40 



THE EXCAVATIONS 



faces of tlie stone rampart, along a line perpendicular to 

 the line of tlie south wall is 398| feet; the breadth, 

 measured along the centre of the Via Principalis, also to 

 the outer line of rampart is 368j feet. The area covered 

 by the fort, making allowance for the irregularity of the 

 shape, but disregarding the rounding off of the comers, 

 is 16,265 square yards, or 3"36 acres approximately. Now 

 that the exact dimensions are known, it will be interesting 

 to compare them with those of other forts, excluding, of 

 course, those that are out of proportion larger than 

 Melandra. These comparisons are more interesting if the 

 forts are taken in groups. Those to which we naturally 

 turn first are the neighbouring forts at Manchester, 

 Brough, and Castle Shaw, and the little earthwork at Toot 

 Hill. The dimensions in feet, as reported, are as 

 follows : — 



The comparison is of course only a rough one, as in two 

 cases an earthwork has been measured. ^^ The fort at 



34. Roeder. Boman Manch., p. 49. Watkin's numbers are 490 and 

 440. Roman Lane, p. 92. 



35. Aikin. Desc. of Country round Manchester. 



36. Proc. Derb. Arch. Soc, 1904. Rom. Brough., p. 10. 



37. Measured by Mr. T. C. Horsfall and myself in 1905. Our measure- 

 ments agreed exactly with those made by Watkin and Earwaker in 1874. 

 The figure is irregular and these numbers indicate greatest length and 

 breadth of vallmn. 



38. In these quotations of areas, I am uncertain in some cases whether 

 the rampart is included. Where this is of clay, the difference may be 

 considerable. Aesioa, mth its earthen rampart, is a case in point. 

 When the above was in type, I found that the areas assigned to Aesica 

 and Vindobala did not quite agree with Mr. Haverfield's figures in his 

 article in Social England. The areas given above are taken from 

 Mr. A. E. Wallis Budge's list in his Roman Antiq. at Chesters. 



