Discussion
Bumble bees were generally less flower constant during pollen collection than honey bees, supporting our prediction that trapline foraging results in a greater proportion of bees visiting more than one plant taxon for pollen collection during single foraging bouts. Likewise, the observation that honey bees tended to forage on a single plant type during a foraging bout supports the hypothesis that dance communication generates cohorts of individuals that forage on a single profitable patch (Visscher & Seeley, 1982). At the colony level, the diversity of pollen collected by bumble bees was greater than that of honey bees, however the richness of pollen collected was temporally variable for both bee species (Fig.3). Similar differences in flower constancy and diversity were also observed between honey bees and bumble bees in agricultural systems in Germany (Leonhardt & Blüthgen, 2012). Interestingly, during period three, we observed an increase in the rate of flower constancy, which was associated with a decrease in the richness of pollen collected by bumble bee colonies, with the opposite pattern observed in honey bees, supporting a link between individual and colony level foraging patterns. These results support the idea that foraging strategies, such as trapline foraging and dance communication help predict differences in pollen foraging patterns by different bee species. Such knowledge not only highlights interspecific differences in foraging patterns, but also informs questions concerning the dynamic responses of each bee species to spatiotemporal ecological challenges present in human altered landscapes. This type of information can be used to facilitate the development of pollinator friendly habitats that also support pollination of bee-dependent crops.
During period three the number of foraging bumble bees declined, while honey bees were regularly foraging for pollen at sites two and three, but not at site one (Appendix Table S1). We might expect changes in resource availability, which are known to mediate distinct responses among bee species, to contribute to the observed temporal foraging dynamics of honey bees and bumble bees (Jha & Vandermeer, 2009). However, at the spatial scale of 500m around each hive site we did not detect any temporal differences in the richness or diversity of plant resources available to bees (Appendix Fig. S2). Interestingly, in period 3, both bee species gathered pollen from plant families found outside the survey areas, such as Brassicaceae by bumble bees or Geraniaceae by honey bees (Table 2; Appendix Fig. S3). It has been suggested that summer is a challenging time for bees, possibly as a result of competition for shared resources (Couvillon, Schürch, & Ratnieks, 2014; Minahan & Brunet, 2018). In July, Bumble bee individuals foraged more frequently from single pollen sources, while more honey bee individuals visited multiple resources, and both bee species exploited new plant families for pollen. For honey bees, it would be interesting to determine if fewer foragers followed the dance communication during this period, which if true would suggest more bees scouting for resources, or visiting previously learned locations (Biesmeijer & Seeley, 2005). Future studies should further examine life history and ecological factors associated with the temporal dynamics of pollen and nectar foraging, and how these are linked to distinct foraging strategies. Understanding these factors will improve the ability of conservation planners to select resource types and plan their spatial and temporal distributions according to the dynamic needs of pollinator communities.
As predicted based on the trapline foraging strategy, bumble bees demonstrated a preference for Fabaceae_Tricolporate but an avoidance of Apiaceae pollen, while honey bees foraged randomly, based on the floral abundance of plant families in the area. The preference for certain plant types when foraging for pollen could depend on differences in the amount of pollen made available by plants, with bumble bees being able to detect such differences (Brunet, Thairu, Henss, Link, & Kluever, 2015). Preference could also be influenced by the ability of a bee species to handle and exploit particular flower morphologies (Raine and Chittka, 2007), or the nutritional complement of pollen relative to other available resources (Somme et al. 2015). The observation of preference by bumble bees but not honey bees suggests a potentially greater ability of bumble bees to respond to floral traits such as fatty acid content of pollen (Ruedenauer et al., 2020) relative to honey bees. This may be necessary when using trapline foraging as opposed to dance communication as a foraging strategy.
There exist some challenges to our approach of determining preference. A first challenge is selecting an appropriate spatial scale and survey method to estimate floral resource availability. Our surveys were limited to 500m, which is likely to encompass many foraging trips, but not all, as bees can forage beyond this range (Beil, Horn, & Schwabe, 2008; Danner, Molitor, Schiele, Härtel, & Steffan-Dewenter, 2016; Nagamitsu, Tsukuba, Ushirokita, & Konno, 2012). A second challenge is that pollen types must be present in both pollen pellets and resource surveys for our analyses. However, it can also be informative when pollen is found in resources but not bee collected pollen, or vice versa. For example, the absence of Linaria vulgaris ,Oxalis, and Phlox pollen in pollen pellets suggests that bees are avoiding these available resources. Plants only present in the pollen pellets, such as Geraniaceae, most likely came from home gardens, which can provide valuable resources for bees (Goulson, Lye, & Darvill, 2008). These results highlight the challenge associated with analyzing preferences in complex human altered landscapes. However, future experiments should explicitly consider the diversity of habitat types present when studying pollinator foraging patterns and their relationship with foraging strategies.