Childhood adversities have been proposed to modify later stress sensitivity and risk of depressive disorder in several ways: by stress sensitization, stress amplification, and stress inoculation. Combining these models, we hypothesized that childhood adversities would increase risk of early, but not later, onsets of depression (Hypothesis 1). In those without an early onset, childhood adversities were hypothesized to predict a relatively low risk of depression in high-stress conditions (Hypothesis 2a) and a relatively high risk of depression in low-stress conditions (Hypothesis 2b), compared to no childhood adversities. These hypotheses were tested in 1,584 participants of the Tracking Adolescents' Individual Lives Survey, a prospective cohort study of adolescents. Childhood adversities were assessed retrospectively at ages 11 and 13.5, using self-reports and parent reports. Lifetime DSM-IV major depressive episodes were assessed at age 19, by means of the Composite International Diagnostic Interview. Stressful life events during adolescence were established using interview-based contextual ratings of personal and network events. The results provided support for all hypotheses, regardless of the informant and timeframe used to assess childhood adversities and regardless of the nature (personal vs. network, dependent vs. independent) of recent stressful events. These findings suggest that age at first onset of depression may be an effective marker to distinguish between various types of reaction patterns.