We evaluated a potential dose relationship between in utero cocaine exposure and scores on the BSID controlling for time independent and dependent covariates. At 6, 12, & 24 months, masked examiners assessed 201 (88% African American) infants (44% not cocaine exposed, 37% lightly, 19% heavily(top quartile) exposed by maternal report/infant meconum assay). The final mixed linear regression model included prenatal exposure to cocaine, marijuana, alcohol, and cigarettes, birthweight (BWT) z score, age at test, caretaker (biologic mother, kinship, foster) and early developmental intervention. Older age at testing was associated with lower MDI & PDI scores (p=.002). For MDI there was a main effect for caretaker (kinship lower, p=.02), but not for cocaine. We found a significant interaction between prenatal cocaine dose & early developmental intervention (p=.04). Heavily cocaine exposed children who received developmental intervention showed mean MDI scores 8.6 points higher than those who did not. There was no main cocaine effect on PDI, but interaction effects such that cocaine exposed children with lower BWT showed lower PDI scores (p=.05). There were also interaction effects for intervention/age, (p=.03) and intervention/caretaker (p=.02). Contrary to earlier predictions, heavily cocaine exposed children respond positively to developmental intervention.