ric shadow determined by lines supposed to be drawn from the cartridge forty feet horizontally away. The bottle was perfectly protected from the shock of the explosion. It was then put in front of the pile. The first shock shivered it into hundreds of fragments. Other bottles, some filled with air and some with

water, were similarly exposed in various directions around the pile, and with the same result—destruction, except when within the protecting shadow. The experiments were varied by immersing stout glass tubes (Fig. 1), incased in thick paper, horizontally across the direction of the sound-rays in water, between two piles which were aligned with the dynamite cartridge. These piles Fig. 2. were twelve feet apart, the nearer one being forty feet from the cartridge. Its shadow, therefore, just covered the second pile, and included the intermediate water, with the middle part of each tube. After an explosion these protected parts were found to be unbroken, while the ends which projected on the two sides beyond the shadow were completely shattered (Fig. 2). The boundary between the regions of shadow and noise was sharply defined on the tubes, even at a distance of twelve feet behind the protecting pile.

To account for the shortness of the sound-waves which were