54 SCIENTIFIC SURVEY OF YORK AND DISTRICT 



was found in 1820 at Milsington, in Roxburghshire. This dates before 

 A.D. 120 and must, says Sir George Macdonald, have been taken from 

 a fortress. There does exist some evidence, then, that York received 

 damage after the destruction of the ninth Legion. 



Third Period. — Our excavations at the north-east corner revealed that 

 the fortress had received great injury towards the close of the second 

 century a.d., and that the fortress wall had been so seriously damaged 

 that it had to be entirely rebuilt. This reconstruction took place in the 

 reign of Commodus or Septimius Severus, ' following upon a serious 

 disaster, which is no doubt to be connected with the troubles that broke 

 out at the accession of Commodus and led to the abandonment of Scotland 

 c. A.D. 182.' This reconstructed wall is still standing at the north-east 

 corner to a height of sixteen feet, and is a most impressive relic of the 

 Roman occupation. From the north-east corner the wall has been ex- 

 posed to rampart level for a considerable distance, and part of an interval 

 tower has been laid bare. The complete excavation of the interval tower 

 cannot be undertaken for another three years, part of it being in a yard 

 leased by the Corporation. Structurally it appears to be of the same 

 character as the corner tower. Its excavation should supply important 

 evidence bearing on the conclusions arrived at respecting the corner 

 tower. 



In the second season of the excavations, 1926, not only was the work 

 at the north-east corner completed, but also an examination was made of 

 the south-west corner. The work here is different in character from that 

 of the north-east corner, the tower projecting outside the fortress wall. 

 It was found that the tower and stretch of wall adjoining belonged to the 

 reign of Constantius at the beginning of the fourth century. The ditches 

 were of the same date and part of the same defence system. We found 

 here no trace of the earlier defences we had discovered at the north-east 

 corner. 



The work, therefore, to be carried out during the third season of digging 

 (1927) was obvious. It was to attempt to define the extent of the earlier 

 and later defence systems. It was already known that the Constantian 

 wall ran from the south-west corner as far as the western gateway of the 

 fortress. We now cut a trench in the Canon's garden half-way between 

 the western gateway and the north-west corner, and once again found the 

 Constantian wall. It plainly continued to the north-west corner. Two 

 trenches were cut in the north line of defence between Monk Bar and 

 the north-west corner, at a wide interval, and in each case remains of 

 the earlier defence system were found. The junction point of the earlier 

 and later defences plainly lies at the north-west corner. Excavation 

 here, unfortunately, is impossible. 



We had now arrived at this — the defences on the north and east sides 

 of the fortress had been found to date from c. a.d. 71 to c. a.d. 200 ; 

 the defences on the south and west sides of the fortress dated from the 

 beginning of the fourth century ; the junctions of the earlier and later 

 systems are almost certainly at the north-west and south-east corners. 

 The further question now demanded an answer — Whereabouts on the 

 west and south lay the line of the early defences ? No trace of them 



