The variance in fitness across population members can influence major evolutionary processes. In socially monogamous but genetically polygynandrous species, extra-pair paternity (EPP) is widely hypothesized to increase the variance in male fitness compared to that arising given the socially monogamous mating system. This hypothesis has not been definitively tested because comprehensive data describing males’ apparent (social) and realized (genetic) fitness have been lacking. We used 16 years of comprehensive social and genetic paternity data for an entire free-living song sparrow (Melospiza melodia) population to quantify and compare variances in male apparent and realized fitness, and to quantify the contribution of the variances in within-pair reproductive success (WPRS) and extra-pair reproductive success (EPRS) and their covariance to the variance in realized fitness. Overall, EPP increased the variance in male fitness by only 0–27% across different fitness and variance measures. This relatively small effect reflected the presence of socially unpaired males with zero apparent and low realized fitness, small covariance between WPRS and EPRS, and large variance in WPRS that was relatively unaffected by EPP. Therefore, although EPP altered individual males’ contributions to future generations, its impact on population-level parameters such as the opportunity for selection and effective population size was limited.