character by the short passage through the rocks that intervene between them and the fault; and hence they insist that the theory of percolation is untenable. There are two methods of securing the mineral springs of this locality: the first is shown at Fig. 2, and consistsFig. 2. in excavating to an extent of twenty or thirty feet square surrounding the spot where indications of mineral water are observed, and extending downward through the various drift-formations until the underlying rock is reached. As the work progresses, a shaft or crib is sunk in order to prevent the sides from caving in; and, to obviate the collection of water and carbonic-acid gas at the bottom of the shaft, powerful steam-pumps are kept in constant operation, which effectually drain the excavation. After reaching the fissured crevices in the rock that environ the fault, and through which the water issues, a pyramidal wooden hopper, about one foot square at the apex, and two or three feet at the base, is placed on the rock directly over that portion of the crevice from which the water issues most abundantly, its position being firmly secured by packing clay tightly around its exterior. As rapidly as the work of filling in the shaft progresses, a wooden tube, about one foot square, is accurately adjusted to the hopper, from which the water gradually rises until it reaches the outlet at or near the top. The depth at which the rock is located from the surface varies from fifteen to fifty-seven feet. The flow of water from springs secured in this manner averages from thirty to one hundred and twenty gallons an hour.

The second method (see Fig. 3) consists in drilling into the rock, in close proximity to the fault, until mineral water is obtained, the drill in the mean time being followed by an iron pipe, which effectually secures the flow, prevents the access of fresh water, and protects the rock through which the drill passes from the combined disintegrating action of both the water and carbonic-acid gas.

Most of the springs secured in this manner are spouting in character; their flow is not, however, continuous, but spasmodic or intermittent. This peculiarity is undoubtedly due to a pocket or cavity in the rock, as represented in Fig. 3. A is the tube leading from the pocket to the surface. As the water flows into the pocket from the surrounding inlets, it gradually rises above the outlet, which results in the compression of