ALLEN: EVOLUTION OF SEXUAL CHARACTERS 9 



complexity, is still, so far as it goes, valid as a formal statement of the case. 

 As will appear, it is now evident that the story is even more complicated than 

 Correns' statement would imply. 



A first suggestion of the complexity to be expected appears in the fact 

 that in dioecious angiosperms as in metazoa there are two general types of 

 genotypic influence upon sex. In one, the more common type in both groups, 

 the male is heterozygous, the female homozygous, for sex-tendency factors 

 as well as for sex chromosomes. In the other, represented by strawberries and 

 possibly by members of a few other genera, the male is homozygous, the female 

 heterozygous. The difference is explicable by descent from hermaphroditic 

 ancestors, different mutations in which have led to opposite results. It adds 

 to the improbability of an assumption of the primitiveness of dioecism. 



A type of mutation observed in a considerable number of hermaphroditic 

 species involves a stoppage at some stage in the development of stamens (or 

 their complete failure to develop), with the result either that no pollen is 

 produced or, if produced, it is nearly or quite functionless. "Male-sterile" 

 mutations of this general nature have been studied, for example, in the sweet 

 pea, shepherd's purse, sorghum, Oenothera, onion, tomato, potato, barley (16). 

 In these and in other plants, the condition in question seems to be due to a 

 mutated gene (or to at least two genes in the tomato), the mutation being 

 always or nearly always recessive. Mutations of a somewhat different sort 

 bring about a replacement of stamens by petals or petaloid structures. It is 

 clear that mutations of both types have occurred on a large scale in the past ; 

 witness the frequent occurrence, previously noted, of staminodia or stamen 

 rudiments replacing some of the stamens in bisexual flowers or all the stamens 

 in flowers which are now unisexual. Notable are the partial petaloid trans- 

 formation of the last-remaining stamen of Canna; the often-observed occur- 

 rence of doubleness in consequence of a transformation of stamens ; and the 

 evidence from the morphological side that petals in many cases represent 

 transformed and sterilized stamens. Mutations tending toward male sterility 

 occur likewise in monoecious species. In maize, the most studied genetically 

 of all plants, at least 27 distinct mutations of this general nature have been 

 observed (7) ; 20 classed as "male-sterile," 5 as "tassel ear," one each of 

 "anthefless" and "pollen lethal." These 27 mutations involve as many distinct 

 gene loci ; all but two are recessive. 



Such mutations in the direction of male sterility might be described as 

 tending toward femaleness. Those of another type, known for example in 

 Silene, Cheiranthus, and Papavcr, in which stamens are replaced by carpels, 

 may be similarly classed. 



Comparable with the mutations which result in or tend toward male 

 sterility are those leading toward female sterility. The striking fact s\\o\\n 



