1VL Gay Lussac's Reflections on Volcanoes . 281 
surface, it is impossible to conceive the existence of water at this 
depth ; for the temperature of the earth having been necessarily 
more elevated at a former period, its fluidity greater, and the 
thickness of its solid crust smaller thap at the present day, 
the water must, by indispensable consequences, have been dis- 
engaged from its interior, and elevated above its surface. 
It would therefore be necessary, in order that the hypothesis 
might retain its probability, and the water its importance as a 
principal agent of volcanoes, that it penetrated to the incandes- 
cent strata of the earth, proceeding from above downwards ; 
but then we must suppose it to have a free communication with 
these beds, see it gradually heating before arriving there ; and 
ask, how its vapour, pressed, moreover, by its whole liquid co- 
lumn, could have an elastic power sufficiently great to raise 
lavas, produce earthquakes, and bring about the other phe- 
nomena of volcanoes ? These difficulties, which might be multi- 
plied, render altogether inadmissible the hypothesis, that the 
heat of volcanoes is owing to the state of incandescence of the 
earth at a certain depth below its surface ; I say more, this in- 
candescence is entirely hypothetical itself ; and, notwithstanding 
the observations on the increase of temperature in mines, I con- 
sider it as very doubtful. 
In the second hypothesis which we have offered, that the 
principal cause of volcanic phenomena is a very energetic affi- 
nity between substances, fortuitously brought into contact, it is 
necessary that the water meet in the interior of the earth, with 
substances to which it may possess a sufficiently powerful affinity 
to be decomposed, and give rise to a considerable disengagement 
of heat. 
Now, the lavas ejected by volcanoes being essentially composed 
of silica, alumina, lime, soda and oxide of iron, bodies all oxi- 
dised, and having no longer any action upon water, it is not in 
this state that they must have originally existed in volcanoes ; 
and from what we now know of their true nature, since the 
beautiful discoveries of Sir Humphry Davy, they must have oc- 
curred, if not all, at least a great part of them, in the me- 
tallic state. We then readily conceive that, by their contact 
with water, they might decompose it, be changed into lava, 
and produce a sufficiency of heat to explain the greater number 
