164 Recent Excavations at Bireens. 



ing- of the bath, indicated by square hatching-, stands on the original 

 foundations, except the piece overlapping the building on the west, 

 which it is evident must be a departure from the original. The walls 

 blocking some of the subsidiary sti'eets probably represent changes 

 also. Nevertheless, the reconstruction works appear to have pro- 

 ceeded practically according to the old lines ; and it is remarkable, 

 considering that the primary footings in the north parts of the 

 camp were undiscovered, that the secondary buildings rose up of 

 the same form and dimension as before, and in point of situation 

 varied only to the extent of the thickness of the wall or less — a 

 circumstance which seems to imply that the station was probably 

 a fixed and constant type. 



It will be observed that the plan is strikingly compact. The 

 south-east and south-west sections, but for the narrow eavesdrops, 

 present each a solid covering of buildings. In the central section, 

 excepting the passages on either side of thepraetorium, the buildings 

 are almost solid, and being turned endways towards the street, 

 frontage is economised. The north sections are less closely built, 

 but nowhere is there redundancy of space ; and the ovens near 

 the east gateway, previously described, and other structures 

 admitting of it, which must otherwise have encroached on the 

 interior, were embedded in the body of the rampart. 



The plan is characterised also by symmetry, exemplified in 

 the uniformity and balancing of the parts. It is believed, and on 

 good grounds, that the Romans rested the proportion of their 

 edifices, not only as regards the elevation and sections, but the 

 plan also, on the squcae; and the method would seem to apply to, 

 and explain, the Dumfriesshire station. 



If the dimensions of the sides, 500 feet and 300, are bi- 

 sected in order to obtain major and minor axes, the smallest 

 number of equal divisions applicable to both is found to be ten 

 an dsix respectively (see fig.), and lines extended along and across 

 the pla.n from these points mark it out into sixty squai'es. 

 The importance of the squares lies in the coincidence of these and 

 of the lines with the divisions of the camp. The station shows five 

 well-defined sections, separated one from another by the main 

 streets, and it is found that each of them contains twelve of the 

 sixty squares, therefore the areas are exactly equal one with 

 another. Four of them correspond also in form and dimensions. 

 In regard to the lines — No. 3, from the east, the major axis of the 



