1910] SHU LL INHERITANCE OF SEX IN LYCHNIS 113 



For convenience, I will designate the four plants successfully 

 used in breeding by the letters A, B, C, and D. When A was self 

 fertilized it produced, as a result of two different operations, 33 

 females and 25 hermaphrodites. When its pollen was used to fertilize 

 four different females, the resultant progenies consisted of 236 females, 

 161 hermaphrodites, and 2 males. When castrated and fertilized 

 with pollen from a normal male, A gave rise to a progeny of 21 females, 

 2 hermaphrodites, and n males. B was also self fertilized and gave 

 a progeny of no females and 95 hermaphrodites. When pollen of 

 B was used to fertilize three different females, it produced 162 females 

 and 144 hermaphrodites. It is thus seen that the two plants, A and 

 B, showed identical behavior and together produced self fertilized 

 offspring consisting of 143 females and 120 hermaphrodites, and 

 when crossed with females gave a total of 398 females, 305 hermaph- 

 rodites, and 2 males. In any explanation of these results the 

 occurrence of these two males will probably have to be left aside as 

 wholly exceptional. Only further breeding will show whether they 

 were true males, or hermaphrodites with pistils suppressed, perhaps, 

 by some cause external to the germ cells. Plants C and D gave quite 

 a different result. Attempts to self fertilize them and to cross them 

 with normal males all proved futile, though more persistent efforts 

 perhaps might have succeeded. Both were used as pollen parents 

 in crosses with normal females. In such a cross C gave a progeny 

 of 39 females and 55 normal males, and D gave 26 females and 18 

 normal males. The details of the several crosses are given in table 

 I, p. 114. 



Considering first plants A and B, and leaving out of consideration 

 for the present the two males occurring in crosses between females and 

 hermaphrodites, and the two hermaphrodites which appeared in the 

 cross between a hermaphrodite and a normal male, it is apparent 

 that the hermaphrodite character belongs only to the males, for in the 

 families in which these hermaphrodites were the pollen parents, the off- 

 spring always showed the same ratios of females and hermaphrodites 

 that would have been expected of females and males, had a normal 

 male been used as the pollen parent. 



It is clear that the hermaphrodite individuals, C and D, belong 

 to an entirely different category from A and B, for in the* families 



