CORAL REEFS, AND ISLANDS. 289 



caus<z> true causes ; in the next place, we should 

 be prepared to show that the assumed causes of the 

 phenomena are competent to produce such phe- 

 nomena as those which we wish to explain by them ; 

 and in the last place, we ought to be able to show 

 that no other known causes are competent to produce 

 these phenomena. If we can succeed in satisfying 

 these three conditions we shall have demonstrated our 

 hypothesis ; or rather, I ought to say, we shall have 

 proved it as far as certainty is possible for us ; for, 

 after all, there is no one of our surest convictions 

 which may not be upset, or, at any rate, modified by 

 a further accession of knowledge." l And again : 

 " Every hypothesis is bound to explain, or at any 

 rate, not be inconsistent with, the whole of the facts 

 which it professes to account for ; and if there is a 

 single one of these facts which can be shown to be 

 inconsistent with the hypothesis, the hypothesis falls 

 to the ground it is worth nothing. One fact with 

 which it is positively inconsistent is worth as much, 

 and as powerful in negativing the hypothesis, as five 

 hundred." Tested by these obligations, it hardly 

 seems clear that the new hypothesis is competent to 

 supersede the old one. At least it is by no means so 



1 " On our Knowledge of the Causes of the Phenomena of 

 Organic Nature." By Professor J. Huxley. London, 1863, 

 P- 135. 



U 



