Heredity in Populatiom and in Pure Lines. 



weights of the daughter classes, a regrer.sion table may be formed 

 similar to the one in Fig. 2. Such a table is shown in Fig. 3. 



The results in this case are markedly different from those 

 secured when the material was studied as a population. Here 

 we do not have a fine example of regression as 13 shown in Figs. 

 1 and 2. It must not be forgotten that both Figs. 2 and 3 repie- 

 sent the facts, but in Fig. 3 the data are s) presented that wo 



Fig. 3. Relation of offspring to parent bean seeds in recr.irJ to weight as studied by 

 Johannsen in pure lines. Compare with Fig. 2. The dotted ;Lie ij, the hne of regresdon. 



see the sequence of the two related means i. e. (means of mother 

 and daughtei classes) in th^ various cases. There is evidently a 

 certain amount of regression expressed in the table, but appar- 

 ently little can be learned from it in regard to the value of selec- 

 tion. 



STUDIES OF RELATIVE WIDTH. 



Studies were also made upon the relative width of bean 

 seeds. The width, in millimeters, was divided by the length 



