Predictable trait variation across environments suggests shared adaptive responses via repeated genetic evolution, phenotypic plasticity, or both. Matching of trait-environment associations at phylogenetic and individual scales implies consistency between these processes. Alternatively, mismatch implies that evolutionary divergence has changed the rules of trait-environment covariation. Here we tested whether species adaptation alters elevational variation in blood traits. We measured blood for 1,217 Andean hummingbirds of 77 species across a 4,600 m elevational gradient. Unexpectedly, elevational variation in hemoglobin concentration ([Hb]) was scale independent, suggesting that physics of gas exchange, rather than species differences, determine responses to changing oxygen pressure. However, mechanisms of [Hb] adjustment did show signals of species adaptation: Species at either low or high elevations adjusted cell size, whereas species at mid-elevations adjusted cell number. This elevational variation in red blood cell number-versus-size suggests that genetic adaptation to high altitude has changed how these traits respond to shifts in oxygen availability.

We spent 2006–2020 collecting specimen-vouchered hematological data from 1,217 wild hummingbirds of 77 species, representing all 9 clades, and spanning ~4,600 meters in elevation. We collected data on six blood parameters known to affect oxygen-carrying functions: three primary indices (Hemoglobin cocentration ([Hb]), Hematocric (Hct), and total red blood cell count (TRBC)), which we used these to calculate three secondary indices (mean cell volume (MCV), mean cell hemoglobin (MCH), and mean cell hemoglobin concentration (MCHC)). We compiled species elevational ranges using published data, field guides, and MSB specimen records. We characterized latitudinal and elevational climatic variation using WorldClim data. Data were processed and filtered in R using described methods (see main paper text) and the following scripts and workflows available on GitHub (https://github.com/jlwilliamson/ComparativeHummingbirdBlood). All data are linked to vouchered specimens housed at the Museum of Southwestern Biology (MSB) at the University of New Mexico, the Centro de Ornitología y Biodiversidad in Peru, and the Pontificia Universidad Católica de Chile in Chile. Specimen records are accessible in the Artcos database (https://www.arctosdb.org).

See README.txt file on GitHub (https://github.com/jlwilliamson/ComparativeHummingbirdBlood) for description of all files (scripts and analyses) included on the GitHub repository for reproducibility and extension of this work. All files were processed and analyzed using R.