Benefits of pair housing are consistent across a diverse population of rhesus macaques

Introducing singly housed rhesus macaques (Macaca mulatta) into isosexual pairs is widely considered to improve welfare. The population of laboratory rhesus macaques is heterogeneous on a variety of factors and there is little literature available to directly evaluate the influence of many of these factors on the benefits of pair housing. Subjects were 46 adult female and 18 adult male rhesus macaques housed at the Tulane National Primate Research Center and the Yerkes National Primate Research Center. Behavioral data totalling 859 h and serum cortisol levels derived from 312 serum samples were analyzed for main effects of housing condition, comparing single housing to pair housing. In addition, a series of analyses were performed to test for interactions between housing condition and seven independent variables: sex, age, prior duration of single housing, presence or absence of a history of self-injurious behavior, and dominance rank, levels of affiliation and agonism in the paired setting. After the collection of 4-8 h of baseline data and three serum cortisol samples, pairs of individuals were introduced to one another and data collection was repeated, no earlier than 4 weeks after introduction. In pair housing both female and male subjects showed decrease in abnormal behavior (females: 54% reduction; P = 0.001; males: 18% reduction; P = 0.0007) and anxiety-related behavior (females: 35% reduction; P = 0.0001; males: 41% reduction; P = 0.0001), and increases in locomotion (females: 41% increase; P = 0.0001; males: 76% increase: P = 0.002). In pair housing, there were no significant sex differences in social behavior. Descriptively, paired females spent 12% of samples engaged in affiliative behavior and 0.5% engaged in agonistic behavior (back-transformed arcsin square root means). The corresponding values for males were 12% and 0.3%. No interaction effects were detected with any of the independent variables tested in this study. Cortisol values varied with sex but did not differ between housing conditions; no differences were detected when any of the above variables were included in the statistical model. Results support the general consensus among those studying the welfare of captive primates that social housing is a potent means for promoting behavioral indicators of the psychological well-being of laboratory primates. These results are of considerable practical significance and include information that refutes common perceptions about the unsuitability of males as socialization candidates, perceived negative.