One of the basic tenets of sexual selection is that male reproductive success should be large in polygynous species. Here, we analysed 6 years of molecular genetic data from a semi–free–ranging population of rhesus macaques (Macaca mulatta), using Nonac's B index, to assess the level of male reproductive skew in the study troop. On average, the top sire in each year produced 24% of the infants, while 71% of troop males sired no offspring at all. Consequently, 74% of infants had at least one paternal half–sibling in their own birth cohort. Reproductive success was greatest for high–ranking males, males who spent the whole mating season in the troop and males of 9–11 years of age. Heterozygosity for major histocompatibility complex (MHC) class II gene DQB1 was the strongest single predictor of male reproductive success. A negative relationship suggestive of female mate choice was noted between the B index and the proportion of extra–group paternities. Reproductive skew was not associated with relatedness among potential sires or with female cycle synchrony. We conclude that reproductive skew in male rhesus macaques is best accounted for by the ‘limited–control’ model, with multiple factors interacting to regulate individual reproductive output.