Variation in pollen limitation, reproductive assurance, and floral traits across small and large populations of a native biennial in a fragmented serpentine grassland system

Background/Question/Methods

Worldwide, species are experiencing rapid changes in their environment as a result of human activities that not only affect abiotic conditions but also ecological interactions that can sustain populations. In particular, plant-pollinator interactions are being disrupted in human-disturbed habitats. Often such effects are mediated by changes in population size and surrounding land use. In a serpentine grassland system in the eastern US affected by habitat loss and fragmentation, I asked how population size influenced pollen limitation (PL), the potential fitness benefit of autonomous self-pollination (reproductive assurance, RA), and floral traits known to influence plant-pollinator interactions across populations of the native biennial Sabatia angularis. In each of 9 S. angularis populations varying in size, we performed pollinator observations and the following experimental manipulations to quantify PL and RA: flowers on each of up to 50 plants/population were either left intact, hand-pollinated with supplemental outcross pollen, or emasculated. We estimated seed set of each and pollen receipt for stigmas of intact and emasculated flowers. We also measured flower size and daily floral display. Finally, we are quantifying the surrounding landscape context of each population using GIS to further evaluate the effect of land use on PL, RA, and floral trait variation. 

Results/Conclusions

All populations suffered from PL for seed set, regardless of their population size. Large populations tended to have lower pollinator visitation rates, on average, potentially leading to strong intraspecific competition for pollinators in those populations, consistent with patterns seen in an earlier study using experimental arrays. Large populations also had significantly larger floral displays, even when accounting for total flower number, than small populations on average.  Despite significant PL, we did not find evidence for RA based on either pollen receipt or seed set data. This suggests that higher self-fertilization seen previously in small populations is most likely pollinator-mediated. Together these results suggest there may have been greater selection in large populations to attract pollinators and/or selection for smaller displays in small populations to avoid geitonogamous pollination, given extreme inbreeding depression in such populations. These hypotheses will be evaluated using phenotypic selection analyses based on additional data collected in these populations. Together, this work can provide insight into the ecological and evolutionary consequences of changes in population size and landscape context and thus how and whether species can adapt to changing landscapes.