SEX DETERMINATION 



evidently the result of error in sampling. The 

 rare exceptions, v/hich were not adequately 

 tested, rr,ay have been the result of contamina- 

 tion, though this is unlikely, or to some com- 

 bination of factors reducing diploid male via- 

 bility, or to mutation in a sex allele itself. 



The nature and evolution of the sex factor, 

 X, the differential segment, may no\7 be con- 

 sidered. Mutant genes are known to cross over 

 both to the left and to the right of x, hence x 

 is interstitial, but so far no crossing over 

 within X has been noted. If, however, crossing 

 over within x did occur, x would be inconstant, 

 xa would not remain distinct from xb, or xb from 

 xc, etc. The work of Horn ( 1943a, b) demon- 

 strates that F2 males segregating from crosses 

 of xa/xb with xe/xf stocks are still xa, xb, 

 xe, or xf, that they all sire diploid sons, and 

 that those that were xa or xb sire diploid sons 

 by xa/xb females and not by xe/xf females and 

 the reverse (as shown by linkage with fused). 

 It may be concluded, therefore, that there is 

 no crossing over within x; or perhaps the rare 

 "mutations" of x, if such actually occur, may 

 be crossovers. 



Whether or not Whiting's multiple allele 

 theory of sex determination is applicable to 

 the Hymenoptera in general is as yet a question. 

 Were it not for parthenogenesis, this comple- 

 mentary scheme might never have been proposed. 

 Diploid males have been shown in two other spe- 

 cies of Habrobracon (Speicher and Speicher, 

 1940; Inaba, 1939), but so far no one has ap- 

 plied breeding and observational tests that 



101 



