VARIABILITY IN LINKAGE OF CHARACTERS OF MAIZE 21 
are more variable than the percentages of the Mendelian characters 
it may mean that crossing over is more sensitive to environmental 
conditions than simple segregation, or that unrecognized genetic 
factors affecting the crossover rate are present. 
At the present time no genetic factors modifying the waxy-non- 
waxy ratio have been identified, and in this respect the waxy char- 
acter is comparable to the crossover rate between C and Wx. s By 
limiting the comparison of variability to the waxy character, the 
objection based on the reduction in variability due to assumed gametic 
compositions for widely aberrant ratios is partially overcome. Com- 
parisons of this nature are shown in Table 14 where the cr 50 of the 
percentage of crossovers is compared with the standard deviation 
of the percentage of waxy for male and female gametes separately. 
All comparisons are limited to reciprocal back-crossed ears where 
the expected percentage of waxy is 50. Although many of the dif- 
ferences shown in this table are not significant, and three instances 
occur where the standard deviation of the waxy percentages is 
greater than the corresponding <r 50 of the percentage of crossing over, 
taken as a whole there is in each sex a significantly greater varia- 
bility in the crossover percentage than in the percentage of waxy 
seeds. Combining the two sexes, the mean difference becomes 
1.93 ± 0.24, clearly a significant difference. 
In general, then, it may be stated that the crossing over of the 
factors C and Wx is more variable than the simple segregation of 
these characters, but whether this greater variability is due to genetic 
or environmental factors or both can not be determined from the 
data thus far submitted. 
VARIABILITY IN THE CROSSOVER CLASS NOT THE RESULT OF DIFFERENTIAL 
DEATH RATE OF ZYGOTES 
In hybrids involving two linked characters it is possible to produce 
F 2 progenies with all four classes in equal numbers. This condition 
results when one parent is homozygous recessive for one of the char- 
acters and heterozygous for the other while the second parent has 
the reciprocal of this factorial composition. When one of the char- 
acters is a dihybrid, hybrids heterozygous for the nonlinked gene 
of the dihybrid character and homozygous for the linked gene also 
will produce the four classes with no opportunity for linkage to be 
expressed. If the increased variability in the apparent rate of 
crossing over is due to a differential death rate of zygotic classes, we 
should find in progenies combined in either of the above ways a 
variability in the percentage of crossover classes comparable with 
that of progenies snowing linkage. 
In Table 15 there has been brought together from published data 
a summary of progenies segregating into all four classes with no op- 
portunity for linkage and progenies of similar parentage combined in 
a way to exhibit linkage. To these have been added four groups of 
ears descended from a self -pollinated plant that was heterozygous 
for both C and Wx. Two of the groups are reciprocals measuring 
8 Brink (4) has presented evidence to show that the ratio of waxy to nonwaxy is modi- 
fied when the Su factor for sweet endosperm is homozygous recessive and states also that 
the I aleurone factor influences this ratio. Neither of these factors was involved in the 
present study, and though other factors present in our experiments may be found to affect 
the proportion of the waxy seeds, they are still unrecognized and theix action appears to 
be uniform. 
