e as _d e_, is not merely
inferred from Dampier's rule, but confirmed by the fact, that,
immediately outside the reef, soundings are either not met with at all,
or only at enormous depths. In short, the ocean is as deep there as
might have been anticipated in the neighborhood of a bold coast; and it
is obviously the presence of the coral alone which has given rise to the
anomalous existence of shallow water on the reef and between it and the
land.

After studying in minute detail all the phenomena above described, Mr.
Darwin has offered in explanation a theory now very generally adopted.
The coral-forming polypi, he states, begin to build in water of a
moderate depth, and while they are yet at work, the bottom of the sea
subsides gradually, so that the foundation of their edifice is carried
downwards at the same time that they are raising the superstructure. If,
therefore, the rate of subsidence be not too rapid, the growing coral
will continue to build up to the surface; the mass always gaining in
height above its original base, but remaining in other respects in the
same position. Not so with the land; each inch lost is irreclaimably
gone; as it sinks, the water gains foot by foot on the shore, till in
many cases the highest peak of the original island disappears. What was
before land is then occupied by the lagoon, the position of the
encircling coral remaining unaltered, with the exception of a slight
contraction of its dimensions.

In this manner are encircling reefs and atolls produced; and in
confirmation of his views Mr. Darwin has pointed out examples which
illustrate every intermediate state, from that of lofty islands, such as
Otaheite, encircled by coral, to that of Gambier's group, where a few
peaks only of land rise out of a lagoon, and lastly, to the perfect
atoll, having a lagoon several hundred feet deep, surrounded by a reef
rising steeply from an unfathomed ocean.

[Illustration: Fig. 119.]

If we embrace these views, it is clear, that in regions of growing coral
a similar subsidence must give rise to barrier reefs along the shores of
a continent. Thus suppose A (fig. 119), to represent the north-east
portion of Australia, and _b c_ the ancient level of the sea, when the
coral reef _d_ was formed. If the land sink so that it is submerged more
and more, the sea must at length stand at the level _e f_, the reef in
the mean time having been enlarged and raised to the point _g_. The
distance between