The Inheritance of Quantitative Characters in Maize 61 
INHERITANCE OF BREADTH OP SEEDS. 
In case of the crosses of California pop with Missouri dent 
and of the latter with Tom Thumb pop, the breadth of twenty- 
five kernels from an ear of each plant was determined. Tables 
23 and 24 contain these data. Of the varieties of corn known to 
us, California pop has by far the smallest seeds. As grown in 
1910 and 1911 (Table 23), the average breadth of twenty-five 
kernels of this variety was only about 85 millimeters, while the 
breadth of twenty-five Missouri dent kernels averaged about 216 
millimeters. The F 1 seeds were distinctly intermediate in size. 
In 1910 the difference between the largest twenty-five seeds of 
California pop and the smallest class of ¥ 1 was 20 millimeters 
and there was the same difference between the largest class of 
F 1 and the smallest of Missouri dent. In 1911 the ¥ 1 families had 
somewhat larger seeds than in 1910. The variation in the F 2 
families was noticeably greater than in the F 1 lots. The range 
of variation, however, was not sufficient in all the F 2 families 
together completely to close the gap between even the inner ex- 
tremes of the parents. The greater variation in F 2 was doubtless 
due to a segregation of size factors, and types like the parents 
would probably have appeared had a sufficient number of F 2 
plants been grown. How many F 2 plants it would be necessary 
to grow in order to have an even chance for the appearance of 
forms like the parents cannot be told. More than enough were 
grown to realize this result if the parents differed by only four 
factors influencing breadth of kernels. Obviously the difference 
between them must be considerably more than four factors. If 
they differ by five factors, over three times as many plants as 
were grown would have been necessary ; if by six factors, nearly 
thirteen times as many; if by seven factors, over fifty times as 
many, and if by eight factors, over two hundred times as many. 
It is out of the question to grow in ordinary pedigree cultures 
any such number of individuals as would be necessary to regain 
the original types of seed size in F 2 if the parents differ by as 
many as seven or eight factors. But, by selecting F 2 plants with 
the smallest seeds and continuing the selection in F 3 and F 4 , if 
necessary, it should be possible to regain the parent types in a 
few generations without growing excessively large numbers in 
any generation. 
The F 3 families of the cross of Missouri dent with Tom Thumb 
pop (Table 24) illustrate the possibility of accomplishing this 
isolation gradually without growing extremely large numbers. 
True the parents of this cross did not differ so greatly in seed 
size as did Missouri dent and California pop and all the inter- 
vening classes were occupied by F 1 individuals. Furthermore 
