While depressive symptoms in adolescent mothers may affect both their own and their babies' development, little research has focused on the mothers. Self-reported symptoms on the Beck Depression Inventory were collected at 1, 6, 12, and 28–36 months postpartum. Concurrent and reciprocal longitudinal relations among symptom levels, stressful life events, and social supports were investigated. Symptom levels declined over the four assessments, with changes in somatic, rather than cognitive affective, symptoms accounting for the decrease. Stressful life events and all sources of social supports predicted concurrent levels of depressive symptoms, but only social supports predicted declines in symptoms in the first year postpartum. Reciprocally, depressive symptoms tended (p = .06) to predict increases in stressful life events over time. Mothers were also categorized as reporting few (50%), intermittent (27.5%), or chronic (22.5%) symptoms in the first 12 months postpartum. Intermittently and chronically depressed mothers perceived their own mothers as less accepting than nondepressed mothers. Compared to nondepressed and intermittently depressed mothers, chronically depressed mothers also reported more stressful life events, were more likely to live alone, and experienced more moves by 28–36 months postpartum. The reciprocal causal relations among depressive symptoms, stress, and attachments to grandmothers and peers are discussed.