[THOMPSON] EARLINESS AND LATENESS IN WHEAT 157 



Last year the cross A x E had not been carried to the second 

 generation. It was hoped that on account of the very wide difference 

 between the parents, this cross would furnish important evidence 

 in regard to the theoretical interpretation of the results as a whole. 

 During the past summer, the F2 plants were raised, and the results 

 are given in Table VII. On account of the influences mentioned the 

 difference between the parents this year is much reduced. Con- 

 sequently, the results are not as valuable as had been hoped. Never- 

 theless, they are exactly the same as were found in all the crosses 

 involving smaller differences. In all probabihty if the difference had 

 been as great as in 1917, the results would still have been the same — 

 a variability extending nearly from the lower extrem.e of A to the 

 higher extreme of E. 



The second hybrid generation of several other crosses were also 

 raised. We now have the results of crosses involving differences as 

 small as three days and as big as 22 days, and various sizes between. 

 Furthermore, the crosses involving small differences were made be- 

 tween parents at various places in the whole range of variabiHty. 



In regard to the length of the period from planting to flowering 

 (heading) records were taken on all the crosses mentioned. In a few 

 cases the records were assembled in tables similar to those which 

 accompany this paper, but as no marked differences from the results 

 in connection with ripening were found, this work was not carried 

 further. 



In discussing (2) the results of the investigation up to and in- 

 cluding the F2, it was pointed out that, assuming the multiple-factor 

 hypothesis of blending inheritance to be correct, each successive pair 

 of parental varieties must differ by a certain minimum number of 

 factors. Summing up the factor differences between the individual 

 pairs we get the number of factors by which the latest differs from the 

 earliest. The number of these factors will determine the relative 

 amount of F2 variation in a direct cross between the earliest and latest. 

 The greater the number of factors, the less the probability of recover- 

 ing the parental types in F2. For example, where the differences be- 

 tween the parents involve 6 factors, the smallest number of F2 plants 

 which must be raised in order to secure the parent type will be 4096 

 (Mendel's 4° where N equals 6); for 7 factors, 16,384; for 8 factors, 

 65,536. It was concluded that the actual recovery of what appeared 

 to be parental types among a small number of F2 plants from direct 

 crosses of earliest and latest, limited the factors to a number less than 

 the sum obtained by adding up intermediate differences. 



