1909.] Two Races of the Moth Acidalia virgularia. 149 



? D (1) in generation F 7 , ? D ^ (3) in generations F 7 and one section of 

 F 8 ;* dark x dark, ex hybrid, viz., C (3) in generation F 8 , c in generations 

 F 9 and Fi ) attained a considerable standard of purity ; and also that a few 

 of the extremest (light x dark) pairings among hybrids (such as b x c, b xf, 

 B * (5), and ? A (2) in generation F 7 ) were the most reliable in producing 

 again genuine intermediates. (GxC was possibly an exception, but the 

 parents here were not so extreme in colour as to render hybridity unthinkable.) 



Another fact that can be stated with certainty is that our experiments have 

 revealed no other decisive " reversion to type " than the kind which Mendelism 

 would demand ; the intermediates have been quite as stable as Mendelism 

 would expect in hybrid pairings. Whatever be the explanation, it would 

 appear that the hybrid form cannot be " bred out " ; except in cases where 

 a selective mating has been employed and the rest of the brood allowed to 

 die out, intermediates have continued to appear through all the generations. 



Without desiring to dogmatise, we feel it is necessary to remark that 

 neither of the points last considered — the obtaining of a comparatively 

 uniform type by selective mating and the persistence of intermediates under 

 other circumstances — belongs exclusively to any one theory of heredity, while 

 such occurrences as those noticed in the footnotes on broods A 2 and D % (3) 

 are harder to reconcile with Mendelism than with, for example, the Galtonian 

 view. On the whole, the apparently large responsibility of direct parenthood 

 suggests to us the idea of some such principle as is involved in the well- 

 known formula of one-half the characters from the parents, one-fourth from 

 the grandparents, etc. 



Over and over again some trifling race-characteristic has interested us in 

 a particular strain, including — besides the tendency for some broods to favour 

 the slightly darker variations and others the lighter — obvious differences in the 

 expression or suppression of the transverse " dot-lines," tendency to develop 

 a dark central shade or a dark marginal area (for instance, brood b (2)), and 

 so forth. Any of these would have been well worthy of minute study, either 

 from a Mendelian or a non-Mendelian point of view, had time and oppor- 

 tunity allowed. We suspect, however, that in large measure they also would 

 be found traceable to direct parentage, for it is certain that in some cases 

 cousin-broods differ quite materially in some of these characteristics, and that 

 a reference to their parents shows how closely these are followed ; see, for 

 instance, some of the references under the statistics of broods c, g, h, etc. 

 The sex-predominance in brood b (1) was another peculiarity which deserved 

 more attention than it received. 



* But how would Mendelism account for the (very variable) other section of F s ? 



