Drawing from transactional models, the authors examined whether attachment security measured at age 3 (a potential source of differential vulnerability) interacts with the course of maternal depressive symptoms over an 8-year period (a potential source of differential exposure) in predicting children's self-reported depressive symptoms at age 11. Participants were from the NICHD Study of Early Child Care (N = 938). Results from growth curve modeling and analysis of covariance suggest that preschool attachment quality moderates the influence of subsequent maternal depression on children. In particular, variability in the course of maternal depressive symptoms predicted offspring depressive symptoms only among those children with an insecure attachment history. A potential protective effect of early attachment security was evident among children exposed to the most chronic levels of maternal depression. Of the children with different patterns of insecure attachments, those with behaviors classified as disorganized appeared most vulnerable to also becoming depressed if paired with a mother experiencing ongoing depressive symptoms.