1.Flower dwelling predators make flowers dangerous foraging sites for pollinators, potentially affecting their anti-predator behavior. Moreover, predation vulnerability often varies among pollinators’ body-sizes with interspecific comparisons showing that smaller species are more vulnerable than larger ones. However, how intraspecific body size variation influences pollinator behavior under predation risk is still unknown, especially under natural conditions.

2.We hypothesized that bumblebee workers of different sizes will exhibit different foraging strategies under predation risk. We predict that (i) small workers should more often exhibit anti-predator behaviors than larger workers. We also hypothesized that the anti-predator behavior should be influenced by predator size and reward availability, therefore we expect (ii) higher avoidance behavior towards larger predator sizes, and (iii) more and longer visits to inflorescences with high nectar availability. Finally, we expect that (iv) nectar availability should overcome the anti-predator behavior in less vulnerable, large, workers.

3.We recorded flower visitation, time spent and rejection behaviors of different sizes of Bombus terrestris (Apidae) workers (large, medium, and small) to inflorescences of Alstroemeria aurea (Alstroemeriaceae) with different treatments of artificial spiders (small and large) and nectar availability (with, without).

4.Anti-predator and foraging behavior of bumblebees was affected by the size of the worker, the presence of artificial spiders and nectar availability. Large and medium size bumblebees, strongly reduced flower visitation and time spent in the presence of artificial spiders, consistently avoiding flowers with spiders, regardless of spider size or nectar availability. Instead, small bumblebees seldom modified their behavior when facing artificial spiders, only increasing their avoidance or decreasing their foraging time in nectarless flowers hosting large artificial spiders.

5.This pattern of larger workers being more sensitive to predation risk than smaller ones at the intraspecific-level in B. terrestris is contrary to the expected and acknowledged trend based on previous interspecific comparisons, but partially consistent with predictions of models of optimal foraging theory. Intraspecific behavioral variability was uncovered only when nectar was available, whereas artificial predator size rarely modified bumblebee anti-predator and foraging behavior. Therefore, our findings suggest that the trade-off between maximizing resource intake and minimizing predation risk strongly varies across bumblebee worker body sizes.

To determine intraspecific variability in bumblebee responses to the presence of artificial spiders, we designed a fully crossed factorial field experiment with three factors: (1) predation risk with three levels (inflorescences without (Control), with a small (S) and with a large (L) artificial spider) and (2) nectar availability with two levels (with and without nectar), using 53 plants per treatment. In addition, (3) bumblebee workers body size was used as another factor with three levels: large, medium and small. We manipulated the amount of nectar in inflorescences by bagging them with a net that prevents insect visitation and enhances nectar accumulation, the day before the assay was carried out. Over the course of the flowering season, we bagged 159 plants from a total of 318 marked plants. The other group of 159 plants was exposed to floral visitors the day before the assay was carried out. If remaining nectar was present we extracted it using microcapillaries, before the behavioral observations were made, to ensure contrasting nectar availability conditions. Each artificial spider was placed in one flower of the inflorescence and the behavior of arriving B. terrestris bumblebees were observed during 10 minute periods.