This replication study revisited conclusions from 2 previous investigations (Gauze, Bukowski, Aquan-Assee, & Sippola, 1996; van Aken & Asendorpf, 1997), which suggested that support from friends buffers against diminished self-esteem arising from poor quality relationships with mothers during the transition into adolescence. The aim of this replication study was to conduct an independent test of these findings with both concurrent and longitudinal data. Concurrent data for replication analyses were drawn from 4 projects, involving a total of 959 boys and 1,119 girls (ages 10 to 14) from Canada and the United States. Three samples reported participant ethnic descent: Africa (12.1%), Asia (5.6%), Europe (65.3%), Latin America (12.1%), and Native North America (0.9%). Child and mother reports of mother-child relationship quality assayed (a) maternal social support (in 3 data sets), and (b) family cohesion and adaptability (in 2 data sets). Main effects were replicated but hypothesized buffering effects were not. Maternal social support and friend social support were independently associated with adolescent self-esteem, concurrently, but not longitudinally. Family cohesion (but not adaptability) was associated with adolescent self-esteem, concurrently and longitudinally. Friend social support did not moderate associations between mother-child relationship quality and adolescent self-esteem, concurrently or longitudinally. The findings are consistent with a cumulative effects model wherein friends uniquely contribute to adolescent self-worth, over and above the contribution of mothers. The findings do not support claims that friends moderate associations between mother-child relationship quality and adolescent self-esteem.