Teasdale's (1988) differential activation hypothesis proposes that a tendency for negative mood to activate latent negative self-schemas characterises people at risk for depression. The current study tested predictions from this hypothesis in a community sample of 102 adolescents who were free from history of psychiatric illness, and who were subdivided according to level of emotionality, a temperamental style as assessed by parental questionnaire. A musical mood induction task was used to induce temporary mild dysphoria, and the effect of mood induction on self-schemas was assessed. There was no difference between high and low emotionality groups in the liability to sad mood induction. However, adolescents with high emotionality endorsed significantly more negative self-descriptors after dysphoric, but not after neutral, mood induction. This was not accounted for by level of self-reported depressive symptoms over the previous week. This suggests that a " dysphoric mood induction challenge" may provide important information about vulnerability to depression that is not identified by routine self-report of mood or cognitions.