DETERMINATION OF SEX DON CASTER. 481 



of Petrunkewitsch, 1 who maintains that the testis of the drone is 

 derived from the fused polar nuclei of the unfertilized egg. Since 

 the second polar nucleus by hypothesis contains the female determi- 

 nant, the spermatozoa may be female bearing and cause the fertilized 

 egg to be female. But there is considerable doubt as to the accuracy 

 of the observation, and in any case it can not be applied to some 

 others; e. g., Silvestri (see above) finds that in Litomastix the polar 

 nuclei are used up in forming the protecting membrane of the em- 

 bryos, and yet the sex phenomena are just as in the bee. An alterna- 

 tive speculation may be offered. It is possible that while the female 

 determinant is extruded in the virgin egg with the second polar 

 nucleus, the presence of a spermatozoon in the egg may cause the male 

 determinant to be eliminated, leaving the egg nucleus female bearing. 2 

 This would fall into line with the explanation suggested above of 

 gynandromorphic bees — that in them the sperm has conjugated with 

 the second polar nucleus. 



It noAv remains to describe work on the determination of sex which 

 has led to similar conclusions to those suggested above, but arrived 

 at from a different starting point. Castle suggested that as sex is 

 inherited as a Mendelian unit, it might at times be " coupled " in the 

 gamete with some other body character. It has been found that 

 something of this kind actually does take place in the case of certain 

 varieties which are inherited differently by the male and female. As 

 an illustration of this we will take some experiments made by the 

 writer, accounts of which have already been published, 3 since the 

 results in that case happen to be simpler than in some other instances 

 which have been worked out. 



In the common currant moth (Abraxas grosstdariata) there is a 

 rare and very distinct variety (" lacticolor ") found in the wild state 

 almost exclusively in the female. Crossing experiments were made 

 between this variety and the type form ; the results were as follows : 



(1) Lacticolor 2 X type $ gave type $, type 2. 



(2) Heterozygous (crossed) type 2 X heterozygous type male gave 



type $ , type 5 , lacticolor 2 . 



(3) Lacticolor 2 X heterozygous type $ gave J type $ - lacticolor $ - 



[ type 2 , lacticolor 2 . 



(4) Heterozygous type 2 X lacticolor $ gave type $, lacticolor 2. 



(5) Lacticolor 2 X lacticolor $ gave lacticolor c? and 2. 



(6) Wild type 2 X lacticolor $ gave type $, lacticolor 2. 



These results at first may seem hopelessly confusing, but there are 

 several points of interest about them. Firstly, the lacticolor charac- 

 ter behaves as a Menedlian recessive, disappearing in the first cross 



1 Zool. Jahrb., vol. 14, 1901, and vol. 17, 1903. 



2 I find that this suggestion has also been made by Prof. T. H. Morgan. 



8 Doncaster and Raynor, Proc. Zoo. Soc. 1906, vol. 1, p. 125 ; and Doncaster, Reports 

 to Roy. Soc. Evolution Committee, vol. 4, 1908, p. 53. 



97578°— sm 1910 31 



