W. M. Davis — Shaler Memorial Study of Coral Reefs. 263 



reef is now forming, must then a second time be uplifted by its 

 amount ; and so on. It is far beyond the present capacity of 

 geological investigation to determine whether such widespread 

 uplifts have or have not taken place contemporaneously with 

 the formation of the earlier and later reefs ; hence this theory 

 is at present not only embarrassing but unverifiable. 



On the other hand, if local submergence is due to local 

 ocean-bottom subsidence, more or less closely compensated by 

 neighboring ocean-bottom uplift, then each island or group of 

 islands is submerged when it subsides and by about the amount 

 of its subsidence. Submergences of different dates do not, 

 under this supposition, involve any complicated difficulties, and 

 the explanation of submergence is greatly simplified. Darwin's 

 theory thus explains the reefs and the embayed shorelines of 

 a group of barrier-reef islands by a simple crustal deformation 

 that need not extend far beyond the area occupied by the island 

 group. This simple theory does not hold extravagant move- 

 ments in extensive regions elsewhere responsible for changes of 

 moderate amount observed in coral-reef areas ; and it does not 

 submerge or disturb all the other coasts of the world, conti- 

 nental as well as insular, every time that a group of coral reefs 

 is formed. Darwin's theory, therefore, has the recommendation 

 of being a simple cause for a simple effect. But recommen- 

 dation is not demonstration, any more than probability is proof. 

 Nevertheless, I believe that the above considerations suffice to 

 advance Darwin's theory of subsidence to so high a degree of 

 probability that it outranks all other coral-reef theories. If 

 general changes of ocean level sometimes take place, depend- 

 ent either on changes of climate or on movements of the ocean 

 floor outside of the coral zone, the submergences or emergences 

 thus produced appear to be subordinate to those produced by 

 subsidences or uplifts of the ocean bottom in coral-reef regions 

 themselves. 



The Problem of Atolls. — The conversion of barrier reefs 

 into atolls by a continuation of the process that converts fring- 

 ing reefs into barrier reefs is a highly probable matter ; for, as 

 already pointed out, it would be unreasonable to suppose that 

 this process, whatever it is, should always have stopped before 

 the central islands of barrier reefs were wholly submerged, 

 and should never have worked in neighboring areas where 

 reefs of identical form but without a central island are given 

 another name. And now that the converting process has, with 

 so high a degree of probability, been shown to be long-con- 

 tinued subsidence of the ocean bottom in the region concerned, 

 and not a change of ocean level during the glacial period 

 or an uplift of the ocean bottom in some other region, it appears 

 reasonable to explain atolls in the same way as barrier reefs. 



