482 ANNUAL. EEPOET SMITHSONIAN INSTITUTION, 1910. 



(No. 1) and reappearing after mating (2). Secondly, starting with 

 a lacticolor female, it is possible to get males of that variety only in 

 one way, viz, by pairing such a female with a heterozygous male, 

 i. e., a male which is typical in appearance, but being of lacticolor 

 parentage bears the recessive lacticolor character. Lacticolor males 

 are also produced from mating lacticolor males and females together 

 (No. 5) ; but from any other form of union all the lacticolor indi- 

 viduals which emerge are females. A third point of great importance 

 is that converse crosses do not give similar results, the most unex- 

 pected case of this appearing in matings of types No. 1 and No. 6. 

 In the first, a lacticolor female paired with a wild (pure) type male 

 gives all the offspring of both sexes perfectly typical, a quite normal 

 Mendelian result, since lacticolor is recessive to the type. But if an 

 apparently pure, wild female is mated with a lacticolor male, the 

 male offspring are typical, but the female are all lacticolor — exactly 

 the same result in fact as when a first-cross female is used instead 

 of a wild one. 



In explaining these phenomena in the first paper this last result 

 was not known; and it was suggested (in accordance with Castle's 

 hypothesis) that the germ cells bore one or the other sex, that fer- 

 tilization was selective, so that all individuals were heterozygous in 

 respect of sex and that in the eggs the lacticolor character was 

 coupled with the female sex determinant. Later Bateson and Pun- 

 nett^ offered a modified hypothesis, which is perhaps more in accord 

 with the facts as known at present. They suggest (1) that the sex 

 determinants behave as Mendelian units, femaleness being uniformly 

 dominant over maleness; (2) that female individuals are heterozy- 

 gous in respect of sex. being of the constitution 2 3 and producing 

 male-bearing and female-bearing eggs in equal numbers; males are 

 homozygous in sex, of the constitution S S , so that they produce 

 only male-bearing spermatozoa ; (3) that there is repulsion in 

 oogenesis between the dominant determinant for femaleness and the 

 dominant grossulariata (type) determinant, in consequence of which 

 all male-bearing eggs bear the type, all female-bearing eggs the 

 lacticolor character. 



This suggestion completely accounts for the facts and has since 

 been greatly supported by the discovery that all females with the 

 type (grossulariata) character are heterozygous and produce lacti- 

 color offspring when paired with a lacticolor male. This fact com- 

 pels us to assume that the lacticolor determinant is present in all 

 females of the species and is only prevented from appearing because 

 typical males bear normally only the type (grossulariata) character, 

 which dominates over lacticolor. If, then, the males are homozygous 



i Science, vol. 27, 190S, p. 785. 



