96 BOTANICAL GAZETTE [SEPTEMBER 
A number of the pedigrees involved rather radical shape differ- 
ences in the original types crossed. Several pure strains of the ordi- 
nary “scallop” or “pattypan” type were obtained (fig. 1) with 
fruit essentially disclike in shape, being from two to four times as 
broad as thick, with the teeth or scallops around the edge varying 
in number, shape, and development. In one race of white discs 
there appeared very early several plants in which the fruit differed 
markedly from that of the disc parent, being nearly spherical in 
shape, approximately as thick as broad, and with a rather weak 
development of teeth. Intermediate forms did not appear, and the 
spheres when inbred never produced any other shape. Several pure 
lines with spherical fruits were thus obtained (fig. 1). 
Cece ey Gl 
A)ist. "SPHERE. 
Fic. r 
A single plant of one of these spherical-fruited lines was crossed 
in 1919 with a plant from each of three disc-fruited lines. These 
three lines differed somewhat in fruit shape, one having a relatively 
flat disc and the others relatively deep ones, one being somewhat 
flatter than the other. The F, generation in each case showed com- 
plete dominance of the disc shape, and the type of disc (deep or flat) 
was essentially like that of the particular disc parent used. In the 
F, generation there was a sharp segregation into approximately 
three-quarters disc and one-quarter sphere. It is evident that the 
chief difference between these two shapes is caused by a single factor, 
and that this shows complete dominance. F, counts for the three 
pedigrees, together with the results expected on the single factor 
hypothesis, are shown in table I. 
The problem is not quite so simple as this, however, for the F, 
segregates do not resemble exactly the original types, it being espe- 
cially noticeable that the sphere segregates differ in shape, those 
