on to the mere gravity of the liquefied lava; I should think
that a channel a little straightly or more open would determine the line
of explosion, like the mouth of a cannon compared to the touch-hole.
If a high-pressure boiler was cracked across, no one would think for a
moment that the quantity of water and steam expelled at different points
depended on the less or greater height of the water within the boiler
above these points, but on the size of the crack at these points; and
steam and water might be driven out both at top and bottom. May not a
volcano be likened to a protruding and cracked portion on a vast natural
high-pressure boiler, formed by the surrounding area of country? In
fact, I think my simile would be truer if the difference consisted only
in the cracked case of the boiler being much thicker in some parts than
in others, and therefore having to expel a greater thickness or depth of
water in the thicker cracks or parts--a difference of course absolutely
as nothing.

I have seen an old boiler in action, with steam and drops of water
spurting out of some of the rivet-holes. No one would think whether the
rivet-holes passed through a greater or less thickness of iron, or were
connected with the water higher or lower within the boiler, so small
would the gravity be compared with the force of the steam. If the boiler
had been not heated, then of course there would be a great difference
whether the rivet-holes entered the water high or low, so that there
was greater or less pressure of gravity. How to close my volcanic
rivet-holes I don't know.

I do not know whether you will understand what I am driving at, and it
will not signify much whether you do or not. I remember in old days (I
may mention the subject as we are on it) often wishing I could get
you to look at continental elevations as THE phenomenon, and volcanic
outbursts and tilting up of mountain chains as connected, but quite
secondary, phenomena. I became deeply impressed with the truth of this
view in S. America, and I do not think you hold it, or if so make it
clear: the same explanation, whatever it may be, which will account for
the whole coast of Chili rising, will and must apply to the volcanic
action of the Cordillera, though modified no doubt by the liquefied
rock coming to the surface and reaching water, and so [being] rendered
explosive. To me it appears that this ought to be borne in mind in your
present subject of discussion. I have