according to Pillbrow, of Roman tiles laid two thick on a bed of brickdust, mortar and concrete, above a layer of flints (Arch. xliii, 154, no. 78). He took this for a hypocaust—why, it is hard to imagine. Hasted ascribes a mosaic to Jewry Lane (iv, 411), but it is that found near the County Hotel (above, 2). At the junction of Stour Street and Hospital Lane a flint-and mortar-wall coursed with Roman tiles was found under the roadway in 1867 (ibid. 156, nos. 85, 86). Pillbrow and Faussett connect this with the Roman town wall, for which view there seems neither proof nor disproof available. The timber recorded by Battely as discovered 15 ft. below Lamb Lane—Stour Street—may be of any date (Somner, op. cit. p. 192).
   (9) St. Margaret’s Street yielded much in 1867—8. Three Roman walls close to High Street are noted above (no. 3). Eighty feet south of them, near the church, another wall emerged, parallel to them and oblique to the street, built of rubble with courses of bonding-tiles; if one may trust Pillbrow’s sketch, it is certainly Roman. A garden near the church contained a burial urn. A few steps to the south, opposite the Fountain Hotel, was a floor of small red bricks, doubtless Roman work, and 15 ft. further on a floor of white tesserae, also doubtless Roman, 2 ft. wide, flanked with walling. Twenty-two feet further south began a series of massive foundations which extended to the end of the street. They ran obliquely to the street-line but not parallel to the other oblique walls and, as described by Pillbrow, are not intelligible. Their age, too, is uncertain, except that the great hardness of one part, 22 ft. thick or long, may assign it to Roman builders. Two remarkable bronze pieces were found here. One is a circular ‘horse-trapping’ enamelled in red and green, Late Celtic in style and probably of the 1st century A.D. The other is a pin 8½ in. long, with two tiny wings pendant from its head, possibly also Late Celtic. For the above, see generally Arch. xliii, 159, nos. 44—5 ; for the horse-trapping, Cant. Olden Time, 5, 47; for the tessellated pavements, ibid. 28, and C. R. Smith, Arch. Cant. xv, 127; for the pin, F. B. Goldney in Proc. Soc. Antiq. Ser. II, xviii, 279, with cut reproduced.
   (10) Under the junction of St. Margaret Street, Beercart Lane, Watling Street, and Castle Street, Pillbrow found a well containing an ‘earthen Roman bottle’ and a coin; close by, in the mouth of Wading Street, was a solid foundation, 13 ft. thick (or long) and 12 ft. deep at bottom, and the mouth, of Beercart Lane yielded another foundation, 7 ft. thick. The well and the first of the foundations may be Roman; the other is too little known to be dated (Arch. xliii, 156, nos. 44—6). Pillbrow adds that he found evidence of a Roman road all along Beercart Lane, 4 ft. deep. But he gives no details; the depth is suspiciously shallow, and the road cannot be accepted without further proof.
   (11) According to Battely, a mosaic floor and a ‘strong piece of stonework, indented so firm that it resisted very strong blows,’ were found in digging cellars somewhere in St. Margaret’s parish, not both apparently in the same spot, 5 ft. below the surface (Somner, op. cit. 191, hence Harris (1719), 196; Hasted, op. cit. (1799), iv, 411; Brayley and Britton, viii, 755, and Brent, Cant. Olden Time, 28). By an error of Brent, the mosaic has sometimes been transferred to St. Martin’s parish (p. 75).
   (12) Proceeding straight south to Castle Street, we have to note that this street yielded, only 20 ft. from the junction of the four streets, foundations, pottery and coins which have not been described in detail, and further on, near St. John’s Lane, very heavy flint masonry, 12 ft. thick (or long), which went down full 12 ft. below the surface; near this latter were black urns, oyster-shells and what Pillbrow calls ‘a piece of asphaltum,’ with coins (Arch. xliii, 158, nos. 41, 42). Close to Castle Street, in Hospital Lane, were found a foundation of rubble and flints with strong concrete 4 ft. wide and 12 ft. off, another wall 12 ft. thick (or long), coursed with Roman tiles, and having some tiles inserted obliquely, almost as if in herringbone fashion—presumably Roman. Here were also flue-tiles, black inside with smoke, indicating the hypocaust of a Roman dwelling (ibid. 156, no. 84). This appears to be the most southerly point of the Roman inhabited area as at present known. Somewhere in Castle Street, according to Somner, a ‘strong and well-couched arched piece of Roman tile or brick’ was found about 1640 (Chartham News, reprinted by Battely in Somner, op. cit. 188; hence Harris, 199; Hasted, iv, 411; Brayley and Britton, viii, 754).
   (13) Along the south end of Castle Street, near the Castle, many Roman cremation burials have been found (p. 78). A hard metalled road, only 4. ft. deep, some ditches and the piers of a bridge were also discovered here by Pillbrow (Arch. xliii, 158). They belong almost certainly to the medieval castle which Hasted describes (iv, 408, note, and 410) in language well suited to Pilibrow’s facts. But one piece of ‘hard concreted wall with courses of Roman bonding tiles’ detected by him is Roman work, unless possibly it is copied from a Roman pattern. It occurred close to Worth Gate, where Castle Street intersects the line of the medieval ramparts and, if Roman, helps to fix the line of the Roman town wall. The Church of St. Mildred, behind the Castle, probably older than the Conquest, but has no claim to he regarded as Roman work.
   (14) At the very end of Castle Street stood the now vanished arch of Worth Gate, piercing the city walls. This gateway is ancient. Apparently a road here passed through the city wall before