J. W. H. Harrison 253 



see, four types are obtained ; but whilst quite in accord with the pre- 

 diction of equality when their paucity is considered this approximation 

 cannot be deemed, in view of the small numbers, as very weighty evi- 

 dence in favour of that equality. 



Genetical analysis of the interspecific cross between dilutata female 

 and Fi latifasciata male need not delay us because the insects are too 

 few to base any argument on. 



There only remains then the F^ generation fo discuss ; to bring it 

 into being an F^ male, producing gametes L(^ and AJ^, was mated 

 with a female of similar origin postulated in our scheme to yield equal 

 numbers of gametes A (/ and A $ . This, as one readily perceives, agrees 

 exactly with the conditions in cross 3 ; naturally therefore the breeding 

 results should be similar, i.e. four zygotic forms, heterozygous latifasciata 

 males, autumnata males, heterozygous latifasciata females and autumnata 

 females, should be obtained in equal numbers, as was indeed the case, 

 the exact figures being 19 latifasciata males, 18 autumnata males, 

 18 latifasciata females, and 21 autumnata females. And the numbers 

 dealt with are great enough to exclude any action due to mere coin- 

 cidence. 



It will be noted that two possible pairings are not included ; these 

 are between heterozygous and homozygous latifasciata males on the 

 one side and latifasciata females on the other ; the reason is obvious ; 

 until latifasciata males had been secured simultaneously with similar 

 females neither pairing was possible, the first becoming so in 1918 and 

 the latter in 1919. 



The complete agreement between predicted and actual results of the 

 experiment thus serves to confirm the view that the gene determining 

 progeny of the latifasciata type shares the distribution of the male sex 

 factor, and furthermore that the original female was heterozygous for 

 both, thereby confirming Doncaster's grossulariata — lacticolor results. 

 Incidentally the fact emerges that a female homozygous for latifasciata- 

 ness can never exist unless some further mutation determining a change 

 in location of the gene for that character takes place. 



We shall now take up the same question in connection with inter- 

 specific crosses between 0. dilutata and 0. autumnata. As was indicated 

 above, reciprocal crosses between autumnata and dilutata exhibited 

 inheritance on a sex-linked basis, apparently involving every gene deter- 

 mining wing markings in the cross between autumnata males and 

 dilutata females, and only some in the reverse cross. But, unfortunately 

 for the genetical analysis of these cases, in the latter all of the females 



