered patches of powdered charcoal and reddish soil, obviously habitation debris from elsewhere. This mixed soil sealed the foundation ; it survived below the footings of the overlying waU of the chapel, though here only 6-8 inches deep. In cutting BB', as in subsequent cuttings, the grey loam antedating the structure produced pre-Flavian pottery, the bulk of it Claudian. Sealed by the foundation were two stamped Samian cups, one by Frontinus (pre-Flavian) with a graffito of Lucanus, the other by the partners Cosius and Vrappus (Claudius-Nero). The sealing layer of mixed soil can be dated on the pottery evidence to the 2nd century with a possible hint in a single fragment of a shghtly later date. SECTION C (FIGS. 11 AND 12). The stratification in this cutting was practicaUy the same as that described for cutting BB'. Some difference in the thicknesses of the yeUow loam and the pale grey loam was apparent but together they constituted the first build-up of about 3 feet. The grey-brown soil above was here rather deeper (12 inches) over the foundation, less of it having been cut away for the modern footings. CUTTINGS D AND E (FIG. 11). These cuttings confirmed the evidence from the previous sections. From them came the only datable fragments actuaUy incorporated in the foundation, three sherds of Vespasianio Samian, c. A.D. 70-76. PLATE V (1). WATLING STREET SITE. Gravel band over yellow loam in cutting A, south end. PLATE V (2). WATLIXG STREET SITE. Late 1st century foundations beneath modern footings in cutting B. INS PLATE VI. WATLING STREET SITE. Decorated Samian ware. CANTERBURY EXCAVATIONS IN 1945 91 MODERN WALL CELLAR FLOOR MODERN BUILD-UP 2zzzz 0 D 0 °o ".J-fl-o A CELLAR FLOOR f e> 0 a ° a ° PLATFORM MODERN BUILD-UP ::|0-:- o o 0 O o ° o ° o 0 o Q o o o a MODERN BUILD-UP 9 p^efto'o MODERN BUILD-UP FEET IO 20 30 40 =3 ROMAN — >>>CREY-BROWN SOIL 1111 DARK BROWN SOIL '// YELLOW LOAM • * ROAD METALLING fittB FOUNDATIONS MEDIEVAL =s •:•:•:• PITS [NUMBERED AS ON PLAN]] UNDISTURBED ^ #t NATURAL CRAVEL Pro. 12. WATLING STREET SITE : SECTIONS. • 92 CANTERBURY EXCAVATIONS IN 1945 Cutting E showed that the foundation continued to the south. Its line, or its angle, was sought in the south-east comer of the ceUar but there medieval digging had penetrated the undisturbed. There is, however, the possibility that if a return waU exists at least its inner face may he just south of cutting E where excavation was at this time not practicable. SUMMARY. The level of the ground hereabouts appears to have been raised anything from 2 feet 6 inches to 4 feet in Claudian times, presumably to combat marshy conditions. This tallies with conditions aheady recognised in Burgate Street. During the third quarter of the 1st century, later than the Claudian buUd-up but earher than the subsequent deposit containing Vespasianio Samian, two bands of gravel, taken to be road metalling, were laid down. The southern road ran more or less on the same line as the modern Watling Street, partly beneath it, partly north of it. It appears to be the first of the three known versions of a street on this line. The second, a hard road of baUast, flints, chalk and concrete, was found below the present street, by PUlbrow in the last century.1 He claimed to have discovered its further course to the west under Beercart Lane. His road declined from 3 feet to 6 feet below the present surface as it approached its termination at a point between the Countess of Huntingdon's chapel and Riding Gate. Beneath it PUlbrow found Roman debris at a depth of 10 feet below the present surface. This would seem to be the late 2nd-3rd century deposit of the present excavation, for this layer, where it was best preserved behind the south waU of the ceUar, ran from 5 feet to almost 11 feet below street level. It foUows that if PUlbrow's road be Roman, which has been doubted,2 it must date late in the period. The north street, to judge by the smaU piece exposed, appeared to be converging on the south street somewhere near the Riding Gate. Without stressing what at the