Behavioural interactions in stream food webs: the case of drift-feeding fish, predatory invertebrates and grazing mayflies

1. We examined direct and indirect behavioural responses of grazing mayfly nymphs (Baetis rhodani) to predation risk from a drift-feeding fish (Europe an minnow; odour manipulated), and two types of invertebrate predators, Diu ra bicaudata (stonefly) and Rhyacophila nubila (caddis larva). We also asse ssed the direct responses of the invertebrate predators to fish chemicals. 2. Based on diel gut content periodicity, D. bicaudata nymphs were strongly nocturnal foragers. R. nubila was also nocturnal, but only in a fish strea m; in a stream lacking fish, their gut contents did not vary significantly on a diel basis. In the laboratory, Diura was nocturnal even in the absence of fish and almost ceased moving when fish was present. Rhyacophila shifte d from aperiodic to nocturnal foraging in the presence of fish. 3. The contrasting behaviours of the two predators may be explained by thei r respective field distributions: D. bicaudata always co-occurs with fish, whereas R. nubila is also found in streams without fish. Therefore, a capac ity for plastic anti-fish responses is beneficial for R. nubila, but not fo r D. bicaudata. 4. Drift of large Baetis nymphs was aperiodic when Rhyacophila was present, but nocturnal in all other treatments. Drift rate was highest when both Di ura and fish were present and lowest in treatments with Rhyacophila. Predat ory stoneflies, but not the caddis larva, induced a night-time peak in the drift of both Baetis size-classes. In the absence of predators, small Baeti s drifted aperiodically. Interaction terms in three-way ANOVAs testing for the indirect effects of fish and invertebrate predators on mayfly drift per iodicity and drift rate were all non-significant, indicating that the respo nse of Baetis to one type of predator was not modified by the other predato r. 5. Our results indicate that the drift periodicity of lotic mayflies may be fine-tuned to variations in the multi-predator environment and that prey r esponses are size-specific. We conclude that the effects of invertebrate pr edators on prey behaviour vary in relation to predator's foraging strategy and generalizations based on studies with only one type of predator should be avoided.