To investigate the effects on infants of the use of marijuana and cocaine during pregnancy and to compare the importance of urine assays with that of interviews in ascertaining drug use, we prospectively studied 1226 mothers, recruited from a general prenatal clinic, and their infants. On the basis of either interviews or urine assays conducted prenatally or post partum, 27 percent of the subjects had used marijuana during pregnancy and 18 percent had used cocaine. When only positive urine assays were considered, the corresponding values were 16 percent and 9 percent, respectively. When potentially confounding variables were controlled for in the analysis, the infants whose mothers had positive urine assays for marijuana, as compared with the infants whose mothers were negative according to both interviews and urine assays, had a 79-g decrease in birth weight (P = 0.04) and a 0.5-cm decrement in length (P = 0.02). Women who had positive assays for cocaine, as compared with nonusers, had infants with a 93-g decrease in birth weight (P = 0.07), a 0.7-cm decrement in length (P = 0.01), and a 0.43-cm-smaller head circumference (P = 0.01). To compare our findings with those of other investigators who did not use urine assays, we repeated the analyses, considering only self-reported use of marijuana (23 percent) and cocaine (13 percent). There were no significant associations between such use as determined by interviews alone and any of the measures of outcome. We conclude that the use of marijuana or cocaine during pregnancy is associated with impaired fetal growth and that measuring a biologic marker of such use is important to demonstrate the association.