266 
AMERICAN JOURNAL OF BOTANY 
[Vol. io, 
results obtained by Richardson (2), who obtained a total of 203 $ and 
173 c? or y in making similar crosses. 
Sex inheritance, in all these crosses, conforms to the theory of a heter¬ 
ozygous condition of the females, the female condition being completely 
dominant, with one possible exception. This was the cross 45/16 in which 
5/15 64-2, a pistillate, derived from the cross 5/15 (table 3) was used as 
the female parent and Glenville as the male parent. This cross resulted in 
32 hermaphrodites and 28 females and approximated the expected 1 : 1 
ratio. The fertility of the seedlings, however, was not what was expected. 
Fifteen of the females set all flowers except an occasional late one but 
produced only a few achenes per berry, and as a consequence all berries 
were nubbins. The remaining 13 females were completely fertile except 
for an occasional late flower. Twelve of the hermaphrodites set some fruit 
and in some cases nearly as many flowers as the females, but all the berries 
were likewise nubbins. Eleven hermaphrodites set nearly all flowers and 
the berries were perfect; 9 others set one half or less of the flowers perfectly. 
The remainder were males. The expectation from this cross was 1 fertile 
female to 1 hermaphrodite varying in degree of fertility. In other words, 
the females should all have been completely fertile and the hermaphrodites 
either fertile or partially sterile depending on whether the H from the 
female parent united with the more fertile determiner of Glenville or with 
the one which has been shown to be practically sterile; and on the extent 
to which the H determiner from the female parent is dominant over the 
sterile H. Instead of the expected ratio, 1 normal female to 1 practically 
sterile female (nubbins) to 1 fertile hermaphrodite to 1 partially fertile 
hermaphrodite was obtained. This and one other are the only cases 
observed in which females have not been nearly completely fertile. There 
has evidently been a decided change in the female determiner in this single 
instance. For an explanation of this peculiar condition it seems we must 
go back beyond the two immediate parents. The male parent Glenville 
is probably not the cause, as, when crossed with other females, it has not 
produced similar results. The female parent was the offspring of a wild 
female X hermaphrodite 778. Two other Fi females of this cross acted 
according to expectation when crossed with Glenville (56/16 and 33/16, 
table 3). It seems, therefore, that we are forced to assume that a change 
took place, probably during the reduction division, in the wild female 
grandparent. A crossing over in the female and male chromosomes between 
the female determiner and the suppressed male determiner and vice versa 
in the male chromosome would give a chromosome in which both sexes are 
suppressed: (mf or h). An egg containing such a chromosome if fertilized by 
a male gamete from 778 carrying an hermaphrodite determiner would produce 
an individual having one determiner for femaleness and one for maleness, 
these two being linked (H). The constitution of this individual would 
then be Hh. We have already seen that a single dose of femaleness is 
