where it abutted directly on a sepulchral chamber B. In the section shown in Fig. 6, it was thickest outside, thinning away outwards to the edge of the talus, and inwards into the cavern.

As we were clearing a passage inwards, along the line of Section No. 5, a thick layer of charcoal, marked A on the Plan, Fig. 4, covered slabs of limestone at a depth of 4 feet from the surface, and marked the site of an old fireplace. There were similar blackened slabs, at various levels, in the strata Nos. 3 and 4, in other parts of the area excavated. There were also numerous burnt and broken bones of domestic animals and fragments of coarse pottery. Intermingled with these were a large quantity of human bones, of various ages, lying under slabs of limestone, which formed a continuous packing up to the roof. On removing these a rubble wall became visible, regularly built of courses of limestone. This turned out to be the west wall of a rectangular chamber, B of Figs. 4 and 5, three outer sides being formed of similar rubble walls, while the fourth was constituted by the inner wall of the cave. They enclosed a space 4 feet 6 inches by 5 feet by 4 feet. Inside was a mass of human skeletons of various ages, more than fourteen in number, closely packed together, and obviously interred at successive times. Along with them were the fragments of a rude hand-made pot, ornamented in the herring-bone pattern of the Bronze age, and showing in its fractured surfaces small fragments of stone sticking out of the paste. A few white quartz pebbles, or "luck stones," two links of Kimmeridge shale, and a carefully polished flint flake were also found, the three last in one group.

C.-The Interments.

The bodies had been interred in a crouching posture, with arms and legs drawn together and folded. In several cases the long bones lay parallel to one another— the left humerus and left fibula, the left ulna, the right tibia, and the right femur, the left humerus, left radius, and right fibula-of the same individuals. Some of the bones were in an oblique position, approaching to the

vertical. It is obvious that so large a number of bodies as fourteen could not have been buried in so small a space at one time, although it is clear from the natural position of the bones, in one case of an ankle, and in the other of a spinal column, that the whole body had been buried. The bodies, therefore, have been buried at successive times, and the sepulchral chamber is to be looked upon as a family vault. When it became full of bones the area A of Figs. 4 and 5 was used for burials, as I found to be the case with the approaches of the stone-chambered tombs on the opposite side of the valley, near Cefn, described in the Ethnological Journal, 1871. In my opinion the access to the sepulchral chamber was on the west side, in the direction of A of Plan, Fig. 4.

D.—Cave used for Habitation, and afterwards for

Burial.

On clearing this portion of the cave, we found the section to be as follows:

The stratum No. 4 extended up to the roof of the cavern, and abutted directly on the sepulchral chamber, while No. 3 passed directly underneath it. We may, therefore, conclude that here, as in the sepulchral caves of Perthi-Chwareu and Rhos digre, near Llandegla, in Denbighshire, the cave was used for purposes of habitation before it was used for burial, while it is an open question whether the accumulation No. 4 belongs to the time of the interments. It is probable, however, that the sepulchral chamber was excavated out of it. It is not likely that the same place would be used by the same tribe for habitation after it had been used as a tomb.1

These are two out of a group of five caves of the Neolithic age, explored by me in 1869-1872, and described in Cave-hunting, Chap. V.

E.-The Pottery.

The fragments of pottery are of types repeatedly met with in interments in Britain belonging to the Bronze age. All are hand-made, coarse, grey in colour, or black, or burnt red, and contain small fragments of stone imbedded in the paste. One specimen found in the refuse heap has a bold overhanging rim, bevelled off on the outside, and adorned with herring-bone marks; below this is a shoulder indented with a single row of circular finger marks, the body of the vessel being in addition ornamented with at least two horizontal lines of small triangular impressions. With the exception of

the last feature it is of the same type as that figured by Hoare in Ancient Wilts., and described by Thurnam in Archæologia, XLIII, 61.

The fragments of pottery found inside the sepulchral chamber belong to an urn with an overhanging rim (Fig. 7), adorned with herring-bone pattern both on the outside and on the inside. A small fragment of the same vase proves also that the body was ornamented with four horizontal bands of oblique lines making two complete herring-bone patterns. The urn to which it belongs is of a type common in interments and refuse heaps of the Bronze age throughout the British Isles.

F.-The Links.

Two oval articles found close to the ground flake inside the sepulchral chamber and resembling links (Fig. 8), are made of jet, or Kimmeridge coal. They are carefully rounded and polished, and each has a large oval perforation in the centre. They are of unequal size, and present the following measurements:

In both the perforation has been formed by scraping, apparently with a flint flake. On neither is there any trace of wear. They were probably intended for dress fasteners. They are of the same pattern as that figured by Thurnam in Archeologia, XLIII, 229, Fig. 206, from a round barrow at Thixendale, in the East Riding of Yorkshire, where it was found under the hip of a doubled-up skeleton, and practically under the same conditions as those under notice. It is assigned by Thurnam to the Bronze age.

G.-Flint Flakes.

Several splinters of flint, and one rough flake of chert, were discovered in the refuse heap, and need no further notice. A flint implement, however, found inside the

sepulchral chamber is of a rare type. (Fig. 9.) It rested close to a doubled-up human femur, tibia, and fibula. It is a smoothly polished flint flake, 71 mm. long, 14 to 19 mm. wide, and only 3 mm. thick. It has been made by grinding down a flake so as to preserve the natural curvature of the flat side, and to remove the rib on the back, and to give it the appearance of the blade of a paper knife. The edges are bevelled bluntly off, and the end is rounded. Similar objects have been met with, as Evans points out (Ancient Stone Implements, 290291), in Yorkshire. Their use is uncertain. The association of an implement of this type with Bronze age pottery in this sepulchral chamber fixes the archæological age of the whole group.

FIG. 9. FLINT FLAKE. (Full size.)

H.-The Animal Remains.

The remains of the animals found in the two upper strata, 4 and 5 of Figs. 4 and 5, consist of the wild and domestic animals usually associated together in prehistoric refuse heaps. All are more or less broken and burnt. The wild animals of the following list need no special notice. It may, however, be remarked that the fox was an inhabitant of the cave up to the time of our