Natural enemy diversity and biological control: Making sense of the context-dependency

Numerous studies have demonstrated that diverse predator assemblages can be more effective at controlling prey populations. Yet, other studies have shown no effect of predator diversity on prey mortality, or even negative effects (for example due to intraguild predation or interference). Much research emphasis has been placed on the traits of predators that maximise functional complementarity. However, comparatively less attention has been paid to the traits of the prey or habitat that may maximise predator diversity effects, even though there must be a variety of prey niches available to be partitioned in order for niche complementarity to occur. Following this logic, we review six hypotheses for when diverse enemy assemblages should be most effective: when 1) prey communities are diverse; 2) prey have complex life cycles; 3) prey are patchily distributed in space or time; 4) studies are conducted at larger spatial and temporal scales; 5) plant structures are complex; 6) prey are abundant. Many of these hypotheses lack direct tests, particularly in agricultural systems, but we find little or no direct or indirect support for hypotheses 1, 4, 5 and 6. However, previous work does provide some support for hypotheses 2 and 3. We discuss methods to test these hypotheses directly, and suggest that natural enemy diversity may only benefit the biological control of arthropods in heterogeneous systems.