Multilocus DNA fingerprinting with oligonucleotide probes (GTG)5, (GATA)4, and (CA)8 was applied in order to determine paternity in one birth cohort (15 infants) of social group (S) from the free-ranging colony of rhesus macaques (Macaca mulatta) on Cayo Santiago. While sires could be identified in 11 cases, all males tested (N = 19) could be excluded from paternity for the remaining four infants. Data revealed marked discrepancies between actual paternity and paternity as inferred from the observation of copulation behavior. Thus, a dominant social rank does not appear to be strongly associated with reproductive success. Furthermore, alternative reproductive strategies were found to yield comparable net benefits in reproduction. A second group of animals (M) was translocated from Cayo Santiago to the Sabana Seca Field Station in 1984. They have continuously resided together in a large outdoor enclosure since then. Here paternity assessment was seriously impeded by a reduced number of discriminating bands, i.e. offspring bands which were unequivocally derived from the sires. This was initially held to be indicative of a smaller degree of heterozygosity in Group M, and was attributed to inbreeding due to a lack of male immigration or extra-group fertilizations. However, a comparison of the DNA fingerprint patterns obtained in Group S and Group M lends only partial support to this idea.