VARIATION AND HEREDITY 223 



mass, the curve for any random lot of individuals is a 

 normal curve. Artificial selection, however, picks 

 out not only extreme individuals, but, at the same 

 time, extreme " pure lines." Since the latter are, 

 themselves, unmodified by selection, a rational 

 explanation is at once offered to account for the 

 puzzling fact that a race, such as the sugar beet, 

 may respond promptly to selection, but refuse to 

 respond at all beyond a certain point. 



Similar results have been obtained for protozoa 

 (Paramecium), in which the habit of asexual re- 

 production enables the experimenter to isolate his 

 pure lines and do away with the disturbing effects 

 of crossing. A mutation or discontinuous variation 

 might be explained, perhaps, as a coming into 

 existence of a new " pure line," though by what 

 method, or from what cause, we at present cannot 

 say. 



Unit Characters and Mendelian Inheritance. 

 The statistical study of inheritance deals only with 

 organisms in the mass, and its conclusions are in- 

 applicable in individual instances. The more pre- 

 cise its results, the more abstract they become. But 

 heredity is, after all, a personal matter, and it is 

 of the greatest interest to discover, if possible, 

 just what are the laws that determine the distri- 

 bution of inherited qualities in individual instances. 

 Within recent years very remarkable progress has 

 been made in the solution of this problem, although, 

 as is so frequently the case, the increase in knowledge 



