The Dynamics of Prey Refuge Use: A Model and Tests with Sunfish and Salamander Larvae

Abstract

Although it has often been suggested that refuges are a crucial factor allowing prey to persist with predators, relatively few quantitative studies have directly addressed the patterns or implications of refuge use. In particular, refuge use is typically studied via a "snapshot" approach. This approach ignores the dynamic nature of refuge use, that is, the importance of rates of emergence from and reentry into refuge. We developed a simple dynamic model of refuge use that predicts patterns over time for prey survival and for the proportion of prey in refuge. Our model yielded some predictions that are, at first glance, counterintuitive. For example, we predicted that it should be possible to have virtually all prey in safe refuges and yet have high predation rates. This occurs if emergence rate is high and if prey are killed at a high rate after they emerge. Although in retrospect this possibility seems obvious, this sort of dynamic approach has not been previously investigated. We tested our model using two species of predatory sunfish feeding on larvae of the small-mouthed salamander, Ambystoma texanum. The addition of either fish resulted in a rapid decrease in the proportion of larvae out of refuge. Despite the fact that fewer than 5% of the larvae were exposed at any given time, no larvae survived 24 hours with bluegill sunfish. In contrast, larval survival with green sunfish was relatively high. The observed patterns in both proportion of larvae exposed and larval survival were accurately explained by the dynamic model. In particular, the apparent paradox of a high predation rate when few prey are exposed could be easily understood only in terms of the dynamics of refuge use. In the Discussion, we address the implications of our results for ecological studies of predator-prey population dynamics and for analyses of adaptive antipredator behavior.