Nitrogen dioxide is an oxidant gas that contaminates outdoor air and indoor air in homes with unvented gas appliances. A prospective cohort study was carried out to test the hypothesis that residential exposure to NO2 increases incidence and severity of respiratory illnesses during the first 18 months of life. A cohort of 1,205 healthy infants from homes without smokers was enrolled. The daily occurrence of respiratory symptoms and illnesses was reported by the mothers every 2 wk. Illnesses with wheezing or wet cough were classified as lower respiratory tract. Indoor NO2 concentrations were serially measured with passive samplers placed in the subjects' bedrooms. In stratified analyses, illness incidence rates did not consistently increase with exposure to NO2 or stove type. In multivariate analyses that adjusted for potential confounding factors, odds ratios were not significantly elevated for current or lagged NO2 exposures, or stove type. Illness duration, a measure of illness severity, was not associated with NO2 exposure. The findings can be extended to homes with gas stoves in regions of the United States where the outdoor air is not heavily polluted by NO2.