Plant species composition alters the sign and strength of an emergent multi-predator effect by modifying predator foraging behaviour.

The prediction of pest-control functioning by multi-predator communities is hindered by the non-additive nature of species functioning. Such non-additivity, commonly termed an emergent multi-predator effect, is known to be affected by elements of the ecological context, such as the structure and composition of vegetation, in addition to the traits of the predators themselves. Here we report mesocosm experiments designed to test the influence of plant density and species composition (wheat monoculture or wheat and faba bean polyculture) on the emergence of multi-predator effects between Adalia bipunctata and Chrysoperla carnea, in their suppression of populations of the aphid Metopolophium dirhodum. The mesocosm experiments were followed by a series of behavioural observations designed to identify how interactions among predators are modified by plant species composition and whether these effects are consistent with the observed influence of plant species composition on aphid population suppression. Although plant density was shown to have no influence on the multi-predator effect on aphid population growth, plant composition had a marked effect. In wheat monoculture, Adalia and Chrysoperla mixed treatments caused greater suppression of M. dirhodum populations than expected. However this positive emergent effect was reversed to a negative multi-predator effect in wheat and faba bean polyculture. The behavioural observations revealed that although dominant individuals did not respond to the presence of faba bean plants, the behaviour of sub-dominants was affected markedly, consistent with their foraging for extra-floral nectar produced by the faba bean. This interaction between plant composition and predator community composition on the foraging behaviour of sub-dominants is thought to underlie the observed effect of plant composition on the multi-predator effect. Thus, the emergence of multi-predator effects is shown to be strongly influenced by plant species composition, mediated, in this case, by the provision of extra-floral nectar by one of the plant species.