1906.] 



BREEDING EXPERIMENTS AYITH LEPIDOPTERA. 



131 



Table X. — Heterozyo-ous g x lacticolor S • 

 (Type DR? X R d' .) 



Table XI. — Lacticolor $ X lacticolor S ■ (Type R $ x R J .) 



It will be noticed that while the results given in the tables are 

 -qualitatively in full agreement with Mendel's Law, yet the 

 numbers depart widely from Mendelian expectation. There is 

 always great mortality in rearing insects, especially in those 

 species which hibernate in the larval state, and the discrepancy 

 is probably accounted for by the greater strength and healthiness 

 of lacticolor, which we have frequently noted, and which has 

 doubtless caused a selective mortality in favour of the variety as 

 compared with the type. 



The following tentative hypothesis is put forward to account 

 for the relations between the variety lacticolor and the sexes. 



Castle * has suggested that the determinants for the two sexes 

 are segregated from one another in gametogenesis like Mendelian 

 characters, and that a male-bearing spermatozoon always meets a 

 female-bearing Qgg or vice versa, so that in respect of sex all 

 zygotes are heterozygous. He has further supposed that somatic 

 characters may occasionally be coupled with one or other sex- 

 determinant, so that of the gametes produced by a heterozygote 

 AB, the male-bearing may all carry one somatic character A, 

 Avhile the female-bearing carry its allelomorph B. 



* Castle, " Heredity of Sex," Bull. Mus. Zool. Harvard, xi. no. 4, 1903, pp. 189, 208. 



