350 The Theory of Natural Selection 



and in the survivors. In this way he hoped to find a mathe- 

 matical index of the effectiveness of natural selection. A 

 great deal of material was thus accumulated in support of 

 Darwin's theory. 



Later, however, more critical studies along the same 

 lines threw some doubt on the interpretations of the facts. 

 We have already seen (page 179) that Villem Johannsen, 

 the Danish botanist, experimented with self-pollinating 

 beans, and showed that selection has no effect whatever in 

 modifying the constitution of the variety. Johannsen suc- 

 ceeded in breaking up a mixed population of beans into 

 several distinct strains or pure lines. Then he showed that 

 although there is individual variation within each pure line, 

 the strain has its own mode and its own limits of varia- 

 tion (Fig. 89). 



The net result of these experiments was to show that 

 it is possible by means of systematic selection to establish 

 a strain having a mode or average quality differing from that 

 of the ancestors, but // is only the average that is shifted, 

 not the essential characteristics of the population. It is 

 possible, for example, by means of selection to obtain a 

 group of larger animals or of smaller animals and to main- 

 tain the extreme size consistently in successive generations; 

 but it is not possible by this process to transcend the original 

 limits of variation of the parental type. Selection, in other 

 words, acts upon the average of a mixed population, but 

 does not affect the genotype of a pure line. This is in agree- 

 ment with the well known experience of breeders that im- 

 provement results very rapidly from consistent selection for 

 a few generations, and then stops almost abruptly. There is 

 a point beyond which selection does not carry improvement 

 further. 



What is Selected 



Darwin assumed, from the well known results of arti- 

 ficial selection, that not only does the mean or average 



