Genotype-environment interaction and the maintenance of genetic variation: an empirical study of (Campanulaceae).

High levels of genetic variation are often observed in natural populations, suggesting the action of processes such as frequency-dependent selection, heterozygote advantage and variable selection. However, the maintenance of genetic variation in fitness-related traits remains incompletely explained. The extent of genetic variation in obligately self-fertilizing populations of (Campanulaceae L.) strongly implies balancing selection. thus offers an exceptional opportunity for an empirical test of genotype-environment interaction (G × E) as a variance-maintaining mechanism under fluctuating selection: is monocarpic and reproduces only by seed, facilitating assessment of lifetime fitness; genome-wide homozygosity precludes some mechanisms of balancing selection, and microsatellites are, in effect, genotypic lineage markers. Here, we find support for the temporal G × E hypothesis using a manipulated space-for-time approach across four environments: a field environment, an outdoor experimental plot and two differing growth-chamber environments. High genetic variance was confirmed: 83 field-collected individuals consisted of 45 distinct microsatellite lineages with, on average, 4.5 alleles per locus. Rank-order fitness, measured as lifetime fruit production in 16 replicated multilocus genotypes, changed significantly across environments. Phenotypic differences among microsatellite lineages were detected. Results thus support the G × E hypothesis in principle. However, the evaluation of the effect size of this mechanism and fitness effects of life-history traits will require a long-term study of fluctuating selection on labelled genotypes in the field.