The social modulation of androgen levels seems to be a widespread phenomenon in vertebrates. It allows the individual to adjust its behavioral output to the context-dependent condition imposed by social constraints. The perception that the individual has of its social environment is needed to activate the androgen response. Animals can use different sensory channels to perceive the key social signals that trigger the response. The transduction of these stimuli into a neuroendocrine response involves the efferent projections from sensory areas of the central nervous system to the preoptic area{plus 45 degree rule}hypothalamus that controls the androgen response through the HPG axis. The development of this response during ontogeny can be primed by early exposure to androgens through maternal effects, and the androgen response to social challenges can be present early in the ontogeny of altricial species. The adaptive value (i.e., function) of having circulating androgen levels open to the influence of the social environment is to allow the individuals to adjust their competitive behavior to the social context according to their relative competitive ability, and this flexibility has advantages over an optimal fixed value of androgen levels, because of the high costs associated with high androgen levels (e.g., mismatched expression of behavior, trade-off with parental care; metabolic costs; immunosupression; survival; etc.). A phylogenetic analysis in birds suggests that the evolutionary scenario in which the social modulation of androgens has evolved was characterized for selective pressures imposed by male-male competition regimes that vary with mating system. In future the social modulation of hormones as an adaptive mechanism to adjust female behavior to social context should also be investigated. So far investigations using the conceptual framework provided by the challenge hypothesis have been mainly limited to androgen responses in males. In females, hormones other than androgens may be relevant in such a role, as is suggested by the less clear relationship between T and agonistic behavior in females (e.g., Floody, 1983) and by the known effects of E2 on female aggression (e.g., Toda et al., 2001). A rise in progesterone has been described in female California mice as a response to a social challenge in a resident-intruder paradigm, while no changes were observed in T levels (Davis and Marler, 2003). These results suggest that in females different hormonal mechanisms may mediate behavioral responses to social challenges. The study of endocrine responses in females to social challenges is therefore a promising avenue for future research in this area. © 2004 Elsevier Inc. All rights reserved.