THOMPSON] EARLINESS AND LATENESS IN WHEAT 149 



ment will be discussed later. It must be remembered, moreover, that 

 the proportion of families given in the table for any particular range 

 of variation does not represent the real facts because only a small 

 fraction of the total possible number of families were actually grown. 

 In F2 the great majority of individuals were intermediate, but those 

 chosen to be carried on to F3 were selected in nearly equal numbers 

 from each day of the ripening time of F2. Consequently, the great 

 mass of intermediate F2 individuals were not chosen. The immense 

 amount of work precluded the study of more families and it was con- 

 sidered desirable to study representatives from the whole range of 

 variation. The small proportion of F3 families grown and the dispro- 

 portionate representation of the extreme families precludes the draw- 

 ing of theoretical inferences from the proportions of early, late and 

 intermediate families. 



4. The degree of earliness of a family corresponds in general with 

 that of its F2 parent. Of course, environmental influences made some 

 F2 plants earlier or later than they should have been with the result 

 that the order of the families is not exactly the same as that of the 

 F2 parents. For example though No. 15 was slightly earlier than 

 No. 17, the F2 parent of No. 15 was recorded as later than that of No. 

 17. Moreover, as pointed out more fully in (5) below, the arbitrary 

 method of arranging the families in the order of their earliest plants, 

 does violence to a true genetic order. Nevertheless, the general 

 agreement is apparent. 



5. There is a considerable difference in the variability of different 

 families. For example, in No. 14 the range of variation extends al- 

 most from the earliest date of the early parent to the latest date of 

 the later parent, while in other cases the range is no greater than in 

 either parent. In some cases, particularly among the later families, 

 the variability is less than in the parents, but this is probably due to 

 the smaller number of individuals recorded or to the rust attack. 

 This difference in variability in different families is evidently due 

 to differences in the genetic constitution of the F2 parents. Following 

 the usual interpretation the great variability of certain families would 

 be due to Mendelian segregation. If only a small number of factors 

 are concerned, all the intermediate families should show the segrega- 

 tion. Therefore, the occurrence of several intermediate families 

 with a small variability demands the assumption of several factor 

 differences. The small proportion of families to the total F2 popula- 

 tion precludes theoretical deductions from the relative numbers of 

 families showing great and small variability. 



