OKIGIN OF LAKE liASINS. 11 



that tlie}" leave but slight if any permanent records. Their waters are 

 so clear that practically no sediments accumulate in them. On conti- 

 nental glaciers, however, such lakes might exist from year to year, and 

 perha[)S receive sufficient deposits to leave recognizable records after the 

 ice disappeared. Certain deposits of exceedingly fine, light colored, clay- 

 like material termed loess, in the upper Mississippi valley, are believed 

 by some persons who have studied them, to have been accumulatt^-d 

 in lakes on the surface of the great ice sheet which formerly covered 

 that region. 



When glaciers flow through valleys surrounded by mountains, they 

 sometimes obstruct the drainage of lateral valleys so as to cause lakes to 

 form. The dams in these instances are formed by the ice in the main 

 valleys. The type of this class of lakes is furnished by ]Marjelen lake, 

 Switzerland. In this instance a lateral valley below the snow line is 

 dammed by Aletsch glacier which flows past its mouth. The lake is 

 variable in area, being sometimes a mile long and at other times completely 

 drained owing to the enlargement of the tunnel beneath the ice dam 

 through which it discharges. 



In Alaska there are many lakes of the IMlirjelen type. About the 

 southern bases of the foot-hills of INIt. St. Elias there are several water- 

 bodies that are held in check by the Malaspina glacier. The largest of 

 these, known as Lake Castani, at the southern end of the Chaix hills, is 

 two or three miles long and a mile broad when at its highest stage, and 

 discharges through a tunnel eight or nine miles long, beneath the ice 

 sheet to the south. The position of this sub-glacial river can be traced 

 by a depression in the surface of the ice, and when above it, the muffled 

 roar of the im})risoned flood can be heard far below one's feet. Of many 

 lakes similar to Lake Castani in the same general region, perhaps the most 

 instructive is one discovered by John ^luir, in Stikine valley, Britisli 

 Columbia, near the Alaskan boundary. In this instance a lake about 

 three miles long and approximately a mile broad, and receiving the drain- 

 age of five or six residual glaciers, is held in a lateral valley by Toyatte 

 or Dirt glacier, which flows past its entrance. The outlet of the lake is 

 through a tunnel in the ice, wliich is sometimes enlarged so as suddenly to 

 empty tlif basin and cause a tlood in Stikine river. 



The lakes formed when glaciers obstruct the drainage, are variable 

 in size, owing to changes in tlieir draining tunnels, and are frequently 

 emptied, as in instances just cited. The surfaces of these lakes are many 

 times covered with lloatintj- u'-v, whicli is left stranded when their waters 



