456 H. J. MULLER 
to this one gene taken by itself would probably not be cumula- 
tive in its effect, for if extreme variants are as likely to occur as 
moderate ones, it is about as easy to get an extreme type directly, 
by one step, from the original form, as to get it by the process of 
piling up small variations in the gene. This should not, how- 
ever, be taken to mean that where a large a group of loci or the 
whole germ plasm is under selection, the latter would be ineffec- 
tive, for in that case, of course, a single change in one locus 
would rarely produce as marked an effect as could be obtained 
from a combination of changes in various genes. Hence we see 
that a misleading picture of the. manner of the occurrence of 
variation is obtained unless the variation is analyzed into the 
separate gene changes of which it is composed. 
Not only may we establish the fact that the magnitude of 
mutations in W are not determined according to the principle of 
simple sampling, but it may also be shown directly that they 
cannot be purely quantitative changes. For, if the eye-color 
differences in question depended merely upon hereditary differ- 
ences in the amount of a certain kind of gene (W), then the 
males, containing only one X chromosome, should have only 
half as much of this material as the females, and their eye color 
should be correspondingly lighter than that of the females, for 
the same reason that the eye color of a female which has one 
gene for écru and one for white is lighter than the eye color of a 
female which has two genes for écru. In all but one (ivory) of 
the mutations of W, however, the male has the same eye color 
as the female,’? and the differences in eye color which do occur 
between the different allelomorphs cannot therefore be due to 
purely quantitative differences between the genes. 
It might perhaps be objected that such comparison between 
male and female is invalid on account of the other genetic differ- 
ences which exist between them. But this would be tanta- 
mount to claiming that the other parts of the male complex 
7 Kosin, the mutant which arose from the factor white, is like ivory, and dif- 
ferent from the other mutations of the normal allelomorph W, in this respect, 
but this is probably because of an influence of sex upon the character eosin, 
since Bridges has shown the latter to be unusually susceptible of modification by 
genes in loci other than that of w° itself. 
