les from the base of the
cone. Assuming that the masses went straight upward at the beginning
of their ascent, and that they were afterward borne outwardly by the
expansion of the column, computations which have a general but no
absolute value appear to indicate that the masses attained a height of
from thirty to fifty miles, and had an initial velocity which, if
doubled, might have carried them into space.

Last of all, we shall note the conditions which attend the eruptions
of submarine volcanoes. Such explosions have been observed in but a
few instances, and only in those cases where there is reason to
believe that the crater at the time of its explosion had attained to
within a few hundred feet of the sea level. In these cases the
ejections, never as yet observed in the state of lava, but in the
condition of dust and pumice, have occasionally formed a low island,
which has shortly been washed away by the waves. Knowing as we do that
volcanoes abound on the sea floor, the question why we do not oftener
see their explosions disturbing the surface of the waters is very
interesting, but not as yet clearly explicable. It is possible,
however, that a volcanic discharge taking place at the depth of
several thousand feet below the surface of the water would not be able
to blow the fluid aside so as to open a pipe to the surface, but would
expend its energy in a hidden manner near the ocean floor. The vapours
would have to expand gradually, as they do in passing up through the
rock pipe of a volcano, and in their slow upward passage might be
absorbed by the water. The solid materials thrown forth would in this
case necessarily fall close about the vent, and create a very steep
cone, such, indeed, as we find indicated by the soundings off certain
volcanic islands which appear only recently to have overtopped the
level of the waters.

As will be seen, though inadequately from the diagrams of Vesuvius,
volcanic cones have a regularity and symmetry of form far exceeding
that afforded by the outlines of any other of the earth's features.
Where, as is generally the case, the shape of the cone is determined
by the distribution of the falling cinders or divided lava which
constitutes the mass of most cones, the slope is in general that known
as a catenary curve--i.e., the line formed by a chain hanging between
two points at some distance from the vertical. It is interesting to
note that this graceful outline is a reflection or con