Robustness of Bergmann’s and Rapoport’s rules to different geographical range estimates in New World pit vipers

Aim Bergmann's and Rapoport's rules are two ecogeographical patterns that predict an increase in species body size and range size, respectively, with increasing latitudes. Here, we evaluated whether using different geographical range estimates affects the assessment of Bergmann's and Rapoport's rules at the cross‐species and assemblage levels in New World pit vipers. Location New World. Taxon New World pit vipers (Crotalinae). Methods We estimated geographical ranges for 135 (i.e. ≈93%) Crotalinae species using ecological niche modelling (ENM) and expert maps and assessed alternative hypotheses that could account for Bergmann's and Rapoport's rules. We evaluated both rules using ‘cross‐species’ (i.e. species as units of observation) and ‘assemblage’ (i.e. sites/grid cells as units of observation) approaches. We used phylogenetic generalized least squares (PGLS) analyses and generalized least squares (GLS) analyses to account for phylogenetic and spatial autocorrelation that could affect the evaluation of these rules at the cross‐species and assemblage levels, respectively. Results Expert maps differed significantly from ENM ranges. The cross‐species approach was robust to these different range estimates when Bergmann's and Rapoport's rules were evaluated. The assemblage analyses were more vulnerable to different range estimates, where contrasting results were observed in the assessment of Bergmann's rule but not in the evaluation of Rapoport's rule. Main Conclusions Different geographical range estimates affected mainly the assessment of ecogeographical rules at the assemblage level, whereas the cross‐species analyses were more robust. However, this increased susceptibility of the assemblage approach to different range estimates was only observed when evaluating Bergmann's rule. Thus, our findings suggest that ecogeographical rules can have different levels of sensitivity to different range estimates, which can especially affect the assemblage‐level assessment of these ecogeographical patterns.