one-half of the natural size, in figs, 87–89. The bulb is shown in figs. 88, a, and 89, and the flake has been worked into a point at the end. The largest flake I am acquainted with is described by M. de Caneto, in the Revue de Gascoyne, for 1865. It was found in the commune of Pauilhaic, and is 1312 inches in length. Fig. 90 is an arrow-shaped flake, chipped away at the base, apparently to adapt it to a handle or shaft.
Figs. 91, 92, 93.—Flakes from a Danish shell-mound, actual size. In my own collection.
Figs. 91—94 represent small Danish flakes: forms exactly similar may be found in any country where the ancient inhabitants could obtain flint or obsidian. In fig. 91 we see that another flake had been previously taken from the same block. Figs. 91, 92 represent flakes of which the points have been broken off, but we see along their whole length the depression caused by the previous removal of other flakes. The section of such a flake is, therefore, not triangular, as in fig. 95, a, but four-sided, as in fig. 95, b. Sometimes, though not often, a wide flake is taken off in such a way as to over-
