russell.] LAVA STREAMS THAT ENTERED WATER. 113 
been seen in single vertical sections, with lake beds or other deposits 
between. The total, however, considering the whole area, is no doubt 
many times this number. The various sheets are not of the same 
extent, but overlap. The thicker portion of the pile is seemingly in 
the broadest part of the basin to the south of Big Butte, but this can 
not be proved, as no canyons exist there, and no drill holes have as 
yet been put down. From the mouth of Salmon River westward, the 
manner in which various sheets terminate unequally is well shown. 
That is, the lower sheets which extend farther south than the southern 
margin of the surface sheet are exposed in the canyon walls. 
The sheets cut through by Snake River are, in general, compara- 
tively thin, the range in thickness being from 20 to 40 feet; but in 
the deeper portion of the canyon, for about 10 miles below Shoshone 
Falls, about 500 to 700 feet of basalt are exposed, in which the thick- 
ness of the individual sheets seem to be greater, in general, than the 
measures just stated, but the various layers are united in such a way 
as to render it difficult to determine their precise number. 
The basalt shows considerable variation in its physical condition, 
being at places vesicular and again compact. Usually, the layers are 
obscurely jointed and columnar, the columns being vertical, but at 
times the columnar structure is pronounced, as is shown on Pis. VIII, B, 
and IX, A. No doubt some variation in mineralogical and chemical 
composition occurs, but enough study has not been given to this phase 
of the subject to permit of drawing general conclusions. 
LAVA STREAMS THAT FLOWED INTO WATER. 
In the north wall of Snake River Canyon, from Bliss upstream to 
above the mouth of the Salmon River, there are abundant exposures 
of the eroded border of a lava sheet and at certain localities near 
Hagerman, of two lava sheets, which present conspicuous contrasts 
between the basal and central or upper portions. In these instances, 
while the upper and central parts of a lava sheet are granular, com- 
pact, or perhaps moderately vesicular, its lower portion is open in 
texture and composed of irregular fragments, some of which are 
extremely rough on the surface and within are open and cellular, 
resembling somewhat a mass of irregular twigs of glassy lava com- 
3ressed into a moderately compact mass. This peculiar structure, 
ncapable of being accurately described, and presenting many varia- 
tions, is imperfectly shown on PI. XIV, A. In many instances this 
naterial forms rounded masses resembling a pillow folded on itself, as is 
ihown on PI. XIX, A. The lower surface of such a "pillow 11 is some- 
imes nearly flat, but more frequently it is concave and when found in 
}lace is seen to have taken the form of the upper surface of a similar 
nass on which it rests. These folded masses clearly indicate a rolling 
Bull. 199—02 8 
