NATURAL SELECTION: I 459 



eration because they continually arose from heterozygous parents. The 

 latter did not exhibit the ebony trait in their own bodies, although they 

 carried the recessive gene for it. Evidently these heterozygous parents 

 {Ee) were at no disadvantage in the competition, as compared to homo- 

 zygous normal parents (EE). Indeed, there is evidence that heterozygotes 

 had one advantage over homozygous normal individuals. In an independ- 

 ent investigation Timofeeff-Ressovsky found that, while homozygous 

 ebony flies had lower viability than did homozygous normal flies, individ- 

 uals heterozygous for the ebony gene had actually higher viability than 

 did homozygous normal flies. This phenomenon is probably a form of 

 "hybrid vigor" (heterosis), by virtue of which hybrids are frequently 

 larger, stronger, and more vigorous than are purebred strains. The hybrid 

 corn so prevalent on modern farms is a familiar example of hybrid vigor, 

 as is the mule, which possesses some superiorities over either of its parents, 

 the horse and the donkey. 



In sum, we see that in the experiment involving competition between 

 ebony and normal flies both negative and positive natural selection (p. 

 357) were at work. Negative selection tended to eliminate the homozy- 

 gous recessives, i.e., the ebony flies. Positive selection tended to increase 

 the proportion of heterozygous flies. The result of the opposing forces was 

 eventual establishment of an equilibrium at a point at which the strength 

 of the negative selection against recessive homozygotes equaled the 

 strength of the positive selection favoring heterozygotes. In the experi- 

 ment cited, the equilibrium was reached when about 85 percent of the flies 

 had normal color (some being homozygous, some heterozygous), and 15 

 percent had ebony color. In each generation this 1 5 percent of ebony flies 

 arose largely as a "by-product" of the matings of the favored heterozygous 

 flies. 



Experiments corroborating the one just described were performed by 

 Reed and Reed (1948). Instead of concentrating on a single mutant gene, 

 these authors used a strain containing a chromosomal constitution which 

 conferred semisterflity and poor viability, amounting almost to lethality. 

 Obviously, flies homozygous for this M-5 chromosome were at an enor- 

 mous disadvantage. When flies possessing this chromosome were placed in 

 competition with substantially normal flies the proportion of individuals 

 having the chromosome declined with great rapidity. By the end of two 

 months natural selection had completed its main task, and approximate 

 equilibrium had been reached. Although the chromosome in question 

 conferred such great disadvantage upon flies homozygous for it, it was 

 not eliminated from the population. Retention was due to the fact that flies 



