Standing variation and the capacity for change: are endocrine phenotypes more variable than other traits?

Abstract

Circulating steroid hormone levels exhibit high variation both within and between individuals, leading some to hypothesize that these phenotypes are more variable than other morphological, physiological, and behavioral traits. This should have profound implications for the evolution of steroid signaling systems, but few studies have examined how endocrine variation compares to that of other traits or differs among populations. Here we provide such an analysis by first exploring how variation in three measures of corticosterone (CORT)— baseline, stress-induced, and post-dexamethasone injection— compares to key traits characterizing morphology (wing length, mass), physiology (reactive oxygen metabolite concentration [d-ROMs] and antioxidant capacity), and behavior (provisioning rate) in two populations of tree swallow (Tachycineta bicolor). After controlling for measurement precision and within-individual variation, we found that only post-dex CORT was more variable than all other traits. Both baseline and stress-induced CORT exhibit higher variation than antioxidant capacity and provisioning rate, but not oxidative metabolite levels or wing length. Variation in post-dex CORT and d-ROMs was also elevated in the higher-latitude population in that inhabits a less predictable environment. We next studied how these patterns might play out on a macroevolutionary scale, assessing patterns of variation in baseline testosterone (T) and multiple non-endocrine traits (body length, mass, social display rate, and locomotion rate) across 17 species of Anolis lizards. At the macroevolutionary level, we found that circulating T levels and the rate of social display output are higher than other behavioral and morphological traits. Altogether, our results support the idea that within-population variability in steroid levels is substantial, but not exceptionally higher than many other traits that define animal phenotypes. As such, circulating steroid levels in free-living animals should be considered traits that exhibit similar levels of variability from individual to individual in a population.