DOMINANCE AND OVERDOMINANCE 285 



Objections to the dominance hypothesis were made largely on two 

 grounds. First, if vigor is not a product of heterozygosity as such, it should 

 be possible by selection to obtain individuals which are homozygous for all 

 the beneficial dominant factors, and hence have the same vigor as hybrids. 

 Secondly, in the Fo of a cross between two inbred strains there should be a 

 skew distribution of the trait being measured — since the dominant and re- 

 cessive loci would be distributed according to the expansion of (3/4 + 1/4)", 

 where n is the number of factors. 



These objections were largely removed when Jones (1917) pointed out 

 that, with linkage, the consequences of the dominance hypothesis were 

 much closer to those postulating superior heterozygotes. If a detrimental 

 recessive were linked with a favorable dominant, the heterozygous chromo- 

 some would be superior to both homozygotes, and the linked combination 

 might not break up readily. Later, Collins (1921) showed that with a large 

 number of factors, regardless of linkage, the skew distribution disappears. 

 The probability of getting all the beneficial dominants into one homozygous 

 strain becomes vanishingly small, so the objections hold only if a small num- 

 ber of factors is assumed. 



Most of the mutations known in Drosophila and elsewhere are recessive, 

 and practically all are in some way deleterious. Even if dominant and re- 

 cessive mutations were occurring with equal frequency, the deleterious mu- 

 tations in a population at any time would be mostly recessive, since the domi- 

 nants would be rapidly eliminated. It is to be expected — and it has been often 

 observed — that at most unfixed loci the recessive is deleterious in compari- 

 son with its dominant allele.^ 



Almost thirty years ago Sewall Wright (1922c) wrote: 



Given the Mendelian mechanism of heredity, and this more or less perfect correlation be- 

 tween recessiveness and detrimental effect, and all the long-known effects of inbreeding — - 

 the frequent appearance of abnormalities, the usual deterioration in size, fertility, and con- 

 stitutional vigor in the early generations, the absence of such decline in any one or all of 

 these respects in particular cases, and the fixation of type and prepotency attained in later 

 generations — are the consequences to be expected. 



It has been shown many times that populations actually contain a large 

 number of detrimental recessives — sufficient to account for a large decline in 

 vigor on inbreeding. In Drosophila pseudoobscttra, Dobzhansky et al. (1942) 

 found that almost every fly examined had at least one concealed lethal. Fur- 

 ther evidence that at least some heterosis is due to dominant favorable genes 

 is provided by the experiments of Richey and Sprague (1931) on convergent 

 improvement in corn. 



2. I consider the statement that a dominant is beneficial and the statement that a reces- 

 sive is deleterious as meaning the same thing. Since a geneticist ordinarily can study gene 

 effects only by substituting one allele for the other, he cannot distinguish what each factor 

 is doing individually or whether it is harmful or beneficial except relative to its allele. That 

 is, he can only tell what the effect of the substitution is. 



