Variation in parental care by child's sex is evident across cultures. Evolutionary theory provides a functional explanation for this phenomenon, predicting that parents will favour specific children if this results in greater fitness payoffs. Here, we explore evidence for sex-biased parental care in a high-fertility, patriarchal and polygynous population in Tanzania, predicting that both mothers and fathers will favour sons in this cultural setting. Our data come from a cross-sectional study in rural northwestern Tanzania, which included surveys with mothers/guardians of 808 children under age 5. We focus on early childhood, a period with high mortality risk which is fundamental in establishing later-life physical and cognitive development. Examining multiple measures of direct/physical care provision (washing, feeding, playing with, supervising, co-sleeping and caring when sick), we demonstrate that fathers favour sons for washing, feeding and supervising, while maternal care is both more intensive and unrelated to child sex. We find no difference in parental care between girls and boys regarding the allocation of material resources and the duration of breastfeeding; or in terms of parental marital and co-residence status. This bias towards sons may result from higher returns to investment for fathers than mothers, and local gender norms about physical care provision.