336 



GORDON E. DICKERSON 



dominance deviations (Fig. 21.2). Here both paternal |-sib correlation and 

 regression of progeny on parent would yield zero estimates of heritability, if 

 only dominance were involved. 



Equilibrium gene frequency actually will be determined by degree of 

 dominance expressed in terms of relative selective values or reproductive 

 rates {k') rather than in terms of relative performance levels {k) of the sev- 

 eral genotypes. Conceivably, k' could be either larger or smaller than k. If 

 culling is mild, difference in reproduction rates will be smaller between Aa 

 and A A and larger between A A and aa than if proportional to phenotypic 

 levels, and effective k' will be smaller and equilibirium qA larger. Conversely 

 if phenotypic selection is intense, differences in reproduction rates between 



V^ = 2q (l-q)[(l+k)^ - 2kq (2+k-kq)] d^ 



2 ,2 



V^= 2q(l-q)[uk (l-2q)] ^ d 



AA-aa 



.1 .2 .3 .4 .5 .6 .7 .8 .9 1.0 

 FREQUENCY OF MORE FAVORABLE ALLELE (q^) 



Fig. 21.2 — Total variance in phenotype (V„) and portion linearly associated with genotype 

 (Vq) in a random breeding population for a single chromosomal unit and heterozygote ad- 

 vantage of jfe = 2, at varying frequencies for the more favorable of two alleles. 



