132 SKAKE KIVEK PLAINS OF IDAHO. [bull. 199. 
and fissures in the lava and enable the stronger streams to advance 
over it and possibly to reach Snake River. As is well known, a .stream 
charged with sediment will seal the openings in its channel, even if 
flowing over a thick stratum of bowlders and loosely compacted stones, 
and form for itself an aqueduct through which it can flow without mate- 
rial loss by percolation. Admirable illustrations of this process are 
furnished by the u lost rivers,' 1 but the difficulties to be overcome are 
great, and the task undertaken is not far advanced. 
The sources of supply of the large number of strong springs referred | 
to on a previous page, which gush out from the north wall of Snake 
River Canyon between American Falls and Bliss, are the streams which | 
come from the mountains to the north and are "lost" on reaching the 
plains, together with the rain and water furnished by melting snow i 
which comes directly to the broken surfaces of the plains themselves. 
The environment of the "lost rivers" which, coming from the west, 
reach the Snake River Plains to the east of the Malade but fail 
to pass them, is not conspicuously different from the combination 
of conditions influencing the lives of the streams to the west of the 
Malade which do reach the Snake during a portion of each year, 
except as noted below. To the east of the Malade the lava sheets 
forming the surface of the plains are in part of recent origin and a 
sheet of soil has not as yet been spread uniformly over them, and 
besides, they are still in the position in which they were spread out; 
that is, their surfaces, although rough, are essentially horizontal. To 
the west of the Malade, however, the lava sheets are comparatively 
old, are covered quite uniformly with soil, and besides have been bent 
into a broad, gentle trough, the axis of which runs in a generally 
east-west direction. The streams from the mountains to the north 
emerge upon a smooth plain, which is tilted gently in their direction 
of flow, and which in a large number of instances they have been 
enabled to cross. Another favorable condition is that the mountains 
northwest of Mountain Home are composed largely of granite, which 
furnishes an abundance of material of such size that it can be trans- 
ported b} r even weak streams, and which are then in a condition to 
upgrade their course across the plain. 
An important factor, how r ever, which determines the existence of 
lost rivers is climate, as is shown by the fact that none of the streams 
to the west of the Malade until Boise River is reached, in spite 
of their favorable environments, are perennial. They all shrink in 
summer to such an extent that they barely emerge from their moun- 
tain valle}^. The Malade is the only stream in the whole of southern 
Idaho from Henrys Fork, within 12 miles of the west boundary of 
the Yellowstone Park, to the Idaho-Oregon line, a distance measured 
along Snake River of fully 450 miles, which, rising in the mountains 
to the north, reaches that river in summer. Even the Malade. on 
