The Inheritance of Quantitative Characters in Maize 109 
length of stalk and therefore, in addition, upon stalk weight 
and even upon total weight of plant. Likewise any factors in- 
fluencing number of nodes, any with an effect upon amount of 
tillering, any factor for diameter of stalk, size of ears, number of 
ears, etc., will all be concerned in the development of the one char- 
acter complex— -total weight of plant. If one argue that total 
weight is not a simple character, he must also admit that neither 
is length of stalk, which of course is determined by the number 
of nodes and the internode length, and must not forget that even 
the length of one internode is the product of the number and 
the length of the cells contained in it. Tho total weight can thus 
be analyzed into numerous sub-characters, it is none the less a 
"character" of the plant or type in question. 
It was shown earlier in this paper that the multiple-factor 
hypothesis furnishes a satisfactory and simple interpretation 
not only of all of the results secured from these maize experi- 
ments but also of the results from experiments previously re- 
ported for other plants and for animals. We are familiar with 
no hypothesis not based upon the Mendelian principle of segre- 
gation and recombination of factors which furnishes a plausible 
explanation of many of the facts regarding the inheritance of 
quantitative characters. We are aware of the suggestion 
(Castle 1912) that a heterogeneous distribution of growth-in- 
ducing substances in the cells of an organism would result in 
variability in its progeny, that this heterogeneity of the proto- 
plasm would be increased by crossing, and that, therefore, the 
variability would be increased in generations beyond F t . This 
hypothesis cannot, however, be said to interpret satisfactorily 
the facts of size inheritance, until it is shown, for instance, how 
one F 2 plant of a size intermediate between the parent varieties 
can produce an F 3 progeny of great variability while another 
F 2 plant of the same intermediate size yields an F 3 progeny 
scarcely more variable than the parents or the ¥ t generation. In 
other words, it must first be shown how some F 2 plants come 
to have the size of F x plants and at the same time have the 
comparatively homogeneous distribution of growth-inducing sub- 
stances characteristic of the parent races while other F 2 plants 
are like F x plants in both size and heterogeneity of these sub- 
stances. And even if increased protoplasmic heterogeneity of 
general growth-inducing substances can be made to account for 
the increased variability ordinarily observed in F 2 , it is still not 
quite clear just how such a condition can bring about an F 2 
range of variation away beyond the outer extremes of the 
parents in respect to one size character, while the range of 
variation of some other size character of the same F L , individuals 
is scarcely enough to bridge tho gap between the inner extremes 
of the parents. 
