PRESIDENTIAL ADDRESS. 251 



eliminated. The vertical line represents the bar to inter- 

 crossing which permits divergence. Do away with that 

 bar and the divergence ceases. The individuals on either 

 side of it interbreed, their characters will be blended, 

 and there will be monotypic evolution. The race will 

 perkaps increase in combined strength and cunning; but 

 there will be no divergence into two races, the one strong 

 and the other cunning. 



Thus to render possible either transmutation or diver- 

 gence, under natural selection through elimination, varia-- 

 tion is essential. For Darwinism variation is a sine qud 

 non. But Darwinism, as such, is not bound to account for 

 the origin of variations. It is sufficient that they occur. 



Evidence of their occurrence is rightly demanded by the 

 student of bionomics. And in answer to this demand Mr. 

 A. R. Wallace has collected a large body of evidence in 

 the chapter on " The Variability of Species in a State of 

 Nature " in his Darivinism. So successful has he been in 

 showing that variations in size '' usually reaching 10 or 20, 

 and sometimes even 25 per cent, of the varying part," and 

 occurring in 5 to 10 per cent, of the specimens examined,, 

 do occur under nature, that one's faith is, perhaps, a little 

 shaken in the inexorable efficiency of the struggle for 

 existence as an eliminating agency. In any case it em- 

 phasizes a fact on which I have elsewhere endeavoured to 

 lay stress, that a variation must reach a certain amount, 

 and be of a fatally deleterious character, before it becomes 

 of elimination value. We have been apt to suppose that 

 a species is so nicely adjusted to its surrounding conditions 

 that all variations from the type, unless of a very insignifi- 

 cant character, would be rapidly and inevitablj^ weeded 

 out. This, it is clear, is not true at any rate for some 

 species. 



