THE ORIGIN OF GYNANDROMORPHS. 79 



of chromosomal elimination applies to von Siebold's and von Enf];cl- 

 hardt's gjmandromorphs. 



In the first place, it is important to understand that there is no 

 conclusive evidence that the racial difference here involved lias 

 anything to do with one particular chromosome, or even with the 

 sex chromosome. In bees the mother transmits her characters 

 directly to her sons, as is the case in sex-linked inheritance of the 

 Drosophila type, but in bees this form of inheritance is obviously due 

 to the fact that the male develops directly from the unfertilized egg, 

 hence must inherit all the maternal characters, whether in the sex 

 chromosome or in the autosomes. The special sex differences are, 

 of course, due to whatever it is that makes the egg a male or a female. 

 In cases where the queen is heterozygous she may produce two kinds 

 of sons, which is expected if the two races differ in one Mendelian 

 gene,^ but this would hold whether this pair of genes is autosomal or 

 in a pair of sex chromosomes. If the two races differ in more than one 

 pair of genes, more than two kinds of males are expected. The clearest 

 evidence that we have in regard to what a hybrid queen produces is 

 furnished by the recent work of Newell (1914) and Quinn (1916). 

 Newell crossed a yellow Italian queen bee to a gray Carniolan drone. 

 The daughters were yellow like the Italian, showing the dominance 

 of that color. In the reciprocal cross, Carniolan female by Italian 

 drone, the daughters were also yellow, but not as completely so as in 

 the last. Whether this is due to modifying factors of some kind is not 

 known. Quinn, as stated above, used Italian and Caucasus races, 

 crossing both ways, and in both the daughters were the same, viz, 

 3"ellow like the dominant Italian race. He also found that the Fi 

 daughters gave two kinds of drones and two only, which indicates that 

 the factor difference is carried by a pair of chromosomes, but this evi- 

 dence alone does not show that the pair is the pair of sex chromosomes, 

 for any other pair would give in the bee the same result. However, 

 when taken in connection with the gynandromorph results, the evidence 

 becomes somewhat stronger that sex chromosomes are involved. 



What, then, is the expectation on the elimination view? It is at 

 once apparent that the elimination must involve a sex chromosome, 

 for, otherwise, there is no reason to suppose that an autosomal dif- 

 ference would at the same time be associated with a difference in sex. 

 In other words, the elimination hypothesis can apply here only if the 

 chromosome that determines sex is the same chromosome that causes 

 this racial difference. 



Elimination of one of the sex chromosomes that carries the factor 

 for mellifica would produce a cell containing only the ligustica-heanng 

 chromosome, and all parts descending from that cell would be both 



^ It has been pointed out that the exceptions recorded by Cu6not may be due to dronea comiDg 

 from hybrid workers. (Morgan, 19()9a, Am. Nat., XLIII.) 



