Argueta-Guzman, M.; Gonzalez, B.; Van Pelt, N.; Dias de Almeida, A. C.; Jimenez Delgado, T.; Pena, L.; Hutchinson, M. C.; Palumbo Gaiarsa, M.

A central challenge in characterizing species' niches is ensuring that foraging data accurately capture both the resources used and their relative importance, and the role of resource abundance in shaping foraging patterns. Most studies infer diet breadth and resource-use patterns from observational records, yet such data can mask resource-specific decisions when animals forage with different goals. Here, we test this experimentally using individually identifiable bees in controlled resource communities to quantify foraging decisions between nectar (for sustenance) and pollen (for offspring provisioning). Combining observations, pollen DNA metabarcoding, and pollen microscopy, we show that observed visitation patterns misrepresent the floral resources most important for offspring provisioning, which ultimately determines offspring survival and population persistence. We further show that interaction patterns are structured from processes beyond resource abundance. Our results demonstrate that commonly used observational approaches can mischaracterize diet breadth, potentially challenging conclusions about species generalization.