Abundance and range relationship in a fragmented landscape: Connections and contrasts between competing models

The specialization and metapopulation dynamics hypotheses are the most prominent alternative explanations of the oft-reported correlation between the distribution and abundance of species. We suggest that mechanisms underlying the two explanations are interacting to produce the correlation. We aim to quantify the relative contribution of the mechanisms underlying the two proposed explanations, should both have some validity. An analysis of 40 species of invertebrates inhabiting 49 miniature erosional rock pools allowed us to explicitly address the relationship between distribution, specialization, and abundance. The analysis involved regressing species abundances against several measures of specialization and distribution. Compound measures of species specialization (niche volumes) correlated less with the observed densities than when the constituent variables were used separately, especially in combination with the distribution data. The latter group of statistical models increased the amount of variance explained compared to the best niche volume derived estimates (53% vs 79%). The comparisons further suggest that specialization is a much stronger determinant of species abundance than is metapopulation dynamics. This appears to be particularly true in our system of discrete habitats. To put these results in a broader context, we propose a conceptual model to explain the relative importance of stochastic processes and specialization constraints in predicting patterns of species abundance. While this model focuses on site-patch isolation and on ecological differences among sites-patches, it permits other dimensions such as habitat resolution differences among species.