84 A. FRANKLIN SHULL 



That inbreeding coupled with selection had produced practi- 

 cally homozygous eggs is further indicated by the negative results 

 of selection in L4 and Hz. Selection for high and low viability 

 within each of these lots of eggs was without effect, as would 

 be expected if the parthenogenetic lines producing the eggs were 

 homozygous with respect to viability. 



The greater viability in eggs produced by crossing two lines, 

 as compared with the inbred eggs produced by the same lines, 

 may be the result of increased vigor due to heterozygosis. I 

 have already shown (Shull '12 b) that inbreeding results in de- 

 crease of vigor in Hydatina, and Whitney ('12), following a single 

 experiment of mine, has proven that crossing increases vigor. 

 It seems probable that the fact that the reciprocal crosses between 

 C and D and between Lo and H-2 show an average viability 

 plainly in excess of the average viability of the inbred eggs of 

 the same lines, is merely another example of the increase of 

 vigor due to heterozygosis. Wliich of the several possible ele- 

 ments contributing to viability is affected by heterozj^gosis can 

 not be determined. 



The inequality of the reciprocal crosses with respect to via- 

 bility of eggs is discussed in a separate section. 



Variability in the duration of the fertilized egg stage is not so 

 certainly explainable because the experiments in which marked 

 differences in this variabilit}" appeared were few. All of the 

 results of the experiments here described accord fairly well with 

 the assumption that this variability is dependent on a single 

 pair of genes, and that low variability of hatching time is domi- 

 nant over high variability. Because of the small number of 

 tests obtained, this interpretation may well be incorrect. 



The inequality of the reciprocal crosses C X D and D X C, 

 together with the equality of the inbred and of the crossed eggs 

 obtained from them, is discussed in another section, in connec- 

 tion with the inequality of reciprocal crosses with respect to 

 viability of eggs. 



The sex-ratio, or its equivalent, the percentage of male-produc- 

 ing females, behaves as if dependent on a number of genes; but 

 it is difficult to formulate a general statement. In the experi- 



