Selection upon Mendelian Character 189 



series still produces one red ear in a thousand, and the red series 

 produces a considerably larger proportion of white ears. 2 



Let us now compare with these series the results obtained by 

 Miss McCracken in selecting for the conditions bivoltinism or 

 univoltinism in a crossed race of silk-moths. 



The original cross (1904) was made between a univoltin female 

 and a male of bivoltin race. Ten of the Fj female moths were 

 tested, and six of these proved to be bivoltin, four univoltin in 

 character. This looks like a Mendelian 1:1 ratio and suggests 

 that one or the other parent in the original cross was heterozy- 

 gous. But the results obtained by Toyama ('06) indicate that 

 univoltinism is dominant over bivoltinism, and there is nothing 

 in the results of Miss McCracken at variance with this idea. It 

 so, the original female was a heterozygote, U (B), and when mated 

 with a bivoltin male produced offspring half heterozygous, U (B), 

 half pure bivoltin, BB. Matings of the Fj, offspring with pure 

 bivoltins would have settled this point, but no such matings were 

 made. We are left therefore with only such information as is 

 afforded by five matings of the Fi individuals inter se and by 

 twenty-four matings of the Fj males with univoltin females said 

 to be of pure race. 



These twenty-four females at any rate, proved to be all uni- 

 voltin, but there is reason to think that not all of them were ho- 

 mozygous m that character. For of thirty tested females ob- 

 tained from this cross, two proved to be bivoltin. Either, there- 

 fore, univoltinism is not always dominant over bivoltinism, or 

 else one or more of these univoltin females of pure race was in 

 reality heterozygous, U (B). There is nothing in this assumption 

 at variance with the statement that bivoltinism did not occur in 

 the race from which these twenty-four univoltin females came. 

 For the results of Doncaster ('08) on Abraxas, of Miss Durham 

 ('08) on canary-birds, and of De Vries ('08), on Oenothera show 

 that in a wide variety of organisms one sex may be regularly heter- 

 ozygous in gametic composition without resulting in the produc- 

 tion of a single recessive individual, except in racial out-crosses. 



- These tables show us what a serious task it would have been for our Puritan ancestors to eliminate 

 from their harvests the occasional red ear which caused such joyous confusion at the New England 

 husking-bees, even had their austere consciences demanded the undertaking. 



