NATURAL SELECTION: I 467 



present in all stocks of chickens. The experimenters attempted by selection 

 and inbreeding to produce a strain of birds free from the gene, but met with 

 only partial success since "the fitness of the . . . line has continually 

 dropped and only a few survivors are available each year for reproduction." 

 On the other hand, selection and inbreeding were successful in establishing 

 an almost pure-breeding Hne having the crooked-toes character; in this line 

 the reproductive capacity was not diminished as it was in the other line. 



The crooked-toes example is illustrative of the evidence accumulating 

 that successful, viable populations, wild or domestic, normally consist 

 of individuals having a high degree of heterozygosity. Experimental 

 evidence even suggests that viability can be increased by increasing the 

 degree of heterozygosity artificially, by inducing new mutations with 

 irradiation (see Dobzhansky, 1959). (Such results may have fascinating 

 implications for the question of the possible genetic effect of irradiation of 

 mankind, from natural and artificial sources, e.g., atomic bomb fallout. 

 But such discussion is outside our present field of consideration.) 



We may note that the view of population structure just described con- 

 trasts with the view formerly held that most wild populations consist of 

 individuals homozygous for most of their genes, which are regarded as the 

 "normal" oenes. If this were true most mutations would be harmful to 

 homozygotes and, if they produced a phenotypic eft'ect in heterozygotes, 

 to heterozygotes also. (We have noted previously that many if not most 

 mutations do have some effect upon heterozygotes, p. 347.) This "classical 

 hypothesis" made possible the separation of "good" genes from "bad" 

 genes. According to the "balance hypothesis" (the terms are Dobzhansky's, 

 1959), on the other hand, "TOod" genes are those which contribute to the 

 fitness of individuals heterozygous for them, almost regardless of their 

 effect in individuals homozygous for them. 



We say "almost" in this last sentence because we must not forget that 

 there are genes which are harmful to both homozygotes and heterozygotes. 

 Thus the majority of lethal (pp. 389-390) mutants in Drosophila are harm- 

 ful in heterozygotes, even though many mutations that stop short of 

 lethality, when individuals are homozygous for them, form adaptively 

 superior heterozygotes. And in cases not involving lethality heterozygotes 

 for a given pair of genes may be inferior to one or both of the homozygotes 

 {Aa inferior to A A and/or aa). What determines whether Aa will be 

 inferior or superior? Many factors are involved, some genetic, some 

 environmental. A gene pair does not operate in a vacuum; it is part of 

 the complete genotype of the individual. Its effects are conditioned by the 

 other genes present. This integrated genotype is the totality that produces 



