TO any one who has examined an extensive collection of the bronze instruments found in this country it will at once be apparent that in the class of celts designed to be fixed in some sort of haft, and not themselves socketed for the reception of a handle, there is a wide range of form. Any attempt, however, to divide them into well-marked classes is soon seen to be futile, as there is found to be a gradual transition from what at first sight appears to be a well-marked form into some other which presents different characteristics. If, for instance, we take the side flanges as a criterion, we find them ranging from a mere thickening on the margins of the flat celts to well-developed flanges, extending along nearly the whole blade; we then find them confined to the upper part of the instrument, and in some cases of great lateral extent, so as to be capable of being hammered over to form a kind of semicircular socket on each side of the blade. In other cases we find that the flanges have some part of their apparent projection due to a diminution in the thickness of the portion of the blade which lies between them. If we take as a criterion the stop-ridge, as it has been termed, a projecting ridge for the purpose of preventing the blade being driven too far into its wooden handle, we find the ridge in a rudimentary form in the blades which taper both ways; next as a slightly raised ridge or bead running across the blade; then as a better-defined ridge, to which, at last, greater development is given by a reduction in the thickness of the blade above it. The presence or absence of a loop at the side is, no doubt, a good differentiation, but as this is a mere minor accessory, and two celts may be identical in other respects with the exception of one being provided with a loop and the other being without it, it does not materially assist in the classification of this group of instruments, although for convenience’s sake it is best to treat of the two varieties of form separately. An additional reason for this may be found in the possibility that the loop was a comparatively late invention, so that the palstaves provided with it may be in some cases of later date than those without it, though the identity in the ornamentation of some of the instruments of the two classes, and the fact of their being occasionally found together, are almost conclusive as to their contemporaneity. [Illustration: Fig. 48. Fig. 49. Icelandic “Palstaves.”] In the present chapter I propose to treat of the celts with a stop-ridge, of the winged celts, and of those of the palstave form.