his, we must determine _experimentally_ how much heat can be
developed by the crushing of a given volume, say a cubic mile, of such
rocky materials as we know must constitute the crust of our globe down
to the bottom of the known sedimentary strata, and extending to such
crystalloid rocks as we may presume underlie these. We must also obtain
at least approximately what are the co-efficients of _total contraction_
between fusion and atmospheric temperature of such melted rocks, basic
and acid silicates, as may be deemed representative of that co-efficient
for the range of volcanic fused products, basalts, trachytes, etc.,
which probably sufficiently nearly coincide with that of the whole
non-metallic mass of our globe.

The first I have determined experimentally by two different methods, but
principally by the direct one of the _work_ expended in crushing prisms
of sixteen representative classes of rock; the specific gravities and
specific heats of which I have also determined.

If H be the height of a prism of rock crushed to powder by a pressure,
P, applied to two opposite faces, which, when the prism has been
reduced to its volume in powder, has acted through a range of H - t,
then

P x (H - t) / 772

is the heat corresponding to the work expended in the crushing,
expressed in British units of heat. The following were the rocks
experimented upon: Caen stone, Portland (both oolites), magnesian
limestone, sandstones of various sorts, carboniferous limestones
(marbles), the older slates (Cambrian and Silurian), basalts, various
granites and porphyries, thus ranging from the newest and least
resistant to the oldest and most resistant rocks. The results have been
tabulated, and are given in detail in my Paper, now in possession of the
Royal Society. The minimum obtained is 331 and the maximum 7,867 British
units of heat developed, by transformation of the work of crushing one
cubic foot of rock. If we apply the results to a thickness of solid
crust of 100 miles (British), of which the upper twenty-one miles
consist of neozoic, newer palaeozoic, older palaeozoic and azoic rocks in
nearly equal proportion as to thickness, and the remaining eighty miles
of crystalloid rocks (acid and basic magmas of Durocher) of physical
properties which we may assume not very different from those of our
known granites and porphyries--and which, in so far as they may differ,
would give a still _higher_ co-efficient of work transforme