The Inheritance of Quantitative Characters in Maize 21 
That individuals of a certain theoretical size have different ge- 
netic possibilities, but is literally overwhelmed with difficulties by 
the knowledge that these theoretical sizes are pushed upivard and 
downward by external causes. One does not know when he 
selects an individual of a certain size ivhether it even belongs in 
the class in which it appears. Besides this trouble, the stimulus 
to development due to heterozygosis also causes confusion. It is 
well known that hybrids often exceed their parents in size. This 
phenomenon is not one of heredity but one of conditions due to 
heredity. There is a greater stimulus to development when a 
character is in the heterozygous condition than when it is in the 
homozygous condition. Moreover this stimulus is in some degree 
cumulative. It increases roughly with the number of hetero- 
zygous factors. For this reason it is a force that aids environ- 
ment in pushing individuals that theoretically should be in one 
size class into a different class. 
But these are not the only troubles. One is not often — if 
ever — able to cross two plants in which auy particular size differ- 
ence can be represented by a notation like (X = 10 inches ) 
+ aabbcc and (X = 10 inches) + AABBCG. A great majority 
of plants are frequently crossed thru insect or other agencies, 
while a comparatively large number have flower structures of 
such a character that they are always cross-fertilized. Hetero- 
zygosis is also favored by bisexuality wherever it occurs. Many 
plant varieties are therefore hybrid complexes from which types 
of different kinds can be isolated by inbreeding. Other plants 
which are especially adapted to self-fertilization may be homo- 
zygous in nearly all their gametic factors, for self-fertilization 
tends to eliminate heterozygosis. But such cases are not common. 
Ordinarily one may presume that the two individuals he wishes 
to cross are heterozygous for a considerable number of factors. 
This reduces the excess of variability which the F 2 generation 
might be expected to show over that shown by the ¥ 1 generation. 
Furthermore, it gives an F x generation in which the zygotic con- 
stitution of each individual is not the same as would be the case 
were both parents completely homozygous. A series of F 2 gen- 
erations raised from different F x individuals, for this reason, may 
be quite different in their variability. 
On the other hand, it may sometimes happen that the varia- 
bility of the F 2 generation exceeds that of the F t generation in a 
degree unsuspected by the difference in size of the parents. East 
(1910) predicted that if the theory of multiple interchangeable 
segregating factors gave a true interpretation of the behavior of 
quantitative characters in crosses, cases would be found when 
parents identical in size would show marked size recombinations 
