J. W. H. Harrison 77 



(e) Thirty (sixteen females and fourteen males) would be correctly 

 described as pale, although all were sprinkled more or less with fine 

 blackish powdering. One alone perhaps if inserted in a series of pure 

 T. crepuscularia would have passed unchallenged. 



Despite this attempt to disentangle their variation, from this family, 

 as from the last, we gain the impression which entirely negatives any 

 occurrence of segregation on Mendelian lines in the gametogenesis of 

 both sets of Fi hybrids. A full examination of the possible causes of 

 this failure and of other results of these experiments will be found in 

 the general discussion below. 



Before this is undertaken one further and extremely interesting fact 

 must be chronicled. In families VII and X, and in one* obtained from 

 a crossing between ordinary types of T. histortata and T. crep^iscularia, 

 parthenogenesis was observed in the case of four females, one from each 

 of the first two broods and two from the last named. Of these only the 

 first put us in possession of facts vital to the present research, although 

 only three imagines (one female and two males) were reared. 



The female would not be out of place either in brood IX or in the 

 nondescript portion of family X, On the contrary, one of the males 

 resembles the blackest forms bred, whilst the second is a beautiful insect 

 of form not previously encountered. Probably it ought to be described 

 as pale, as its ground colour is white, heavily " peppered," and shaded 

 with dove grey scales, the shadings becoming more intense terminally. 

 On both anterior and posterior wings it exhibits a strengthening of the 

 median black line, but all of the other transverse lines, in addition to 

 the usual white subterminal band or line, are obsolete. It strongly 

 reminds one of some of the caesious species of the genus Gnophos in all 

 of its characteristics. 



IV. Discussion of the Results. 



We have discovered a case in which one and the same unit character 

 displays inheritance in absolute conformity with Mendelian conceptions 

 at one time, yet at another behaves so aberrantly as to suggest the 

 workings of a scheme of inheritance fundamentally different from the 

 former, the character in question being introduced by insects from 

 the same brood. Let us, however, be careful to explain that the phe- 



1 And later in a female of parentage {crepuscularia 5 x bistortata /J ) $ x {bUtortata 2 

 X crepuscularia (J) <J. 



