MEANS AND METHODS OF EVOLUTIONARY CHANGE 347 



ual will, of course, exhibit the new characteristic. We see, then, that muta- 

 tions may remain hidden for many generations following the time of actual 

 change in the gene concerned. 



Evidence accumulates, however, that many, if not most, mutations are 

 not completely recessive but rather exhibit some degree of dominance (see 

 Muller, 1950). That is, they produce some effect in heterozygous individ- 

 uals. In terms of our example: Aa individuals are not exactly like A A 

 individuals in characteristics. The o gene in the heterozygote modifies the 

 action of the A gene. When this is the case a new mutation will make 

 its presence felt at once, even before any aa individuals appear. If the ef- 

 fect of the a gene, in combination with the A gene in a heterozygote, is 

 beneficial, natural selection will favor the heterozygotes; if the effect of the 

 a gene is harmful, natural selection will tend to eliminate the heterozy- 

 gotes. In the former instance the frequency of the a gene in subsequent 

 generations will tend to increase; in the latter instance the frequency will 

 tend to decrease. 



As we shall discuss more fully later (pp. 457-469), evidence is being ob- 

 tained that in a state of nature organisms are heterozygous for many pairs 

 of genes, and that in many cases the phenotype (p. 333) of heterozygotes 

 is superior to the phenotype of homozygotes (e.g., Aa is superior to AA 

 and aa). When this is so, what gene a produces when combined with A 

 is more important than what it produces when combined with another gene 

 of its own kind (a). Thus completely recessive genes, "covered up" in 

 heterozygotes as mentioned above, may turn out to be the exception rather 

 than the rule. 



Natural selection can act on new mutations which have phenotypic ef- 

 fect in heterozygotes. This includes not only the partially recessive genes 

 just mentioned but also changes of normally present recessive genes to 

 completely dominant ones (e.g., change of gene a to gene A). In our pres- 

 ent state of knowledge such mutations seem to be rarer than do mutations 

 of dominant genes to genes having some degree of recessiveness or, to ex- 

 press it differently, somewhat lessened dominance, as described above. 

 The Ancon ram (p. 337) may have constituted an example of a completely 

 dominant mutation, since all offspring seem to have exhibited the short- 

 ened less. 



Returning to our recessive, or largely recessive, mutation a, we may 

 now inquire: After the gene a has become widely enough distributed in 

 the population so that aa individuals occasionally arise, what will happen 

 to these individuals and to the "new gene" itself? 