moment is mere conjecture, it should be mentioned that this hne projected west runs direct to the West Gate (Fig. 10, 2). As has been aheady mentioned the Vespasianio layer over the south gravel was absent above the north street which was sealed by the late 2nd-3rd century deposit. This points to the north street having been in use to a later date than the south street, though laid down at the same time. The width of the waU foundation running north-south across the area between the two streets indicates a buUding of considerable size. Sunk into the Claudian make-up and incorporating in its material fragments of Vespasianio pottery the waU may be dated to the late 1 V.O.H., Kent, p. 71. 2 Ibid., p. 72. CANTERBURY EXCAVATIONS IN 1945 93 1st century. It would, therefore, have been in course of erection at about the time the south street was being masked with yeUow loam. Its latter end coincides with that of the north road ; by the 2nd-3rd century it had been reduced to its foundation and obliterated. It had disappeared when the recently excavated buUding in Burgate Street was being erected (p. 74 above). The arguments marshaUed against the Roman origin of the Riding Gate include the facts that the date of the street leading to it, as recorded by PUlbrow, is doubtful and that the gate itself lay some distance from the inhabited area. The present site, meagre and therefore inconclusive as the excavation results are, hints strongly that these two points may, when further work has been done, need reconsideration. B. MEDIEVAL. The medieval cavities, occurring throughout the site, numbered thirty one (Figs. 11 and 12). They aU ran up to the make-up for the modern ceUar floor and therefore represented only the lower parts of excavations from a higher level destroyed in the digging of the ceUar. Theh original depth cannot be known. A few (Nos. 1, 9, 14, 16, 23, 26 and 31) penetrated the undisturbed gravel. It has been suggested that the verticaUy-sided shafts (1, 10, 16 and 26) may represent weUs but this seems unlikely in that none goes deep enough into the gravel to reach water and, secondly, no linings of any kind, surely essential in this sub-soU, were observed. More probably both shafts and bagshaped holes were normal rubbish pits. Theh homogeneous filling of dirty soU and household debris attested theh having been quickly filled and abandoned. The intermittent patches of clean soil noted in simUar pits in St. George's Street1 were missing. Pit 21 (section BB') had a post-hole, just over 12 inches in diameter, dug in the bottom but there was no trace of the post in the filling of the pit proper. It had evidently served its purpose and been removed before the pit became a midden. The frequency with which new pits were cut into old ones (7-9 and 13-15) shows the popularity of this area, hke that previously examined in St. George's Street, for the disposal of rubbish. The sherds from these pits (Fig. 16) represent chiefly cooking pots of the local long-necked type (Nos. 1-6). There are a few fragments from jugs (Nos. 7-9) and part of one bowl (No. 10). The ware generaUy is hard-fired and of sandy texture. Its colour graduates from red or buff to grey or black. Very few glazed pieces were found. Thumbing and stabbing occur and one jug (No. 9) has incised decoration of zig-zag lines and arcs. The group is consistently of late 13th century date with the exception of the unusual bowl (No. 10) from Pit 15 which dates to the 14th-15th century. 1 Arch. Oant., Vol. LIX, p. 66 and Boman Canterbury, No. 2, p. 6. 94 CANTERBURY EXCAVATIONS IN 1945 THE FINDS. I. FROM THE MID-1ST CENTURY DEPOSITS. (a) Samian. From the grey loam beneath the foundation. 1. Form 30. Style of MASCLVS of La Graufesenque. Cf. his ovolo and large leaves on 30, London, MASCLJVS F (Stansfield CoU.) and his ovolo and trifid bud on 30, Vindonissa, MASCLJVS F (Knorr 1919, 87 G). Claudian. DD' (PI. VI, 1.) 