252 Genetical Studies in Moths 



what was expected from the last but certainly not in the third case. 

 The exact composition of the broods, with that of the ^i generation to 

 complete the series, is appended in tabular form. 



To anyone acquainted with the work of Morgan \ Doncaster^ and 

 others and the ratios obtained in their sex-linkage work, this table 

 adequately demonstrates that here we are dealing with a case where 

 the male is sexually a homozygote and the female a heterozygote, and 

 that the gene for latifasciata-ness is coupled with, or follows the dis- 

 tribution of, the male sex gene, or is located in the X chromosome, to 

 give all of the various expressions used by different workers according 

 to the views they accept. 



In the conventional fashion, by the use of symbols, we can render 

 this clearer by allowing L to stand for the latifasciata factor and A for 

 its absence, or rather for the ordinary autumnata-ness. Then, since the 

 original wild female was not pure for the character concerned, for its 

 ova gave rise to both autumnata and latifasciata, we shall have to indi- 

 cate its genotypic composition as AL^ % producing the two kinds of 

 gamete A ? and L^ in equal numbers, and that of autumnata male 

 as AA^^ yielding gametes all A^. This would ensure that on the 

 fusion of the gametes producing the F^ generation equal numbers of 

 two zygotic types AL^ ^ and AA^ % would appear. The latter ap- 

 peared as typical females and the former as the heterozygous latifasciata 

 males. At gametogenesis the -^k/c/ produce two kinds of gamete, 

 A^ and L^^, and the females likewise two, A^^ and A %. Now how 

 does this constitution of the F^ gametes square with the results in 

 the F^ generation and in crosses 2, 3 and 4 ? In cross 2 we mate a 

 pure autumnata male, therefore yielding gametes all of composition A (/ 

 with an F^ ? ; these meet in fertilisation on our hypothesis with gametes 

 constituted A </ and A $ and thus develop zygotes of the forms AA ^ ^ 

 and A A </ % which are ordinary autumnata — and precisely what the 

 experiment yielded. Again in cross 3 a heterozygous latifasciata male 

 was paired with an autumnata female. In this instance the male gives 

 gametes of the two types J.^/' and Z,/ as does the autumnata female, 

 but in this case they are of composition A </ and J. $ ; all four gametes 

 are generated in equal quantities. Theoretically from this mating 

 zygotes of four kinds, AA^ ^ {autumnata males), AL^^ (^ (heterozygous 

 latifasciata males), AA(^% {autumnata females) and AL(^% (hetero- 

 zygous latifasciata females) should appear in equal proportions. As we 



1 Morgan. Heredity and Sex. Columbia Univ. Press, New York (1913). 



2 Doncaster. The Determination of Sex. Univ. Press, Cambridge (1914). 



