Affiliation is often used as an intervening variable in behavioral studies of nonhuman primates. Variables used to measure affiliation should be strongly correlated if it is a valid intervening variable. Social context also should not strongly influence relationships between variables used to measure affiliation. Few studies have, however, reported either the correlations between variables used to assess affiliation or the influence of social context on relationships between the variables. Correlations between affiliative variables were therefore calculated on data from two groups of rhesus {Macaca mulatta)and influences of social context on these correlations were assessed. Affiliation was measured with 7 variables. Two methods were used to investigate the influence of social context: Analyses were made of interactions between several age/sex categories of individuals. Comparisons were made between an experimental group and a matched control group. There were higher rates of sexual behavior in the experimental group. The mature males in the experimental group were vasectomized so females did not conceive. In this group females had repeated nonpregnant estrous cycles. Males were intact in the control group. In this group the mature females conceived and were pregnant during data collection. The variables were significantly correlated across all social contexts. Affiliation may therefore be a useful intervening variable. The magnitude of the correlations between variables did vary considerably across social contexts. Sometimes the sign of the correlations between measures changed as a function of social context. Analyses of individual variables and their interrelationships may therefore be necessary for detailed understanding of the meaning of affiliative interactions in nonhuman primates.