WHITE— STUDIES OF INHERITANCE IN PISUM. 557 



In F2, the usual result is a complete or almost complete inter- 

 grading series with occasional small breaks. The classification of 

 such a series into early, intermediate and late is generally arbitrary, 

 though often based on the blooming period of the two parents and 

 the Fj when these are grown under the same or similar conditions. 

 With such a method of classification, Tschermak obtained from 

 crosses involving seven diflierent varieties, the following results : 



Actual, 60 early: 190 intermediate : 88 late, 

 Expected, 63.3 early: 190 intermediate : 84.4 late, 

 Ratio, 3 : 9 : 4. 



Keeble and Pellew from crosses involving two varieties obtained 

 63 early: 128 intermediate:! late. Lock (54), classifying them in 

 three 5-day frequency classes, obtained 63 early: 186 intermedi- 

 ates : 279 late. 



Lock (54), Tschermak and Hoshino (40.5) have noted an Fo 

 association between colored flowers and lateness on the one hand 

 and white flowers and earliness on the other. The modifying rela- 

 tion or coupling, whichever it may be, is only partial, as the follow- 

 ing Fo results show : 



Class. Early. Intermediate. Late. 



Flower color white purple white purple white purple 



Tschermak 25 22 48 94 4 64 



Ratio I : .88 : i : 1.96 i : 1.16 



White flowers : purple flowers jy : 180 or i : 2.34. 



Lock 29 34 79 107 104 175 



Ratio I : 1. 17 I : 1.35 i : 1.68 



Purple flowers : white flowers 383:123 or 3.13:1. 



The expected relation of the purple- to the white-flowered class, 

 providing there was no couphng, is of course 3 : i in each of the 

 classes — early, intermediate and late. 



Tschermak (85) and Keeble and Pellew (49) have obtained some 

 curious results regarding the relation of tallness and dwarfness to 

 the time of flowering. In the one case (Fig. 3B) given by 

 Tschermak the F^ is tall and almost as late flowering as its late- 

 flowering parent. In F. 32 tails and 10 dwarfs result. Classifying 



