Eight hundred-five blood samples obtained from women throughout pregnancy and at delivery were examined for fetal erythrocytes by the Kleihauer technic. The incidence of positive smears postpartum was 33.5%, considerably higher than the 11% incidence observed in antepartum samples. If a correction is made to compensate for the apparent 10% incidence of false positive smears, as was observed in blood samples from 100 healthy adult males, the difference in the frequency of antepartum and postpartum transplacental hemorrhages becomes even more pronounced. Large fetal-maternal hemorrhages (> 0.4 cc) were 9 times as frequent in postpartum samples. These findings indicate that most sensitizing fetal-maternal hemorrhages occur at delivery and that postpartum injection of anti-Rh gamma globulin to destroy fetal cells and suppress Rh isoimraunization is a rational approach. The incidence of detectable postpartum transplacental hemorrhages was increased when the ABO grouping of the fetal erythrocytes was compatible with the maternal serum. No large fetal-maternal hemorrhages were observed in the absence of ABO compatibility. Manual separation of the placenta from the uterine wall in vaginal deliveries was associated with an increased frequency ol detectable transplacental hemorrhages.