Assessing variability in affiliative maintenance behaviours in captive coppery titi monkeys, Plecturocebus cupreus

Abstract

Listen

Translate article

Coppery titi monkeys maintain pair bonds through affiliative behaviours (e.g. tail twining, maintaining social proximity). Findings from previous studies suggest that the amount of time pairs spend in affiliative contact is not stable over time within pairs and differs across pairs. The present study investigates several predictors (temperament, age at pairing, pregnancy status and presence of infant) that may help explain observed changes in affiliation over time and between pairs. We hypothesized that affiliation varies for individual dyads and that time-varying and dyad level predictors explain variability in affiliation. We analysed pair affiliation using scan samples collected five times per day for 1 year for 29 pairs. We used Bayesian nonlinear mixed-effects effects modelling to test the effects of predictors. Pregnancy predicted higher affiliation levels (ϒ 02 = 0.40, 95% highest posterior density interval (HPDI): [0.21, 0.58]). The presence of an infant and aggression did not predict variability in affiliation. Models would not converge when we included age at pairing, so we cannot draw conclusions about the role of age in affiliation. Findings from this study suggest that time-varying predictors like pregnancy status help explain the large degree of variability observed in affiliative interactions between partners in this species. It is possible that hormonal changes associated with pregnancy are driving changes in affiliative behaviours in pairs. An alternative explanation is that pairs that are more affiliative to begin with are more likely to get pregnant. Future studies will assess whether these patterns persist in subsequent pairings. • Titi monkey dyads varied in their levels of affiliation. • Pregnancy predicted higher levels of affiliation. • Aggression and infant presence did not significantly explain affiliation patterns. • Effect of age at pairing could not be evaluated using our models.