Print PageSelf-fertilization in the absence of pollinators (autonomous autogamy, “AA”) should be favored in unpredictable pollination environments, especially when selfing occurs after opportunities for outcrossed pollen receipt have passed (‘delayed’). In this case, individuals can garner reproductive assurance without sacrificing opportunities for outcrossing. AA via delayed selfing has been shown previously to provide reproductive assurance in Collinsia verna (Plantaginaceae) when pollinators are scarce. However, we have little empirical evidence on how AA and thus mixed mating evolve in general. We clonally replicated C. verna genotypes from three populations (BT, EF, TMC) where field selfing and pollinator failure rates are known. We estimated AA rates (proportion fruit set) of clones in a controlled, pollinator-free environment and measured a suite of floral traits related to morphology and timing of male and female phases across floral lifespan. Specifically, we (1) evaluated whether observed pollinator conditions may be driving divergence in AA and developmental floral traits associated with high AA rates across the three C. verna populations and (2) estimated the heritability of AA rate and the measured floral traits. In addition, we estimated family-level (genotypic) inbreeding depression to assess whether offspring of genotypes with higher AA rates have lower inbreeding depression, as predicted by general theory on the evolution of selfing.
Results/Conclusions
We found that AA rates vary significantly among C. verna genotypes (range 0.0-0.78) and populations (range 0.20 to 0.35) and show significant broad sense heritability (mean H2 across populations 0.36). Bivariate analyses reveal that genotypes with smaller flowers had significantly higher rates of AA in the BT population only. Mean inbreeding depression per population is low (< 0.5) for all traits examined, but we detected significant variation among families. In contrast to our prediction, higher AA rates are not associated with lower levels of inbreeding depression for fitness traits. Together, these results demonstrate that C. verna meets several necessary conditions for the evolution of AA and mixed mating (i.e. genetic variation, heritability). We also saw evidence of population differentiation in AA rates. Greater selfing rates and pollinator failure in population BT in the field is consistent with differential selection for AA among populations. Our results further suggest that ecological context and the timing of selfing may be important for the purging of deleterious alleles and the maintenance of mixed mating.
