VARIABILITY IN LINKAGE OF CHARACTERS OF MAIZE 61 
The method of measuring crossing over does not permit the sepa- 
ration of causes operative on the actual rate of crossing over from 
those affecting the gametic and zygotic proportions subsequent to 
the reduction divisions. 
The field can be narrowed somewhat by measuring the gametic 
proportions at different periods, on separate inflorescences and on 
different parts of the same inflorescence. 
A check on gametic death rates can be had by subjecting the 
gametes to adverse conditions, or death rates can be measured indi- 
rectly through correlating the apparent crossover rate with fertility. 
With changes in the apparent rate of crossing over so small as 
those here considered, it is impossible to distinguish with certainty 
gametic changes from differential death rates of zygotic classes. To 
compare coupling and repulsion ratios the comparison necessarily is 
made between different progenies, and it has been shown that even 
closely related progenies may have different rates of crossing over. 
It is apparent from the data considered in this bulletin that many 
factors, both environmental and genetic, affect the crossover rate. 
The results obtained with male and female gametes, with first and 
second ears, with homozygous and heterozygous i?, and with butts and 
tips leave no room for doubt that each of these factors has a slight 
though significant effect upon crossing over. On the other hand 
such factors as vitality of the pollen, length of silks, and age of 
plants seem to bear no relation to the crossover rate, though in these 
cases the populations were not large enough to establish definitely 
their lack of effect within 3 or 4 per cent. However, for all these 
cases it seems reasonable to conclude that the factors considered are 
of minor importance, at most causing a difference of but 2 or 3 per 
cent and thus failing to provide an explanation for the occasional 
large nonheritable variations from the mean rate. The explanation 
for these variations must be looked for earlier in the ontogeny of 
the plant. 
It is not inconceivable that environmental conditions of any two 
adjacent plants might differ sufficiently at the critical period when 
reduction and segregation divisions were in progress to cause the 
observed differences in the percentage of crossing over. Such factors, 
being no longer operative by the time the flowers had matured, 
could not be associated with the effects which they might produce. 
As the work has progressed it has become increasingly evident that 
the distinction between factors that produce a measurable effect 
and those that do not depends not a little on the numbers involved 
in the several experiments. In other words, there appears to be a 
positive correlation between the degree of certainty assigned to the 
influence of a factor and the number of individuals involved in the 
experiment. Like other biological phenomena, crossing over appears 
to be not a thing apart but to react in a complex manner and in a 
varying degree to practically all changes in environment. There 
can be little doubt that, if more refined tests were made, many of 
the factors dismissed as producing no significant changes in the rate 
of crossing over would have to be transferred to the list of significant 
factors. 
The present study seems to have eliminated most of the natural 
factors that could affect the gametic proportions subsequent to the 
