Bees in trees: How do forest bees use floral resources?

Background/Question/Methods

Bee species that require forest habitat were likely dominant in the northeastern United States until the widespread conversion of forest to agriculture during the 19^th and early 20^th centuries. Despite this, efforts to restore pollinators in the northeast have largely neglected forest-associated species, instead focusing on restoring meadow-type habitats. Floral resource use by forest bees is not well understood both because sampling animal-pollinated trees is difficult and because forest bees emerge in the spring, a time of rapid floral turnover. To overcome these sampling limitations, we examined the record of floral visits contained in the pollen female bees carry to their nest. We answered the following questions: 1) What proportion of forest bee diet comes from each of the main vegetation strata, herbs, shrubs and trees? 2) How does the diversity of pollen used by forest bee species, and also by individuals within species, change over time? From March to May 2016 we collected bees from five forest sites, removed pollen from female specimens of forest-associated species, and identified and quantified the pollen using microscopy. We used linear models to determine how aggregate and individual floral resource diversity changed over time.

Results/Conclusions

We analyzed 413 scopal pollen loads from females of 54 species of forest-associated bees. 36% of pollen was from trees, 6% was from shrubs, 56% was from herbaceous plants, and 3% was unknown. This balance remained the same even when our analysis did not count the pollen from the plant on which each bee was collected, which could be biased towards understory plants which are easier to collect from (33 % trees / 6% shrubs / 55% herbaceous / 2% unknown). Shannon diversity of pollen used by all forest bees as a group increased strongly over the course of the season (slope = 1.14, p = 0.002, R² = 0.85), and Shannon diversity in pollen loads from individual bee foraging bouts also increased, though with a weak effect size (slope = 0.14, p < 0.001, R² = 0.09). Our results suggest that trees are important floral resources for forest-associated species and that a few key early blooming plant species are critical to supporting forest bee populations. Further, the fact that individual pollen-load diversity increases over the course of the season suggests a possible advantage for early blooming plants as they may receive more conspecific pollen from pollinators.