remains of a stone parapet surmounting the ramparts, and the lowering of the level
of the interior by excavation, usually at the entrance. As affording some sort of
analogy to the fort at Crawthat Cottage in Tundergarth Parish, attention may be
directed to the fort at Over Cassock, Eskdalemuir, which is unlike the generality
of the forts in the dale, being formed on a promontory and separated into two divisions
by a cross rampart and trench; but in this case the upper enceinte into which the
main entrance opens appears to be the more important enclosure, and not, as in the
Crawthat Cottage fort, an outer bailey. A fort on the Loch Hill (No. 211) in Ewes
Parish, though now much worn away, appears to be distinct from the general type in
the neighbourhood. A remarkable construction (No. 175), which, though an earth-
work, can hardly be deemed either a fort or an enclosure, is situated on the right
bank of the Esk some 3/4 mile above the mansion-house of Castle O'er. It is a plat at
the top of a steep bank rising from the river but at the bottom of a semicircular
hollow in the hillside, that towers above it, and from which it is overlooked at all
points. It has around it certain lines that are of a defensive character, but its
purpose is inexplicable. A Roman fort at Raeburnfoot in Eskdale has already been
noted, and its presence there may find a likely explanation in the group of forts
referred to as indicating a considerable population, which the Romans may have
found it necessary thus to overawe.
Of stone-walled forts in the region of upper Eskdale or of Ewesdale there is not
one. A single specimen, however, crowns the Craig Hill in Westerkirk Parish (No. 637),
some 3 miles above Langholm.
It now remains to consider those defensive constructions which we have
classified under the name of "Enclosures." In form they are as a general rule
circular, or oval, protected by a single rampart with a ditch in front of it, and
having the entrance giving into an excavated hollow in the interior. But one
remarkable feature distinguishes the whole class, that is the hollowing or lowering by
excavation of almost the whole interior surface, so that in some cases the floor actually
lies at a depth of from 4 to 5 feet below the surface of the surrounding ground. Their
close resemblance to certain of the forts, especially to those earthworks which carry
the remains of a stone parapet on their respective ramparts, and on the sides of whose
trenches rock-cutting is visible, renders it a matter of no small difficulty to distinguish
between the two kinds of constructions and also indicates that they are of late date.
Not a single example is recorded in Nithsdale, thirty-seven appear in Annandale, being
in the proportion of somewhat more than one-third to the number of forts, whereas
in the Eskdale and Ewesdale districts they number forty, exceeding the forts by nearly
two to one. The situations which many of these enclosures occupy are not in them-
selves highly defensible, but, set back from the edge of some high bank which margins
the river valley, they are such as would easily escape the notice of marauders on
the roadway through the haugh-land below, while the depression of the interior would
further tend to the concealment of stock herded within.
A small enclosure showing an excavated interior is one of the group of construc-
tions which lie on the flanks of Birrenswark Hill. This particular entrenchment is
situated at the west end, and was examined by the Society of Antiquaries of Scotland
when they made a partial exploration of the Roman camps there in 1898. Within it
were found a broken quern and a piece of bracelet of opaque glass, ¹ the latter an
object whose probable date is in the 1st or 2nd century of our era. Though some of

1 Proc. Soc. Ant. Scot., vol. xxxiii. p. 235.

-- lv

  Transcribers who have contributed to this page.

valrsl- Moderator

  Location information for this page.