Associations between social behaviour and adrenal activity in female Barbary macaques: Consequences of study design

Faecal glucocorticoid metabolite (fGCM) concentrations have been used to evaluate adrenal activity in a variety of species; including as an indicator of the physiological response to social stress. However, across studies, the relationships between dominance rank, social behaviours and adrenal responses can be inconsistent. Differences in the relationship between rank and glucocorticoids may be due to the relative costs of social status, and the relative frequencies of social stressors and potential coping mechanisms. However, the differences in observed relationships between specific social behaviours and glucocorticoids may be partly explained by sampling frequency, as studies often use average fGCM concentrations collected over a period of weeks or months, rather than fGCM concentrations that are temporally-matched with behavioural data. In this study, we directly compared long-term average and temporally-matched data to determine whether particular social behaviours were related to adrenal activity in female Barbary macaques (Macaca sylvanus) at Trentham Monkey Forest, UK; and whether observed relationships were consistent using these two approaches. Average rates of autogrooming were positively correlated with average fGCM; however, this relationship was not robust in temporally-matched samples. Instead, specific social behaviours associated with agonism were associated with fGCM in temporally-matched samples within individuals. These results indicate that analyses of relationships using long-term average fGCM and temporally-matched samples do not necessarily provide comparable results, highlighting that study design is critical in determining associations between an individual's social behaviour and the relative physiological costs involved.

► Faecal glucocorticoids (fGCM) can be used to assess adrenal activity in Barbary macaques. ► fGCM were elevated 1 day following biological validation. ► Agonistic behaviours were associated with time-matched fGCM the following day. ► Average and time-matched fGCM provided inconsistent results with behavioural data.