1 88 ROMAN BROUGH : = ANAVI0. 



The wall was not on this occasion followed all round, but it 

 seemed probable that the area was defined' by the masonry 

 indicated on the plan. The two portions of the wall on the 

 south-west side do not align, and at their juncture was an unusual 

 feature. About the centre of that side a four-walled enclosure 

 was found to descend to a depth of 8 feet, with steps leading 

 down from the north-east against the side. The masonry of 

 this cellar or well hardly seemed to be contemporary with that 

 of the walls above, and certain differences were also noted 

 between the opposite portions of the south-west wall of the 

 praatorium. This curious feature is described in the next 

 section. (Plates IV. and V.) 



Adjoining the piaetorium towards the east was found the 

 indication of another strong building, which, however, was not 

 explored. A water channel lay between the adjoining walls, 

 while another fronted the latter building. Other similar stone 

 drains were found in the field. There were indications also 

 of a roadway of cobble-stones passing from side to side, in 

 front of the praetorium. 



This preliminary excavation, then, has shown the Roman fort 

 to be of normal size for the smaller class of square fortresses 

 in the north. It is, however, more strongly built than could 

 have been anticipated. The stone wall which surrounded the 

 whole was six feet in thickness, faced on either side. In this 

 respect it offers, at first glance, a marked contrast with its 

 neighbour at Melandra Castle, though, unfortunately, no 

 sufficient excavation has yet been made at either place to 

 render a scientific comparison possible. At the latter site a 

 laudable beginning was made some years ago, but the work 

 has not yet been completed, and nothing more than a 

 temporary sketch-plan has yet been published.* In the 

 praetorium itself, though hardly delimited, there is again a 

 remarkable difference, which will lend additional interest to 

 the future excavation. But the importance of the present work 

 will not end here. This prastorium is not merely unlike that of 



* See vol. xxiii., p. 90, of this Journal. 



