Discusses the prediction of response to lithium in the treatment of mania and depression, and the prophylactic effects of lithium in recurrent affective disorders. Different studies have pointed out the clinical, biological and psychological factors that can predict the response in bipolar and non-bipolar affective disorders. Issues discussed include diagnosis and clinical features, the frequency and sequence of episodes, early and acute response, comorbidity and biological predictors. It is concluded that the search for predictors of outcome has not been particularly rewarding, although multivariate analysis has shown that variation in prophylactic response is accounted for by 3 factors: diagnosis, quality of symptoms-free interval and the recent frequency of episodes. A trial of lithium is seen as the most powerful predictor of outcome