Abstract.-The number of ovarioles of the Drosophila melanogaster ovary is a trait thought to be associated with female fecundity, and therefore is expected to be under strong natural selection. This hypothesis may be tested by examining patterns of genetic and environmental variation for ovariole number in natural populations, and by determining the association between ovariole number and fitness in isogenic lines derived from a natural population. We measured ovariole number, and competitive fitness and its components, for 48 homozygous chromosome 3 substitution lines in a standard inbred background; and body size in a sample of 15 chromosome 3 substitution lines. We found significant segregating genetic variation for ovariole number, with a broad-sense heritability (EP) of 0.403 and correspondingly high coefficients of genetic variation (CV a = 20.8) and residual variation (CV R = 25.3). Estimates of quantitativegenetic parameters for body size (EP = 0.191, CV a = 2.15, and CV R = 3.87) are similar to those previously reported for this trait. Although the isogenic chromosome 3 substitution lines varied significantly for components of fitness, there was no significant linear or quadratic association of ovariole number and body size with fitness. There was, however, highly significant sex X genotypeinteractionfor fitness amongthese lines. This specialcase of genotype X environment interaction for fitness may contribute to the maintenance of genetic variation for fitness in natural populations.Key words.e-Drosophila melanogaster, ovariole, variation.Received April 14, 1996. Accepted March 13, 1997.Understanding the nature of the evolutionary processes leading to the maintenance of quantitative genetic variation within populations and the divergence of trait means between populations and species remains one of the leading unanswered questions in evolutionary biology today. Several mechanisms could, in theory, maintain variation for quantitative traits: mutation-drift balance of selectively neutral alleles (Lynch and Hill 1986), overdominance (Robertson 1956), mutation-selection balance (Barton and Turelli 1989), antagonistic pleiotropy (Rose and Charlesworth 1981), and environmental heterogeneity (Gillespie and Turelli 1989). Different categories of quantitative traits (e.g., morphological and life history) are characterized by different magnitudes of, and proportional contributions of, additive genetic variation and environmental variation (Houle 1992;Falconer and Mackay 1996). This implies that the evolutionary mechanisms maintaining variation and the nature of genetic variation are likely to be different for morphological and lifehistory traits. However, progress in evaluating which mechanisms are responsible for maintaining genetic variation for any quantitative trait has been hindered by tests that rely on relative magnitudes of segregating genetic and mutational variance, genetic correlations between traits, inbreeding depression, the genomic mutation rate and average selection against new mutations (Crow an...
