286 ShuU. 



In full haiiuoiiy with this In-pothesis are two crosses in which the 

 same hermaphroditic plant was crossed upon other females. In the first 

 of these (Ped. Nos. 11335 (19) X 11266 (51) = 12 262) the mother was 

 a homoz\'g:ous sib of the female plant which was used as mother of the 

 preceding family. The offspring were all broad-leafed as expected, and 

 the 96 plants which bloomed were all female! In the second (Ped. 

 Nos. 11286 (2) X 11266 (51) = 12 339) the mother was likewise a pure 

 Melandrium album, a pure first cousin of the hermaphroditic specimen 

 under consideration. All of the 85 plants bloomed and all were female! 

 unfortunately these were the only crosses made with tliis hermaphrodite, 

 and repeated attempts to get self-fertilized seeds of it failed because of 

 the ease with which the Inids, flowers, and immature capsules were 

 caused to drop off, — a feature more strongly developed in M. album 

 than in any of my other strains of Lychnis dioica, and also generally 

 more prominent in the hermaphrodites of any family than in the females 

 of the same family. This greater tendency to cut off the buds and 

 flowers b}' an absciss-layer in the hermajjhrodites as compared with the 

 females, is not strange, as the formation of such absciss-layers is uni- 

 versal in the males. 



The failure of this remarkable hermaphrodite to jiroduce any her- 

 maphroditic offspring renders it impossible to derive from it any infor- 

 mation concerning the state of normal males and genetic hermaphrodites 

 with respect to the gene B. Doxcaster (1913 a) has reported a some- 

 what similar experience with Abraxas grossulariata in which certain 

 females produced only female offspring, and he was able to show that 

 these females were characterized by an abnormal chromosome-number. 

 No cytological studies were made on my female-producing hermaphro- 

 dite Lychnis, but I do not think it probable that such an examination, if 

 made, would have discovered a cytological basis for its peculiar behavior^). 



IV. Couceriiing geuotypic formulae of the sexes. 



Various attitudes may be taken with regard to genetic formulae. 

 Some writers api)ear to hold (though ai)i)earances may be misleading in 

 this case) that such formulae i)resent a ])ictui'e of the actual constitution 



') These wholly female progenies may suggest to some the possibility of par- 

 thenogenesis, but against the plausibility of such a suggestion may be cited the fact that 

 in hundreds of the crosses which 1 have made among various strains of Lychnis dioiea, 

 the occurrence of an individual having only the characters of the mother would have 

 strikingly differentiated such individual from its sibs, but I have never seen a single 

 case of this kind, and do not believe that parthenoppnesis occurs in this species. 



