The relationship between maternal smoking and infant respiratory distress syndrome (RDS) was investigated among 550 premature (36 weeks or less) births delivered at the University of Washington Hospital from 1977 to 1980. Forty-five percent of the mothers were smokers. To avoid bias due to the reduced birth weight of infants of smokers, infants of smokers and nonsmokers were compared within small gestational age categories (two-week intervals) and not by birth weight categories. Infants of mothers who smoked had a reduced incidence of RDS for their gestation compared with infants of nonsmokers. The probability of RDS (adjusted for gestational age and method of delivery) was 25% for the infants of smokers versus 38% for the infants of nonsmokers (odds ratio = 0.55, P = .005), equivalent to approximately a 1.5-week acceleration in lung maturity for infants of smokers. The smoking effect was not explained by demographic differences between smokers and nonsmokers, nor by differences in the incidence of pregnancy complications between the two groups. This study adds support to the theory that adverse pregnancy conditions may lead to an acceleration in pulmonary maturity to allow earlier extrauterine adaptation.