GENIC INTERACTION 



183 



extremes is likely to lead to multiple peak genotypes in the neighborhood of the 

 threshold or ceiling on the underlying scale, on the hypothesis that the pleiotropic 

 effects of the favored genes are largely deleterious. 1427 



Fig. 30. Mean selective values of 14 homallelic populations. 

 1 



1.25 



1.00 abCD 



A\ 



0.75 



Based on the selective values of figure 29. 

 and abed (W = 0) are omitted. 



Population homallelic in ABCD ( W = 0.25) 



The effects on the underlying scale need not be uniform. Figure 31 illustrates a 

 case in which a mutation (M') has an effect that offers the promise of a major step in 

 advance but has other effects that make it highly deleterious. By introducing modifiers 

 A', B', C", D', E', F' that neutralize these other effects of M' and have only very slightly 

 deleterious effects otherwise, the direction of effect of M' may be reversed in 

 a combination that is far superior to the initial one. 



THE IMPORTANCE OF INTERACTION IN EVOLUTION 



There may be some genes that have such unequivocally injurious effects that these 

 cannot be reversed by any combination with other genes. In general, however, it 

 seems safe to conclude of any pair of alleles that one is more favorable in some combina- 

 tions, the other in others. The primary condition for an effective evolutionary process 

 would seem to be that selection operate to increase the frequencies of favorable multi- 

 factorial genotypes (or even systems of genotypes) rather than of single genes because 

 of their favorable net effect in the population in question. 



The least effective evolutionary process is the one that is still, perhaps, most 

 widely accepted : evolution as a succession of rare favorable mutations, each gradually 

 displacing its type allele in what is at each stage essentially a single type genotype. 



