The effect of maternal smoking during pregnancy on birth weight was studied in 12,068 births, the mothers in 1819 of which were regarded as smokers. The children of the smokers were compared with those of controls of similar age, parity, marital status and place of residence. Maternal smoking reduced birth weight in a dose-related manner. Birth weight was least affected among young, primiparous mothers who smoked only slightly, a difference which was, however, entirely caused by those mothers who stopped smoking for the last 3 mth of pregnancy, and this subgroup showed similar figures for postneonatal mortality and morbidity up to the age of 5 to those of their controls. Postneonatal mortality was higher in the total group of the smokers than among their controls, the primiparous or young mothers not differing in this respect from the others. Morbidity up to the age of 5 was significantly higher (p less than 0.001) among the children of the smokers, the children of the primiparas and young mothers being affected in a similar way to the others. The low birth weight infants of the smokers had a higher mean birth weight and lower perinatal mortality than the low birth weight infants of the controls, but morbidity up to age of 5 and postneonatal mortality were higher among the smokers in respect of both low birth weight infants and others.