24 CALIFORNIA ACADEMY OF SCIENCES. 



acting l)elow 100, and so isolate the effects due to the 

 mere withdrawal of selection. By the conditions of our 

 assumption, all variations above 100 are eliminated, 

 while below 100 indiscriminate variation is permitted. 

 Thus, the selective premium upon variation 99 being no 

 greater than upon 98, 98 would have as good a chance 

 of leaving offspring which would inherit and transmit 

 this variation as Avould 99; similarly, 97 would have as 

 good a chance as 98, and so on." He then shows how 

 there would be a constant tendency toward reduction in 

 the part, but that the greater the reduction the less pos- 

 sibility of future reduction would remain. ''Thus," he 

 says, "theoretically the average would continue to 

 diminish at a slower and slower rate, until it comes to 

 50, where the chances in favor of increase and diminu- 

 tion being equal, it would remain stationary." 



Prof. Romanes then gives examples of parts which he 

 thinks have degenerated through cessation of selection. 

 He considers the cases where the phylogenic stages are 

 omitted in the developing embryo to be instances in 

 point, and argues that such omissions cannot be ex- 

 plained by economy of growth, for in allied forms where 

 economy would be'equally operative the structure per- 

 sists. Neither can the absence of such parts be due to 

 disuse. Prof. Komanes contends, for they were not gen- 

 erally produced by use. The case of hard coverings, 

 which are developed by natural selection as a protection 

 to certain animals, and afterwards lost when their period 

 of usefulness is past, is also cited as an example of de- 

 generation without disuse. 



Pammixis, then, is not new as a theory, nor is it 

 the only available explanation of degeneration. Ro- 

 manes himself suggested three alternatives. Besides 

 reversal of selection, he has stated two other factors, as 

 follows: "The first of them is inheritance at earlier 



