Predator foraging success and habitat complexity: quantitative test of the threshold hypothesis.

Numerous studies have demonstrated a negative relationship between increasing habitat complexity and predator foraging success. Results from many of these studies suggest a non-linear relationship, and it has been hypothesised that some "threshold level" of complexity is required before foraging success is reduced significantly. We examined this hypothesis using largemouth bass (Micropterus salmoides) foraging on juvenile bluegill sunfish (Lepomis macrochirus) in various densities of artificial vegetation. Largemouth foraging success differed significantly among the densities of vegetation tested. Regression analysis revealed a non-linear relationship between increasing plant stem density and predator foraging success. Logistic analysis demonstrated a significant fit of our data to a logistic model, from which was calculated the threshold level of plant stem desity necessary to reduce predator foraging success. Studies with various prey species have shown selection by prey for more complex habitats as a refuge from predation. In this stydy, we also examined the effects of increasing habitat complexity (i.e. plant stem density) on choice of habitat by juvenile bluegills while avoiding predation. Plant stem density significantly effected choice of habitat as a refuge. The relationship between increasing habitat complexity and prey choice of habitat was found to be positive and non-linear. As with predator foraging success, logistic analysis demonstrated a significant fit of our data to a logistic model. Using this model we calculated the "threshold" level of habitat complexity required before prey select a habitat as a refuge. This density of vegetation proved to be considerably higher than that necessary to significantly reduce predator foraging success, indicating that bluegill select habitats safe from predation.Implications of these results and various factors which may affect the relationships described are discussed.