148 THE ROYAL SOCIETY OF CANADA 



Caporn suggests that the results indicate a three-factor difference 

 between the parents. But this interpretation is admittedly incapable 

 of proof from the available data. 



Table II shows the length of the period from planting to ripen- 

 ing of those varieties for whose crosses in various combinations the 

 F3 records are somewhat detailed. It will be observed that the length 

 of this period is much greater than in 1917 (compare with Table I). 

 This is largely due to the fact that on account of the cold spring the 

 seeds did not germinate until nearly a month after planting. It will 

 also be observed that the ripening times overlap to a greater extent 

 than in 1917. This is particularly true of the later varieties for the 

 reasons already mentioned. 



Table III gives the data obtained on the F3 plants of the cross 

 between the two earliest parents, together with the range of the pure 

 parental varieties for comparison. The second column gives the 

 length of the growth period of the F2 parent plant of each family in 

 1917. The F2 plants, whose offspring were to be grown, were selected 

 in such a way that there would be a few representatives of each day 

 of the F2 variation. This column should be compared with Table I. 

 For convenience in comparison, the families are arbitrarily arranged 

 in the order in which their earliest plants ripened. As will be explained 

 more fully below, this does not mean that the order is a genetic one. 



In connection with this table, the following points should be noted. 

 Some of them are much better illustrated in later tables, but it is 

 desirable to note the uniformity in all crosses. 



1. One family (No. 1) began to ripen before the earlier parent, 

 and all its members were ripe before the later individuals of this 

 parent. The difference is not very great and may not be due to 

 hereditary causes, but as a very large number of individuals of the 

 parental variety were recorded in order to determine the extremes of 

 variation, it seems probable that there is a true hereditary basis for 

 the difference. In the same way a few families began to ripen after 

 the later parent. In these late families the range of variation was less 

 than that of the parent, partly because not nearly so many individuals 

 were grown, and partly because the rust attack pushed forwa:rd some 

 individuals. 



2. In several families the range of variation in the ripening time 

 nearly coincides with that of the earlier parent, and in several other 

 families it nearly coincides with that of the late parent. 



3. Many families are intermediate between the parents. Ap- 

 parently, all possible degrees of earliness are shown in different 

 families. To what extent this appearance is produced by environ- 



