Plant responses to herbivory shift the contributions of different pollinator taxa to plant reproduction

Both mutualistic and antagonistic interactions affect plant survival and reproduction. Plant responses to antagonists such as herbivores are expected to negatively impact plant interactions with mutualistic pollinators and plant fitness, but there is mixed support for this idea. Bee taxa vary in their effectiveness as pollinators and their preferences for floral traits, both of which may be affected by herbivory. Therefore, the behavioral responses of specific pollinators to herbivory could determine if and how herbivory exerts reproductive costs for plants. We manipulated herbivory on Physalis philadelphica (tomatillo) using Lema daturaphila beetles and simulated the plant's physiological response to herbivory using jasmonic acid (JA). We evaluated how herbivory and JA treatments affected overall pollinator importance to plant reproduction by different bee taxa (generalists Apis mellifera, Bombus, Augochlorini, Lasioglossum and the specialist Colletes latitarsis ). We calculated pollinator importance by combining per‐visit effectiveness (fruit weight and percent fruit set) with bee abundance. Relative to the control, herbivory reduced the overall pollinator importance of C. latitarsis bees but no other taxa. Herbivory affected fruit weight following visits by Augochlorini but had no overall effect on Augochlorini pollinator importance, demonstrating that the pollinator importance metric describes a distinct net effect of herbivory on pollination services. Nectar foraging was positively associated with fruit set for all taxa. Herbivory treatment affected foraging only for C. latitarsis, which had higher rates of nectar foraging on herbivore‐treated compared with control plants. Treatment had inconsistent effects on taxon‐specific visit duration, with JA increasing Lasioglossum visit duration in one of two study years. Our results suggest that variation in pollinator community composition could alter the strength and direction of herbivore effects on pollination, possibly explaining the inconsistent support for hypothesized herbivore‐driven pollination deficits. Therefore, the identity of interaction partners may determine whether the expected tradeoffs between managing mutualists and antagonists occur.