232 Natural History of Volcanos and Earthquakes, 
thinks himself justified in supposing the caverns beneath the 
Solfatara of Puzzuoli to have a subterranean communication 
with Vesuvius, because whenever the latter is in action, the for- 
mer is in repose. A slip of paper which Davy threw into the 
mouth of the Solfatara, during an eruption of Vesuvius, was 
not rejected, from which he concluded that there must be a de- 
scending current of air. The subterranean thunder, which is 
heard at such great distances from beneath Vesuvius, seems to 
him to indicate the existence of great subterranean caverns, filled 
with gaseous substances, and that the same caverns which, du- 
ring the activity of the voleano, continue for a long time to eject 
enormous quantities of aqueous vapor, must be filled, during its 
repose, with atmopheric air. In proof of the existence of exten- 
sive caverns, he mentions those in the limestone of Carniola. 
Now, as the metals of the earths in the supposed voleanic ¢av- 
erns are not only exposed to the action of the air but also to that 
of aqueous vapor, they will be oxidized at the expense of both, 
and be converted into lava. He thinks his hypothesis capable of 
explaining all the phenomena which he observed. 
* Davy also touches upon the circumstance, often mentioned by 
geologists, that almost all great volcanos are situated near the sea.* 
Supposing their first eruption to have been caused by the action 
of the sea-water upon the metals of the earths, and the metail 
oxides, ejected from the craters in the form of lava, to have left 
vast caverns, the succeeding eruptions would be effected by the 
oxidations which would ensue in those caverns. Davy is of 
opinion that when volcanos lie at a distance from the sea, 
those of South America, the water may be furnished from subtel 
ranean lakes; for Von Humboldt asserts that some of these vol- 
canos cast up fish. ; 
If we wish to ascribe voleanic phenomena to chemical actos 
“says Davy, the oxidation of the metals of the earths and alkaties 
eee 
in the centre of Asia, which is 260 geographical miles distant from an 
and from which st f lava have issued within the period of our history: 
the opinion that the vicinity of extensive lakes operates on the volcanos of Ce ‘a 
Asia, in the same manner as the ocean, is unfounded. The volcano of pt ne 
surrounded by very inconsiderable lakes, and the Lake of Temartu OF Ts ‘tie 
which is not twice as large as the Lake of Geneva, lies fully 25 geographical ra ' 
from Pesckan. See also Girardin in opposition to Davy’s hypothesis in JamesoP 
Phil. Journ. vol. ix, p. 136. 
