The extent to which species ranges reflect intrinsic physiological tolerances is a major, unsolved question in evolutionary ecology. To date, consensus has been hindered by the limited tractability of experimental approaches across most of the tree of life. Here, we apply a macrophysiological approach to understand how hematological traits related to oxygen transport shape elevational ranges in a tropical biodiversity hotspot. Along Andean elevational gradients, we measured traits that affect blood oxygen-carrying capacity—total and cellular hemoglobin concentration and hematocrit—for 2,355 individuals of 136 bird species. We used these data to evaluate the influence of hematological traits on elevational ranges. First, we asked whether the sensitivity of hematological traits to elevation is predictive of elevational range breadth. Second, we asked whether variance in hematological traits changed as a function of distance to the nearest elevational range limit. We found that the correlation between hematological sensitivity and elevational range breadth was slightly positive, consistent with a facilitative role for sensitivity in elevational range expansion. We further found reduced local variation in hematological traits near elevational range limits and at high elevations, patterns consistent with intensified natural selection, reduced effective population size, or compensatory changes in other cardiorespiratory traits. Our findings suggest that constraints on hematological sensitivity and local genetic adaptation to oxygen availability promote the evolution of the narrow elevational ranges that underpin tropical montane biodiversity.

From 2006 to 2020, we measured total blood hemoglobin concentration ([Hb]) and the percentage of hematocrit (Hct) for birds captured during collaborative fieldwork in Peru by the Museum of Southwestern Biology (MSB) in New Mexico, USA, and the Centro de Ornitología y Biodiversidad (CORBIDI) in Lima, Peru. As quickly as possible after capture, we punctured the brachial vein on the underside of each bird’s wing and collected whole blood using heparinized microcapillary tubes (for Hct) or spectrophotometer cuvettes (for [Hb]). We centrifuged microcapillary tubes for 5 minutes at 13,000 rpm to separate red blood cells from plasma and used digital calipers to quantify the relative volume of each, averaging two samples per bird to arrive at our final estimate of Hct as a percentage. We measured [Hb] (g/dL blood) for a ~5 μL blood sample with the Hemocue HB201 analyzer and associated hemoglobin photometer. As this proprietary method generates values approximately 1.0 g/dL higher than measures from a direct cyanomethaemoglobin spectrophotometer (Simmons & Lill, 2006), we corrected the resulting estimates by subtracting this quantity (DuBay & Witt, 2014). Specimens and tissues are housed at the MSB and CORBIDI and specimen data are detailed in a separate data publication (Witt et al., in prep).

Raw data are stored as comma-separated values (.csv) files. All data processing, analyses, and figures can be produced by sequentially running three R Markdown (.Rmd) notebooks found in the associated Zenodo repository (this procedure is documented in the README).