272 Genetical Studies in Moths 



worn examples of that sex were captured. Roughly, the same period 

 covered the emergence of my sound domesticated brood ; on the contrary, 

 the bacteria infected insects failed to show themselves until October 15th 

 and straggled on, most unusually, until the middle of November. This 

 dilutata brood was not alone in thus exhibiting the phenomena of 

 infection and retardation, for both of the dilutata-autumnata crosses of 

 that year likewise contracted the disease, and they too were late in 

 emerging. Despite the infection, as is customary, the females of the 

 autumnata $ x dilutata </ lot were obtained in June and July, proving 

 thereby that the presence of bacteria did not inhibit somatic develop- 

 ment ; on the other hand the males of both broods, as well as the odd 

 female of the reciprocal cross, instead of accompanying the alderwood 

 insects only started to emerge on October 13th, subsequent to which 

 date they yielded small numbers over a protracted period. 



From what has been related it is manifest that further work with 

 the hybrid females was hopelessly excluded, but matters with the males 

 were different. Genitalia, primary and secondary, external and internal 

 were alike perfect, the testes containing an abundance of well developed 

 spermatozoa. Arrangements were therefore made to confine the males 

 in cages with autumnata and dilutata females to secure the apparently 

 certain back crosses. But matters did not proceed smoothly ; ova were 

 not forthcoming immediately, and it was only after a fairly long interval 

 had elapsed that eggs were obtained. In the two crosses no marked 

 differences in the fertility of the ova were noticed, some 30 °/^ in both 

 cases being fertile. Spring, however, told a different tale, for Svery ovum 

 from the pairing between autumnata ^ and the I\ </ collapsed after 

 development had been initiated, whilst every fertile egg in the cross 

 between dilutata $ and F^ (/ delivered a sturdy little larva which fed up 

 speedily and, until the last moult, well. These larvae, in direct opposi- 

 tion to their ostensible three-quarters dilutata blood, could neither in 

 design nor habits be distinguished from those of the F^ generation of 

 the autumnata- dilutata crosses. There was not the slightest sign of 

 passage toward dilutata — a totally different phenomenon from the almost 

 instant return, on back crossing, of the Fi autumnata x jiligrammaria 

 insects to their parent forms. This undoubtedly depends directly on 

 the different courses pursued in the gametogenesis of the two F-^ broods. 

 In the filigram^maria $ x autumnata ^ cross nearly all (sometimes 

 actually all) of the chromosomes find mates previous to the first matura- 

 tion division, so that a reduction division does occur accompanied, in my 

 opinion, with an "averaging up" of the values attachable to the various 



