Darwin's Theory of Coral Reefs. 175 



all the better if it is found in both of these ways. New kinds 

 of facts may be discovered after a theory has been invented ; 

 if such facts, as well as the original facts, can be reasonably 

 accounted for by the theory, the probability of its correctness 

 is greatly increased. Or new consequences may be deduced 

 from the theoiy, thus indicating the occurrence of specified 

 kinds of facts not before noticed ; if such facts are then found 

 in their designated positions, the probability of correctness of 

 the theory that thus gives its inventor the power of prediction 

 is immensely increased. 



. The possibility of the development of new atolls by the wear- 

 ing down of older uplifted atolls, or of new barrier reefs by the 

 wearing down of uplifted older reefs, is one of the interesting 

 complications of the simple theory of subsidence that Darwin 

 touched upon briefly,* but that Agassiz' abundant observations 

 have emphasized. This possibility, however, in no wise affects 

 the origin of the earlier atoll or reef, whatever that may have 

 been. True, Agassiz doubted or denied that the uplifted 

 masses of "Tertiary coralliferous limestone" which he repeatedly 

 found in the Paumotus were atolls of an earlier date, but no 

 other adequate explanation of the origin of these great trun- 

 cated conical masses, rising several thousand feet over a sub- 

 marine plateau, has been suggested. Agassiz likewise denied 

 the barrier-reef origin of certain uplifted coralliferous lime- 

 stones of the Fiji islands, and thus differed from Gardiner,f 

 who on the grounds of his own observations accepted this 

 interpretation of their formation. If the uplifted limestones 

 of the Paumotus and the Fijis are eventually proved to be 

 atolls and barrier reefs of an earlier period than the present, a 

 general theory that explains them will have all the more merit, 

 in that it will then apply to more remote as well as to recent 

 geological time. 



Murray's theory of the formation of atolls by upgrowth from 

 submarine banks of proper depth is eminently possible, if the 

 banks can be provided in sufficient number ; but possibility is 

 not proof. The depth of the ocean is not supposed to be 

 changed by subsidence in this theory ; hence if subsidence is 

 shown to have taken place in the production of barrier reefs, 

 the theory that accounts for neighboring atolls without subsi- 

 dence will be rendered less probable. 



The development of a foundation for atolls by the trunca- 

 tion of a volcanic island in the manner indicated by Wharton 

 is eminently possible, provided no change of level takes place 

 during the progress of the work ; but it is inconceivable that 

 floating coral larvae should abstain from establishing themselves 



* Coral Eeefs, 1842, 55, 146. 



fProc. Phil. Soc. Cambridge, ix, 1898, 417-503. 



