Page:Archaeological Journal, Volume 29.djvu/378

31S GAULISH F0KTRESSE3 ON THE COAST OF BRITTANY. sometimes regarded as places of reriii:;e for the inhabitants anterior to Roman times. In tliis diilicuh}' I thought that the spade and }>ickaxe Youhl answer such questions more satisfactorily than the most ingenious conjectures ; and liaving been provideil with the means hy a small grant from the Council-General of the Department, with the assistance of M. Grenot, I commenced o{)erations in September, 1809. During fifteen days' digging, the houses (B, C, E, among others) were proved to be rectangular, ^ith one exception, ^Yllich, abutting against the rock, was semicircular. Tiie ordinary dimensions of the rectangular sites were 5 J yards in length and 3 yards 1 ft. in breadth. They had not the carefully-executed masonry of the houses 13, C, E, but were simply hollows sunk in the ground down to the bare rock, varying in de])th fi-om 2 ft. to 2^- ft., and the out- line marked by a low ridge of earth. The walls are simply the sides of the Jiatural rock, except where, in some instances, occurs a kind of dry masonry of small stones roughly put together. In the majority of cases the fire-])laces are of circular form, and of little more than a yard in diameter, and placed about 6 in. under the level of the ground, being sunounded b}-^ pointed stones placed upright in the ground. At the time of the discovery, one of these hearths still retained some cinders mixed with animal bones and a great nundjcr of shells of the common limpet. A second lire-i)lace was also found in tlie house E, with cinders and charcoal upon it. (See fig. 9.) It is formed of a large stone placed against the wall on the south side. The fire-place in the house C differed altogether from those of the other houses, (Eig. 10.) It occupied the south-west an<rle of the chamber, and was of rectanirular form. Two blocks of unwrought stone boundccl it on [v north and east sides. A third stone inserted in the wall near the angle projected about a foot above the k-vi'l of the floor, as if the hearth or lire-j)lace had been originally covered en- tirely with stones. In the interior its breadth is 4 ft. 8 in., its length 3 ft. 3 in., and the opening between the two stones one or two inches more than the length. No traces of dfjorway or staircases were foun<l in any of the houses exaniincil ; the open sjiriee in (lie eastern wall of the chandler (., and wliieli, at first sight, might Ix; taken for a doorway, has been caused by the lalling of the stones.