the junctions, small crevices, the width of
which varied with the nature of the rock. It was ascertained that
fine-grained granite expanded with 1 degrees F. at the rate of
.000004825; while crystalline marble .000005668; and red sandstone
.000009532, or about twice as much as granite.

Now, according to this law of expansion, a mass of sandstone a mile in
thickness, which should have its temperature raised 200 degrees F.,
would lift a superimposed layer of rock to the height of ten feet above
its former level. But, suppose a part of the earth's crust, one hundred
miles in thickness and equally expansive, to have its temperature raised
600 degrees or 800 degrees, this might produce an elevation of between
two and three thousand feet. The cooling of the same mass might
afterwards cause the overlying rocks to sink down again and resume their
original position. By such agency we might explain the gradual rise of
Scandinavia or the subsidence of Greenland, if this last phenomenon
should also be established as a fact on farther inquiry.

It is also possible that as the clay in Wedgwood's pyrometer contracts,
by giving off its water, and then, by incipient vitrification; so, large
masses of argillaceous strata on the earth's interior may shrink, when
subjected to heat and chemical changes, and allow the incumbent rocks to
subside gradually.

Moreover, if we suppose that lava cooling slowly at great depths may be
converted into various granitic rocks, we obtain another source of
depression; for, according to the experiments of Deville and the
calculations of Bischoff, the contraction of granite when passing from a
melted or plastic to a solid and crystalline state must be more than ten
per cent.[788] The sudden subsidence of land may also be occasioned by
subterranean caverns giving way, when gases are condensed, or when they
escape through newly-formed crevices. The subtraction, moreover, of
matter from certain parts of the interior, by the flowing of lava and of
mineral springs, must, in the course of ages, cause vacuities below, so
that the undermined surface may at length fall in.

_The balance of dry land, how preserved._--In the present state of our
knowledge, we cannot pretend to estimate the average number of
earthquakes which may happen in the course of a single year. As the area
of the ocean is nearly three times that of the land, it is probable that
about three submarine earthquakes may occur for one exclusively