Skip to main content

Global patterns in bumble bee pollen collection show phylogenetic conservation of diet

Cite this dataset

Wood, Thomas et al. (2021). Global patterns in bumble bee pollen collection show phylogenetic conservation of diet [Dataset]. Dryad.


1. Bumble bees (Bombus) are a group of eusocial bees with a strongly generalised feeding pattern, collecting pollen from many different botanical families. Though predominantly generalists, some bumble bee species seem to have restricted dietary choices. It is unclear whether restricted diets in bumble bees are inherent or a function of local conditions due to a lack of data for many species across different regions.

2. The objective of this study was to determine whether bumble bee species displayed specific patterns of pollen collection, and whether patterns were influences by phylogenetic relatedness or tongue length, a trait known to be associated with structuring floral visitation.

3. Bumble bee pollen collection patterns were quantified from 4,132 pollen loads taken from 58 bumble bee species, representing 24% of the pollen-collecting diversity of this genus.

4. Phylogenetic trait mapping showed a conserved pattern of dietary dissimilarity across species, but not for dietary breadth. Dietary dissimilarity was driven by collection of Fabaceae, with the most similar species collecting around 50–60% of their diet from this botanical family. The proportion of the diet collected from Fabaceae also showed a conserved phylogenetic signal. Greater collection of Fabaceae was associated with longer tongue lengths, with shorter tongued species focusing on alternative botanical families. However, this result was largely driven by phylogenetic relatedness, not tongue length per se.

5. These results demonstrate that, though generalists, bumble bees are still subject to dietary restrictions that constrain their foraging choices. These dietary constraints have implications for their persistence should their core resources decline in abundance. 02-Jun-2021 --


Pollen diet quantification. Bumble bee pollen diets were quantified following established methodologies (Kleijn & Raemakers, 2008; Wood et al., 2019; see Supplementary Methods). Briefly, dried pollen balls were removed from selected pinned bumble bee specimens taken from museum and private collections in order to achieve the greatest possible taxonomic representation. Use of museum material is necessary for species such as B. cullumanus that have declined to extinction across large parts of their range (Williams et al., 2013), and therefore cannot easily be studied in contemporary landscapes. These balls were rehydrated, stained with fuchsin jelly, and pollen grains were identified using light microscopy, predominantly to genus. The proportion of each plant taxon was assessed, with taxa representing <2–5% excluded (depending on identifier) as these may have arisen from contamination (Müller, 1996). These proportions were weighted by the overall size of the pollen ball relative to a full load, and weightings were used to calculate final dietary proportions. A total of 4,136 pollen loads from 58 bumble bee species belonging to 13 subgenera (representing approximately 93% of subgenera and 24% of species that collect pollen, Dataset S1, Dataset S2) from 34 countries were analysed (Table S1), 1,743 of which were published previously (Kleijn & Raemakers, 2008; Wood et al., 2019). Specimens for which label dates were available (n=4,089) were collected between 1874-2020 (median=1992). Data of capture was not considered or controlled for in this study, as there is no evidence to suggest that bumble bee diets change meaningfully over time (Kleijn & Raemakers 2008). The most important museums were the Laboratory of Zoology collection, University of Mons (1,221 specimens, Mons, Belgium), the Naturalis Biodiversity Center (552 specimens, Leiden, the Netherlands), the J.B. Wallis/R.E. Roughley Museum of Entomology (369 specimens, Winnipeg, Canada), the Royal Belgian Institute of Natural Sciences (341 specimens, Brussels, Belgium), the Bavarian State Collection of Zoology (323 specimens, Munich, Germany), the A.J. Cook Arthropod Research Collection (296 specimens, East Lansing, USA), and the Natural History Museum (239 specimens, London, UK). Full details are available in Dataset S2.

Usage notes

Information explaining the relevant sections for each datafile can be found within each file on the 'Metadata' tab.