Children in late elementary and middle school tend to form friendships with same-race peers. Yet, given the potential benefits of cross-race friendships, it is important to understand the individual and contextual factors that increase the likelihood of cross-race friendship over time. Guided by contact hypothesis and systems theory, we examine the student and classroom predictors of change in same-race friendships over 1 school year using a sample of 553 African American and European American students in 53 classrooms. Results suggest that same-race friendships increase over time, with greater increases among European American and older children. Youth externalizing behavior predicted a greater increase in same-race friendships; classroom support predicted less of an increase in same-race friendships from fall to spring. Lastly, African American students in classrooms with greater differential teacher treatment were more likely to engage in crossrace friendships over time. Findings are discussed in light of psychological and educational theories and prior research.