J. W. H. Harrison 259 



later. Although the variation in wing pattern cannot be studied 

 statistically and therefore so accurately as size, the degree of variation 

 was obviously less than in the F^ lot, for the extreme banded forms of 

 that generation showing some approach to filigrammaria were non- 

 existent. 



Now it has become the fashion to explain the absence of any appear- 

 ance of segregation of the grandparental form in cases like this on the 

 ground that in place of few clear cut unit characters depending on 

 similarly definite genes we are dealing with hosts of factors segregating 

 and recombining independently, some with and some without dominancy. 

 Under these circumstances, it is urged, it would be futile except in 

 thousands, perhaps millions, of cases to expect a perfect grandparental 

 segregate ; what we are to look for, we are told, is a smooth variational 

 curve of more extended spread than in the ^i brood. But we have 

 pointed out that the variation exhibited is most markedly less than that 

 shown by the F^ insects, which cannot be explained on other grounds 

 than the ordinary fluctuating and continuous variation due to the play 

 of the more or less plastic germinal material of the pure species. Why 

 then should we invoke multiple factors to explain F^ variation and not 

 that in F^ ? To drag in two distinct explanations for one and the same 

 phenomenon seems absurd. 



The balance of probability lies in favour of the view that we are 

 rather concerned with contamination of the gametes, or with permanent 

 gametic blending ; this I strongly suspect is the rule in interracial and 

 interspecific crosses, if any weight at all attaches to their generally 

 uniform evidence. 



Again the insects proved completely fertile in both sexes and a 

 number were confined for ova; these were successfully obtained and 

 hatched in due season in the following year. 



The F3 generation. 



As ova (hatched March 8, 1918)' and pupae this generation offered 

 no obvious peculiarities, but the case was different with the larvae, which 

 in about one half of the instances tended to exhibit a massing of the 

 darker dorsal suffusions on each side of the paler lines edging the dorsal 

 vessel, particularly just before the anterior trapezoidal tubercles, in this 

 fashion developing a pattern involving a series of dorsal lozenges. This 

 was unusual enough for me to select a single larva for description but 

 still not extraordinary enough to be considered beyond the possible range 

 of variation. Nevertheless, in the light of subsequent knowledge, it was 



Joum. of Gen. ix 17 



