A selective prevention design was applied to 238 recently separated families. Of these, 153 mothers randomly assigned to the experimental (E) group participated in 14 group sessions focused on Parent Management Treatment (PMT). Prior analyses showed that, over time, the group of families in the untreated group deteriorated in both parenting practices and in child outcomes. In keeping with the classic prevention pattern, families in the E group showed modest improvements in parenting and in child outcomes. Improvements in parenting were associated with significant reductions in problem behavior. The data showed that those mothers who improved their parenting skills during the first 12 months also showed significant reductions in maternal depression during that same interval. A cross-lagged panel analysis showed that a reduction in maternal depression during the first year of the study was a significant predictor of maintenance or improvements over the next 18 months. The findings are consistent with the concept of the family as a system.