The repeatability of glucocorticoids: A review and meta-analysis

Glucocorticoids are highly conserved hormones that mediate a suite of responses to changing conditions in vertebrates. Recent work has focused on understanding how selection operates on glucocorticoid secretion in natural populations. Because heritability is rarely estimated and difficult to measure in the wild, many studies report within-individual repeatability as an estimate of stable between individual differences in glucocorticoid secretion. We conducted a systematic review and meta-analysis on estimates of within-individual glucocorticoid repeatability to elucidate general patterns of repeatability, and to test for relationships between covariates and estimates of repeatability. To this end, we collected 203 estimates of within-individual glucocorticoid repeatability drawn from 71 separate studies and 55 species. Overall, we found moderate levels of repeatability (0.29). We also found that repeatability varied by sample type. Long-term measures (e.g., fecal and feather samples) and acute stress-induced plasma glucocorticoids had higher repeatability (long-term: 0.44, stress-induced: 0.38), than baseline glucocorticoid levels (0.18). Repeatability also decreased with increasing time between repeated sampling events. Despite significant overall repeatability, there was substantial heterogeneity in estimates from different studies, suggesting that repeatability of glucocorticoid secretion varies substantially across systems and conditions. We discuss the implications of our results for understanding selection on glucocorticoid traits and suggest that continuing work should focus on evaluating the repeatability of within-individual glucocorticoid reaction norms.