NICARAGUA AND THE ISTHMIAN ROUTES 265 



that amount throughout the summit level. Hence to maintain 

 a depth of 30 feet it will he necessary to secure and hold a depth 

 of 32 feet at the beginning of every dry season. There are very 

 strong indications, however, that whole years sometimes occur 

 in which the evaporation exceeds the rainfall by two feet, so 

 that amount should be stored against the exceptionally dry 

 years, and a depth of 34 feet secured at the opening of each dry 

 season to secure a permanent depth of 30. 



Only two streams carried any considerable quantity of water 

 into the lake last April. These were Rio Frio, which empties 

 at Fort San Carlos, and Rio Zapote, coming in at the extreme 

 south end. A few springs and brooks also flow the year round,' 

 but most of the tributaries are, during the dry season, stagnant 

 estuaries, from 10 to 25 feet deep. Aquatic vegetation grows in 

 great profusion in these streams, choking the channel, unless a 

 passage for navigation is kept open by canoes. This vegetation 

 is not rooted to the soil, but floats freely and shows no disposi- 

 tion to flow out toward the lake. The water is dark-colored, 

 foul, and stagnant. When the rainy season sets in in earnest, 

 however, these wide, deep channels are filled with a rushing 

 current so swift that it is very difficult to paddle a canoe against 

 it. The floating vegetation is carried out in great quantities, 

 forming floating islands several hundred feet across. As the 

 current of the stream often extends far out into the lake, the 

 floating vegetation forms a large crescent around the mouth, and 

 in some cases constitutes quite an annoyance to navigation. 

 Some of these islands find their way to Fort San Carlos, and 

 pass down the San Juan river. The inflow to the lake is some- 

 times as great as 300,000 cubic feet per second, and to prevent 

 injurious submergence of adjoining property it is necessary to 

 provide a spillway of large capacity, and to control the discharge 

 S< i as to waste the surplus and yet to store the large amount of 

 water necessary to provide for the evaporation from this im- 

 mense lake surface. 



As illustrating the problems solved by this survey, the extreme 

 flood discharge of Rio San Juan at Ochoa was estimated by the 

 company at 63,000 cubic feet per second. A discharge of about 

 1.05,000 cubic feet per second has been already observed, and a 

 study of the floodplain in connection with the observations at 

 the same point shows that the maximum discharge cannot be 

 less than 21)11,1)1)0 cubic feet per second, or more than three times 

 the estimated maximum. The San Carlos and Serapiqui rivers 

 to 



