IN MILITARY OPERATIONS, 375 In a discoveiy of various instruments and weapons at Notre Dame d'Or, in France, of which I shall hereafter make further mention, there were found eight celts of this kind (" haches a bords recourbes"), and also the half of a bronze mould for casting them. The cavities in the mould were deep in pro- portion to the projection of these extraordinary ridges. It is evident that, after the upper part of the celt had been placed in the handle, these ridges were beaten down upon it with a hammer, and they must have embraced it so tightly, that even without an}- bandage, a strong crowbar or lever would be produced. Bandages were, however, employed as the ordinar}'- means of attachment, and those who have examined the modern stone celts used by the South Sea Islanders and the Indians of North America, will be at no loss to perceive that in this manner the handle might be fastened as firmly as possible. The annexed Avood-cut is the imaginary section of a handle, designed to explain the manner in which the cleft or cavity of the handle was some- times adapted to the wedge-shaped pal- stave. That palstaves were sometimes cast in this form is proved by a bronze mould intended for casting them, which is now in the British Museum, and of which I shall hereafter give an account. When the upper part of the palstave was inserted into this cavity, and the wood and bronze, so dove-tailed together, were tightly bound with cords or thongs, it is manifest that the two would be in- separable. They might be broken, but they could not be dislocated. It has been supposed that the hollow celt of Mr. Du Noyer's fifth class was likewise attached to its handle b}'^ cords. (See the preceding wood-cut, p. 373.) Although there is the highest probability that this mode of attachment was often used in the case of these celts, as well as of palstaves, yet I have no doubt that another method was also adopted, which appears still more effectual. If I cut a stick, more especially of green wood, so as to adapt it to the socket of a celt, and dip it into water, the swelling of the wood will make it fill the celt so completely, that it will require very great force to separate them. If I wish to separate them, I have only to
