Variance in male mating success and female oviposition rate was studied on a lifetime and daily scale in a nonterritorial libellulid dragonfly. While controlling for weather, population density, and sex ratio, we investigated whether phenotypic variation could explain a significant amount of the observed variation. By collecting most of the data in a large outdoor field enclosure, we had access to (1) real lifetime data as emigration was prevented, (2) exact values of density and sex ratio, and (3) phenotypic data that are virtually unavailable under natural conditions. Variance in LMS was higher in males than variance in lifetime number of ovipositions in females. Lifespan was of major importance in explaining the variation. Males readily remated on the same day whenever possible (maximum 6 times), whereas in females each oviposition bout usually was followed by a nonreproductive period of approximately 4 days. An analysis of the sources of variance in daily reproduction rate showed that, besides the expected effects of weather, an excess of one sex resulted in a negative effect on the individual mating probability of members of that sex. Even after correction for weather and population parameters, phenotypic traits explained a significant amount of the remaining variation: males with a high mating rate were large, agonistic, active mate searching, and long-living individuals. Females were more likely to oviposit when they showed higher flight activity and when more days had elapsed since the previous oviposition. The presence of a strong size effect on male mating probability, and its absence in females, was confirmed at a pond (native habitat). We propose a model on the causal relationships between the relevant parameters and the variation in reproductive success. Finally, we briefly discuss why variance in reproductive success is not a very good measure of the opportunity for sexual selection in this species.