is compact snow--which has, indeed, taken on the
form of ice--has been revealed to view, it has been quarried and
conveyed to the towns upon the seacoast. It is likely that there are
many such masses of ice inclosed between the ash layers in the upper
part of the mountain, where, owing to the height, the climate is very
cold. This curious fact shows how perfect a non-conductor the ash beds
of a volcano are to protect the frozen water from the heat of the
rocks about the crater.

The furious rains which beset the mountain in times of great eruptions
excavate deep channels on its sides. The lava outbreaks which attend
almost every eruption, and which descend from the base of the cinder
cone at the height of from five to eight thousand feet above the sea,
naturally find their way into these channels, where they course in the
manner of rivers until the lower and less valleyed section of the cone
is reached.

Such a lava flow naturally begins to freeze on the surface, the lava
at first becoming viscid, much in the manner of cream on the surface
of milk. Urged along by the more fluid lava underneath, this viscid
coating takes a ropy or corrugated form. As the freezing goes deeper,
a firm stone roof may be formed across the gorge, which, when the
current of lava ceases to flow from the crater, permits the lower part
of the stream to drain away, leaving a long cavern or scries of caves
extending far up the cone. The nature of this action is exactly
comparable to that which we may observe when on a frosty morning after
rain we may find the empty channels which were occupied by rills of
water roofed over with ice; the ice roofs are temporary, while those
of lava may endure for ages. Some of these lava-stream caves have been
disclosed, in the manner of ordinary caverns, by the falling of their
roofs; but the greater part are naturally hidden beneath the
ever-increasing materials of the cone.

The lava-stream caves of AEtna are not only interesting because of
their peculiarities of form, which we shall not undertake to describe,
but also for the reason that they help us to account for a very
peculiar feature in the history of the great cone. On the slopes of
the volcano, below the upper cindery portion, there are several
hundred lesser cones, varying from a few score to seven hundred feet
in height. Each of these has its appropriate crater, and has evidently
been the seat of one or more eruptions. As the greater part of these
c