Habitat selection in slowly regenerating environments

Theories of density dependent habitat selection (isoleg theories) have been used to deduce community phenomena from the behavior of the individual. Previous theories assume either that at any point in geographical space representatives of a single patch type have the same resource density, or that any spatial variation in resource density is not perceived by potential foragers. We extend these theories by allowing patches of a particular type to vary perceptibly in resource density in space (as well as time). We analyze a model incorporating distinct-preference community structure: each species does best in a different habitat type. The results differ greatly from similar models with no spatial variance. First, so long as some patches achieve their maximum resource density, the isolegs (the lines in state space separating regions of qualitatively different behaviors) have negative or vertical slopes. Previous models have had positive slopes. Second, opportunistic foragers are only semi-opportunists: they exhibit partial preferences by accepting only a fraction of their secondary patches. Third, isolegs are isoclines (lines of constant per capita growth rate). Fourth, although the habitat selection promotes the coexistence of species (as it does in zero variance models), there is little likelihood of it producing "the ghost of competition past" (i.e. interaction coefficients of zero in the neighborhood of the density equilibrium). Fifth, niche shifts can behave perversely: increase of a competitor's density may actually increase niche breadth. Or it may cause niche breadth to decrease, in agreement with other theories. As might be expected, models which allow intermediate amounts of spatial variation produce conclusions intermediate between the two extremes of: (1) zero variation and (2) sufficient variation to produce some patches at resource carrying capacity no matter how low the mean resource density.