62 BULLETIN 1463, U. S. DEPARTMENT OP AGKICULTUBE 
reduction division, and it is left for future investigations to de- 
termine the effect on crossing over of factors operative at an earlier 
j)eriod in ontogeny. 
Finally, it is apparent from these studies that the percentage of 
crossing over, that is the degree of association of parental char- 
acteristics, is variable and can be altered by numerous factors, a 
condition which facilitates the dissolution of character combina- 
tions and permits flexibility in breeding. 
SUMMARY 
A comparison of the variability of the linkage between C and 
Woe, genes for aleurone color and 'endosperm texture, with that for 
other linked genes in maize shows that the G—Wx linkage is no 
more unstable than that for other linked pairs. 
The variability of the crossover ratio in some progenies is greater 
than that of the ratios of the Menclelian characters involved, but 
this is not invariably so. 
Crossing over between linked genes is not always more variable 
than the proportion of independent genes, indicating that the vari- 
ability is not associated completely with linkage. 
In a study of several progenies it is apparent that some are much 
more uniform with respect to the rate of crossing over than others. 
A comparison of the variability in male as opposed to female 
gametes in 13 progenies shows that the rate of crossing over is as 
variable in one sex as in the other. 
A comparison of the variability in upper and lower ears for three 
progenies shows that the rate of crossing over is equally variable 
in both ears. 
The variability of the rate of crossing over was measured sepa- 
rately in the bases and tips of the ears of two progenies and was 
found to be alike for both sections. 
The variability of the rate of crossing over in several samples of 
single individuals was found to be less than that in single samples of 
several individuals. 
In 13 cases where pollen was divided into two lots, one being 
applied at once and the other being held for seven hours, the 
variability of the crossover ratio in the fresh pollen was as great 
as that in the old. 
In seven cases where two samples of pollen from the same plant 
were used, one lot representing the first pollen shed and -the other 
taken several days later, it was found that the pollen produced at the 
later date was no more variable with respect to the crossover per- 
centage than that produced early. 
A comparison of the observed variability of the crossover rate 
with that expe cted from the operation of chance factors, as calcu- 
lated by Tj'pq/H, shows clearly that the rate of crossing over in some 
progenies is no more variable than is to be expected, though in most 
cases it seems certain that factors other than chance are affecting 
the rate. In a series of 17 progenies the observed variability exceeds 
the expected (a 50 -^pq/H) by 3.53 + 0.19. 
From the results obtained with IT related progenies it is clear 
that progenies differ widely in the closeness of the linkage between 
the genes C and Wx. 
