1911] SHULL—REVERSIBLE SEX-MUTANTS 335 
Particular attention should be given to only two points in this 
table until after the results secured in the second generation have 
been considered. The assumptions which form the basis of the 
first section of the table lead to the expectation (a) that females 
derived from hermaphrodite families, whether they be fertilized 
by normal males or by their hermaphrodite sibs, will yield families 
in which the male offspring are hermaphrodite and normal male in 
equal numbers; and (6) that the hermaphrodites of the second 
generation when used to fertilize females from normal male families 
will produce no hermaphrodites, but only females and males. The 
alternative assumptions involved in the second and third sections 
of the table, on the other hand, lead to the expectation that, re- 
gardless of the origin of the female, no hermaphrodites will be pro- 
duced normally, except when fertilization is brought about by 
sperms from a genetic hermaphrodite, and then the result will always 
the same whether this hermaphrodite was a mutant or whether it 
was derived from an antecedent hermaphrodite. 
We may now proceed to examine the results of the crosses. 
This will be most easily accomplished by considering each type 
of cross separately in the following fourteen cases. In the model | 
pedigrees, illustrated under each case, the oldest ancestors entered © 
in the diagrams are females and males both of which came from 
normal families, whose matings had been controlled during at 
least three still earlier generations, and which are known to have 
been in each such previous generation the result of crosses between 
females and normal males, and to have belonged to families in 
which no hermaphrodite mutants appeared. In the diagrams 
all male and hermaphrodite individuals which appeared as mutants 
are indicated as such, and it should be understood that any male 
or hermaphrodite not so marked was a member of a family which 
consisted of a normal proportion of its own type, that is, either 
male or hermaphrodite. : 
CASE I 
CROSSES OF GENETIC HERMAPHRODITE MUTANTS WITH FEMALES 
Only 2 of the 8 plants recorded as hermaphrodites in 1909, in 
otherwise normal male families, were successfully used for breeding. 
One of these, bred to 2 different unrelated females, produced 72 
