THE FITNESS CONSEQUENCES OF MULTIPLE-LOCUS HETEROZYGOSITY: THE RELATIONSHIP BETWEEN HETEROZYGOSITY AND GROWTH RATE IN PITCH PINE (PINUS RIGIDA MILL.).

Positive correlations between measures of "fitness" and the number of electrophoretic loci for which an individual is heterozygous have been observed in many species. Two major hypotheses have been proposed to explain this phenomenon: inbreeding depression and overdominance. Until recently, there has been no way to distinguish between these hypotheses. The overdominance model devised by Smouse (1986) is used here in a reanalysis of Ledig et al.'s (1983) study of heterozygosity and growth rate in eight populations of pitch pine and is contrasted with an inbreeding-depression analysis. Ledig et al. (1983) regressed mean growth rate per heterozygosity class on the number of heterozygous loci, a method of analysis which, although it points to general trends in the data, does not differentiate between hypotheses. The correlations they obtained in four populations were significant only because regressing on the means eliminates most of the sum of squares for error and does not weight the unequally sized heterozygosity classes. Reanalysis of Ledig et al.'s data using individuals, not means, showed no significant correlations between heterozygosity and fitness. A major assumption of Smouse's overdominance model is that genetic polymorphism is in part a reflection of selection for heterozygotes at genetic equlibrium. The homozygote for the most frequent allele at a locus should be more fit than a homozygote for a less frequent allele, with the heterozygote superior to both homozygotes. Smouse's model predicts a negative, linear relationship between fitness and "adaptive distance," a variable that for a heterozygote is zero and for homozygotes is equal to the inverse of the frequency of the corresponding allele. The adaptive-distance model accounted for between 15% and 50% of the variation in growth rate within eight P. rigida population samples by accounting for genotypic differences at eight polymorphic loci. This is over twice as much of the variation in growth rate accounted for by Ledig et al.'s (1983) analysis using individuals. Significant correlations were found between adaptive distance and growth rate in four of the eight populations, but in only two of the populations were more of the partial coefficients negative than positive, as would be predicted by the overdominance hypothesis. The remaining two populations in which correlations were significant did not lend themselves to such clear-cut interpretation, as the majority of the partial coefficients were positive. Positive partial coefficients indicate that the growth rate of the heterozygote is inferior to that of at least one of the homozygotes. The adaptive-distance analysis provides evidence that specific genotypes do play a role in determining growth rate in pitch pine. The correlation between growth rate and adaptive distance increased significantly with the age of the population, possibly reflecting competition subsequent to crown closure. © 1987 The Society for the Study of Evolution.