Fig. S7.

The present-day diversity–productivity relationship across space in extant large mammals. (A) Spatial variation of present-day genus richness for 861 extant species (267 genera) of large mammals (orders Artiodactyla, Carnivora, Perissodactyla, Primates, and Proboscidea) in 1° grid cells across the globe. Insets show genus richness within our focal regions (44 genera in total). (B and C) Results of generalized linear models with Poisson error distributions on a scatter plot (B) and corresponding map (C) investigating the spatial relationship between present-day genus richness (A) and NPP (adjusted for human appropriation; Fig. S6D) across different sets of 1° grid cells: within our focal regions (blue) and across the globe (gray), where we used the full NPP gradient or the fossil gradient only (darker shading, i.e., cells where the present-day NPP values lie within the limits of the NPP gradient estimated from the Neogene paleobotanical data, shown by dashed vertical lines in B). Symbols in B are semitransparent to show overlapping data points. Generalized linear models with Poisson errors showed significant and strong positive spatial relationships across all grid cells globally (continuous gray line in B: n = 12,730, residual deviance = 46,883, slope estimate = 0.0006, z = 169.5, P < 0.001) and across all cells falling within the fossil NPP gradient (black line in B: n = 4,531, residual deviance = 26,289, slope estimate = 0.0005, z = 75.1, P < 0.001). In contrast, an equivalent model across only the grid cells within our focal regions (red line in B) showed a significant but much weaker relationship (n = 1,460, residual deviance = 1,118, slope estimate < 0.0001, z = 3.2, P < 0.01). All these models were refitted accounting for spatial autocorrelation in the residuals; the results of these models were virtually identical to the models presented here (SI Materials and Methods).