RESULTS
The number of pollinator visits by the three pollinator groups (honeybee, bumblebee, and other pollinators) varied considerably among observed flowers (ranges: honeybees = 0–7, bumblebees = 0–38, other = 0–54; χ2 = 16.12, df = 2, p = 0.0003). Honeybees had fewer visits compared to bumblebees (p = 0.0003) but did not differ from combined other pollinator visits (p = 0.15; Figure 1a; Appendix S1: Table S3, S5, and S6). On the other hand, the time bees spent on flowers per visit (seconds) did not differ among honeybees, bumblebees, and other pollinators (Figure 1b; χ2 = 4.95, df = 2, p = 0.08), despite substantial variation in total duration per visit among flowers observed (ranges: honeybees = 0–219.25, bumblebees = 0–222.75, other = 0–484 seconds). We further explored how bee species may differ in how much time per visit they spend interacting with different aspects of the flower, including the petals, pollen, and simultaneously contacting the pollen and nectar (pollen+nectar). Bumblebees spent less time per visit on petals compared to honeybees (p < 0.0001) or other pollinators (p < 0.0001; Figure 1c). Other pollinators spent more time per visit on pollen-only visits relative to honeybees (p = 0.0025) and bumblebees (p < 0.0001; Appendix S1: Fig. S1). On average, all three pollinator groups spent similar amounts of time per visit in contact with pollen+nectar (Figure 1d; χ2= 3.60, df = 2, p = 0.17), though there was a wide range of visit times for pollen+nectar (ranges: honeybees = 0–197.75, bumblebees = 0–218.75, other = 0–295 seconds). Overall, each pollinator group differed in the number of visits and duration of time spent per visit interacting with different aspects of the flowers, which could contribute to variation in the likelihood of bees depositing or picking up parasite spores during floral visits.
V. ceranae was highly prevalent in both honeybees and bumblebees at all six field sites. In total, 68.0% (95% CI: 56.7–77.9%) of honeybees and 64.0% (95% CI: 52.9–73.6%) of bumblebees had V. ceranae detected in their midguts. V. ceranae prevalence did not significantly differ between host species (χ 2 = 0.14, df = 1, p = 0.71). Among different sites, V. ceranaeprevalence ranged from 57.1% to 81.3% in honeybees and from 40.0% to 93.8% in bumblebees (Appendix S1: Fig. S2, Table S9).
To determine whether floral visitation behaviors were linked withV. ceranae prevalence, we explored how the number of pollinator visits and the duration of time per visit spent interacting with certain parts of the flower correlated with V. ceranae prevalence in honeybees and bumblebees. Despite a lower number of honeybee visits compared to bumblebee visits (Figure 1a), the number of honeybee visits was the only factor that had a significant impact on V. ceranaeprevalence. V. ceranae prevalence in bumblebees was positively linked with honeybee flower visits (p = 0.005; Table 1, Figure 2a) but not bumblebee flower visits (p = 0.98, Figure 3a). In contrast, V. ceranae prevalence in honeybees was not linked with honeybee flower visitation (p = 0.57, Table 1, Figure 2a) or bumblebee flower visitation (p = 0.47, Figure 3a). V. ceranae prevalence in honeybees and bumblebees was also not linked with flower visitation by other bee genera (both p > 0.54, Figure 3b; Table 1, Appendix S1: Table S7 and S8).
We also expected that greater amounts of time bees spent per visit on flowers and interacting with different aspects of the flower (e.g., petals, pollen, and pollen+nectar) would increase the V. ceranaeprevalence by increasing the chances of parasite transmission. For the duration per visit models, we used a Bonferroni-corrected significance threshold of 0.0125 because the durations per pollinator visit to each part of the flower were analyzed separately. V. ceranae in honeybees and bumblebees was not associated with the total duration per visit of honeybees, bumblebees, or other pollinators (Table 1). However,V. ceranae prevalence in bumblebees was marginally, but not significantly, higher the longer that honeybees spent interacting with flowers per visit (p = 0.014, Figure 2b). We further explored this result by breaking down the total floral visit duration by the duration of time that bees spent interacting with different flower parts, including petals, pollen, and pollen+nectar to determine which specific behaviors contributed most to V. ceranae prevalence in each host species (Table 1; Appendix S1: Table S7 and S8). V. ceranaeprevalence in bumblebees was higher the longer honeybees interacted with pollen+nectar (p = 0.009; Figure 2c), despite no overall differences in time spent on pollen+nectar per visit among honeybees, bumblebees, and other pollinators (Figure 1d). V. ceranae in both host species was not impacted by bumblebee or other pollinator duration spent per visit on pollen+nectar (Figure 3c,d; Table 1). V. ceranae in honeybees and bumblebees was not correlated with the time per visit that bees spent on petals or pollen, regardless of bee species (Table 1).