14 BULLETIN 1489, U. S. DEPARTMENT OF AGRICULTURE 
Of every 16 F 2 kernels, therefore, 9 will have red aleurone and 7 
will have colorless aleurone. If the different classes are grown and 
continuously self -fertilized the following results will be obtained in 
later generations: Of the colored kernels, group a is homozygous 
for both A and R (A A, R R) and will produce nothing but red 
kernels in later generations. Group b is homozygous for A btit 
heterozygous for R r (A A, R r). All of the progeny of this group 
will be A A, whereas the factor pair R r will segregate as a simple 
Mendelian pair into R i?, R i% and r r, producing 3 colored kernels 
to 1 colorless in the F 3 generation. Group c will behave like group b, 
producing 3 colored kernels to 1 colorless in the F 3 generation and 
giving the expected simple Mendelian ratios in later generations. 
In group c, however, it is the R factor that is homozygous and the 
A a pair that is heterozygous. Group d is similar to the F ± cross 
and will produce red and colorless kernels in a ratio of 9 to 7 in the 
F 3 . Groups e, % g, h, and i are homozygous for a a or r r or both 
(a a, r r). They therefore will breed true for colorless aleurone 
under self-fertilization or under any other system of mating that 
does not bring A and R into the same individual. 
ALTERNATIVE AND BLENDING INHERITANCE 
In the preceding analysis it was assumed that the stocks used were 
homozygous pr pr. As already noted, Pr converts what otherwise 
would be red aleurone into purple. Factors A< C ', and R must be 
present if Pr is to have any effect, and the kernels must be i i, as 
the presence of / inhibits the formation of any aleurone color. It 
follows that if the parents in the cross considered had been Pr Pr 
all of the colored kernels would be purple instead of red. 
If, on the other hand, one parent were Pr Pr and the other pr pr, 
the I\ kernels would all be purple (Pr pr) , and the Pr pr pair would 
segregate and recombine in later generations as a simple Mendelian 
pair. This would be independent of any segregation of the A a 
and R r factors. 
If 64 instead of 16 kernels are taken as a basis, 36 (or nine-six- 
teenths) would be colored and 28 (or seven-sixteenths) would be color- 
less. Of the 36 colored kernels one-fourth would be purple (Pr Pr) , 
two-fourths would be purple (Pr pr) , and one-fourth would be red 
(pr pr). Accordingly the F 2 kernels would consist of 27 purple, 
9 red, and 28 colorless. Here, then, we have three major color 
types — colorless, red, and purple — instead of two. 
It has been stated that, owing to incomplete dominance, kernels 
receiving the R factor from the male parent (r r R) are less deeply 
colored than those receiving R from the female parent (R R r). 
This provides two classes each of red and of purple, and a more 
continuous series is obtained, beginning with white and extending 
through light red, darker red, and light purple to dark purple at the 
other extreme. 
This is a relatively simple example of the complexities arising from 
the interaction of only three factor pairs affecting a single character. 
Formerly it was supposed that there were two distinct phenomena 
of inheritance: (1) That in which two or more characters are dis- 
tinct alternatives of each other with little intermediacy and (2) that 
