256 Heredity and Environment 



was made to increase the egg-laying capacity of hens by selecting 

 for breeding in each generation only those which had a high 

 record for egg production. It was found that certain "blood 

 lines" produced a larger number of eggs than other lines, but 

 by no amount of selection was it possible to increase the egg pro- 

 duction within any line. 



On the other hand Castle has strenuously defended the im- 

 portance of the selection of fluctuations in "building up" a char- 

 acter. From among the descendants of piebald rats and rab- 

 bits he has selected through many generations individuals show- 

 ing the largest and also the smallest extent of color in the 

 coat and he has thus produced one line which is nearly all-black 

 and another nearly all-white. He maintained that not only in- 

 herited characters, but also their factors are variable and that by 

 means of selection of plus or minus variations the mode, or mean, 

 of a character may be shifted in one direction or the other. This 

 is the old view which flourished before a distinction was made 

 between fluctuations and mutations. Breeders have long been 

 acquainted with similar results of selection from a mixed popu- 

 lation containing different hereditary lines ; however Castle has 

 been careful to employ as pure a race of rats and of rabbits as he 

 could obtain, but it is not possible to get as pure a race of these 

 animals or of any organisms in which cross fertilization occurs 

 as in the case of self-fertilizing plants such as beans. Johannsen 

 defines a "pure line" as "all individuals which are derived from a 

 single, absolutely self-fertilizing, homozygous individual" and 

 within such a pure line he maintains that selection is unable to 

 change any character. 



Adherents of the "pure line" hypothesis explain Castle's re- 

 sults in one of two or three ways : either his material may not have 

 been genetically pure, or mutations may have occurred during the 

 course of his experiments and his selection served merely *o iso- 

 late distinct hereditary lines already present; or, more probably, 

 extent of color in his animals may depend upon multiple factors, 

 or modifying factors, which are more numerous in some individu- 



