98 BOTANICAL GAZETTE [SEPTEMBER 
From a study of these few individuals, however, certain points stand 
out clearly. The F, shows complete dominance of the disc shape. 
In fact, in every case the F, is even a little flatter than the parent 
disc. This is notably the case in pedigree 1, but there is reason to 
believe that the pure type of this line is probably a little flatter than 
is shown by the individuals here portrayed. 
In pedigree 15, which involves the deepest disc, both the sphere 
and disc F, segregates are essentially like the original types, being a 
little deeper in each case. A single factor difference between disc 
and sphere is apparently sufficient to account for these facts. It will 
be noted that there is a very sharp segregation between the two 
shapes in the F,. In pedigree 19, however, which involves the 
flattest disc, the F, spheres are decidedly flatter than the parent 
sphere type (1.50 as compared with 0.97), and the F, discs distinctly 
deeper than the parent disc type (2.94 as compared with 3.28). 
Pedigree 1 is somewhat intermediate between these, the F, spheres 
being somewhat flatter than the original sphere type, and the F, 
discs being about the same as the original disc type, but consider- 
ably deeper than the F, discs. 
In pedigrees 19 and 1 there is evidently something more than a 
single disc or ‘‘flattening” factor at work. This coincident flatten- 
ing of the spheres and deepening of the discs may readily be ex- 
plained by assuming that there is a second dominant flattening 
factor, considerably weaker in its effect than the major one already 
discussed, and segregating independently of it. The parent disc 
type would possess both of these and the parent sphere type lack 
them both. The major difference between disc and sphere would 
still be caused by the larger factor, the ‘‘spheres” lacking it and the 
“discs’’ possessing it. In the F,, however, three-quarters of the 
spheres would possess the smaller flattening factor, and the average 
shape of the whole sphere group would thus be flatter than that of 
the pure type; and a quarter of the discs would lack this smaller 
factor and thus be less flat than the original discs in which both 
flattening factors are present. This would tend to bring the two 
shape types somewhat nearer together in the F, than in the pure 
types, a condition which evidently obtains in these two pedigrees. 
