One hundred and seventy–two low–income women were interviewed prenatally to determine risk for later mistreatment of their children. High–risk and low–risk women were then randomly assigned at delivery to either limited or extended postpartum contact with their newborns over the first two days after birth. Mother–infant interaction observations were performed at 48 hours and at one, three, six, twelve and eighteen months postpartum. Infants were tested at nine months with the Bayley Scales of Infant Development. Results indicated that outcome following extended mother-infant postpartum contact varies with maternal risk status and measures employed for evaluation. Low–risk extended–contact mother–infant pairs differed from low–risk controls in observed interaction while high–risk extended–contact and controls did not differ from each other in interaction. High–risk extended–contact infants were more advanced in motor development than control infants at nine months, however, while low–risk extended contact and control infants did not differ in development.