Iitfluence of Selection 205 



down by the rivers that became greater in amount during these 

 years. He tested this view by keeping crabs in water contain- 

 ing the same mud in suspension. The crabs that died were on 

 the whole broader than those that survived. Weldon suggests 

 that the frontal breadth is correlated with the breadth of the 

 opening into the gill chamber, and the latter determines the 

 amount of mud that enters, and in this way the ehmination is 

 produced. 



Should these conclusions of Weldon be estabhshed, they show 

 that a change in the external conditions may cause the ehmina- 

 tion of certain individuals of the species, and in this way affect 

 the mean of the survivors. In other words, natural selection 

 occurs within the hmits of the species. But it does not follow 

 from this, as Weldon takes for granted, that if the same process 

 of ehmination were continued, a new species would be evolved. 

 Raising the mean to the highest point attainable within the hmits 

 of the species does not necessarily lead to the formation of a new 

 species. This point has already been discussed in other connec- 

 tions. 



Crampton has studied variation and ehmination in the moths 

 of Philosamia Cynthia. In the first generation that he studied 

 there was a high rate of mortahty, and since the pupae were col- 

 lected in their normal environment, before they had been sub- 

 jected to the winter frost, the death rate must have been con- 

 nected with some inherent weakness or with conditions that they 

 met with during their normal hfe. Out of 942 pupae collected, 

 there were 628 that were dead, so that only 329 "selected" 

 individuals remained. The causes of the death of the 628 

 pupae are not known. One might suspect that they had become 

 parasitized ; but Professor Crampton tells me that he examined 

 them to see if this was the case, and found no evidence of the 

 sort. It does not seem probable that the external conditions 

 to which the pupae were exposed had anything to do with the 

 high death rate, and its causes are probably to be found in the 

 organization of the animal. Whether bacterial disease, or dis- 

 ease of some other kind, is responsible for the results, was not 



