BASIS OF THE PKOBLEM 69 



problem of the manner in which germinal changes arise. A refer- 

 ence first to what are known as ' pure line ' investigations and 

 afterwards to Mendelian phenomena will illustrate what is known 

 as to the nature of existing germinal differences ; for until we 

 have some such information, we cannot profitably ask how exist- 

 ing germinal differences arise. The best known ' pure line ' 

 experiments are those carried out by Johannsen with beans. 

 The flowers of beans fertilize themselves ; the offspring, there- 

 fore, have approximately the same germinal constitution. The 

 offspring constitute a ' pure line ', for by a ' pure line ' is meant 

 a group of children which are the offspring of a single parent. 

 The character selected for investigation was weight and it was 

 found that, if beans were collected from a bean-field and weighed, 

 every gradation occurred between a minimum of about 20 centi- 

 grams and a maximum of about 90 centigrams. When the beans 

 were separated into three classes heavy, medium, and light 

 sown and plants raised from them, the average weight of the beans 

 produced by the plants derived from the heavier seed was greater, 

 though not proportionately greater, than the mean weight of all 

 the beans, and that the average weight produced by the plants 

 derived from the lighter seed was less, though not proportionately 

 less, than the mean weight of all the beans. There was, in fact, 

 a certain regression on the part of the heavy and light classes to 

 the mean weight. A similar tendency to regression to the mean 

 can be observed when other characters are similarly studied. 

 The average stature, for example, of the offspring of tall parents 

 is slightly nearer the mean stature of the race than the average 

 stature of the parents. 



This tendency to regression to the mean has long been known, 

 but it was not until Johannsen proceeded farther and investi- 

 gated inheritance within a ' pure line ' that it was understood. 

 In the experiment described, no attention was paid to the ' pure 

 line '. When Johannsen separated the beans produced by the 

 self-fertilization of a single plant, divided them into heavier, 

 medium, and lighter classes, sowed them and weighed their 

 progeny, he discovered a very interesting fact. The average 

 weight of the progeny of a heavy bean and of a light bean belong- 

 ing to the same ' pure line ' was the same. The conclusion to be 

 drawn from this result is that the differences in weight between 

 the offspring of a single self-fertilized plant are due to differences 



