Many contemporary theorists believe serotonin (5-HT) neurotransmitter functioning plays a role in the regulation of human aggressive behavior. We argue that the evidence supporting this 5-HT hypothesis of human aggression is less compelling than commonly assumed, due to (a) conflicting study results, and (b) significant methodological limitations of existing studies. Recent models that integrate the role of psychological and contextual variables in 5-HT--associated aggression are reviewed. The need to incorporate psychometrically sound measures of aggression in 5-HT studies, to use experimental and longitudinal designs, and to test hypotheses drawn from multifactorial models in future research is advocated.