SUBMERGED CLIFFS OF REEF-ENCIRCLED ISLANDS 537 



fices^ as above noted^ that subsidence should proceed so slowly as not to 

 inter nq^t cliff -cutting, but rather to aid it by gradually and coiitinually 

 deepening the water on the abraded platform, and thus allowing the. at- 

 tacking waves to reach the cliff face with much less diminished force than 

 if no subsidence took place. Thus the platform might be more inclined 

 and the cliff lower than if they had been cut during a long still-stand, 

 or the platform and cliff might be subdivided in many little treads and 

 rises. But in order to explain the growth of a wide-lagoon barrier reef, 

 R, offshore from a steep-sloping island, AB, after a series of small treads 

 and rises had been cut during intermittent subsidence instead of a broad 

 platform during a stationary period, abrasion must have begun farther 

 down the volcanic slope, as at C, and ceased at D, so that reef growth 

 beginning at D could rise to R; and this involves a greater total subsi- 

 dence than the preceding supposition requires. 



It is possible that occasional more rapid subsidences during the abra- 

 sion of the benched slope, CD, permitted the temporary establishment of 

 reefs, Q, which grew upward until they were overwhelmed by outwashed 

 detritus, whereupon abrasion would again set in. But eventually, when 

 the discharge of detritus was decreased by reason of the advanced stage 

 of island erosion then reached, subsidence, especially accelerated subsi- 

 dence, w^ould gain the upper hand ; the reefs thereupon established v^ould 

 thereafter persist, and thus the present condition of a reef-encircled, em- 

 bayed, non-clift island would be attained. 



This evidence for Darwin's theory is manifestly of a theoretical nature ; 

 but the evidence supplied by unconformable reef contacts is also theoret- 

 ical. The difference between the two lines of evidence is not so much that 

 one is theoretical and the other is not, as that the principles and processes 

 involved in one are here applied in a novel manner and are not yet gen- 

 erally accepted, while those involved in the other have long been familiar 

 in geological science and are universally accepted, even though they have 

 not been usually utilized in the coral-reef problem. 



Absence of Reefs on Coasts of Emergenck 



reefless coasts in the australasian archipelagoes 



Conditions favorable to reef establishment are not, as a rule, provided 

 on coasts of recent emergence, because the simple shoreline and the smooth 

 adjacent sea-bottom are occupied for scores of miles together by uncon- 

 solidated sediments on which corals can not grow. In the absence of 

 reefs such coasts are attacked by the waves and cut back, except that 

 where large rivers mouth their deltas may be built forward. Stationary 



