12 T O R R E Y A 



and that, as shown by the persistence of rudimentary structures, very many 

 of them have become fixed as part of the specific genotype. 



The conclusion indicated by genetic evidence hence agrees with that most 

 strongly suggested by morphological study ; the general tendency in angio- 

 sperm evolution has been from a primitive hermaphroditism toward dioecism. 

 In many species various intermediate stages have been reached ; in something 

 less than 5000 according to available counts (20). the final step to dioecism 

 has been taken. 



The mutations that have been chiefly concerned in the evolution of sexual 

 conditions in angiosperms have invoh-ed a diminution or loss of the power 

 of spore-production ; commonly also a loss or reduction of the organs con- 

 cerned. The mutations of this nature which are appearing at present are with 

 rare exceptions recessive. It is reasonable to assume that similar mutations 

 in the past history of angiosperms have, at their origin, likewise been chiefly 

 recessive. Those mutations which have played the major role in floral evolu- 

 tion agree, therefore, in two respects — in involving a loss or diminution of 

 potentialities and in being originally recessive — with the general run of 

 observed mutations in all organisms. So far. then, as concerns one important 

 group of structures and functions within one subdivision of the plant kingdom, 

 evolution has proceeded by means of the t}'pe of mutation A\-hich genetic study 

 has shown to be the prevalent type. In connection with this particular ph}-]o- 

 genetic problem, the familiar difficulty of reconciling "progressive evolution" 

 with genetic results does not arise. 



It may be added that the mutative changes here shown to have been 

 important are in harmony with the tendency toward the sterilization of sporo- 

 genous tissue which has characterized the evolution of br}'ophytes. pterido- 

 phytes. and seed plants. 



The succession of steps in the changes from primitive hermaphroditism 

 must remain for the present speculative. Obviously male sterility and female 

 sterility may appear in different plants of a single species, as has happened 

 in Rubiis (4). In Riibiis, however, dioecism is not yet reached, for matings 

 of certain male and certain female plants produce some hermaphrodites and 

 some "neuters" (without functional stamens or pistils j. The species may at 

 present (not considering the neuters) be classed as trioecious. At least two 

 additional genetic changes would seem to be necessary (11) in order for 

 the ultimate goal to be reached. Since mutations are likely to occur inde- 

 pendently, it is to be expected that in the transition from hermaphroditism 

 species now dioecious have passed through several intermediate stages. 



It is possible to imagine the early steps to have been b}- wa}- of gyno- 

 or andromonoecism. trimonoecism. or monoecism. Any of these conditions 

 could conceivably be reached by the establishment in homozygous condition 



