Strong Preferences for Apparently Dangerous Habitats? A Consequence of Differential Escape from Predators

Abstract

Predators may affect prey populations not only through the direct removal of individual prey (e.g., Taylor 1984), but also through their prey's anti-predator behavior (e.g., Abrams 1984, Sih 1987, Kotler and Brown 1988). Several recent studies (see Lima and Dill 1990) demonstrate that these behaviorally-mediated effects are particularly important in determining the spatial distribution of prey in the environment: animals avoid habitats with abundant predators and correspondingly high risk of attack. This is a reasonable result, and ruling-out between-habitat differences in food abundance, mating prospects, etc., it might appear entirely general. However, consider the fact that habitat choice may be influenced by the probability of escape as well as the probability of attack itself. Studies of habitat use under the risk of predation have focused almost exclusively on habitat-specific attack rates, predator abundance, etc. (see Lima and Dill 1990), whereas habitatspecific escape has rarely been considered (or essentially held constant between habitats). When considering both attack and escape simultaneously, the above avoidance of habitats with high risks of predatory attack may no longer hold in general. This paper explores some simple ideas on why animals may not only prefer apparently dangerous habitats, but also exhibit such strong preferences for them that a very high risk of attack may not drive them away. Consider a generic animal with a choice between two habitats, designated 1 and 2. Let ai and ej represent the probability of attack in habitat i, and probability of escape given attack in habitat i, respectively. The probability of being killed by a predator in habitat i (Pi) is thus Pi = ai(1-ei). For simplicity, assume that the habitats differ only in Pi values, thus habitat 1 will be chosen when Pi<P2, or ai(1-ej)<a2(1-e2), which rearranges to