landmarks which had survived unnumbered years were swept 

 away in the last century, they left traces on the map of Bath 

 which will not be obliterated while the present city lasts, and so 

 a study of local topography must necessarily begin with the walls. 



The earliest record of the form of the city dates from mediaeval 

 times,* but in the absence of evidence to the contrary it is 

 concluded that the Roman city was of the same shape. If so it 

 was pentagonal, instead of the set square described by Polybius. 

 From this fact Sir R. C. Hoare concludes that " we are still 

 enabled to trace the irregular form of the British town ; " and this 

 view is admitted by an investigator for whose opinion I have the 

 highest respect — Mr. J. T. Irvine, t Still it seems to me that 

 there are so many circumstances here which might induce a 

 modification of plan, that I think further evidence of the inflexi- 

 bility of the Polybian rule is required before we are altogether 

 justified in making this deduction. 



Collinson describes the wall of the Roman city with detail of 

 its material, gates and towers, and gives the dimensions of the 

 city in feet. J It must be understood however that he is only 

 arguing from analogy, and helpful as this may be it is always 

 desirable to distinguish it from positive fact. Dr. Stukeley's 

 vivid imagination saw in the Roman sculptures which stood in the 

 walls in 1724 mere reparation of the actual Roman work, || but 

 the fragments of wall that stiU survive for calmer and more criti- 

 cal examination are distinctly modern. Governor Pownall, who 

 got down to the foundations of the wall during some excavations 

 on the Borough Walls opposite the Hospital, in 1795, pronounced 

 the masonry undoubtedly Roman.§ Mr. Irvine however does 

 not accept his evidence, and holds that we have not traced the 



• Dr. Jones's map, publiNhed 1572. 



t The Bath Herald, 17 Oct., 1874. 



J History of Somerset, 1,8. 



II liin. Cur., p. 138. 



§."A Particular Description," p, 27.. 



