Floral and mating system divergence in secondary sympatry: A geographic and phylogeographic analysis of Clarkia xantiana ssp. parviflora

Background/Question/Methods: Mating system shifts in plants can lead to reproductive isolation and eventual speciation between diverging taxa.  Reproductive isolation may be due to evolutionary responses of taxa to the pollination environment and consequently different strategies that produce reproductive isolation as a by-product of adaptation or direct selection in secondary contact due to low hybrid fitness.  We tested the prediction that mating system evolves in response to the pollination environment using the native wildflower Clarkia xantiana ssp. parviflora. Previous work in C. xantiana has shown that two subspecies, xantiana and parviflora diverged in allopatry ~10,000-65,000 years ago and have migrated to their current ranges, which has resulted in a secondary zone of sympatry. Phylogeographic analyses of C. x. parviflora populations have also defined population clusters supported by genetic data. Within C. x. parviflora, we tested for divergence between sympatric and allopatric populations and between phylogeographic clusters in floral traits using greenhouse-obtained floral measurements from populations throughout the range of C. x. parviflora.  To test for geographic and phylogeographic mating system divergence, we performed pollen supplementation and emasculation treatments in natural and reciprocal translocation populations across the geographic range of C. x. parviflora to attain estimates of pollen limitation and reproductive assurance. 

Results/Conclusions: There were significant differences in floral characteristics between geographic and phylogeographic clusters.  Allopatric populations and allopatric phylogeographic clusters had significantly larger flowers.  There was also significant mating system divergence between geographic regions and phylogeographic clusters.  In experiments using natural populations, sympatric populations and more sympatric phylogeographic clusters experienced greater pollen limitation (PL), measured as the difference between supplemented fruits and either unmanipulated fruits (overall PL) or emasculated fruits (pollinator-mediated PL).  Sympatric populations and clusters also exhibited greater reproductive assurance (RA), measured as the difference between unmanipulated and emasculated fruits, indicating autonomous selfing further elevates fecundity relative to pollinator-mediated reproduction in these populations.  In reciprocal translocations, there was no significant difference between allopatric and sympatric destination sites in PL or RA. However, across all translocation sites genotypes from sympatric populations had significantly greater RA and pollinator-mediated PL then genotypes from allopatric populations.  This suggests that mating system divergence across the range of C. x. parviflora is the product of divergence in floral size and pollination environment is not the sole selective force driving floral and mating system divergence.