The Evidence from Variation. 163 



why variation is itself hereditary, for the variation of 

 any particular cell will cause adjacent or related cells to 

 throw off gemmules, and thus to produce variation in 

 successive generations. 



We can also understand why specific characters should 

 be more variable than generic characters; why the spe- 

 cies of large genera should vary more than the species of 

 small genera; why a part developed in an unusual way 

 or to an unusual degree should show a marked tendency 

 to vary, and why secondary sexual characters should ex- 

 hibit a similar tendency. 



Unless our theory is true, what possible reason can 

 there be why a part which is excessively developed in 

 males should vary more than a part which is similarly 

 developed in females alone, or why the males of our 

 higher domesticated animals should be more variable 

 than the females? Its power to deal with and interpret 

 special cases of this kind separates our theory from all 

 other attempts to explain the phenomena, and seems to 

 show that there can be but one choice between it and 

 any other explanation which has ever been proposed. 



If we accept Darwin's view that variations are purely 

 fortuitous, there are certain grave difficulties which must 

 prevent us from giving the theory of natural selection 

 unqualified acceptance as an adequate and complete ex- 

 planation of the origin of species. 



Natural selection can rarely lead to permanent modi- 

 fication unless many individuals tend to vary in nearly 

 the same way at about the same time, and if variation is 

 fortuitous the chance against this is very great indeed. 

 While there is no reason to doubt that natural selection 

 might bring about all the changes which have led to the 

 formation of a complicated organ, by the preservation of 

 fortuitous variations, if time enough were given, there is 



