In 1849, Berthold demonstrated that testicular secretions are necessary for aggressive behavior in roosters. Since then, research on the neuroendocrinology of aggression has been dominated by the paradigm that the brain receives gonadal hormones, primarily testosterone, which modulate relevant neural circuits. While this paradigm has been extremely useful, recent studies reveal important alternatives. For example, most vertebrate species are seasonal breeders, and many species show aggression outside of the breeding season, when gonads are regressed and circulating testosterone levels are typically low. Studies in birds and mammals suggest that an adrenal androgen precursor-dehydroepiandrosterone (DHEA)-may be important for the expression of aggression when gonadal testosterone synthesis is low. Circulating DHEA can be metabolized into active sex steroids within the brain. Another possibility is that the brain can autonomously synthesize sex steroids de novo from cholesterol, thereby uncoupling brain steroid levels from circulating steroid levels. These alternative neuroendocrine mechanisms to provide sex steroids to specific neural circuits may have evolved to avoid the "costs" of high circulating testosterone during particular seasons. Physiological indicators of season (e.g., melatonin) may allow animals to switch from one neuroendocrine mechanism to another across the year. Such mechanisms may be important for the control of aggression in many vertebrate species, including humans.