L. DONCASTER 3 



from the unisexual family '10. 'J8, three ('11. 15, '11. /.9, 7i. 2^) gave only 

 female offspring, and that of these families one was by a wild male, the 

 other two by males unrelated to the female. Half the matings of 

 females from '10.2H were nearly or totally infertile, in two cjises ('11.27 

 and '11.29) when the male was unrelated to the female. Similar 

 infertility was found in two of the three matings of females from *10. 21, 

 and in four matings (not included in' the table) of females from '10. 10 

 and '10.17 (families in which there was a irn-jit pn'ponderance of 

 females). 



A fresh point of importance also appeared. In three matings {'11. ,?/7, 

 '11. 9 and '11. Jf) of which the female parent in each case was daughter 

 of a female from '09. 18, and the male in two cases a son of a female 

 from '09. 18, and in the third a member of the preponderatingly female 

 family '10. 17, only females were produced, although both parents were 

 themselves members of bisexual families. 



It is clear, therefore, not only that the tendency to produce purely 

 female families may be transmitted direct from mother to daughter, 

 even when the male parent is unrelated, but also that females of 

 bisexual families directly descended from unisexual families may produce 

 only female offspring, at least when mated with males of similar origin. 

 Further, of the females of unisexual families, about half have only 

 female offspring, and the remainder have offspring of both sexes. 



Several points of importance remain to be determined, and it is to 

 be hoped that the results of the pairings made in 1912 will clear them 

 up. The more important are (1) whether females of bisexual families, 

 whose mothers belong to unisexual families, can have only female 

 offspring when mated with unrelated males ; (2) how the families with 

 great preponderance of females are related to the completely unisexual 

 broods ; and (3) whether mortality in the egg or larval stages has any 

 relation to the production of only female offspring. Experiments to 

 test all these points are in progress. It was hoped that the third 

 would be settled by the 1911 matings, but the exceptionally hot dry 

 weather at the time of hatching caused great mortality in the very 

 early stages. Although in some cases a large proportion of the eggs 

 failed to develop, in other batches from which only females were pro- 

 duced almost every Qgg hatched, and the case of '10. 28, in which 62 

 females were reared from 91 eggs, strongly suggests that the production 

 of only female offspring is not due to the dying off of the male larvae. 

 Further evidence in the same direction is provided by the preliminary 

 account of the chromosomes in unisexual broods which follows. Another 



1-2 



