THE EXCAVATIONS 31 



is that of Hard Knott.^'' Lastly, reference is made to 

 Mr. Haverfield's mention of a similar construction in some 

 of tlie smaller Eoman forts of Northern Africa.is 



Let it be said clearly that, as far as the eastern and 

 western gates are concerned, the question is still an open 

 one, which may yet be settled by a fuller excavation of the 

 former. Fortunately, we discovered part of the central 

 pier at the north gate, and there is little doubt that the 

 arches at that entrance were equal. At all events, we 

 have there the exact width of one span, and, assuming 

 that the door jambs (if such existed) rested on the first 

 course (and this is rendered probable by the fact that the 

 road seems to have been made up to this level), the exact 

 width of the opening would be 7 ft. 10 in. Neglecting 

 the door jambs the space might be 8 ft. 6 in. This is almost 

 precisely the width assumed by Mr. Garstang for his wider 

 arch,i9 the calculation being made from one of the 

 voussoirs found, which indicated a span of eight feet. We 

 are then left with a little over 13 feet for the other span 

 and the central pier. As the pilasters are exactly equal on 

 both sides, it is difficult to see why we should assume that 

 the other span was smaller. Of course one arch may have 

 been built up, leaving only a small arched door for 

 entrance, but in that case the whole idea of adducing 

 Lincoln and Hard Knott as parallels falls to the ground, ^o 

 In both those cases the inequality is shown by foundations. 



17. The inequality of the arches there worked out in one instance to 

 3 mches! (9ft. llin. and 9ft. 8in.). In two other gates, however, 

 Mr. Dymond reports as much as 2ft. llin. and 3ft. 7in. respectively. 



18. In his own very interesting account of Melandta {The Victoria 

 History of Derbyshire, vol. i.), Mr. Haverfield states that the arches 

 were reported to be unequal at the western gate also. Here western has 

 evidently been prmted for northern. (The northern arches were at first 

 supposed to be unequal). Mr. Hamnett, who excavated the western 

 gate, tells me (March, 1906), that he found no such indications at that 

 entrance. 



19. See drawings. Proc. Derb. Arch. Sac, vol. xxiii., p. 93. 



20. It is clear, however, from Mr. Garstang's plan (Some Features of 

 Boman Military Defensive Works, Plate iv.) that he did not intend this 



