Introduction
As one of the most valuable pollinators, bumble bees are vitally important to productivity for both natural and agricultural systems (Hegland & Totland, 2008; Garratt et al., 2014). They are distributed world-wide in temperate climates (Williams & Osborne, 2009). An increasing number of papers have suggested that high priority should be given to their conservation due to the serious risk of declines (Goulson et al., 2008a, 2015; Williams et al., 2009; Roulston & Goodell, 2011). A main cause for bumble bee decline is the decrease of floral resources (Roulston & Goodell, 2011; Goulson et al., 2015; Thomson, 2016); there is strong evidence that floral resources significantly influence abundance of bumble bees (reviewed in Roulston & Goodell, 2011 and Goulson et al., 2015) and that responses to floral resources can be species-specific (Ogilvie et al., 2017). However, the interspecific interactions of bee species in response to changing floral resources still need to be explored, as they are fundamental for understanding the patterns and mechanisms of bee decline in this changing world (Thomson, 2016).
It was suggested that two or more bumble bee species coexisting in a plant community may have to obey rules for partitioning the limited floral resources through competitive interactions (Inouye, 1978; Heinrich, 1976; Ranta & Lundberg, 1980), and niche differentiation in bumble bee proboscis length and flower use could be be an effective way of partitioning resources (Heinrich, 1976; Inouye, 1980; Pyke, 1982). Matching of proboscis length and flower depth was identified as a key component for the resource partitioning among bumble bee species (Inouye, 1977; Ranta et al., 1981; Pyke, 1982). Therefore, differences in proboscis length may influence niche relations and community structure for bumble bee species (Ranta & Vepsalainen, 1981; Pyke, 1982; Pyke et al., 2012). On this basis, a classic hypothesis was postulated that coexistence would be difficult for bumble bee species with similar proboscis length in a community (Inouye, 1977, 1978, 1980; Pyke, 1982; Ishii et al., 2008). However, studies also indicated that the structure of bumble bees may also depend on the community’s temporal and spatial factors rather than niche differentiation in proboscis length and flower use (Ranta & Lundberg, 1980; Ranta & Vepsalainen, 1981; Pyke et al., 2012; Miller-Struttmann & Galen, 2014). However, studies revealed that multiple bumble bee species with similar proboscis length can coexist in a community with few differences in their abundance (Ranta, 1982, 1984; Ranta & Tiainen, 1982; Williams, 1989; Goulson & Darvill, 2004). This indicated that the interspecific interactions within a community of bumble bees are more complex than previously thought.
The intensity of interspecific competition among bumble bee species was thought to be highly environment-dependent (Ranta & Lundberg, 1980). For example, Ranta et al. (1981) showed that the niche overlaps of the bumble bees were smaller in a northern (unstable environment) than in the southern community (more stable environment) in Fennoscandinavia. Moreover, the foraging strategy of bumble bees displayed a high level of plasticity in response to changes in landscape factors and floral resource diversity (Jha & Kremen, 2013). However, whether and how interspecific competition influences the foraging strategy of bumble bees in a scenario of changing floral resources remains unclear, but is fundamental for understanding the coexistence of multiple bumble bee species. Thus it may be important to incorporate a phenological component to study of their resource use, as abundance of bees in the annual colonies changes markedly during the course of the summer.
Nutritional requirement was proved to be a key force in shaping foraging strategy and floral preference for bumble bees (Vaudo et al., 2016), which is not surprising given the impact of diet on their colony health (Roger et al., 2017). Different bumble bee species may vary in nutritional requirements, which would force them to prefer different plant species for comprehensive nutrition intake (Pyke, 1982; Goulson et al., 2005; Hanley et al., 2008; Moerman et al., 2016; Woodard, 2017). Given that pollen and nectar of flowering plants provide different nutrients for pollinators (reviewed in Baker & Baker, 1983; Roulston & Cane, 2000), bumble bees may differ in preference for pollen or nectar, and thus preferentially forage on different plants since some plants do not provide nectar (Konzmann & Lunau, 2014). For the bumble bees that forage preferentially for pollen rather than nectar, niche differentiation in proboscis length and flower use may not be important in partitioning of floral resources, although such mechanisms are not well studied. Detecting the influence of intensity of interspecific competition for floral resources on foraging strategy of mainly pollen-seeking bumble bees should provide a good perspective to understand their community structure.
In this study, we hypothesized that the intensity of interspecific competition may influence foraging strategy of bumble bees. On one hand, the niche differentiation in proboscis length and flower use might not be stationary, but depend on the degree of competition. Alternatively, the foraging preference for pollen or nectar might also be plastic when coexisting species reduce the available floral resources. To understand better the factors influencing community structure and the coexistence of bumble bees with similar proboscis lengths, it is necessary to consider study sites where the level of spatiotemporal variability is minimal, or sites are sufficiently isolated that immigration should not be significant (Pyke et al., 2012). To test the hypothesis, we conducted field investigations on a relatively isolated alpine meadow in the eastern Himalayan of Yunnan Province, China, a world top biodiversity hotspot. Three Bombus species, namely, B. friseanus,B. lepidus and B. festivus were the most abundant species, and their proboscis lengths (workers and males) vary from about 6 to 10 mm. Specifically, the main objectives of the study are: (i). what is the pattern of partitioning of floral resources by the three bumble bees? (ii). whether and how the niche differentiation in proboscis length and flower use is influenced by intensity of interspecific competition for the three bumble bees; and (iii). whether and how the intensity of interspecific competition influences foraging preference between pollen and nectar for the three bumble bees.