if the stones had had the earthwork packed close up against
them and into the crevices. So far, the structure might seem to
resemble, in all essentials, the previous sections, with the sole
difference of a double core of stone added for strength's sake.
But one feature exists which it may be difficult to reconcile witli
that inference. From the face of the northmost core to the
scarp, the distance is 2G feet. At a point 14 feet from the scarp,
there is a stone which may be a kerb, and the relation of this to
the wall gives considerable scope for speculation. Between the
two points, the quasidcerb and the northmost stone core, there
are, at one special portion, layers as in a vallum. Except at this
particular portion, the soil outside the core shows no signs of
layers, and, indeed, is quite different in character from the soil
between the cores. In this excepted portion, which extends
6 feet out from the stone core and there ceases, the layers are six
1. 3 inches up, 1 inch thick.
— 1 2 >> f >>
3. 8f „ I „
4- Hi „ i „
5- 15 „ i
6. 17 „ J-
What the nature of this portion may be it would be hazardous
to suggest. Perhaps, the vallum was here of greater width than
ordinary. Perhaps, some upper portion got detached, and lies
here outside the face of the true vallum. Until a new series of
excavations is made in Roughcast! e, it would be premature to
express an opinion. The fosse and outer mound here answer
very closely to the reference to them made in connection with
the preceding section, Roughcastle No. 1.
ROUGHCASTLE.—SECTION No. 3.
[Section of rampart only.]
This long cutting exhibits to the fullest extent the characteristics
of the vallum as shown in previous sections. It is the west