2. Form 29. By BILICATVS of La Graufesenque. The leaves in winding scroU in upper frieze, with radiate rosette, are identical with the upper frieze of 29, OFIC BILICATI, Mainz (Behn, VHI, 2). The cock is somewhat simUar to that used by AQUITANVS (0.233) COSIVSVTO V , / COlrrEONTp , 3 . FIG. 13. WATLING STREET SITE. Samian. ware. No. 1 (£) j potters' stamps and graffito (J). \lPPI«AK> but the legs are different. Evidently a new type. Claudian. DD'. (PI. VI, 2.) 3. Form 27. Stamped 0 | IFRONTI (unusual form). Graffito round lower part of body reads IV CA NI TX, probably LVCANVS cf. VCANIA M at Leicester. The groove on the footring is a pre- Flavian characteristic. Probably Neronic. BB'. (Fig. 13, 1.) . 4. Form 27. Stamped COSIVS VRAP. By Cosius and Vrappus of La Graufesenque. Claudius-Neronic. CC. (Fig. 13, 2.) (b) Coarse Ware. From the earlier yellow loam in cutting AA' (Fig. 14, 1-5). 1. Jar with bead rim ; black ware burnished on rim and deoorated with coarse irregular combing. Belgic type; of. Richborough I I I , 244, Claudian; but here the rim is more definitely beaded. 2. Jar with shghtly concave neck and vertical rim ; red-buff ware with coarse obhque combing. Belgic type. 3. Rim fragment from jar of Belgic type; grey ware, burnished black on rim and neck, decorated horizontal combing. CANTERBURY EXCAVATIONS IN 1945 95 4. Rim fragment from Belgic jar, simple form ; pale buff-red ware pohshed on rim and neck, decorated obhque combing. Cf. Richborough I I I , 247. 5. Large jar showing Belgic characteristics, cordoned and grooved below sUghtly everted rhn ; smooth buff ware. From the grey loam below the foundation (Fig. 14, 6-7). 6. Jar of Belgic type; coarse black ware decorated obhque combings now badly worn. Cf. Richborough I, 29. CC. ? 1 ^ y FIG, 14. WATLING STREET SITE. Coarse ware from the mid-lst century deposits (£). From the same spot came a fragment from the heavy recurved rim of a Belgio storage jar. - • 7. Plate, imported terra nigra. Cf. Hofheim, Taf. XXXXI, 97Ab. Claudian. CO'. This layer had also a quantity of indeterminate fragments of combed Belgic ware. II. FROM THE LATE 1ST CENTURY DEPOSITS. (a) Samian. From the later yellow loam over the south gravel in cutting AA'. Three smaU sherds, not iUustrated, comprise two bases of form 18, Flavian (probably Nero-Vespasianic), and a base of form 27 of similar date. 96 CANTERBURY EXCAVATIONS IN 1945 From the foundation. . 1. Form 29; lower frieze, blurred impression. By IVCVNDVS of La Graufesenque. Cf. 29, London, of IVCVNDA (London Mus.), Vespasianio, c. A.D. 70-75. DD'. (PI. VI, 3.) 2. Form 29; upper frieze, with pendants of heart-shaped leaves characteristic of MANDVILVS of La Graufesenque. Cf. 29, Leicester, MANDIL. Vespasianio, c. A.D. 70-75. EE'. (PI. VE, 4.) 3. Form 37. SmaU plain rim with zone of pinnate leaves below smaU ovolo. Ovolo not identified. Probably Vespasianio. EE'. (PL VI, 5.) (b) Coarse Ware. From the foundation. Indeterminate sherds of combed Belgic ware. BB' and DD'. IH. FROM THE LATEB 2ND CENTURY DEPOSITS. (a) Samian. From the grey layer in cutting AA'. 1. Form 37. Ovolo of CLNNAMVS, with hare in festoon. Antonine, c. A.D. 140-50. (PI. VI, 6.) 2. Form 30. Panel marked by beadrows ; ring and pinnate leaf in field ; stag used by PATERNVS (O.1770). Antonine, c. A.D. 140-50. (PL VI, 7.) 3. Form 18/31. Stamped GJIPPI-M. By GIPPVS of LEZOUX. Cf. similar stamp on 33, Leicester (Oswald CoU.). Antonine, c. A.D. 170. (Fig. 13, 3.) Not iUustrated are four further Antonine fragments (forms 31 (2), 38 and 45). (b) Coarse Ware. From the grey layer in cutting AA' (Fig. 15, 1-10). 1. Jar with Belgic type, bead rim and high round shoulder; black combed ware with tooled lines below neck. 2. Dish with bead rim and obhque sides ; smooth grey ware. A simUar dish in burnished black ware came from the same spot. 3. Flanged bowl with ledge for hd ; sandy black ware. This and No. 4 are progressively debased examples of a late 1st century type at Richborough (see Richborough I, 22). 4. SimUar bowl but neck more concave ; thin sandy black ware. 5. Flanged bowl or jar straight sided with ledge for hd; "sandy grey ware. 6. Carinated flanged bowl; sandy dark grey ware. 7. SmaU jar, outbent rim, cordoned body; sandy black ware. For general type cf. Richborough III, 313, A.D. 50-80. 