Herbivores are expected to select for defensive traits in their hosts, but hosts may not be able to respond in an unconstrained manner. Investment in defense may come at a fitness cost, and this trade-off might lead to stabilizing rather than directional selection. There also may be trade-offs between different types of defenses. There may even be constraints imposed by selection on other genetically correlated traits that may be seemingly unrelated to defense. Comparisons across populations in the Mimulus guttatus species complex indicate the evolution of selfing is associated with a reduced investment in trichome production. We have hypothesized that reduced investment in herbivore defenses may be adaptive in short-lived selfing populations if individuals are able to reallocate resources from defense to fitness components (rapid development time and greater flower and biomass production). To evaluate this hypothesis we tested for fitness trade-offs and constraints in a M. guttatus population from Napa County, CA. We used three full-sib families from each of 35 sires to estimate genetic variation and covariation for defense, fitness, and floral traits.
Results/Conclusions

Upper and lower trichome densities and all floral traits showed significant heritabilities. Performance traits (time to flowering, flower production, and aboveground biomass) also showed significant heritabilities. Although genetic correlations among floral traits were very strong, they typically did not show significant genetic correlations with defensive traits. The exceptions to this pattern were highly significant negative genetic correlations between trichome density (both upper and lower) and stigma-anther separation, a trait strongly associated with mating-system evolution in this genus. Interestingly, this trend is opposite the direction of the phenotypic correlations among populations within the species complex. Our study suggested a potential cost to trichome production in M. guttatus. Genetic correlations between upper and lower trichome density and date of first flower were strongly positive, suggesting that high trichome densities were associated with delayed flowering. Lower trichome density showed significant negative genetic correlations with above-ground biomass and total flower production, suggesting a cost to trichome production in size and fecundity. These findings are consistent with our hypothesis that the less defended selfing populations in the M. guttatus complex benefit from a reduced cost of defense and that these savings could be reallocated into growth and reproduction. Taken together, these results are consistent with the hypothesis that selection, rather than genetic constraints, is responsible for the differences in defensive strategies among Mimulus populations with divergent mating-systems.