In a previous study, we demonstrated the importance of social interactions in the development of non-human primate infants. These results confirmed the social network concept. According to this concept, it is assumed that socialization processes would differ in various social environments. However, much variability remained to be explained. In the present study we investigated especially the influence of infants' gender on socialization processes. In relation to the previous results, the influence was tested within groups having the same social structure. At the Wisconsin Regional Primate Research Center, four mother-peer groups of rhesus macaques (Macaca mulatta) were studied. Therefore no adult male model was available to infants. Twenty infants, including eight males, eight females and four prenatally DES-treated females, were the subjects of this study. We considered six comprehensive developmental parameters to account for the two main socialization processses, the acquisition of the social behavioral repertoire and the establishment of a network of social relationships. We analyzed the variability of these parameters using a new multivariate technique previously described in a companion paper. This new technique is derived from the Principal Components Analysis and Multivariate Analyses of Variance. The model used in this analysis included two intrinsic features of infants, sex and age and two variables related to the social organization of the groups, mother's rank and social group. This 4-variable-model significantly accounted for most of the variability of the developmental parameters. The results showed that the social group and the infant's sex were the most influential variables when infants initiated interactions: male and DES female infants displayed a greater behavioral diversity than non-treated female infants. When infants were recipients in interactions, social organization variables (social group and mother's rank, respectively) had the greatest influence on the socialization processes while infants' gender was less influential. In this case the effect of sex was confounded with the social group effect; male infants then differed from both kind of female infants in being looked at more frequently, in being contacted by a larger variety of partners and in receiving a greater variety of social behaviors. This new example of applying multivariate methods to the study of individual social development confirms the potential of this method to provide new insights into behavioral development.