Alcohol consumption by pregnant women and birth outcome were studied in 9953 livebirths, 3309 fetal deaths and 5332 infant deaths from the 1988 National Maternal and Infant Health Survey. In crude analyses, race, age, mother's education, prenatal care, parity, low birthweight, gestational age, smoking and alcohol consumption during pregnancy were significantly related to the occurrence of fetal deaths and infant deaths. Among women having livebirths, race, age, mother's education, prenatal care, prematurity, gestational age, smoking and alcohol consumption during pregnancy were significantly related to having a low birthweight baby (< 2500 g). In this group, women who drank more during pregnancy also smoked more, were younger and less educated than women who drank at lower levels or not at all. The relationship of alcohol consumption during pregnancy and infant birthweight for those women having livebirths was studied using multivariable linear regression. The results indicated that race, mother's education, baby's sex, parity, mother's height, mother's body mass index and smoking, but not alcohol consumption, were significantly related to birthweight. Multivariable logistic regressions were performed for the occurrence of low birthweight, fetal death and infant death. The effect of alcohol was significant in all these analyses. These results indicate that alcohol has an important relationship with birth outcome, but that for the drinking reported in this study, the alcohol effect on mean birthweight is small relative to that of other risk factors, accounting for the non-significant result in the multiple linear regression.