480 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1910. 



. The essence of Wilson's hypothesis is that the sex determinants 

 behave as Mendelian characters, segregating from one another in 

 gametogenesis (at the reduction division), that femaleness is domi- 

 nant, and that there is selective fertilization ; so that a male-bearing 

 egg is fertilized by a female-bearing spermatozoon. Suggestions 

 closely similar to these were put forward by Castle 1 on quite different 

 grounds in 1903. Castle collected a quantity of evidence from breed- 

 ing experiments and from what is known with regard to partheno- 

 genetic reproduction. He supposed that every individual arising 

 from a fertilized egg is heterozygous (hybrid) in respect of sex, and 

 that segregation takes place at the second maturation division, so 

 that half the gametes bear maleness, half femaleness. Male-bearing 

 eggs conjugate with female-bearing spermatozoa and vice versa ; but 

 dominance is alternative, so that roughly half develop into each sex. 

 In most parthenogenetic animals only one polar body is produced in 

 eggs which will not be fertilized ; in these cases femaleness is always 

 supposed to dominate. Since with only one polar division no segre- 

 gation takes place, the offspring are females. If in a parthenogenetic 

 species two polar bodies are produced the offspring are commonly 

 males, since the female determinant is supposed to be eliminated with 

 the second polar nucleus. A further valuable suggestion was that 

 the male and female determinants might be " coupled " with certain 

 body characters, either invariably — so explaining sexual dimor- 

 phism — or frequently, by which is explained the general association of 

 one variety with one sex, another with the other, in the offspring of 

 certain crosses. Wilson has since- observed coupling of ordinary 

 with idiochromosomes, which may be connected with this phenom- 

 enon. 



Castle's suggestive paper has stimulated much work on the matura- 

 tion of parthenogenetic species, but his hypotheses do not always 

 hold good. For example, it is now known that parthenogeneticalty 

 produced males in the Aphides arise from eggs which have only one 

 maturation division; 3 some of his explanations of other exceptional 

 cases, although ingenious, will not now bear critical examination. 

 One of the most difficult is that of the hive bee and those insects re- 

 sembling it, in which all eggs have two polar divisions and when 

 fertilized yield females, when parthenogenetic, males. Castle sup- 

 posed that the female determinant is extruded with the second polar 

 nucleus, leaving the egg male bearing, and accepted the observations 



1 " The Heredity of Sex," Bull. Mas. Comp. Zoo. Harvard, vol. 40, p. 189. 



2 Science, May 17, 1907, p. 779. 



3 B. g. Stevens, Journ. Exp. Zoo., vol. 2, 1905, p. 313. But Morgan (Proc. Soc. Exp. 

 Biol, and Med., vol. 5, 1908, p. 5G) finds in Phylloxera that the females have six chromo- 

 somes, the males five. And Stevens (Journ. Exp. Zoo., vol. 6, 1909, p. 115) susrsests that 

 in Aphis also one complete chromosome is extruded in the maturation of male partheno- 

 genetic eggs, but not in female eggs. [This has since been confirmed. Biol. Bulletin, 

 vol. 18, p. 72, 1910.] 



