Infants born to mothers who smoke are a few hundred grams smaller, on average, than the infants of nonsmokers. This effect on fetal growth is regarded as evidence of the reproductive toxicity of cigarette smoking. In this paper, data from nearly 260,000 births in the state of Missouri (1980-1984) were analyzed using a method based on adjustment to relative birth weight. Two additional effects of smoking are demonstrated with this analysis; i.e., smokers are at higher risk of delivering very small preterm infants, and their infants have higher perinatal mortality at every relative birth weight. The latter is not apparent on an absolute birth weight scale and thus is not generally recognized. A supplementary analysis of births at high altitude is carried out to suggest that effects on fetal growth (whether from smoking or other factors) can occur independently of effects on mortality.