ALLEN: EVOLUTION OF SEXUAL CHARACTERS 13 



of one or (more probably) more than one mutant gene. How dioecism may 

 arise from monoecism is illustrated by the success of Jones (8, 9) and of 

 Emerson (6) in the production of dioecious races of maize through the selec- 

 tion of appropriate mutations. In each case two mutations were involved ; 

 and in each of the three dioecious races obtained, one pair of chromosomes 

 differed with respect to a mutant gene which is epistatic to the mutant allele 

 of the other selected pair. 



Another conceivable transition from hermaphroditism to dioecism is by 

 way of gynodioecism, w^iich has been considerably studied, or of andro- 

 dioecism, about which nothing is known genetically. If within a hermaphro- 

 ditic species male sterility becomes a fixed character of one strain, evidently 

 other strains of the species must usually retain functional pistils if the species 

 is to persist — that is, a condition of gynodioecism must ensue. An alternative 

 would be the development (by an additional mutation) of a structurally 

 female but functionally parthenogenetic species. This seems to have happened 

 in Hieracimn excellens (14) ; but such mutational coincidences must be rare. 

 A mutation (or mutations) transforming the hermaphrodites of a gyno- 

 dioecious species into males would lead to dioecism. 



The difficulty of explaining the behavior of gynodioecious species by any 

 simple genie scheme led Wettstein (18) to the assumption of a cytoplasmic 

 influence — an idea tentatively accepted by Correns (3) and recently empha- 

 sized and generalized by Lewis (10, 11). Apart from the inadequacy of an 

 explanation based upon one or two mutations, the argument for a cytoplasmic 

 inhibition in the female upon the functioning of male-tending genes rests 

 upon the demonstration of such an apparent influence in several typically 

 hermaphroditic plants, including forms of Liniim, Nicotiana, Geranium, 

 Epilohiuui, and Strcptocarpiis. With the exception of one case in maize (15), 

 the known phenomena of this nature are limited to interspecific hybrids. On 

 the other hand, also, gene mutations leading to male sterility are, as has been 

 seen, of frequent occurrence. It is entirely possible that, when the variable 

 behavior of gynodioecious species becomes better known, a (perhaps com- 

 plicated) Mendelian explanation may be found possible. 



Nearly twenty years ago Emerson (5), pointing out that "there are at 

 least nine pairs of genetic factors which influence the expression of sex in 

 maize," suggested that "the genetic situation in maize .... may perhaps 

 afford some help toward a solution of sex problems." The prophecy has 

 been abundantly confirmed. Today more than 40 genes are known in maize, 

 borne on at least 9 of the 10 chromosomes. Avhose presence in the "normal" 

 or usual condition is directly essential to the sex-expression typical of the 

 species as it exists at present. There are others, likewise essential in this 

 regard, whose more conspicuous influence is upon the form, size, or vigor 



