386 Inbreeding, Selection, and Heterosis 



homozygous lines containing all the positive or all the negative 

 size genes will segregate out, only extremely slowly. Even with 

 linked genes, we could finally arrive at homozygous strains of 

 maximum qualities as a result of crossovers. The process would 

 be merely slowed by linkage. 



This theory has been accepted by a number of geneticists but 

 has not yet been proved to the exclusion of other possibilities, 

 and there are some definite objections especially by some that 

 approach the problem from the physiological side. Jones's hy- 

 pothesis does not necessarily demand that the dominant allele 

 of each pair be dominant within the usual meaning of the term. 

 It is adequate merely to assume that the ''dominant" gene in a 

 heterozygote has more than half the effect that it would have 

 in a homozygote. East has modified this theory further by 

 assuming that at most loci there are at least three alleles. One 

 is deleterious and the other two are contributing, but not to the 

 same extent. One parent might be a'^a^ and the other a^a^. There 

 is no dominance of either of these alleles over the other, and- 

 both act in the hybrid, which is a^a^. The hybrid has the size 

 increase of each and therefore is larger than either parent. The 

 really deleterious recessives, however, East considers of no im- 

 portance in heterosis. 



Some interesting ideas on heterosis have come out of the re- 

 cent studies by Dobzhansky and others on natural populations. 

 They have pointed out that most mutations are harmful, vary- 

 ing from lethals at the one extreme to only mildly deleterious 

 genes at the other. Harmful dominant mutations are readily 

 eliminated by natural selection, but recessives accumulate in a 

 population as they usually exist in a heterozygous condition. If 

 the species regularly reproduces by self-fertilization, as species 

 of wheat, the recessives quickly become homozygous and are 

 eliminated unless they are only slightly harmful. The same 

 would be true if the effective population size is very low — that is, 

 if the population is so small that a union of gametes with the 

 same harmful genes occurs frequently. If a species has an inter- 

 mediate or moderately large effective population size and nor- 

 mally is cross-fertilized, as Drosophila pseudoobscura, probably 

 most domesticated animals, and possibly man, deleterious re- 

 cessives accumulate in natural populations. If inbred strains 



