Abstract
Two large experiments designed to detect allozyme-associated heterosis for growth rate in Drosophila melanogaster were performed. Heterosis associated with allozyme genotypes may be explained either by functional overdominance at the allozyme loci, or closely linked loci; or by genotypic correlations between allozyme loci and loci at which deleterious recessive alleles segregate. Such genotypic correlations would be favored by consanguineous mating, small effective population size, population mixing and strong natural or artificial selection. D. melanogaster is outbred, has large effective population size and there is little evidence for genotypic disequilibria. Therefore it would be unlikely to show allozyme heterosis due to genotypic correlations. In the first experiment I estimated the genotypic values of 97 replicated genotypes. In the second experiment, 500 individuals were raised in a fluctuating, stressful environment. In neither experiment was there any consistent evidence for allozyme heterosis in size or development rate, fluctuating asymmetry for size or in tendency to deviate from the population mean. In the first experiment, heterosis explained less than 5.6% of the genetic variance in growth characters. In the second, heterosis explained less than 0.1% of the phenotypic variance in growth characters. Outside of the molluscs, species which show allozyme heterosis have population structures or histories which tend to promote genotypic correlations. There is little evidence that functional overdominance is responsible for observations of allozyme-associated heterosis.