broken into a number of small fragments at any point in its path, these fragments immediately spread out after the manner of shrapnel, covering a cone whose base is approximately one-third of its height; moreover, the distribution of the fragments in space—that is, within the conical surface—will be almost uniform. Such a distribution is almost ideal from the point of view of the work in hand. A desirable solution to the problem would appear to lie in the direction of a bullet composed of pellets or shot embedded in a matrix of only just sufficient strength to hold together, so that on comparatively light impact the component pellets will be released, and each will follow its individual direction of motion.
In every case the degree of expansibility requires, finally, to be determined by experiment, though with sufficient previous experience, and a proper comprehension of the conditions, it is usually possible to hit off the right thing without much difficulty. For the destruction of wooden struts, spars, etc., it is clearly necessary to obtain the most rapid expansion possible (corresponding to instant disintegration), since the bullet has to do its work in a distance rarely exceeding some 3 in. or 4 in. Owing to the fact that the "tissue" penetrated is of very low density—about 30 lb. or 40 lb. per cubic ft.—it is probable that good results would be obtained by drilling the nose with a conical aperture of large diameter, at the entrance about half the diameter of the bullet, after the manner illustrated in Fig. 14. The author is at the present time making a series of experiments with expansive bullets on a target made up of wooden scantlings of cross-section comparable to aeroplane struts or wing members. We may form some estimate of the latent destructive power of the military Service bullet from the fact that it will penetrate some 4 ft. of deal or
92
