russell.] STREAM-CUT LAVA SHEETS. 123 
The general conclusion in this connection is that the Snake River 
lava has suffered but slight changes, far too slight in fact to account 
for the presence of even the thin layer of soil that forms the actual 
surface over the several thousands of square miles of the lava-floored 
plains. The comparatively slight alterations referred to seem to mean 
either a very slow rate of change in response to atmospheric agencies, 
or else but a short exposure of the rock. The first of these reasons is 
not certainly true, but the second is plainly evident in a large number 
of instances. 
In reference to progress made in disintegration and decay a com- 
parison of the Snake River lava with the Columbia River lava is of 
interest. The lower sheets of lava beneath the Snake River Plains, as 
previously stated, are of approximately the same geological age as 
the main mass of the Columbia River lava, but the upper sheets of 
the Snake River lava are very much younger than any part of the 
basalt included in the Columbia River lava. In harmony with this 
conclusion is the fact that the extensive sheets of basalt in southern 
Idaho, although exposed to essentially the same climatic conditions that 
exist on the plains of the Columbia, are but slightly decayed; ^while 
the Columbia River lava at many localities and over areas embracing 
hundreds of square miles is so decayed to a depth of 40 to 60 or more 
feet that it is in the condition of clay, and can be molded in the fin- 
gers. a These facts sustain the conclusion derived from fossils, that 
the Columbia River lava is far older than the main portion of the 
similar formation in southern Idaho. 
CANYONS AND WATERFALLS. 
The streams which rise among the mountains bordering the Snake 
River Plains and flow across them bring silt, sand, gravel, and other 
deposits from their upper courses, and with these have cut deep chan- 
nels even in the hard and nearly level lava sheets. The master stream 
is Snake River. As that strong stream lowered its bed the rate of 
flow of its tributaries was increased and they deepened their chan- 
nels also, but could not cut below the level of the controlling stream. 
While the stronger tributaries, such as Salmon and Bruneau rivers, 
were enabled to deepen their channels at the same rate as the mas- 
ter stream, and at their mouths have essentially the same surface 
level as the Snake, and enter it as slack-water streams, the weaker 
tributaries, like the numerous creeks in Elmore and adjoining counties, 
could not work so fast, and now in several instances enter the master 
stream with swift currents, and have falls near their mouths. The 
weaker streams, however, which flow through soft material near their 
a The surface changes of the Columbia River lava are described in the following reports: I. C. Rus. 
sell, Water-supply and Irrigation Paper U. S. Geol. Survey No. 4, 1897, pp. 57-68; Water-supply and 
Irrigation Paper U. S. Geol. Survey No. 53, Washington, 1901, pp. 42-51. 
