AIMS:

Deleterious mutations are an unavoidable part of life. In sufficiently large populations, all such mutations are removed by natural selection, though always with a concomitant loss of population fitness equal to the mutation rate. A natural question to ask is then: What happens if at another locus a rare allele appears (selectively neutral in itself) that decreases, but does not abolish, the negative phenotypic effect of mutations at the first locus ? or, in other words, makes the organism more canalized against the mutational damage? In the long run this new allele will not affect the mean fitness of the population, since the mutation load will always equal the mutation rate, but will this modifying allele be favoured in some indirect way? The answer is yes, but it is interesting to note how recombination affects this process: With less linkage between the two loci, the easier it becomes for the modifying allele to spread. Thus, recombination promotes mutational canalization in sexual haploids, in a manner that is impossible in asexual haploids. This result is easy to derive but has been surprisingly overlooked, probably because the underlying question was originally discussed in diploids and then phrased in terms of "the evolution of dominance". The secondary selective forces involved are, however, easier to grasp in haploid organisms, where the process instead becomes a question of the "evolution of canalization". That the outlined secondary selective force may be of evolutionary importance is shown by studying a modifying allele that acts on the trait-output of many loci. The force of secondary selection favouring canalization does then depend on the sum of all the mutation rates involved, which gives the process a chance to become evolutionarily effective.