a b s t r a c t This essay describes recent research that has treated testosterone levels as an individual difference var-iable. We argue that this and similar work broadens our understanding of the person Â situation interaction. Personality psychologists have long been interested in the ways that behavior reflects an interaction of person and situation. More recently, the field has begun to integrate insights from neurobiol-ogy and affective neuroscience (cf. Canli, 2006), in attempt to broaden our understanding of these interactions. This brief essay describes recent research along these lines that has treated testos-terone levels as a personality variable. First and foremost, testosterone meets the criteria for a good personality variable (cf. John & Benet-Martinez, 2000). Levels of testosterone are stable over time (Sellers, Mehl, & Josephs, 2007), demonstrate convergent and discriminant validity (Sellers et al., 2007) and have high predictive validity in specific situations— namely, those in which one's status or dominance is uncertain. In multiple studies, we have demonstrated that high testosterone men and women are motivated to maintain high status, and both-ered when they lose it. When placed in a lower status position, high testosterone individuals become aroused, distracted, and determined to regain status. (Josephs, Sellers, Newman, & Mehta, 2006; Newman, Sellers, & Josephs, 2005). Importantly, no testos-terone differences emerged in control conditions in which status was neither won nor lost. In fact, several authors have suggested that testosterone only relates to behavior when status is threa-tened (Sapolsky, 1991; Wingfield, Ball, Dufty, & Hegner, 1987). Personality psychology has an historical reluctance to measur-ing hormone levels and to treating them as a stable individual difference variable. Few attempts have been made to link testoster-one to self-reports of dominance or status—and those that have find little or no relationship (e.g., Archer, Birring, & Wu, 1998). This may be because most questionnaire measures ask about ''typical" behaviors or traits, whereas a growing body of data suggests that testosterone's relationship to behavior appears to be moderator-dependent; that is, the relationship only appears in certain individ-uals and under certain conditions, and thus oftentimes bears no di-rect, first-order relationship to behavior. Indeed, recent work has demonstrated the importance of con-sidering the interaction of other biologic systems when assessing the link between testosterone and behavior. For example, Mehta and Josephs (2008) and Popma et al. (2007) showed that when sta-tus is threatened, high testosterone men became dominant and aggressive, but only if they also possessed low circulating levels of cortisol. Interactions between testosterone and genetic poly-morphisms have proven important as well. Sjoberg et al. (2008) showed a link between high testosterone and anti-social personal-ity and lifetime aggression, but only among men who showed the low activity MAO-A genotype. A growing body of data suggests that testosterone is an impor-tant player in a complex system of biologic and situation variables that is just starting to be understood. In a recent paper (Josephs et al., 2006), we reported that the testosterone by situation interac-tion was a much better predictor of dominance behavior than were self-report measures of dominance. This and more recent findings suggest an important role for testosterone in studies of dominance, aggression, and other outcomes. More generally, this research ar-gues for increased integration of biological variables into personal-ity psychology (cf. Canli, 2006).