The association between parental cigarette smoking and children's pulmonary function was investigated in 8,706 nonsmoking white children, followed annually by questionnaire and spirometry between 6 and 18 yr. Exposure to maternal and paternal smoking was each divided into three components: exposure in the first 5 yr of life, cumulative exposure between age 6 and the year prior to each visit; parental smoking status reported at each visit. Regression splines were used to assess the effects of parental smoking on the level and growth rate of pulmonary function adjusting for age, height, city of residence, and parental education. Models best predicting pulmonary function level included current maternal smoking and exposure to maternal smoking in the first 5 yr of life but did not include the other measures of parental smoking. After adjusting for early exposure, current maternal smoking for each pack/d was associated with a reduced level of FEV1 (-0.4%, 95% CI: -0.9, 0.1), FEV1/FVC (-0.6%, 95% CI: -0.9, -0.4), and FEF25-75% (-2.3%, 95% CI: -3.6, -1.0) in children 6 to 10 yr. These effects were slightly smaller in children 11 to 18 yr. Adjusting for current maternal smoking, those exposed to maternal smoking in the first 5 yr of life had higher FVC (+0.5%, 95% CI: -0.1, 1.2) and lower FEV1/FVC (-0.7%, 95% CI: -1.1, -0.4) and FEF25-75% (-2.8%, 95% CI: -4.4, -1.2) than those not exposed in children 6 to 10 yr.(ABSTRACT TRUNCATED AT 250 WORDS)