ded, and of the former
inhabited, while the other sides are from twenty to thirty feet under
water; where, however, they may be perceived to be equally _narrow_ and
well defined. It is on the leeward side also that the entrances into the
lagoons occur; and although they may sometimes be situated on a side
that runs in the direction of the wind, as at Bow Island, yet there are
none to windward." These observations of Captain Beechey accord with
those which Captain Horsburgh and other hydrographers have made in
regard to the coral islands of other seas. From this fortunate
circumstance ships can enter and sail out with ease; whereas if the
narrow inlets were to windward, vessels which once entered might not
succeed for months in making their way out again. The well-known
security of many of these harbors depends entirely on this fortunate
peculiarity in their structure.

"In what manner is this singular conformation to be accounted for? The
action of the waves is seen to be the cause of the superior elevation
of some reefs on their windward sides, where sand and large masses of
coral rock are thrown up by the breakers; but there is a variety of
cases where this cause alone is inadequate to solve the problem; for
reefs submerged at considerable depths, where the movements of the sea
cannot exert much power, have, nevertheless, the same conformation, the
leeward being much lower than the windward side.[1128]

"I am informed by Captain King, that, on examining the reefs called
Rowley Shoals, which lie off the north-west coast of Australia, where
the east and west monsoons prevail alternately, he found the open side
of one crescent-shaped reef, the Imparieuse, turned to the east, and of
another, the Mermaid, turned to the west; while a third oval reef, of
the same group, was entirely submerged. This want of conformity is
exactly what we should expect, where the winds vary periodically.

"It seems impossible to refer the phenomenon now under consideration to
any original uniformity in the configuration of submarine volcanoes, on
the summits of which we may suppose the coral reefs to grow; for
although it is very common for craters to be broken down on one side
only, we cannot imagine any cause that should breach them all in the
same direction. But the difficulty will, perhaps, be removed, if we call
in another part of the volcanic agency--subsidence by earthquakes.
Suppose the windward barrier to have been raised by the mech