A test of the 'challenge hypothesis' in cichlid fish: Simulated partner and territory intruder experiments

In male birds, the responsiveness of androgens to sexual and territorial behaviour is predicted to vary with mating system and the degree of paternal investment ('challenge hypothesis', CH; Wingfield et al. 1990, American Naturalist, 136, 829-846). The CH predicts a higher and longer lasting 'breeding baseline' androgen level in males of polygynous species with no or only short-term paternal investment than in males of monogamous species with a high degree of paternal investment. Since the applicability of the CH to nonavian vertebrates has been unclear, we experimentally tested its predictions in several cichlid fish (Neolamprologus pulcher, Lamprologus callipterus, Tropheus moorii, Pseudosimochromis curvifrons and Oreochromis mossambicus) using a simulated territorial intruder protocol. Androgens (11-ketotestosterone: 11-KT; testosterone: T) were measured from fish-holding water. In all species sampled, the 11-KT patterns confirmed the predictions of the CH originating from the avian literature, but T patterns did not. Males of all species sampled were highly responsive to territorial intrusions; however, the magnitude and duration of this response, that is, the rapid return to baseline 11-KT levels, could not clearly be explained by the degree of paternal care. 11-KT responses to interactions with ovulating females were observed in maternal mouthbrooders but not in biparental species (e.g. Lamprologini). At the interspecific level, androgen responsiveness was greater among males of monogamous species, as predicted, but also in species with more intense pair bonding (e.g. Tropheus moorii). Thus, this study confirms the predictions of the CH in cichlid fish at both the intraspecific and the interspecific levels.