EFFECT OF HETEROZYGOSIS ON D^OSOPHILA 465 



In four of the six latter cases the number of outcrossed flies 

 (2, 4, 1, and 3) was too small to be of significance. In the other 

 two cases the number was somewhat larger than in the above, 

 although still rather small, but here there is very little or no 

 difference between outcrossed and stock ( — 0.2 and 0). 



By comparison of the lots that were started on different days 

 in either stock or outcrossed, it is evident that there is variation 

 in the character (presumably due to environment) quite apart 

 from any difference between the stock and the outcrossed. The 

 differences between the latter two types, however, are too con- 

 sistent to be due to chance operations of the same cause. 



Neither is the difference due to contamination from the nor- 

 mal allelomorph, otherwise the outcrossed would be nearer the 

 normal than the stock is. We find, on the contrary, that the out- 

 crossed type is really better balloon than the stock. Moreover, 

 the standard deviation of the s^ock is 1.55 grades, and of the 

 outcrossed is 1.37. Hence there has been no increase in the 

 variability in the outcrossed. The greater vairiability of the 

 stock constitutes additional evidence that there has been no 

 contamination or fractionation. 



The difference, then, must be due either to variation in the 

 balloon factor itself (multiple allelomorphs) or to a difference 

 in modifying factors (multiple factors). Assuming that varia- 

 tion of one type or the other could occur at all, natural selection 

 would tend to cause the races finally to differ in the direction 

 found. For the wings that made for increased balloon, being 

 less adequate for efficient flight, would tend to be eliminated. 

 Also the watery blister, which usually bursts, often causes the 

 fly to adhere to the side of the bottle and perish. The inbred 

 stock has had an excellent opportunity to become changed in 

 the direction of normal in the period of three years, but the out- 

 crossed did not have a chance for such selection, since balloon 

 is recessive and was kept hidden while in the state of hetero- 

 zygosis. 



Tests could be made to determine definitely whether the change 

 was due to multiple allelomorphs by the use of linkage experi- 

 ments, since the knowledge which has already been obtained of 



THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 22, NO. 2 



