tion of water. Yet after
having thus succeeded in removing the principal objection once so
triumphantly urged against Davy's hypothesis, Bunsen concludes by
declaring that the hydrogen evolved in volcanic regions cannot have been
generated by the decomposition of water coming in contact with alkaline
and earthy metallic bases. For, says the Professor, this process
presupposes the prevalence of a temperature in which carbonic acid
cannot exist in contact with hydrogen without suffering a partial
reduction to carbonic oxide; "and not a trace of carbonic oxide is ever
found in volcanic exhalations."[773] At the same time it will be seen,
by consulting the able memoirs of the Marburg chemist, that he supposes
many energetic kinds of chemical action to be continually going on in
the interior of the earth, capable of causing the disengagement of
hydrogen; and there can be no doubt that this gas may be a source of
innumerable new changes, capable of producing the local development of
internal heat.

_Cause of volcanic eruptions._--The most probable causes of a volcanic
outburst at the surface have been in a great degree anticipated in the
preceding speculations on the liquefaction of rocks and the generation
of gases. When a minute hole is bored in a tube filled with gas
condensed into a liquid, the whole becomes instantly aeriform, or, as
some writers have expressed it, "flashes into vapor," and often bursts
the tube. Such an experiment may represent the mode in which gaseous
matter may rush through a rent in the rocks, and continue to escape for
days or weeks through a small orifice, with an explosive power
sufficient to reduce every substance which opposes its passage into
small fragments or even dust. Lava may be propelled upwards at the same
time, and ejected in the form of scoriae. In some places, where the fluid
lava lies at the bottom of a deep fissure, communicating on the one hand
with the surface, and on the other with a cavern in which a considerable
body of vapor has been formed, there may be an efflux of lava, followed
by the escape of gas. Eruptions often commence and close with the
discharge of vapor; and, when this is the case, the next outburst may be
expected to take place by the same vent, for the concluding evolution of
elastic fluids will keep open the duct, and leave it unobstructed.

The breaking out of lava from the side or base of a lofty cone, rather
than from the summit, may be attributed to the hy