8. Jar, clubbed rim, vertical neck with offset at base ; sandy black ware with burnished lines. CANTERBURY EXCAVATIONS IN 1945 97 9. Jar with out-turned rhn and straight sloping neck; sandy black ware. A later version of Richborough III, 256, dated A.D. 50-75. ( fe PS* =^L f=, i P 1 7 FIG. 15. WATLING STREET SITE. Coarse ware from the later 2nd century deposits (-J). S 10. Jar; recurved rim and high round shoulder; graffito (probably incomplete) PIIII; smooth grey black ware. Cf. Richborough I, 26 ; there mid-lst century. The flattened profile here indicates a later date. 98 OANTEBBUBY EXCAVATIONS XS 1945 From the brown soil over the foundation (Fig. 15, 11-16). 11. Dish with bead rim and obhque sides; burnished black ware with tooled trellis pattern. Cf. No. 2 above. DD'. 12. Dish of similar form and ware, but undecorated, DD'. 13. Similar dish. DD'. 14. Carinated bowl with ledge for hd; sandy black ware with girth grooves. EE'. 15. Poppy-head beaker ; fine pohshed black ware. A body fragment with stud decoration, probably from the same pot, came from the same spot. Cf. Richborough I, 54, Antonine. DD'. 16. Rim of cooking pot; burnished black ware. Probably Antonine here but in the north can be as late as the early 3rd century. Cf. Birdoswald, Fig. 14, 18 K. EE'. B. MEDIEVAL.1 (FIG. 16, 1-10.) FROM LATE 13TH CENTURY PITS. 1. Cooking-pot, flat topped rim with slight internal flange, long neck, sides very shghtly convex ; hard sandy buff ware. A common local type difficult to parallel elsewhere. Pit 2. 2. Rim of cooking pot of similar type ; sandy buff ware. Pit 25. 3. Rim of cooking pot of simUar type and ware. Pit 25. 4. Rim of cooking pot of simUar type but without internal flange ; sandy black ware. Pit 22. 5. SmaUer cooking-pot of similar long necked type but rim squarish in section with internal bevel; sandy black ware. Pit 21. 6. Cooking-pot of same general type but rim flanged and beveUed internaUy ; sandy buff-grey ware. Pit 23. 7. Rim and handle of jug ; sandy black ware, decorated on handle with thumb prints. Pit 23. 8. Jug-handle, with thumbing and stabbing ; buff ware with yeUowish-green glaze. Pit 10. 9. Fragment from upper part of body of jug ; red-buff ware with traces of brownish glaze ; incised decoration of zig-zag lines and arcs. Pit 20. 1 The medieval pits contained Roman sherds in quantity but none of intrinsic interest. They also produced the only Roman coins from the site. These are here recorded in case numismatic statistics for Roman Canterbury are desired later. 1. Vespasian A.D. 69-71 Denarius 2. Victorinus A.D. 268-70 Antoninianus 3. Carausius A.D. 287-93 Antoninianus 4. Allectus A.D. 293-96 Antoninianus 5. Valentinian I A.D. 364-75 3 AE. M. and S. M. and S. M. and S. M. and S. 5 71 983 38 Pit 8. Pit 14. Pit 4. Pit 6. Pit 26. OANTEBBUBY EXCAVATIONS IN 1945 99 10 y T > N «r FIG. 16. WATLING STREET SITE. Medieval pottery (£). 10 100 OANTEBBUBY EXCAVATIONS IN 1945 FROM A 14TH-15TH CENTUBY PIT 10. Carinated bowl, coUared rim with narrow thumbed moulding ; hard reddish-buff ware with traces of yeUow-brown glaze internaUy. An unusual form. Pit 15. Acknowledgments. Thanks are again due to Mr. Martyr Smith for making the preliminary surveys of the sites and to Mr. Mann for taking levels. The Samian ware has been reported upon by Dr. Fehx Oswald, the glass by Mr. D. B. Harden, the coins by Mr. B. H. St. J. O'NeU, some of the coarse ware by Mr. P. Corder, and the two mortarium stamps by Mr. E. Birley. For this help I am sincerely grateful. A.W. The thanks of the Canterbury Excavation Committee are due to the foUowing for permission to dig on theh property : Burgate Street— Mr. Percy Court; Mr. C. C. Craik ; Miss Hawkins ; Mrs. Howden ; Mrs. J . D. Langmead. Watling Street—Mrs. Broad ; Mr. C. W. Clinch ; Miss K. Hamblin Smith; Mrs. H. C. Smith ; Mr. A. J. SneUer; Trustees of the Countess of Huntingdon Connexion. Also to Frank Cooper (Canterbury) Ltd., for accommodation for tools etc. F.W.T.