Recent reports indicate a relation between prenatal tobacco exposure (PTE) and offspring smoking. Many of these reports have been retrospective or have not included important variables such as other prenatal substance exposures, maternal and child psycho-social characteristics, mother's current smoking, and friends' smoking. No prior study has examined the timing of PTE. In this prospective study of a birth cohort of 567 14-year-olds, we examined the relation between trimester-specific PTE, offspring smoking, and other correlates of adolescent smoking. Average age of the adolescents was 14.8 years (range: 13.9-16.6 years), 51% were female, 54% were African-American. Data on maternal tobacco and other substance use were collected both prenatally and postnatally, 51% of the mothers were prenatal smokers and 53% smoked when their children were 14 years. PTE in the third trimester significantly predicted offspring smoking (ever/never, smoking level, age of onset) when demographic and other prenatal substances were included in the analyses. PTE remained a significant predictor of the level of adolescent smoking when maternal and child psychological characteristics were added to the model. When more proximal measures of the child's smoking were included in the model, including mother's current smoking and friends' smoking, PTE was no longer significant. Significant predictors of adolescent smoking at age 14 were female gender, Caucasian race, child externalizing behavior, maternal anxiety, and child depressive symptoms. Although direct effects of PTE on offspring smoking behavior have previously been reported from this study and by others, by early-adolescence, this association is not significant after controlling for the more proximal covariates of adolescent smoking such as mother's current smoking and peer smoking.