more strength in the loop attached to the blade at both ends than in the mere neb or projection. Some Italian socketed celts have similar projecting nebs, one on either side. In the case of the palstaves and celts with two loops, it seems probable that the handle must have been somewhat prolonged beyond the side branch, which received the palstave or went into the socket of the celt. It has been stated that some of the Spanish palstaves[559] with two loops were, when first discovered, attached to a straight handle of wood. But this opinion may have been formed from the grain of the wood impressed on the upper part of the blade running along and not across it. In the first account[560] given of the discovery, these palstaves were regarded as having been used for _picking_ out the strata of coal, and one of them is said to have been firmly attached to a wooden handle by means of thongs interlaced and held by notches in the wood. This handle was described as having been straight, so that the instrument was fitted to be used as a crowbar and not as a hatchet. But inasmuch as the groove for the handle is only 2¼ inches long and ½ inch wide, while the length of the blade projecting beyond the handle is nearly 5 inches, it is almost impossible for it to have served in this manner. Axe-heads of bronze of the modern form with an eye through them to receive a straight helve have not been found in this country, though, as already observed, they are not uncommon in Hungary, Southern Germany, and Italy. That the form was already known in Greece in the Homeric Age is evident from the feat of skill in shooting an arrow through the shaft holes of a number of axe-heads, arranged in a row, recorded in the Odyssey.[561] I have in my collection a fine double-edged axe, or πέλεκυς, from Greece, 8½ inches in length, with a round shaft-hole ⅞ inch in diameter. I have also two from Salamis.