A majority of U.S. children attend some type of child care before entering kindergarten. The quality of child care environment and of teacher-child interactions appear to influence children's development, but little attention has been paid to the influence of child-care peers. Using data from the Early Childhood Longitudinal Study, Birth Cohort, the authors identify patterns of peer group racial-ethnic composition in center-based child care. Bivariate statistics show the relationships among child care quality and racial-ethnic peer-group composition, and a series of regression models assesses the unique contribution of peer-group composition to children's literacy, receptive and expressive language, math skills, and eagerness to learn. The authors find that center-based child care is highly segregated, with children attending care with predominantly same-race-ethnicity peers. This segregation is associated with differences in quality. Hispanic- and white-dominated peer groups have similarly high quality across domains, whereas black-dominated groups have somewhat lower structural quality and caregiver warmth. Other race-ethnicity-dominated groups have the lowest levels of caregiver warmth. Peer segregation is also associated with differences in child development. Children attending Hispanic-dominated care are found to demonstrate significantly worse outcomes on four out of five outcomes. Possible reasons are explored, and implications for social work research, practice, and policy are discussed.