RECENT CRITICISMS AND HYPOTHESES. 561 



Here, and throughout the arguments of those who accept 

 the hypothesis of Panmixia, there is an unwarranted assump- 

 tion — nay, an assumption at variance with the doctrine in 

 support of which it is made. It is contended that in such 

 cases as the one given there will, apart from any effects of 

 disuse, be decrease in the disused organs because, not being 

 kept by Natural Selection up to the level of strength pre- 

 viously needed, they will vary in the direction of decrease; 

 and that variations in the direction of decrease, occurring in 

 some individuals, will, by interbreeding, produce an average 

 decrease throughout the species. But why will the disused 

 organs vary in the direction of decrease more than in the 

 direction of increase? The hypothesis of Natural Selection 

 postulates indeterminate variations — deviations no more in 

 one direction than in the opposite direction: implying that 

 increases and decreases of size will occur to equal extents 

 and with equal frequencies. With any other assumption the 

 hypothesis lapses; for if the variations in one direction 

 exceed those in another the question arises — What makes 

 them do this? And whatever makes them do this become? 

 the essential cause of the modification : the selection of 

 favourable variations is tacitly admitted to be an insufficient 

 explanation. But if the hypothesis of Natural Selection 

 itself implies the occurrence of equal variations on all sides 

 of the mean, how can Panmixia produce decrease? Plus 

 deviations will cancel minus deviations, and the organ will 

 remain where it was.* 



* In a letter published by Dr. Romanes in N'ature, for April 26, 1894, he 

 alleges three reasons why "as soon as selection is withdrawn from an organ 

 Iho minus variations of that organ outnumber the pliis variations." The first 

 is that " the survival-mean must descend to the birth-mean." The interpre- 

 tation of this is that if the members of a species are on the average born 

 with an organ of the required size, and if they are exposed to natural selec- 

 tion, then those in which the organ is relatively small will some of them die, 

 and consequently the mean size of the organ at adult age will be greater 

 than at birth. Contrariwise, if the organ becomes useless and natural selec- 

 tion does not operate on it, this difference between the birth-mean and the 

 survival-mean disappears. Now here, again, the plus variations and their 



