Published online by Cambridge University Press: 14 April 2009
The experiments reported in this paper are primarily addressed to test the hypothesis of overcompensation; i.e. whether polymorphic populations exploit limiting environmental resources better than populations uniform for a single genotype. Overcomposition is an ecological consequence of some models of frequency-dependent selection. Secondarily, the experiments investigate whether overdominance exists at the Mdh-2 locus in Drosophila pseudoobscura.
Two types of experimental populations are established: ‘low-variability’ populations, in which all flies in a culture are offspring from only two laboratory strains; ‘high-variability’ populations, in which the flies in a culture are derived from 20 different strains. However, the overall degree of individual heterozygosity is the same in both types of populations. Three kinds of populations with respect to the Mdh-2 locus are established within each type; two are homozygous for either the 100 or the 112 allele, the third is heterozygous. A fourth kind of population exists among the high-variability populations; namely, populations in which all three Mdh-2 genotypes are present. The experiments are done at two densities; one quasi-optimal, the other highly competitive.
Populations with high overall levels of genetic variation consistently produce more flies than low-variability populations. The differences are significant at the low, but not at the high, density. Moreover, populations polymorphic for the Mdh-2 locus generally produce more flies than populations having only one Mdh-2 genotype. At high density, the Mdh-2 polymorphic populations have greater productivity than populations with anyone of the three genotypes, and the differences are statistically significant when the polymorphic populations are compared with either one of the two homozygotes or with the average of all three genotypes. In brief, overcompensatory effects – which may account for frequency-dependent selection – are observed in the experiment and may be a common phenomenon in nature.
Populations in which all individuals are heterozygous at the Mdh-2 locus produce in every case more flies than populations with only homozygotes for one or the other allele. The superiority of the heterozygotes is statistically significant for all comparisons at low density, but at high density it is statistically significant for the comparison with the homozygote rarer in nature and only in low variability populations.