Experimental comparison of the genetic component of pollinator effectiveness in a shrub pollinated by birds, non-flying mammals and European honeybees.

The range and frequency of pollinator movement, influences patterns of pollen dispersal with consequences for plant mating. It is predicted that pollination by birds promotes outcrossing and multiple paternity. Alternatively, pollination by insects or non-flying mammals (NFMs) is expected to increase inbreeding and correlation of paternity, leading to less genetically diverse and less vigorous seedlings. We tested these predictions in Banksia catoglypta (Proteaceae)-an obligately outcrossing heathland shrub pollinated by honeyeater birds, NFMs (honey possums and rodents) and insects (mostly introduced Apis mellifera). Previous research employing selective pollinator exclusion determined that pollination by NFMs was associated with reduced fruit set, compared to flying pollinators. Here, we used microsatellite genotyping of seedling families and a common garden experiment to compare the effectiveness of bird, NFM, and insect pollinators in terms of the genetic quality and vigour of the resulting seedlings. Despite floral traits of B. catoglypta suggesting adaptation to mammalian pollinators, pollination solely by NFMs led to lower multilocus outcrossing rates (tm, indicating biparental inbreeding) and higher correlation of paternity (rp), compared to pollination by flying pollinators (birds and insects combined). Otherwise, there were no significant differences between pollination treatments and open-pollinated controls. Overall, seedling survival was positively correlated with individual heterozygosity, with seedlings resulting from pollination solely by NFMs least likely to survive, and exclusion of NFMs associated with increased seedling survival-although these effects were slight. Our results highlight the importance of quantifying plant mating and seedling fitness for understanding the consequences of pollination by different animals.