Forest-associated native bees are important pollinators of spring-blooming plants. They are particularly vulnerable because land-use change is driving the loss and reconfiguration of forest habitat. The loss of forest-associated bees could have cascading effects on biodiversity through loss of pollination function to spring wildflowers and flowering trees. Polemonium reptans, a spring-blooming wildflower, is critically endangered in New Jersey; it likely suffers from habitat loss, but the specific drivers of its near-extirpation are unknown.
Here, we investigated (1) what proportion of bee visitors to P. reptans are forest-associated, (2) to what extent forest-associated bees contribute to P. reptans pollination and (3) how the richness and abundance of forest-associated bees vary with forest area and forest edge length, which were uncorrelated metrics in our data. We established 25 experimental arrays of P. reptans across a 5x5 km grid in central New Jersey, situating each adjacent to forested land. We measured forest area and the length of forest edges at a 200m radius surrounding each array using a GIS, and bee abundance and richness by capturing and identifying visitors to P. reptans to species level. We also estimated per-capita pollination efficiency of each type of bee by conducting single-visit experiments with virgin flowers.
Results/Conclusions
We collected 1,726 bee specimens representing 80 species foraging on Polemonium reptans. Of these, 19.1% of individuals (331) representing 14 species were previously known to be forest-associated. Two species, Osmia pumila and Augochlora pura, accounted for most (35.0% and 28.1%, respectively) of the forest-associated bee visitors to P. reptans. Forest-associated bees deposited an estimated 27% of the total pollen grains on P. reptans. We detected a significant positive relationship between the abundance of forest-associated bees and both forest area and forest edge length (R2=0.387, p=0.003, and p=0.002, respectively). We also detected a significant positive relationship between richness and forest area, but not forest edge length (R2=0.305, p=0.02 and p=0.771, respectively). Our results suggest first that the forest-associated bee species remaining in these fragmented landscapes are sensitive to forest loss and may benefit from the increasing forest edge density found in fragmented landscapes and second, that other bee species that are not known to be forest-associated, play an important role in pollinating P. reptans. These findings indicate that the near-extirpation of P. reptans from New Jersey may be due to factors other than the loss of its original pollinators.