The role of testosterone in female fitness is poorly known in most vertebrates. It has been proposed that the dynamics of female testosterone may represent a correlated response to selection acting on male testosterone. However, several costs and benefits of enhancing the circulating levels of testosterone in females have been documented in a number of species, suggesting that female testosterone may be the product of direct selection. Despite this, no study has tried to quantify the long-term fitness consequences of manipulating female testosterone. We report the results of an experiment in which we implanted females with testosterone (T-females) and recorded the consequences of this manipulation on investment and reproductive success throughout their lifetime in relation to control females (C-females). Our results show that T-females bred more years in the colony than C-females. The onset of laying was delayed in T-females compared with C-females, but this effect decreased in the years following manipulation. The T-females laid fewer eggs per year, but eggs were heavier than those laid by C-females. The T-females fed nestlings at a lower rate than C-females in the 3 years in which feeding rates were measured and raised fewer fledglings than C-females in the manipulation year. However, since T-females accomplished more breeding attempts than C-females, as a result of their longer stay in the colony, there were no significant differences between the lifetime reproductive success of T- and C-females. There were no overall differences in body size of nestlings raised by T- and C-females, but the body mass of fledglings reached a peak 2-3 years after manipulation for C-females, a tendency not recorded in offspring of T-females. Also, young that were fledged in broods of C-females tended to recruit into the study colony more frequently than those fledged in T-females' nests. All these results indicate that our testosterone manipulation improved the ability of females to acquire and maintain a breeding site but interfered with their reproductive biology and parental investment. Thus, our main conclusion is that under the ecological conditions of our experiment, the addition of exogenous testosterone reduced female fitness. However, differences in population density and female-female competition are likely to alter the fitness landscape of this trait and change the adaptiveness of increased testosterone as a component of female reproductive strategies.