sed may deprive such
subterranean air of its oxygen. In this manner he explains the great
quantities of nitrogen evolved from volcanic vents and thermal waters,
and the fact that air disengaged from the earth in volcanic regions is
either wholly or in part deprived of its oxygen.

Sir H. Davy, in his memoir on the "Phenomena of Volcanoes," remarks,
that there was every reason to suppose in Vesuvius the existence of a
descending current of air; and he imagined that subterranean cavities
which threw out large volumes of steam during the eruption, might
afterwards, in the quiet state of the volcano, become filled with
atmospheric air.[769] The presence of ammoniacal salts in volcanic
emanations, and of ammonia (which is in part composed of nitrogen) in
lava, favors greatly the notion of air as well as water being deoxidated
in the interior of the earth.[770]

It has been alleged by Professor Bischoff that the slight specific
gravity of the metals of the alkalies is fatal to Davy's hypothesis, for
if the mean density of the earth, as determined by astronomers, surpass
that of all kinds of rocks, these metals cannot exist, at least not in
great quantities in the interior of the earth.[771] But Dr. Daubeny has
shown, that if we take the united specific gravity of potassium, sodium,
silicon, iron, and all the materials which, when united with oxygen,
constitute ordinary lava, and then compare their weight with lava of
equal bulk, the difference is not very material, the specific gravity of
the lava only exceeding by about one-fourth that of the unoxidized
metals. Besides, at great depths, the metallic bases of the earths and
alkalies may very probably be rendered heavier by pressure.[772] Nor is
it fair to embarrass the chemical theory of volcanoes with a doctrine so
purely gratuitous, as that which supposes the entire nucleus of the
planet to have been at first composed of unoxidated metals.

Professor Bunsen of Marburg, after analyzing the gases which escape from
the volcanic fumeroles and solfataras of Iceland, and after calculating
the quantity of hydrogen evolved between two eruptions, affirms, in
contradiction of opinions previously entertained, that the hydrogen
bears a perfect relation in quantity to the magnitude of the streams of
lava, assuming the fusion of these last to have been the result of the
heat evolved during the oxidation of alkaline and earthy metals, and
this to have been brought about by the decomposi