Genetics 53 



case, equal to genetic variance S^. plus environmental variance 

 S%. Thus S-^- 5-;^= S|. This procedure for estimating the variance 

 components rests on the assumption that the environmental variance 

 is independent of genotype, an assumption that is often incorrect. 



Even if the genetic variance can be determined, further com- 

 plexities exist. It cannot be assumed for all characters that the effects 

 of genes are additive in a simple cumulative fashion. The genetic 

 variance itself must be broken down into components. There is the 

 additive component representing the differences between the homo- 

 zygotes and heterozygote(s) for each locus. Also to be taken into 

 account are a component resulting from interactions of alleles, i.e., 

 a dominance component, and a component resulting from inter- 

 actions of nonalleles, an epistatic component. In many situations the 

 additive genetic component is the only one that may be estimated 

 conveniently. Then the phenotypic variance is partitioned into addi- 

 tive genetic and remainder variances. The latter is a catchall term 

 for the nonadditive components of the genetic variance plus tlie 

 environmental variance and gene-environment interactions. The 

 proportion of the phenotypic variance attributable to additive 

 genetic effects is known as the heritability {h^ = S-^/Sj,). Some- 

 times heritability is defined in a narrower sense as Sy'S},, where 

 S^ is the additive genetic variance. Heritability is a good estimator 

 of the degree of resemblance between offspring and parent and as 

 such is of great value to the plant and animal breeder. 



The evolutionist must deal with these complexities since the great 

 majority of traits found to be variable in organisms vary quantita- 

 tively and are under the control of multiple factors. \\'here crosses 

 can be made between races, species, or even genera, the F] offspring 

 generally prove to be more or less intermediate and the Fo show the 

 continuous variation characteristic of polygenic inheritance. (This 

 is an overgeneralization of a complex situation; those wishing further 

 information should consult Falconer, especially the sections on in- 

 breeding depression and heterosis in chap. 14.) 



Control of a characteristic by many genes provides a stability of 

 phenotypic expression that may not occur when only single genes 

 are involved. For instance, a single mutational event is unlikely to 

 disturb seriously the expression of a character dependent upon, say, 

 the additive effects of 35 loci. However, a single mutational event 

 f -> / in the color-inhibitor gene of an onion will result in a white 

 onion rather than a red or yellow one. In view of the possible drastic 

 effects of changes in "major" genes, it is not surprising that most 

 characteristics of organisms are controlled multigenically. Selection 

 would have favored the development of such systems since they tend 



