OBGANIC EYOLUnON — ^THE FACTORS 105 



" But while I cannot admit my failure to understand 

 Weismann, I confess that I do not understand Dr. 

 Romanes. How when natural selection, direct or 

 reversed, is set aside, the mere cessation of selection 

 should cause decrease of an organ ii-respective of the 

 direct effects of disuse, I am unable to see. Cleai-er 

 conceptions of this matter would be reached if, instead 

 of thinking in abstract terms, the physiological processes 

 concerned were brought into the foreground. Beyond 

 the production of changes in the sizes of parts by the 

 selection of fortuitously-arising variations, I can see 

 but one other cause for the production of them — the 

 competition among the parts for nutriment. This has 

 the effect that active pai-ts are well supplied and grow, 

 while inactive parts are ill supplied and dwindle. This 

 competition is the cause of ' economy of growth ' ; this 

 is the cause of decrease from disuse ; and this is the 

 only conceivable cause of that decrease which Dr. 

 Romanes contends follows the cessation of selection. 

 The three things are aspects of the same thing. And 

 now, before leaving this question, let me remark on 

 the strange proposition which has to be defended by 

 those who deny the dwindling of organs by disuse. 

 Their proposition amounts to this — that for a hundred 

 generations an inactive organ may be partially denuded 

 of blood all through life, and yet in the hundredth 

 generation wiU be produced of just the same size as in 

 the first" (pp. 67-8). 



I have already set forth my reasons for believing 

 that, as regards any structure, retrogression follows 

 Cessation of Selection, because there is a greater 

 tendency in every organism to vary ancestor-wards than 

 in any other direction, and that therefore retrogression 

 in such a case is nothing other than a continued lapsing 

 of inborn variations, the more recent in the phylogeny 

 first, the more ancient in succession later, till there is 

 at length a reversion to the remotest condition, the 

 condition when there was no such structure. But as 



