Ecogeographic rules provide a framework within which to test evolutionary hypotheses of adaptation. Gloger’s rule predicts endothermic animals should have darker colors in warm and rainy climates. This rule also predicts animals should be redder in warm and dry climates, the so-called “complex Gloger’s rule.” Empirical studies frequently demonstrate that animals are darker in cool and wet rather than warm and wet climates. Further, sensory ecology predicts that, to enhance crypsis, animals should be darker in darker light environments. We aimed to disentangle the effects of climate and light environments on plumage brightness and redness in the large Neotropical passerine family Furnariidae. Birds in cooler and rainier climates had darker plumage, even after controlling for habitat type. Birds in darker habitats had darker plumage, even after controlling for climate. The effects of temperature and brightness interact so that the negative effect of precipitation on brightness is strongest in cool temperatures. Finally, birds tended to be redder in warm and dry habitats but also, surprisingly, in cool and wet locales. We suggest Gloger’s rule results from complementary selective pressures arising from myriad ecological factors, including crypsis, thermoregulation, parasite deterrence and resistance to feather abrasion.

We gathered 23,588 georeferenced locality records (mean = 79.4 records/species, range = 1:786) representing all furnariid species. We obtained the locality records from three general sources: specimens, audio recordings, and observational records. Specimen records were obtained from ORNIS (www.ornisnet.org). Audio records were obtained from Macaulay Library of Natural Sounds (Cornell Lab of Ornithology) and Xeno‐Canto (www.xeno-canto.org). The coordinates of all documented records (both specimens and audio recordings) included in this study were vetted for accuracy using gazetteers. The third group of records came from observational data gathered by the eBird citizen science initiative (May 2013 release; Sullivan et al. 2009), which are extensively vetted by expert review (www.ebird.org). To further ensure accuracy, we applied additional filters to the observational records. For each species, we then thinned all localities so that no two localities occurred within 1 km of each other, the resolution of the climatic data, except for species represented by 10 or fewer localities.

For each locality record, we extracted elevation and 19 bioclimatic variables from the BioClim database of present‐day climatic conditions (Hijmans et al. 2005) and obtained each variable's mean value for all species.