THE EXCAVATIONS 53 



appear to me absolutely exact ; the only doubt possible to me 

 is about his conclusion as to the sections north and south of 

 the east gate, where to him (p. 45 )the clay-mound " seems to 

 end in a vertical face " towards the outside of the camp. I 

 am not quite convinced that the face may not once have been 

 a sloping, and not a vertical front. On the other hand, in 

 several sections of the southern rampart the outline of the 

 whitish-brown clay seems to me fairly distinct, sloping 

 outwards beneath a mass of darker-coloured rubble. From 

 what now is visible I find it difficult to understand the 

 sketch provisionally given by Mr. Garstang (in his paper 

 on Eoman Defensive Works) of the rubble (i.e., the stones 

 and earth outside the clay rampart and inside the 

 facing of the wall) as thickest at the ground level. I am 

 at least certain of this much, that in no single spot of the 

 rampart now exposed Avill the yellowish clay be found above 

 any rubble ; while, as I have said, I can point to more than 

 one place in the section of the southern rampart where the 

 rubble seems, to me at least, to have been superimposed upon 

 the clay. I cannot help, therefore, inclining to the belief 

 ti , '^he wall and all that belongs to it was later than the 

 clay rampart ; but I am far from thinking that the evidence 

 is clear enough to make this provable. 



R. S. C. 



THE ANGLE TUERETS. 



Mr. Garstang reported (p. 92) that as the outer wall 

 was stripped from the corners, it was not possible to ex- 

 amine the exact connection between it and the corner 

 towers. The excavations last year, however, practically 

 settled this point. All four corners have now been cleared. 

 At both ends of the northern wall the dressed stones re- 

 main, and the rounding of the comers is distinctly shown, 

 as well as the fact that the side walls of the turrets ran 

 up to the outer wall. Whether there was an outer pro- 

 jection, as at the Saalburg,®^ cannot now be determined. 

 At the latter fort no foundations of corner towers were 

 met with. The curve of the wall at Melandra proved (as 



65. Op. cit., p. 25. 

 E 



