Infants from low socioeconomic status families were randomly assigned to an instructional day-care program designed to prevent socioculturally caused mental retardation or to a nonintervention control group. This assignment procedure resulted in an equal distribution of full-term, full-birth-weight, fetally malnourished babies in 2 environments varying in intellectually supportive characteristics. The condition of fetal malnourishment was defined by infants having low ponderal indices (PI). At 3 months of age low-PI infants showed lower Bayley Mental Development Index (MDI) scores than normal-PI infants, independent of the environment. In the control group low-PI infants still had lower MDI scores than normal-PI infants at 18 months of age. However, at that time in the day-care group, low-PI infants scored as well as normal-PI infants. These findings were replicated when the infants were 24 months of age with Stanford-Binet intelligence tests. Observations of mothers' involvement with their infants showed that, although all groups had similar amounts of maternal involvement when the babies were 6 months of age, the mothers of low-PI infants in the control group showed less involvement with their infants at 18 months of age than the other mothers. We suggest that this longitudinal study provides experimental evidence for a transactional model of development which emphasizes both newborn infant characteristics and environmental quality as cocontributors to the process of development.