Control of Heredity: Eugenics 255 



and that thereafter selection serves only to maintain the charac- 

 ter at its high level but not to advance it. The probable explana- 

 tion of this fact has been found only in recent years. The re- 

 searches of deVries, Johannsen, Jennings, Tower, Pearl, Mor- 

 gan and others have shown that in some cases at least selection 

 merely isolates mutants or distinct hereditary lines which are al- 

 ready present in a mixed population but that it does not "build 

 up" characters nor produce new mutations; in short it does not 

 create the variations on which it acts. 



Johannsen found that from a single species of beans he was 

 able, by keeping the progeny of each individual bean separate from 

 the others, to isolate 19 different "pure lines," each differing in 

 certain respects from every other line. These lines were not 

 created by selection but were merely sorted out of the general 

 species where they existed already. He further found that when 

 extremely large or small individuals from any pure line were se- 

 lected and propagated, none of the progeny showed that charac- 

 ter in a still more extreme degree but all merely fluctuated within 

 the original extremes of that line. He concludes therefore that 

 selection within a pure line is absolutely without, effect in modify- 

 ing any character in the offspring of that line. 



Jennings found that different races or pure lines of Paramc- 

 cium differ in size, structure and rate of division, and that these 

 differences are "as rigid as iron." With respect to average length 

 of body he was able to isolate eight pure lines which constantly 

 differed more or less from one another. Within each of these 

 lines there was considerable fluctuation in size, but he was un- 

 able by selecting extremes to increase these fluctuations, the pro- 

 geny of any pure line always fluctuating about the mean of that 

 line (Fig. 22). 



Similarly Tower found in his studies on the potato beetle that 

 he was unable to shift the mean or the extremes of any character 

 by selection of extreme forms of an inbred line. 



Pearl also made an extensive study of the records of breed- 

 ing experiments extending over many years in which the attempt 



