29

feet next to the head gate and only 5 feet thick for the rest of the distance. In excavating for this extension wall, alternate layers of hard and soft limestone were encountered, as shown in Pl. VI, B. The bottom layer of the concrete filling (HK, fig. 5) was laid on one of these hard strata, but it was fully demonstrated that a current of water was running underneath. Several holes were drilled, through which the water welled up in jets several inches high. However, it was thought safe to plug these holes and to ignore the stream below. The bulkhead masonry originally extended to the level RS, fig. 5, 36 feet below the top of the dam; but as an extra precaution a tunnel (RSTU) 6 feet by 6 feet square and 60 feet long was cut under the bulkhead masonry back to the end of the dam proper, and the space was filled with concrete. The space below the 42-foot level under the tunnel was not disturbed.

The foundation for the power house was then excavated to a depth of more than 80 feet below the crest of the dam. The original contractors, after a dispute in regard to the excess of water flooding their work, surrendered their contract. The new contractors succeeded in controlling this water by the use of a cement chamber at D, fig. 5, and a 10-inch horizontal pipe, which projects from the wall of the power house at the point C, about 54 feet below the crest of the dam. This is often referred to as "the spring." A view of it is shown in Pl. VII, A, where it is being discharged through a 10-inch pipe in the third course of granite above the toe of the dam. Measurements taken in October, 1895, showed a range, on the horizontal surface of water in the tailrace, of 5.1 feet and a fall of 5.8 feet, giving a discharge of 4.6 feet per second. A 3,000,000-gallon pump entirely exhausted this "spring." In May, 1897, the flow from the so-called "spring" suddenly increased. An average of several measurements gave a range of 11 feet and a fall of 5.8 feet, and there- fore a discharge of about 10 second-feet. Measurements recently taken give a range of 8.8 feet and a fall of 5.8 feet, and therefore a discharge of 8 second-feet.

In the spring of 1899 it was discovered, from the behavior of drift, that water was disappearing. from the lake at point A, fig. 5. This source of leakage was stopped by filling the lake at that point with clay, loose and in bags. While the filling in was going on the discharge of water from the 10-inch pipe almost ceased for a few hours, but soon reached its normal amount. The filling kept the water discolored and muddy.

In the fall of 1899 it was noticed that water was disappearing at point B, fig. 5, only a few feet from the shore. A cofferdam of sheet piling was constructed around this point and it was filled with hay and earth. A view, taken from point P, fig. 5, of the two leaks and of the method of filling by clay and sheet piling is shown in Pl. VII, B. The clay filling is the pile of dirt adjoining the upstream face of the dam.