In group-living species, a dominant male’s ability to monopolize reproduction, and the cost of doing so, are expected to vary with a group’s gender composition. We used spawning observations of a group-living cichlid, Neolamprologus pulcher, to test this expectation. We constructed groups that contained a dominant breeding pair and either two male subordinates, one male and one female subordinate or two female subordinates. Parasitic spawning by male subordinates was more common in groups with two male subordinates than in groups with one male and one female subordinate. Female subordinates were never observed laying eggs in dominant females’ clutches, but three female subordinates laid independent clutches. During spawning, frequencies of dominant male aggression towards male and female subordinates were similar. Dominant males were less aggressive during non-reproductive periods. The declines were greater for female subordinates, such that, during non-reproductive periods, dominant males were more aggressive towards male subordinates. Aggression towards each subordinate was also affected by the second subordinate’s gender; the direction of that effect differed for large and small subordinates. Male subordinates approached breeding shelters less often than female subordinates, and both male and female subordinates approached shelters more frequently when the second subordinate was male. Collectively, these patterns suggest: (1) that male subordinates impose higher costs on dominant males than female subordinates do and (2) that the presence of a second male subordinate imposes additional costs beyond those of the first male subordinate. We discuss the implications of these effects for dominant and subordinate group members.