In this study of infants with isolated birth defects, 69 cleft palate only cases, 114 cleft lip with or without cleft palate cases, and 284 controls with noncleft birth defects (all born in Maryland between 1984 and 1992) were examined to test for associations among maternal exposures, genetic markers, and oral clefts. A significantly higher frequency of positive family history of birth defects among both groups of oral cleft cases compared with controls was seen in these data. While there was a modest increase in the less common C2 allele at the TaqI site in the transforming growth factor alpha (TGF alpha) locus among cleft palate only infants compared with the birth defect controls, the association appeared to reflect an underlying interaction between maternal smoking and infant genotype. This apparent gene-environment interaction was also found among those reporting no family history of any birth defect. Infants carrying the rarer C2 allele who were exposed to maternal smoking of 10 or fewer cigarettes per day showed a 6.16-fold increase in risk for cleft palate only (95% confidence interval 1.09-34.7), while similar infants whose mothers smoked more than 10 cigarettes per day showed an 8.69-fold higher risk (95% confidence interval 1.57-47.8). However, the dose-response relation was not significant.