Latitudinal patterns of New World species diversity and range size

The latitudinal gradient in species diversity is one of the best-documented patterns in biogeography. Explanations for the latitudinal gradient in species diversity (LGSD) ranging from null models to evolutionary and ecological hypotheses rely on relationships between range size and species richness. I test Rapoport's rule, where mean range size of tropical organism is predicted to be smaller than for temperate and boreal organisms, using New World woody plants. Geometric constraint models (GCMs) posit that barriers to dispersal, in the absence of ecological and evolutionary gradients can generate gradients in species diversity. I describe an analytic solution for the most common geometric constraint model and test expected and observed richness patterns for New World palms. I show that domain truncation will inevitably and spuriously increase the predictive power of GCMs while potentially decreasing the predictive power of other competitor variables. The LGSD is often considered a general pattern driven by general processes. Assuming that generality of pattern may reflect generality of process, I examine the LGSDs for New World mammal orders and families, testing for generality in two components of the LGSD, shape and peak. The overall LGSD is not a general pattern for across mammal orders and families and therefore it is unlikely that the LGSD is generated by a shared general mechanism or historical process.