164 EVOLUTION 



of natural selection by actual evidence tliat 

 the survivors survive and the eliminated are 

 eliminated because of some differentiating 

 peculiarity or peculiarities. Hence a few 

 more examples may be given. 



Poulton fastened 600 pupse of the tortoise- 

 shell butterfly to nettles, tree-trunks, fences, 

 walls, and so on. At Oxford, the mortality 

 was 93 per cent., and the only pupse that 

 survived were on nettles, where they were 

 least conspicuous. In the Isle of Wight, the 

 elimination was 92 per cent, on fences, as 

 against 57 per cent, among nettles. Here, 

 again, there was definite evidence of dis- 

 criminate elimination. 



Professor Crampton's very careful research 

 on the pupae of a Saturnid moth, proved dis- 

 criminate elimination, and yielded also this 

 interesting result, that the selected characters 

 (for the most part concerned with dimensions 

 and proportions) were not such as would have 

 appeared to be directly or indirectly "useful" 

 to their possessors, though they were de- 

 monstrated to have the high utility of deter- 

 mining survival — which is indeed, for the 

 evolutionist, the final criterion of utility. 



Professor Bumpus took 136 benumbed 

 house-sparrows into his laboratory, where 

 72 revived and 64 succumbed. There were 

 general differences — of a somewhat subtle 



