Social modulation of androgens in male vertebrates: meta-analyses of the challenge hypothesis

The challenge hypothesis (Wingfield et al. 1990, American Naturalist, 136, 829–846) predicts varying androgen responses to mating, breeding or territorial behaviour in avian males. At the interspecific level, the highest androgen responsiveness has been observed in males from monogamous species with paternal incubation, and the lowest in males from promiscuous, nonpaternal species. Studies of a number of vertebrate species have discussed the extension of the challenge hypothesis predictions to nonavian vertebrates, but a general ‘vertebrate consensus’ has not been achieved. For this quantitative review we included data from all vertebrate species available in the literature into several meta-analyses. We distinguished between the effects of androgens on sexual, aggressive and paternal behaviour and the effects of behaviour on androgen levels and compared these effects between taxa, mating systems and types of parental care. We observed large variations between taxa in all data sets. Nevertheless, at the vertebrate level the challenge hypothesis predictions originating from the avian literature were confirmed for the modulation of androgen responsiveness (1) to sexual behaviour by paternal care but not by mating system, and (2) to paternal behaviour by mating system but not the degree of paternal care. In contrast, our results provide (3) no support for the predicted modulation of androgen levels in response to agonistic interactions by mating and parental care system at the vertebrate level. Furthermore, our meta-analyses suggest that the effects of exogenous testosterone on sexual and agonistic behaviour, as a rule of thumb, may be expected to be larger in nonpaternal than in paternal systems.