water as raised by the dam will be not less than 34 feet. Between the lake and Toro dredging to an average depth of 4 feet will be required throughout an aggregate distance of 24 miles, the material to be removed being gravel, clay, and loose stones. Below Toro no excavation will be needed in the bed of the river, except at Castillo. Between the rapids the depth of water attained will vary from 30 to 40 feet, and from Machuca to the dam from 60 to 130 feet. The width of the navigable channel where no excavation is required will average 1,000 feet, and in excavation 125 feet at the bottom. The surface width will at no point be less than 1,200 feet, expanding in places to 2,500 between the banks, and in the flooded adjacent valleys to one mile or more. A fall of three-fourths of an inch to the mile has been allowed from the lake to the dam as the necessary slope to discharge the surplus waters, and, consequently, the level of the river at the dam is estimated at 106 feet above sea level, or 4 feet below the lake. For the purpose of navigation, however, that portion of the river may be re

are close together, and may be regarded as one. Still above these are the Rapids of the Castillo. For the whole distance between the Machuca and the Castillo, the banks of the river are rocky; the bottom is also rough and rocky. The depth is very uneven, varying from 5 to 20 feet within the space of hardly as many rods. The current is rapid, and all upward navigation difficult. The Rapids of the Castillo are the shortest of the series, and almost deserve the name of falls. Here considerable ridges come down to the river on either side. . . . . The rapids, without artificial modifications, would present nearly an insuperable obstacle to all kinds of navigation deserving the name. . . . . . Five miles above the Castillo are the Rapides del Toro. The banks are firm and high."-Vol. ii. p. 223.

.....

garded as an extension of the lake, in which the maximum current will probably never exceed half a mile an hour.

A work of some magnitude, but presenting no unusual difficulty, is the rock-blasting under water at Castillo and Toro rapids (amounting to about 400,000 cubic yards), the quantity to be determined by the side slopes found necessary. This work can be more economically done before the water is raised to the assumed summit level, but not before the lower section of the river has been raised by the Ochoa dam to the level of the upper rapids. Otherwise, the excavation in the upper rock ledge might cause an undue fall in the lake level, which would greatly interfere with navigation and the progress of the works in river and lake.

THE SAN CARLOS.

A short distance above the dam the river San Carlos debouches into the San Juan from the south. This stream drains a large area in Costa Rica and possesses in a marked degree the general characteristics of a tropical torrential river; namely, extreme fluctuations in volume, from a nearly dry bed, with barely enough water to float a canoe, to a discharge of, possibly, 3,000 cubit feet per second. Its upper channel and tributaries, confined by high banks, and flowing from mountain slopes, gradually broaden and flatten as they approach the lowlands near the San Juan, and the flanking hills recede from the banks, so that, for a few miles above the confluence, the San Carlos flows through a wide valley, elevated but a