Broad-Scale Hypotheses do not Account for Species Richness Patterns of Central American Mayflies

Abstract

I performed an exploratory analysis of four broad-scale hypotheses (area availability, energy availability, habi- tat heterogeneity and geometric constraints) for changes in species richness over Central American mayfly species, and compared their outcomes at different taxonomic levels. I employed an eigenvector-based spatial filtering to control spatial autocorrelation effects and conducted multiple spatial eigenvector regressions to identify the strongest predictors of spe- cies, genus and family richness. The usefulness of higher-taxa as surrogates for species richness was assessed by Pearson correlations. Mayfly species richness is characterized by a patchy pattern. None of the variables accounted for observed patterns. Only area availability was an effective predictor of mayfly genus and family richness, although both habitat het- erogeneity and energy presented marginal effects on genus richness. I did not observe any relationship between species and higher-taxa richness. Broad-scale hypotheses did not explain species richness patterns of mayflies, which instead can be explained by interactions among studied variables and spatial structure.