174: Davis — Dana's Confirmation of 



made only of a few leading contributions. During the voyage 

 of the " Challenger," Murray saw the reefs of the Fiji islands ; 

 he could not explain them by Darw,in's theory of subsidence, 

 and he therefore afterwards* replaced that theory by what may 

 be called the theory of outward growth, with its provisos of the 

 organic upbuilding of submarine banks to serve as foundations 

 for atolls, and of the production of lagoons by the removal 

 of the inner part of barrier reefs by solution, no movement 

 of subsidence being included in this theory. Still later, A. 

 Agassiz, in the reports of his world-wide investigation of coral 

 reefs, f emphasized the occurrence of uplifted coralliferous 

 limestones, which might be worn down and dissolved away 

 while new fringing reefs grew around them, thus producing 

 barrier reefs and atolls in association with elevation instead 

 of with subsidence ; he suggested farther that even volcanic 

 islands might be worn down and obliterated, so that their 

 barrier reefs would survive as atolls. At the same time, he 

 reintroduced the idea of wide sea-cut platforms on the margin 

 of volcanic islands, and the growth of a comparatively thin 

 veneer of coral on the outer edge of the platform, thus pro- 

 ducing a barrier reef without subsidence, elevation, or solution. 

 Wharton went farther^ in suggesting that a volcanic island 

 might be worn down to a depth of 20 or 25 fathoms by marine 

 agencies, thereby producing a submarine bank on which 

 upgrowing corals could form an atoll, and thus, like Agassiz, 

 accounted for atolls without postulating either subsidence, 

 elevation, or solution. 



Each of these theories has been more or less favorably 

 received, for each one has the same kind of recommendation 

 as that which sufficed to lead the scientific world to adopt 

 Darwin's theory ; that is, each theory explains the group of 

 facts that it was made to explain, provided its postulates are 

 accepted. Such a measure of success is of course commenda- 

 tory, but it is not sufficient for the establishment of that high 

 degree of probability which is recognized in geological science 

 as demonstration. Even when only a single theory 7 is under 

 consideration, it must evidently do something more than explain 

 the things that it was made to explain, before it deserves 

 unquestioned acceptance. When several theories are advanced 

 to explain a single group of facts and each theory succeeds in 

 explaining what it was made to explain, the need of some inde- 

 pendent means of verification is still more manifest. Indepen- 

 dent verification of a theory is found usually in one of two ways, 



*Proc. Eoy. Soc. Edinb., ix, 1880, 505-518. 



f Bull. Mus. Comp. Zool., xxxiii, 1899; Mem. Mus. Comp. Zool., xxviii, 

 1903. 

 % Nature, iv, 1897, 390-393. 



