Several risk factors for sudden infant death syndrome (SIDS) have been consistently reported, while results regarding seasonality and age at death of SIDS victims are conflicting. In the present population-based cohort study, single births in Sweden from 1983 through 1990 were used to estimate the relative and absolute risks for SIDS associated with season at death, age at death, and maternal smoking. In the winter period, 283 SIDS deaths occurred, while only 98 infants died during summer (winter/summer ratio = 2.9). Taking person-time at risk into account and restricting the analysis to infants aged 7-180 days, the authors determined the relative risk for SIDS to be 3.5 times higher in winter than in summer. When comparing incidence rate differences, they found a more noticeable seasonal variation for early SIDS (7-90 days at death) than for late SIDS (91-180 days at death). For early SIDS, the incidence rate was 0.6 cases per 100,000 person-days higher among smokers than among nonsmokers; for late SIDS, the corresponding difference was 0.3. The effect of smoking on SIDS was not associated with seasonality. Since exposure to passive smoking is likely to vary by season, the results suggest that the effect of smoking on SIDS is prenatal rather than the result of passive smoking after birth.