walls, nor can any such material be seen in the few fragments now visible. The Roman wall appears to have been backed by an earthen bank, a section of which was seen and recorded by Payne near the southern corner.¹⁹  There is no record of internal towers, and the only hint of an external tower is the mysterious ‘buttress’ (noted below) at the northern corner. Nor is there any direct evidence as to the Roman ditch or ditches except that, when the town defences were extended towards the south-west in 1225, the new wall was built at a distance of 120 ft. from the old Roman wall, and must certainly have stood clear of the Roman ditch system.²⁰
   The walls enclose an irregular area of about 23½ acres, which measures some 475 yds. along its greatest length and 350 yds. in its greatest width. Except along the north-west front which overlooks the river, their outline can be traced with reasonable certainty (fig.14). The antiquary who wishes to follow it may start from the enceinte of the castle, opposite the pier and public baths, where the Roman core can be seen underlying the medieval curtain-wall at the top of the present escarpment. Some 50 yds. to the south the Roman core turns inward and disappears, but is found between that point and the south-western wall of the Norman keep, which partly overlaps it.²¹  Thence the Roman wall proceeds in a straight line to the Bishop’s Palace, where a large section of it was laid bare about 1905 ‘during the extension of the drawing-room of the now remaining wing of the Palace.’²²  Further again to the south-east it is recorded under Precinct Road, and it forms the basis of the ‘north-eastern wall of the Norman refectory now partially covered by the Choir School. In a basement beneath the school the unfaced Roman core is still partly visible, and a further stretch can be ‘seen in the adjacent deanery garden.²³  Another 80 yds. brings the investigator to the rounded southern corner of the Roman enceinte. Although partially masked by the abutment of Norman and medieval extensions, the Roman work is here unusually well preserved and retains the only original facing now visible throughout. its length.²⁴  To the north-east of the site of East Gate a stretch of the Roman wall is recorded,²⁵  and the rounded eastern corner which bounds the playground of the Mathematical School, although medieval above ground, is continued northward, for 30 yds. or more, by a visible stretch of Roman core, and doubtless stands on Roman foundations.²⁵
   The north-eastern wall runs parallel to the road known as The Common and at the backs of the houses which front upon it, and the Roman masonry was seen in 1905 during the widening of Northgate (formerly Cheldergate or Pump Lane).²⁷  Further north, in 1902, demolitions immediately to the west of Brooker’s forge on The Common revealed a large portion of the city wall still partly faced.²⁸  In 1910—11, during excavations made by the South-Eastern and Chatham Railway viaduct in the rear of the Guildhall, the wall was cut and its exact width (6 ft. 10 in.) ascertained.²⁵  To the north of the more northerly railway viaduct a further stretch of Roman core can be seen on both sides of a narrow alleyway known as Parr’s Head Lane. Near
   ¹⁹  Arch. Cant. xxi, 4; xxviii, lxxxix.                    ²⁰  Ibid. xxiv, 15
   ²¹ Ibid. xxi, 26, and pi. and iv; xxvii, 188—9.
   ²²  Ibid. xxvii, lxx.                                               ²³  Ibid. xxi, 5.
   ²⁴  Ibid. XX1, 4—5, 25 note, and pl. i, 5 and 6.
   ²⁵  Ibid. xxi, 9, 16                                               ²⁶  Ibid. 9.
   ²⁷  Ibid. xxvii, lxix.                                              ²⁸  Ibid. xxvii, lxvi.        ²⁹ Ibid. xxix, lxxxiv.