GLASGOW ARCHAEOLOGICAL SOCIETY.
huge rampart on the north bank of the great ditch," and he adds
that " this rampart without question has originally been all along
its track."1 As regards the fosse, Gordon repeatedly expresses his
admiration of its " vast breadth and depth/' giving dimensions at
various points as 40 feet broad and 35 feet deep,2 37 feet broad
and about 20 deep,3 25 feet broad and 18 or 20 deep,4 37, 40, and
43 feet broad and 20, 23, and 25 deep,5 43 feet broad and 20
deep,6 66 feet broad and about 25 deep.7 Of the shape of the
fosse he has said nothing in his text, but his " profile or section "
gives it a broad flat bottom not far short of 20 feet in width.
John Horsley, in his great work "Britannia Romana," published
in 1732, devoted a substantial and valuable section to "the
ancient and present state of the Roman Wall in Scotland." He
has, however, little original to offer on its structural aspects
except in his ideal profile in the general map of the wall.8 He
7/it/iro/jk yU/c\uman wall in Scotland
commonly call'd Graham's dike
^-' Jljmk<y$pacts or /5yards
Profile from Horsley's '' Britannia Romana."
there amplifies Gordon's observation that the foundation was of
stone by picturing the vallum as 15 feet in height, of which the
lowermost three feet consist of five courses of squared stone. As
1 Arguing against the allegation that the north agger of the North English
vallum was of an unmilitary character, he says (p. 85)—"As for the first part of
the objection of an agger or rampart to the north of the Fossa, Graham's Dike in
Scotland has had the same : so that we find it was usual with the Romans in the
like great works to have such a eespitious (sic) rampart placed after that manner.
It is likewise certain that the more ramparts placed on either side of the ditch
made the work the stronger, and the Romans, of course, more able to defend the
limits of their province."
2 Itin. Sept., 55 (at Barr Hill).
3 Itin. Sept., 55 (at Barr Hill).
4 Itin. Sept., 56 (east of Croy camp).
6 Itin. Sept., 56 (at Westerwood).
6Itin. Sept., 57 (at Netherwood, west of Castlecary).
7Itin. Sept., 58 (near Bonnybridge). These depths are much overstated, but
the breadths cannot have been far wrong. The 66-foot example, however, was
probably due to a subsidence of the bank.
''Inserted at the end of Book I., chapter 10.