90 SNAKE RIVER PLAINS OF IDAHO. [bull. 199. 
which causes lava to rise, and the other that the lava is forced to the 
.surface by pressure on the reservoirs from which it is derived. In 
the lava streams among the Cinder Buttes, there is no evidence of a 
diminution in the amount of steam occluded in the molten rock 
toward the end of an eruption, but rather the reverse, if account is not 
taken of the greater freedom of its escape from lava that flowed far 
as compared with that which cooled without motion, the lava at the 
immediate source of a stream being, as already stated, in all cases 
highly vesicular, and in part even highly scoriaceous. There is thus 
no reason for assuming a decrease in steam tension within the lava 
itself. On the other hand, the hypothesis that there was a gradual 
diminution in the pressure on the lava still remaining below the sur- 
face meets all the requirements of the facts observed. There still 
remains another alternative, namely, that the supply of material in a 
condition to be extruded gradually diminished. This condition need 
not be considered at this time, however, as it does not seem to have a 
bearing on one's choice between the steam and pressure hypotheses. 
The facts now to be observed indicate that the decrease in the rate of 
flow was due to decrease in volume, and that the more slowly moving 
material retained its steam and gases so as to become highly vesicular, 
while that which flowed far congealed to a more compact rock. 
On the rate of flow, or, more definitely, on the ratio of rate of flow 
to rate of surface cooling, depend certain marked contrasts in the 
resultant surface features. When motion was slow and continued 
after the lava had become viscous, the stiffened crust was left with 
either a generally smooth, flat surface, or acquired oval, stream-like 
ridges, bulging mounds and dome-like swells, while the crust formed 
where motion was more rapid or had continued after the surface had 
passed to a rigid condition became broken and the blocks were variously 
displaced and heaped up so as to produce excessive roughness. These 
two leading varieties of surface features correspond with what Dana a 
has termed pahoehoe and aa in the case of the lavas of the Hawaiian 
Islands. His description of what may be termed the generic charac- 
teristics of these two strongly contrasted lava surfaces is as follows: 
Lava streams are of two kinds. (1) There is the ordinary smooth-surfaced lava 
of volcanoes. It is the pahoehoe of Hawaii, the term signifying "having a satin- 
like aspect." The surface of the lava shows, by the fine and coarse flow lines over 
it, that it cooled as it flowed. Through one means and another the surface is usually 
uneven, being often wrinkled, twisted, ropy, billowy, hummocky, knobbed, and 
often much fractured. * * * The streams have sometimes a firm, glassy exterior 
half an inch or less in thickness. When lava overflows from a boiling lava lake it 
carries along a surface scum 1 to 3 or 4 inches thick, which is a glassy scoria, usu- 
ally easily separable from the more solid and chief part of the lava stream. 
The crusting over of a stream while it is still flowing, owing to contact with the 
air above, results in the leaving of empty tunnel-like caverns, which are sometimes 
hung with stalactites. 
a J. D. Dana, Characteristics of Volcanoes. New York, 1890, p. 9. 
