It is now evident that the genetic mating system can be very different to the observed mating system. However, it is less well known what makes particular individuals more (or less) successful than expected from the observed system. In this study the observed territorial structure of a field population of the agamid lizard, Ctenophorus ornatus, was compared with the mating system as evidenced by microsatellite parentage assignment. This study also investigated whether any male traits predicted reproductive success. Sixty-five per cent of clutches were sired at least partially by a male other than the main territory-holding male and 35% of clutches were sired by a male with no overlap of the female's territory. Multiple paternity was moderately frequent at 25% of clutches. Male chest patch size predicted territory size and the number of females in the territory, but did not predict reproductive success. Instead, male head depth and body size were independently related to the number of offspring sired. As male head depth also predicted the number of females in a territory, these males are likely to be gaining increased reproductive success as a consequence of the higher number of females in their territories. Larger body size males, however, did not have a greater number of females in their territory and instead had more extra-territorial copulations. Whether these extra-territorial copulations are due to female choice or success in male competition is unknown.