UC Berkeley

Global declines of pollinators have highlighted the lack of information on pollinator species and communities necessary to improve conservation efforts. Specifically, significant information is missing on resource use of native pollinators and factors at both the species and community level that influence resource selection. My dissertation work uses a variety of methods to examine resource use and patterns at the species level and community level for native bumble bees (Bombus) and how preference, nutrition and competition affect foraging and community assembly.

Identifying resource and habitat preference is considered a key component to conserving declining species and is commonly used by conservation biologists for birds and mammals. Although identifying resource preference is also considered critical to conserving pollinators, to date no comparable work has been done on pollinator species. In chapter one, pollen use of the yellow faced bumble bee (Bombus vosnesensekii) was conducted using a Dirichlet multinomial model (DM) by comparing three different models of collection: 1) Use=Availability, 2)Use=Preference* Availability and 3) Use. Although B.vosnesenskii was thought to be a generalist forager, it does not collect pollen with respect to its availability and demonstrated significant preferences for some resources when accounting for availability. Additionally, models that excluded availability were typically superior to those that included availability. This suggests that estimates of preference and resource reliance can be adequately estimated without measuring availability which may be critical for cryptic species and those difficult to identify available resources such as pollinators.

To better understand the patterns of resource preference found in chapter one, analysis of features that may influence resource use was conducted. Morphological, nutritional (protein and amino acid composition) and availability (pollen per inflorescence) features of plants were measured for 12 species available to B. vosnesenskii. Plants were divided first by origin (native or non-native) and then by collection (collected or not-collected) and compared across groups. Although it is assumed native species should be superior for native pollinators no significant differences were found when compared based on origin. However, when grouped based on collection, collected species were significantly different for amino acid content, protein, and availability. This suggests that not all native plants are adequate resources to support pollinator communities and that not all invasive species are inadequate and that care should be taken when restoring plant communities for pollinators.

Lastly, phylogenetic methods were used to understand patterns of community assembly and competition in bumble bee communities. Community assembly theory would suggest that communities should attempt to limit similarity, either morphological or phylogenetic similarity, to maintain coexistence. Bumble bees have long been considered to limit overlap in a community by dividing niche space by tongue length and thus communities were thought to be comprised of species with different tongue lengths. To test this, the Nearest Taxon Index (NTI) and Net Relatedness Index (NRI) were calculated for 118 co-occurring communities in Nearctic Areas of North America using a phylogenetic distance matrix and a trait distance matrix. Although, long believed to limit similarity by having different tongue lengths this method found bumble bee communities were both phylogentically closely related and had more similar tongue length when compared to randomly generated null communities. This could suggest that patterns of community assembly and resource use may be driven by nutritional requirements of bumble bees and require them to share resources rather than divide the niche space as expected.

Although further work is needed to support these findings, the patterns found here have implications for conservation of native pollinators. The observed non-random patterns of resource use, both for preference and floral traits, suggest that significant effort should be made to determine plant species that support native pollinators particularly species that are declining. It is commonly assumed that by restoring native plant communities native bee communities will return but this work highlights that not all native plants can provide necessary protein or amino acids to support bee communities. Lastly, the high level of relatedness of bumble bee communities may suggest that whole communities may respond similarly to threats such as disease or habitat fragmentation and thus may be more susceptible to whole community losses. Efforts should be made to monitor populations and limit damage to communities. While all species have intrinsic value, pollinators are also invaluable to maintaining wild flower and crop plants that support higher trophic level diversity and food security and thus particular efforts should be made to conserve them.