OTHER THEORIES OF SPECIES-FORMING. 247 



being so similar and simple that any bent nail is able to pick 

 them. 



Seebohm 18 criticises Romanes's theory of physiological 



selection, which should better be called physiological 



Seebohm's isolation, as demanding an almost impossible 



criticism of coincidence of conditions to make it work. 



physiologic se- 

 lection. Formulated as nearly as it can be in a single 



sentence, Romanes has defined physiological selection as 

 follows : "Wherever, among all the possible variations of 

 the highly variable reproductive system there arises toward 

 any parent form any degree of sterility which does not 

 extend to the varietal form, there a new species must neces- 

 sarily take its origin." Seebohm points out that this is 

 exactly a condition that can rarely, if ever, occur, for to 

 bring it about we must presuppose : 



"ist. The special variation of the reproductive organs 

 must occur in two individuals, otherwise the possible an- 

 cestor of the new species would leave no descendants. 



"2d. It must occur at the same time in both. 



"3d. It must occur at the same place. 



"4th. The two individuals must be of opposite sexes. 



"5th. They must each of them possess some other varia- 

 tion, or their progeny would not differ from that of the rest 

 of the species. 



"6th. The variation must be the same in both or appear 

 simultaneously in the majority of their children, otherwise 

 it would be swamped by interbreeding within the physio- 

 logically isolated family." 



Romanes's theory has also been strongly criticised by 

 Wallace 19 and by Karl Jordan. 20 Wallace contends that 



Other criticisms " no ^ orm * in f ertmtv or sterility between the 

 ofEomanes's individuals of a species can be increased by 



eory> natural selection unless correlated with some 



useful variation, while all unfertility not so correlated has a 

 constant tendency to effect its own elimination. But the 



