The effect of indoor nitrogen dioxide on the cumulative incidence of respiratory symptoms and pulmonary function level was studied in a cohort of 1,567 white children aged 7-11 years examined in six US cities from 1983 through 1988. Week-long measurements of nitrogen dioxide were obtained at three indoor locations over 2 consecutive weeks in both the winter and the summer months. The household annual average nitrogen dioxide concentration was modeled as a continuous variable and as four ordered categories. Multiple logistic regression analysis of symptom reports from a questionnaire administered after indoor monitoring showed that a 15-ppb increase in the household annual nitrogen dioxide mean was associated with an increased cumulative incidence of lower respiratory symptoms (odds ratio (OR) = 1.4, 95% confidence interval (95% Cl) 1.1-1.7). The response variable indicated the report of one or more of the following symptoms: attacks of shortness of breath with wheeze, chronic wheeze, chronic cough, chronic phlegm, or bronchitis. Girls showed a stronger association (OR = 1.7, 95% Cl 1.3-2.2) than did boys (OR = 1.2, 95% Cl 0.9-1.5). An analysis of pulmonary function measurements showed no consistent effect of nitrogen dioxide. These results are consistent with earlier reports based on categorical indicators of household nitrogen dioxide sources and provide a more specific association with nitrogen dioxide as measured in children's homes.