Themedionigragene in the mothPanaxia dominulcr: the case for selection

Abstract

Analyses of changes in frequency of the genemedionigrain colonies of the mothPanaxia dominula, begun by R. A. Fisher, E. B. Ford and P. M. Sheppard, have long been regarded as a model study of natural selection under field conditions. Recently, their conclusions have been criticized, on the grounds that phenotypes have been improperly scored and that population structure has been misunderstood. The results are re-examined here, including recent unpublished collections. It is argued that the colonies studied are distinct populations, as usually defined, and that the results could not arise as a result of migration. Fluctuation in population size from year to year, large variance in fecundity and some features of mating behaviour probably reduce the effective number to less than half the estimated population size. Variable expressivity and consequent subjective variation in scoring, casts some doubt on earlier claims that selection fluctuates significantly from generation to generation, and on a reported case of increase in frequency from a very low starting point. However, the one natural and three artificial colonies studied are consistent in providing estimates of selection of 7 per cent or more against medionigra, despite the variation introduced by small population size and scoring difficulties. There appears to be an equilibrium at a low frequency, and the rate of approach to it suggests that the selection is frequency dependent. One artificial colony, at West Kirby on the Wirral, Merseyside, U. K. , does not show an equivalent decline. Selection cannot be as strong as indicated elsewhere, but since the population size is small some disadvantage to medionigra cannot be rejected.