Abstract
Aim Interspecific patterns of Bergmann's rule may be concealed by species extinctions within clades. We evaluate whether incorporating extinct species and paleoenvironmental information improves the detection of Bergmann's rule in the family Canidae, and test whether temperature and net primary productivity (NPP) explain interspecific variation in canids' body mass. Location Global. Time Period Neogene to present. Taxon Canidae. Methods We analysed 81 canid species, including 36 extant and 45 extinct species, using an interspecific approach. We combined body size and occurrence records for both extant and fossil species. We linked the records to the environmental conditions—mean annual temperature and net primary productivity (NPP)–at their corresponding geological ages through a multitemporal calibration. We assessed relationships of body mass with temperature and NPP using phylogenetic generalized least squares (PGLS) regressions. Results With extant species only, body mass was unrelated to temperature. In contrast, the inclusion of extinct species substantially increased body mass variation and revealed a significantly negative relationship between temperature and body mass. Body mass decreases by approximately 2.8% for each 1°C increase in temperature. No significant relationship was detected between body mass and NPP. Main Conclusions Bergmann's rule in canids is valid, but it is only detectable when fossils and paleoenvironmental data are incorporated. Disregarding extinct species may obscure the detection of interspecific ecogeographical patterns. Forwarding the integration of paleontological and neontological data helps improve understanding of current macroecological patterns.