330 BOTANICAL GAZETTE [NOVEMBER 



al organs of both sexes, and because (2) self-fertilized hermaphro- 

 dites produce dimorphic progenies, consisting of females and 

 hermaphrodites. 



In my first paper on the inheritance of sex in Lychnis (SHULL 26), 

 I represented the sex genes by the conventional signs for the sexes 

 (?, $, and $). As these signs were used in my tables with two 

 different meanings to represent sometimes the character of the 

 genes and at other times the character of the soma I suspect that 

 readers may have experienced some difficulty in comprehending 

 the tables. I shall therefore adopt here the plan usually followed 

 by students of genetics, of representing the genes by letters, letting 

 Ff be respectively the presence and absence of a female determiner, 

 Mm a male determiner, and Hh a hermaphrodite determiner. 

 The conventional signs for the sexes will be used in this paper 

 only in their more usual signification, referring to the nature of 

 the soma, that is, the sporophyte. 



If CORRENS' view of sex determination is correct, and the males 

 are heterozygous, the females must be homozygous. CASTLE (5) 

 suggests that in such a case the females will always be positive 

 homozygotes, having a pair of sex genes (FF) corresponding with 

 a single equivalent gene (Ff) in the male. I do not believe that 

 this view can be substantiated, as there seems no good reason why 

 females should not be negative homozygotes in some plants and 

 animals, "neutral" homozygotes in others, and positive homozy- 

 gotes in a third class. If the females are positive homozygotes, the 

 somatic formula of the two sexes may be represented thus: FF=$, 

 and Ff= $ ; if the females are negative homozygotes, the correspond- 

 ing symbols will be FFmm=%, and FFMm=$>\ and if the female 

 is a "neutral" homozygote, the formulae of the two sexes will be 

 FF=$, and FM=$.. Only the first two of these assumptions con- 

 cerning the nature of the females were considered in my earlier 

 paper, and either was found capable of explaining the results secured 

 in the first generation, provided the presence of a partially inde- 

 pendent hermaphrodite factor (H) might also be assumed. 



Whether there was any genetic relationship between the her- 

 maphrodites A and B which produced hermaphrodite offspring, and 

 C and D which produced males, could not be determined in the first 



