Shadows of predation: habitat-selecting consumers eclipse competition between coexisting prey

Ecologists have made substantial progress evaluating the influences of adaptive behaviors on population dynamics and communities. But no-one has examined the joint influences of stochastic variation, predators, and density-dependent habitat selection on our interpretations of species coexistence. I begin the search with simulation models of habitat isodars (lines along which the fitness of individuals is identical in two or more habitats) assuming ideal-free habitat selection by two prey species exploited by a habitat-selecting parasitoid predator. The models include both regulating and non-regulating stochasticity. The intriguing results include the following: (1) all three species often achieved a true ideal-free distribution; (2) predators reduced prey population sizes and increased the frequency of local habitat extinctions; (3) despite the predator's differential reduction of prey densities, there was no evidence of apparent competition; (4) all species exhibited pulses of dispersal associated with donor–receiver population dynamics; (5) isodars produced valid estimates of competition between prey only in constant environments lacking habitat-selecting predators; (6) habitat-selection by predators forced prey into their preferred habitats; (7) the resulting ghost of competition obscured the prey species' competitive interaction; (8) isodars correctly revealed density-dependent habitat selection by the predator. Overall, the results appeared to depend primarily on the predator's habitat choice, rather than on prey trade-offs between competitive ability and reduced value (handling time) to the predator. Habitat selection theory, and its revelation via isodars, can thus provide considerable insight into processes affecting real communities, and most especially if ecologists assess carefully the constraints for their analysis and interpretation. Nevertheless, isodars designed to measure competition are likely to be most reliable in donor-controlled or experimental systems where regulating stochasticity has relatively little influence on prey dynamics.