254 Genetical Studies in Moths 



emerge months before their brothers, whilst in the former only one 

 female was bred and was killed for dissection before its unique nature 

 had been observed. Nor was the position of the males much better ; 

 their fertility was not greatly inferior to that of the pure species, but, 

 on the contrary, the viability of any zygotes in the production of which 

 their gametes took part was so slight that of the four possible back 

 crosses only two gave larvae proceeding beyond the first instar. More- 

 over, of these 'two sets, members of only one managed to pupate, and it 

 gave only five male pupae which died shortly after their maturation 

 divisions commenced. Progress with these back crosses seemed impos- 

 sible ; nevertheless, in the hope that success might crown further efforts, 

 they were repeated in 1915, when the families repeated the conditions 

 of the previous season in every respect, one isolated female, wholly of 

 autumnata facies, appearing in the autumnata ^ x dilutata % brood. 

 This was paired immediately with an ordinary autumnata male and in 

 due course 120 fertile ova agreeing exactly in size, colouration and 

 sculpturing with genuine autumnata ova were deposited. They hatched 

 in 1916 and gave rise to a brood of 87 imagines (50 males and 37 females) ; 

 these proved in all respects and at all stages indistinguishable from 

 pure autumnata. Crossed back with autumnata of pure wild blood and 

 paired inter se they produced the same insect in 1917. This type of 

 behaviour is susceptible of analysis and explanation as employed in the 

 symbolical consideration of cross 2 above. It is probably a genuine 

 case of sex-linked inheritance ; still, a warning in this case is obligatory. 

 The case is complicated by the absolute failure of reduction division in 

 the gametogenesis of ^i autumnata x dilutata hybrids, reinforced by the 

 fact itself that only one female appeared in each of two broods. 



Anomalies in chromosome behaviour other than the typical me- 

 chanism of sex-linkage may be at work. Furthermore, in 1917, when 

 once more a single female was reared from several broods of the same 

 hybrid, the female was half- sized and on pairing yielded no ova ; subse- 

 quent dissection proved it to lack any trace of ovaries. 



« 

 (e) Hybridisation Experiments. 



(1) Hybrids between O. autumnata and 0. filigrammaria. 



The Fi generation. 



As we have insisted at several points above, filigrammaria begins 

 to emerge from the pupae toward the end of August and continues to 

 do so in diminishing numbers until mid September. Add to this the 



