A novel approach to estimating the cost of male fertility restoration in gynodioecious plants

In many gynodioecious plants, sex is determined by cytoplasmic male sterility genes (CMS) and nuclear male fertility restorers (Rf). Models predict that the costs of restoration are important determinants of population sex ratios. However, current approaches to the estimation of these costs require prior identification of CMS genotypes, information that is available for few species. We tested a novel approach to estimating the cost of restoration in natural populations without determining CMS or Rf genotypes. We used estimates of pollen viability and offspring sex ratios from open- and hand-pollinated families of Lobelia siphilitica to test whether the cost of restoration, expressed as low pollen viability, is higher in populations with more females. Among populations with CMS, we found that variation in pollen viability was higher in small populations with more females, as expected if the proportion of females within populations increases with the maximum cost of restoration. In controlled crosses, families with low pollen viability also produced fewer females, suggesting that variation in viability is primarily determined by the number and frequency of Rf alleles carried. This approach to estimating the cost of restoration can be applied to other cytonuclear gynodioecious species, offering new opportunities for testing gynodioecy models in the wild.