8 TORREYA 



rarely viable pollen. The appearance of such variably developed but func- 

 tionless or chiefly functionless organs is difficult to explain save by reduction 

 in the course of descent from a hermaphroditic ancestor. A different con- 

 ception may be based upon those species whose unisexual flowers show no 

 trace of organs of the opposite sex. A\'hen, however, such plants belong to, 

 or are obviously closely related to. families among whose members are some 

 with bisexual flowers or with unisexual flowers containing staminodia or 

 pistillodia. the obvious explanation of unisexuality is still that of descent 

 by reduction from a bisexual condition. There remain the relatively few 

 species that supply no indication, through either structure or relationship, 

 of such descent. It was this condition in Casuarina which made it. in A\'ett- 

 stein's phylogenetic scheme (19j, the starting-point for angiosperms. 



The other set of facts with a similar bearing is the variability of sexual 

 conditions in those angiosperms whose flowers are typically unisexual. In 

 many monoecious and dioecious species, bisexual flowers now and then appear. 

 Even more frequently, staminate replace pistillate flowers and vice versa : 

 flowers of either sex appearing on the dioecious plant or on the part of the 

 monoecious plant which regularly bears flowers of the opposite sex. W^hatever 

 its explanation, such lability of sex-expression in the sporophyte contrasts 

 sharply with the rigid separation of sexual characters in the gametophyte. 

 Comparable lability seems to characterize gyno- and andromonoecious, gyno- 

 and androdioecious species. 



In this connection, too, cases may be found which could be thought to 

 point in an opposite direction. Among the dioecious species that have been 

 extensively studied, three (two of Lychnis and one of Bryonia) present a very 

 sharp sex-separation. Doubtless, when other less well-known species are 

 studied, similar instances will be found. But in Bryonia dioica staminate 

 flowers, in L\chnis dioica and L. alba, both staminate and pistillate flowers, 

 contain rudiments of organs of the opposite sex. The change in sex-expression 

 in pistillate flowers or Lychnis under the influence of the anther smut is 

 well known, although human ingenuity has yet found no means of accom- 

 plishing a like result. No modification of sex-expression is known to have 

 been induced in Bryonia dioica, though there is one old report of a female 

 plant bearing some bisexual flowers. 



The common variability of the unisexual condition is among the facts 

 which long ago led to the conclusion that in all angiosperms genotypic bases 

 are present for both femaleness and maleness. Xo reason has appeared to 

 question this conclusion ; indeed, all later-adduced evidence has but served 

 to confirm it. Correns (2) postulated for dioecious species an additional gene 

 or gene-complex for sex tendencies superposed upon those representing sex 

 potentialities. This formulation, recognizing a certain degree of genotypic 



