e must usually be destroyed
by denudation as soon as a reef begins to rise; since it is immediately
exposed to the action of the breakers, and the large and conspicuous
corals on the outer rim of the atoll or barrier are the first to be
destroyed and to fall to the bottom of vertical and undermined cliffs.
After slow and continued upheaval a wreck alone can remain of the
original reef. If, therefore, says Mr. Darwin, "at some period as far in
futurity as the secondary rocks are in the past, the bed of the Pacific
with its atolls and barrier reefs should be converted into a continent,
we may conceive that scarcely any or none of the existing reefs would be
preserved, but only widely spread strata of calcareous matter derived
from their wear and tear."[1136]

When it is urged in support of the objection before stated (p. 767),
that the theory of atolls by subsidence implies the accumulation of
calcareous formations 2000 or 3000 feet thick, it must be conceded that
this estimate of the minimum density of the deposits is by no means
exaggerated. On the contrary, when we consider that the space over which
atolls are scattered in Polynesia and the Indian oceans may be compared
to the whole continent of Asia, we cannot but infer from analogy that
the differences in level in so vast an area have amounted, antecedently
to subsidence, to 5000 or even a greater number of feet. Whatever was
the difference in height between the loftiest and lowest of the original
mountains or mountainous islands on which the different atolls are
based, that difference must represent the thickness of coral which has
now reduced all of them to one level. Flinders, therefore, by no means
exaggerated the volume of the limestone, which he conceived to have been
the work of coral animals; he was merely mistaken as to the manner in
which they were enabled to build reefs in an unfathomed ocean.

But is it reasonable to expect, after the waste caused by denudation,
that calcareous masses, gradually upheaved in an open sea, should retain
such vast thicknesses? Or may not the limestones of the cretaceous and
oolitic epochs, which attain in the Alps and Pyrenees a density of 3000
or 4000 feet, and are in great part made up of coralline and shelly
matter, present us with a true geological counterpart of the recent
coral reefs of equatorial seas?

Before we attach serious importance to arguments founded on negative
evidence, and opposed to a theory which so adm