10 T O R R E Y A 



bv a review of the literature is that mutations of this type appear, in both 

 hermaphroditic and monoecious species, to be far less frequent than are 

 those leading toward male sterility. Female sterility, pistils being more or 

 less aborted, seems to be recessive in mutants of Phleum pratcnse, Anfirrhi- 

 lutiu, rice, and raspberries. In calycanthcuia forms of Campanula and of 

 Rlwdodcndron. the mutant condition (pistils developed but sterile) is domi- 

 nant. In Geranium, pistils functional as such but showing structural transitions 

 toward the staminate character have appeared in interracial crosses. The 

 behavior of the character in back-crosses, while not entirely clear, suggests 

 a ]Mendelian segregation. In a cross between species of Geum, the results of 

 further matings are likewise not clear-cut. It is possible that in these crosses, 

 as apparently in a few interspecific crosses which have resulted in male 

 sterility, cytoplasmic influences are involved. 



In the same list of mutations in maize which shows 27 genes involved in 

 male sterility, only 6 mutations leading toward female sterility are reported : 

 two "anther ear," 2 "barren stalk," 1 "lethal ovule," 1 "silkless." 



To the story of observed mutations in this direction must be added the 

 known cases of pistillodia, which represent the result of past mutations ; 

 possibly the reduction in number of ovules in certain lines ; and the relatively 

 few cases of doubleness which have involved the transformation of pistils 

 as well as of stamens into petals. 



To be mentioned also are a few known mutations which, like two observed 

 in intervarietal crosses in Oryza (12), tend simultaneously toward both male 

 and female sterility. Obviously mutations of this class can hardh- have played 

 a direct part in evolution. In general they seem to result in monstrosities 

 which, even apart from the accompaming sterility, would probably not be 

 favored by selection. Chromosomal changes may be involved. 



It is not yet clear why mutations toward male sterility are much more 

 frequent than those toward female sterility. This difference holds not only 

 for observed mutations. As to the past, it is evident that petals in a large 

 proportion of instances represent sterilized stamens ; only rarely can they 

 be considered sterilized carpels. 



The mutations thus far cited involve changes in the general direction 

 from hermaphroditism toward dioecism. As already mentioned, very many 

 variations (to be distinguished from mutations) occur in the opposite direc- 

 tion—involving the appearance in monoecious or dioecious species of bisexual 

 flowers or, in dioecious species, of both pistillate and staminate (and some- 

 times bisexual) flowers on the same plant. Such variations are in large 

 measure shown to be reactions to environmental conditions. They may be 

 considered expressions of genot3-pic possibilities present from a remote 



