GENERAL MEETING. 89 
process would come to a standstill. But there would be no voleano 
in this case, for the supposed condition is evidently statical and 
stable. For the pressure which is supposed to force the vapor in is 
that due to the hydrostatic pressure of a column of water. The 
pressure which keeps it from blowing out is that due to an equally 
high or even higher column of rock, the density of which is at least 
two and a half times greater. 
od. The analogy of the geyser thus fails to become a true ho- 
mology, or an epitome of the voleano. For the geyser is due to the 
access of cold water to a cavity walled by hot rocks and its vapor- 
ization ; the volcano, if due to the penetration of water, is due to pen- 
etration in the form of vapor in the first instance; and the difference 
is radical. 
4th. The proximity of voleanoes to large bodies of water does 
not necessarily imply a logical and causal relation, and is not nec- 
essarily the true law of distribution. Another and perhaps a more 
rational law of distribution may be given. As a matter of fact all 
active volcanoes are not situated near seas or lakes, though in truth 
the exceptions are at the present time few, as for instance, Sangay, 
in the eastern Cordilleras- of Peru, and the volcanoes of Central 
Asia. It seems as if Darwin had acutely divined the true associa- 
tion, viz: that volcanoes are situated in areas which are undergoing 
elevation. So far as we know this rule is without exception, but 
there are many cases where the verification of the elevation is want- 
ing. So far, however, as the test has hitherto been applied it has 
approved the rule. This is especially conspicuous in the western 
half of our own country when applied to the late Tertiary and Post 
Tertiary volcanoes, and it is true, so far as known, of the Andes, 
Java, Phillippines, and Mediterranean, and I have recently been 
able to verify it in the case of the Hawaiian volcanoes. It happens 
that elevations, as well as subsidences, are much more frequent and 
extensive near coast lines than in continental interiors, whence the 
proximity of volcanoes to the sea becomes a secondary rather than 
a primary relation. But elevations also occur in continental in- 
teriors, though less frequently. And when they do occur, we find 
associated phenomena of volcanism as abundant and forcible as in 
littoral regions. This has been the case in the great Tertiary ele- 
vation of the Rocky Mountains, of the Alps, and of the Himalayan 
plateau. Darwin’s law of the distribution of volcanoes is as thor- 
oughly sustained by geological history as by modern instances; 
10 
