286 CORALS AND CORAL ISLANDS. 
Florida Bank, it does not appear to suffice for the Eastern 
Bahamas, with their steep submarine slopes toward the ocean, 
or for the Grand Cayman and other reefs southwest of Cuba, 
or for the great majority of the atoll and barrier reefs of 
the Pacific. It is, however, not yet proved to be true for 
Florida. 
e. By Darwin’s theory, the growing reef increases its 
thickness as the slow subsidence progresses; and the inside ~ 
channel, so common a feature, is a consequence in the way 
that has been explained. But by the Murray method, after 
the outer edge of the reef has a thickness equalling the 
range in depth of reef-corals for the region, the reef, as it 
further enlarges, keeps extending out into deeper and deeper 
water. But the width cannot be doubled without using three 
times as much material as for the preceding part in case the 
slope of the bottom is unchanged and the coast is a straight 
line, and much more than three times if the coast line is con- 
vex like that of Tahiti, and still more if the slope of the 
bottom increases. With the feeble amount of debris to be 
had, the rate of growth would therefore be extremely slow, — 
not over one hundredth of that for the first one thousand feet. 
During all the long era of this extension the waves and tidal 
movements and winds would be at their usual work, throw- 
ing the chief part of the debris on the reef and beyond into 
the channel; and thus the channel, the process being aided 
by its own growing corals, would be sure to become filled 
and nearly obliterated; for it is ever receiving, and has no 
chance to increase its capacity or relieve itself of the debris 
received. 
Among examples of a reef undergoing such an extension 
where there was little or no subsidence, is that of the north 
side of Upolu, of the Samoa Group, which the author’s Expe- 

