Both retrospective studies of idiopathic aborters, as well as prospective studies of normal couples, have shown reduced fertility among couples sharing HLA antigens. However, the effects of maternal-fetal histocompatibility on surviving embryos are largely univestigated. We thus prospectively studied 53 healthy, fertile women whose timed pregnancies were verified within 21 days of conception. Maternal-fetal histocompatibility status was determined for HLA-A,-B, and -DR locus antigens. Fetal growth rates were monitored by ultrasound at 8, 12, and 20 weeks gestation. Neonates were weighed, measured (birthlength, chest circumference, head circumference), and examined within 72 h of delivery (116 major and minor anomalies) in standardized fashion by one of two geneticists. Although no significant differences were found between infants compatible and incompatible at the HLA-A or HLA-B locus, significant differences were observed between HLA-DR compatible and incompatible infants for sex ratios (p less than .003) and minor anomaly rates (p less than .05). Although differences in mean birthweights between HLA-DR compatible and incompatible infants were not significant in this sample, HLA-DR compatible infants were on average 200 grams smaller than HLA-DR incompatible infants. We interpret these findings as evidence for selection against histocompatible fetuses throughout gestation, particularly with respect to HLA-DR compatibility. Potential immunologic and genetic mechanisms are discussed.