266 Shull. 



produces a dimorphic progeny. As it is the male of the dioecious species 

 that produces in each case a dimorphic progeny, there can be no reason- 

 able doubt of the correctness of CORRENS'S inference that the male is > 

 in Bryonia dioica, the heterozygous sex. 



The discovery of hermaphrodite mutants in Lychnis dioica L. which 

 is usually strictly dioecious, has enabled me to verify in this species 

 the hypothesis that the sexes are related to each other as Mendelian 

 homozygotes to heterozygotes. I have shown (SHULL 1910, 1911) that 

 these hermaphrodite mutants result from a modification of the male 

 sex-determiner or its homolog and that this modification can be reversed, 

 so that the male type again appears with great rarity among the off- 

 spring of the hermaphrodites. In crosses of either males or "genetic'* 

 hermaphrodites upon females, the type of the non-female offspring is 

 determined by the type of the pollen-parent, regardless of the ancestry 

 of the latter; in other words, crosses between females and males produce 

 females and males, while crosses between females and hermaphrodites 

 produce females and hermaphrodites. From the fact that the male can 

 become modified into a functional hermaphrodite, I inferred that the male 

 of this species is heterozygous with respect to the sex-determiner. The 

 crucial tests of this hypothesis were found in self-fertilizations of the 

 hermaphrodites, and in crosses between hermaphrodites and males, for 

 both of these combinations ought to yield mixed progenies if the males 

 and hermaphrodites are heterozygous, and uniform progenies if both are j 

 homozygous. The actual results in every case were in accord with my 

 hypothesis, in so far that in each a mixed progeny was produced, but 

 the proportions in which the different sexual types appeared, presented 

 some difficulties which could not be cleared up at that time, and which 

 are still only partially understood. I have some further data bearing 

 on these difficulties and am continuing the search for a complete solu- 

 tion. The new results will be presented in detail in another connection ' 

 and it need only be stated here, that it has become certain that my 

 earlier inference that the egg-cells of the hermaphrodites are unable to 

 produce hermaphroditic offspring unless fertilized by a hermaphrodite- 

 bearing sperm, was erroneous, though it appears that the hermaphrodite- 

 bearing eggs are usually rare. Since such eggs are now known to be 

 produced, even though usually in relatively small numbers, it should be 

 possible, unless selective fertilization or non-viability contravene, to 

 produce homozygous hermaphrodites and so to reconstruct, out of the 

 dioecious race, a pure-breeding hermaphroditic race. Although this goal 

 has not yet been attained, the irregularities in the genetic behavior of 



