Context‐dependence in an ant–aphid mutualism: direct effects of tending intensity on aphid performance

1. In ant–hemipteran mutualisms, ants receive carbohydrates in the form of honeydew, while hemipterans receive protection from natural enemies. In the absence of natural enemies, however, the direct effects of tending are generally less well known. We hypothesised that with increasing tending intensity (ant to aphid ratio), aphid performance would increase initially, then decrease at high tending levels due to the metabolic cost of producing high quality honeydew. 2. We tested our hypothesis in a greenhouse experiment by manipulating Argentine ant (Linepithema humile Mayr) colony size while holding constant the initial size of aphid (Chaitophorus populicola Thomas) aggregations. The two parameters associated with survival, aphid survivorship to maturity and longevity, declined with increasing tending intensity, whereas per capita birth rate and time to first reproduction showed no relationship to attendance. The intrinsic rate of increase declined only at relatively high tending levels, suggesting a nonlinearity in the effect of tending intensity. 3. Tending intensity measured in the experiment was similar to that observed in free‐living aggregations of C. populicola. Furthermore, the per capita recruitment rate of ants to free‐living aphid aggregations was negatively density‐dependent, indicating that small aggregations tend to experience the highest levels of tending intensity. This finding suggests that the aphid's intrinsic rate of increase may be positively density‐dependent, mediated by the aphid's mutualistic interaction with the ant. 4. In the Argentine ant–C. populicola interaction, experimental manipulation of colony size revealed a direct cost of ant attendance that was conditional upon tending intensity. Experiments that manipulate only ant presence or absence may yield an incomplete understanding of the mutualistic interaction if underlying nonlinearities exist.