Animal coloration serves many biological functions and must therefore balance potentially competing selective pressures. For example, many animals have camouflage in which coloration matches the visual background that predators scan for prey. However, different colors reflect different amounts of solar radiation and may therefore have thermoregulatory implications as well. In this study, we examined geographic variation in dorsal patterning, coloration, and solar reflectance among Horned Larks (Eremophila alpestris) of the western United States. We found that plumage brightness was positively associated with soil granularity, aridity, and temperature. Plumage redness—both in terms of saturation (i.e., chroma) and hue—was positively associated with soil redness and temperature, while plumage patterning was positively associated with soil granularity. Together, these plumage-environment associations support both background matching and Gloger’s Rule, a widespread ecogeographic pattern in animal coloration. We also constructed thermoregulatory models that estimated cooling benefits provided by solar reflectance profiles of the dorsal plumage of each specimen based on the collection site. We found increased cooling benefits in hotter, more arid environments. Finally, cooling benefits were positively associated with residual brightness, such that individuals that were brighter than expected based on environmental conditions also had higher cooling benefits, suggesting a tradeoff between camouflage and thermoregulation. Together, these data suggest that natural selection has balanced camouflage and thermoregulation in Horned Larks and illustrate how multiple competing evolutionary pressures may interact to shape geographic variation in adaptive phenotypes.

This Dryad repository includes various data types and code to execute analyses associated with the manuscript titled "Plumage balances camouflage and thermoregulation in Horned Larks (Eremophila alpestris)."

The scripts included in this Dryad repository are to be run with R and Python.