GABRIELLE L. C. HOWARD. 65 



practically constant is further evidence that, within the same 

 type, height and number of leaves are not correlated. This 

 is a somewhat surprising result. It coincides, however, with 

 the conclusion of Hayes, that the correlation between height 

 and number of leaves was less than +.5. 



The data as regards this character in the cross Type 9 / Type 

 51 are given in Table VIII. The Fj generation was intermediate 

 between the parents. The F^ generation has a range far outside 

 those of the parents, one of the plants having as many as 

 forty leaves. That this is not due to accidental circumstances 

 is shown by the reappearance of this phenomenon in one of the 

 cultures in the F4 generation, where the number of leaves varies 

 from twenty-eight to forty. The eight cultures of the F3 

 differed greatly in their range of variation ; none were 

 uniform. Two of these were continued in the F^ generation, 

 i.e., Nos. 738 and 694. From 738 two cultures were obtained, 

 Nos. 738-81 and 738-19, which from their range of variation 

 and the form of the curve appear to be uniform. The first 

 has an average of nineteen leaves per plant, and a variation 

 of seventeen to twenty-two. The latter culture, 738-19, is 

 a replica of Type 9 as regards number of leaves. In the culture, 

 derived from 694 only one can possibly be uniform. No. 694-1, 

 but as the range of variation is somewhat large it may have 

 originated from a heterozygote between two forms differing 

 in one small factor. In either case it points to the occurrence 

 of intermediate homozygotic combinations. Thus in the F4 

 generation there have been isolated a culture (probably uniform) 

 with a smaller number of leaves than either parents 

 a culture with the same number of leaves as one parent, 

 a culture which is either uniform with a number of leaves 

 intermediate between both parents or a heterozygote between 

 two intermediate forms, and a culture in which the majority of 

 plants possess more leaves than either parent. 



In the cross between Type 16 and Type 35 (Table IX), 

 the Fi is again intermediate, but approaches the parent with 



