160 THE ROYAL SOCIETY OF CANADA 



The F3 results reported in the present paper have revealed the 

 genetic constitution with respect to earliness of many of the F2 plants. 

 They have shown that in every case, no matter how wide the parental 

 difference hybrids have been obtained which are apparently at all 

 grades between the parents. That the spread of the families in this 

 way is not entirely due to environmental influences even in the crosses 

 involving only a small difference is indicated by the general agree- 

 ment between the Fo plants and their F3 families in the order of their 

 arrangement according to earliness. As one looks over a plot contain- 

 ing a large number of third generation families, one cannot fail to notice 

 the marked differences between the different families and the fact 

 that the means of the different families apparently occupy all possible 

 positions between the parental means. There appears to be no group- 

 ing of the families at any particular centres as in cases of simple 

 segregation. Of course it may be that there are several centres separa- 

 ted by slight distances about which the families group themselves, 

 and that these centres are obscured by environmental influences. 

 But the general agreement, already referred to, between the order of 

 the F2 plants and F3 families would argue to the contrary. 



If we assume that the multiple-factor hypothesis of blending 

 inheritance is correct, it is obviously extremely difficult to state what 

 are the exact factorial conditions involved. The uniform distribution 

 of the families — the lack of grouping — supplies no basis for interpre- 

 tation. Only a small proportion of the F2 plants could possibly be 

 carried on ; therefore the proportion of the total possible families which 

 fall at any particular point (for example, along with one parent) or 

 which show great variability (indicating heterozygosity in the F2) 

 cannot be determined. Apparently, the only conclusion possible on 

 the multiple-factor basis is that even slightly different parental vari- 

 eties differ by several factors. Segregation will then produce races 

 whose genetic differences are so slight as to be obscured by environ- 

 mental influences. 



If this interpretation is put upon the results, then the sum of the 

 differences between the successive pairs of varieties must be great. 

 The number of factors by which the members of each successive pair 

 differ must be considerable because if only one or two, then the parental 

 types should be recovered among four or sixteen F2 plants respectively, 

 whereas really, the great majority of F2 plants are intermediate. 

 Moreover, all the F2 intermediates should give a very variable F3, 

 including the parental types, though it must be admitted îhat the 

 numbers are rather small to reveal the whole range of variation. 



