In some palstaves, but more especially in those of the South of Europe, there is at the butt-end of the blade a kind of dovetailed notch, which appears to have been formed by hammering over a part of the jets or runners of the original castings, which were left projecting a short distance instead of being broken off short at the blade. Whether the hammering over was for the purpose of rounding the angles or for that of forming this dovetailed notch is somewhat uncertain; it is, however, possible that one or more pins or rivets may have been driven through the handle, so as to catch the dovetails and retain the blade in its place. It is not often the case that this portion of the blade is so long that it would have gone through the handle and have allowed of a pin beyond it, as suggested by Mr. Dunoyer[555] in the case of a long palstave, with a rivet-hole near the butt-end of the blade. A palstave, found in a tomb in the department of Loir et Cher,[556] by my friend the late Abbé Bourgeois, is provided with a rivet-hole near the top, counter-sunk on either side so as to guide a pin into the place intended for it; and it seems probable, as the Abbé suggests, that this was connected with the securing of the blade, which is destitute of a loop, to the helve. Of six thin flat bronze celts, 7 or 8 inches long, from the Island of Thermia,[557] or Cythnos, in the Greek Archipelago, which are now in the British Museum, three that are broad are provided with square or lozenge-shaped holes towards the upper end of the blade, and three that are narrower are without. A flanged celt from Italy,[558] 6 inches long, has a circular hole in the same position, which may have received a pin. Some contrivance for keeping blades of smooth bronze fast in their handles must have been necessary or desirable from the earliest times. With stone celts we often find that the butt-end destined to be let into the wooden or horn socket was purposely roughened. With bronze, however, such a process does not seem to have been adopted to any extent; and probably with blades of bronze, so much less tapering than those of stone, the difficulty of keeping them in place was surmounted by attaching them with some sort of resinous or pitchy cement. A safe remedy against slipping out was no doubt found in the addition of the ring or loop to the side, which there can be but little doubt served for a cord to pass through, so as to hold the blade back to the handle. In a socketed celt, 5½ inches long, found in the Seine, at Paris, and now in my own collection, not only is the wood preserved in the socket by saturation with some salt of copper, but within the upper part of the loop there are distinct traces of a cord which was apparently formed of vegetable fibre. The Irish palstave, Fig. 105, with the curved projection instead of the usual loop, seems to show that it was only against the upper part of the loop that the strain came. No doubt, however, there was