The hypothesis that maternal alcohol consumption affects the development of structures possibly derived from a common embryonic cell population, the cranial neural crest, was explored using data collected by a case-control surveillance program of birth defects in greater Boston, Philadelphia, and Toronto, Ontario, Canada, and in five counties in Iowa from 1983 through 1987. Maximum and average alcohol consumption during the first four lunar months of pregnancy were compared between the mothers of 1,464 infants with malformations of the ear, face, anterior neck, and upper heart (cranial neural crest cases) and 1,427 infants with other malformations (controls). For maximum number of drinks in a day and average number of drinking days per week, relative risks approximated unity across levels of exposure. For average number of drinks per drinking day, the relative risk (and 95% confidence interval) for heavy intake (5 or more drinks) was 1.8 (0.8-4.4). When the largest defect subgroup of cases, infants with cleft lip with or without cleft palate, was considered separately, the relative risk for an average of 5 or more drinks per drinking day was 3.0 (1.1-8.5). These findings suggest that maternal alcohol use is less related to overall malformations derived from cranial neural crest cell than to one specific defect among them--cleft lip with or without cleft palate. However, the latter association has not been reported previously in humans and remains to be confirmed in other studies.