118 SNAKE RIVER PLAINS OF IDAHO. [bull. 199. 
Farther downstream, just below the mouth of Rattlesnake Creek, 
this sheet of lava thickens abruptly to at least 175 feet — the river 
having not as yet cut through its entire depth — and south of the 
river it thins abruptly to a feather edge and disappears. This lava sheet 
must have advanced over a generally plane surface in which there was 
a depression at least 135 feet deep. The depression was tilled by the lava 
which flowed beyond it, and an essentially level surface was produced. 
The lava where it is thin has a glassy structure and bears other evi- 
dence of having flowed into water, but where it is thick such charac- 
teristics are wanting in the part exposed. From the facts observed, it 
is surmised that the lava entered a lake, in the bottom of which there 
was a marked depression which became tilled with lava, or else flowed 
over an eroded land surface the hollows in which were occupied by 
water. Where the lava sheet was thin the changes produced on com- 
ing in contact with water affected it throughout its entire depth, but 
where it was thick the upper portion cooled slowly and assumed the 
appearance of lava which has cooled on land. 
The study of the lava sheets that advanced over surfaces having 
hollows or depressions in them has not been carried far enough to 
show whether they entered lakes in all instances or not. Sometimes 
the lava rests on waterworn gravel, and this and other facts suggest 
that it advanced over a previously stream-eroded land surface and 
entered the valleys where it became thicker in case the depressions 
were filled, and also came in contact with the waters of streams. The 
results of these two sets of conditions would be much the same, so far 
as the appearances to be observed in the walls of canyons subsequently 
excavated through the lava and associated beds are concerned. In 
the case of a lava sheet entering a broad lake, however, one would 
expect that the beds beneath it would be fine clays or sand, and in the 
case of a lava-filled stream valley coarse gravel and bowlders would 
be apt to be present. But such a correlation can not be relied upon 
implicitly, since lakes may have broad areas of pebbles and bowlders 
adjacent to their margins and bars of similar material extending far 
from shore, while rivers may flow through broad valleys floored with 
fine flood-plain deposits which are indistinguishable in most ways from 
even the finest sediments of lakes. 
CHANGES PRODUCED IN EAVA SHEETS AFTER COOLING. 
OPEN FISSURES AND FAULTS. 
The surfaces of the recent lava sheets about the Cinder Buttes are 
crossed in several instances by open fissures, from a fraction of an 
inch to several inches and in some cases a foot or more wide, which 
have opened since the lava hardened. Possibly these are due to 
stresses generated by the contracting of the lava on cooling, but at 
