EXPERIMENTS IN HYBKIDIZATION AND TEMPERATURE. 77 



phenomena closely related to gynaudromorphism. On the other 

 hand, it is obvious that the individuals are not hermaphrodites. 

 The females remain females, only their fertility has suffered by 

 the experiment, or has been destroyed by the same. 



At this opportunity it would be well to mention, that female 

 germs seem to be much more sensible to influences than the 

 male. 



Firstly, all were females in the above-mentioned temperature 

 experiments, at least as far as our present experience goes. 



Secondly, the gynandromorphic individuals of the secondary 

 hybrids are predominantly female in their character. 



Thirdly, the fact that the female hybrids, with rare ex- 

 ceptions, are not capable of reproduction, whereas the male 

 hybrids are always fertile. 



Fourthly, must here be mentioned the inclination of female 

 individuals to sterility when badly fed, if this want of food is 

 caused by man (I have mentioned such instances in my book 

 ' Paliiarctische Gross-Schmetterlinge,' pp. 194, 195), or from 

 the habits of certain social insects — for instance, Hymenoptera 

 (ants and bees) and Termites. Prof. Emery (Bologna) has 

 pul)lished many facts about this alimentary castration, as he 

 calls it, of ants (' Compte-rendu des seances du troisieme congres 

 international de Zoologie,' Sept. IGth to '21st, 1895 : Leyde, 

 pp. 395-410). 



But now to return to the continuation of the hybridization 

 experiments in 1897. Among the primary hybrids obtained in 

 the first crossings between the Saturnias spini, pyri, and pavonia, 

 neither in my own experiments nor in those made by others was 

 a single primary hybrid obtained which was capable of repro- 

 duction, even in the slightest degree, when crossed with one 

 another, only one secondary hybrid. No. 18, was fertile to a 

 slight degree. 



Moreover, no other primary hybrid bred by me had the 

 necessary qualifications, with the exception of the crossing 

 between Zygcena trifolii 3 and Z. Jilipendalce $ ; and that cross- 

 ing was, at that period, unfortunately not investigated with 

 regard to this question ; nor has such a capability of reproduc- 

 tion been noticed in any of the other Lepidopterous hybrids which 

 have been reared up to the present. 



It was not to be expected, according to present experience, 

 that this blank in the experiments could be filled by related 

 species which exhibited marked differences in physiognomical 

 aspect ; and, on the other hand, nearly related types, with 

 habitats either entirely or nearly entirely separated, could not 

 be used for the experiment. In the latter case, there was always 

 the possibility that these were only local races of the same 

 species. It is well known that local races of the same species 

 often exhibit highly important physiognomical differences. 



