An experimental test of the nature of predation: neither prey- nor ratio-dependent

<b>1</b>. There is a current debate about the appropriateness of prey-dependent vs. ratio-dependent functional responses in predator–prey models. This is an important issue as systems governed by these models exhibit quite different dynamical behaviour. However, the issue is not yet resolved on a theoretical basis, and there is a lack of experimental evidence in natural systems. We used a paper wasp–shield beetle system in a natural setting to assess the validity of either approach. <br> <b>2</b>. We manipulated the abundance of herbivorous insect prey on thistle plants and of predatory paper wasps in the immediate environment of the prey by opening or closing cages containing wasp nests. <br> <b>3</b>. The number of wasps foraging at the site increased when cages were opened, but rapidly reached an asymptote, indicating predator interference. The predation rate per predator decreased with the number of wasps in the environment. Thus, the functional response depended on both prey and predator density.<br><b>4</b>. Neither a pure prey- nor a pure ratio-dependent model fitted perfectly our observations. However, the functional response of the paper wasps towards shield beetle larvae was closer to ratio-dependence. To our knowledge, this is the first experimental evidence discriminating between ratio- and prey-dependence in a natural setting with unconfined predators and prey.<br><b>5</b>. Predator interference was most probably responsible for the specific form of the functional response found. We found indications that both direct (e.g. aggression) and indirect interference mechanisms (e.g. depletion of easy-to-find prey) were at work in our system. We conclude that predator density cannot be ignored in models of predator–prey interactions.