Respiratory symptoms, atopy, and bronchial reactivity were measured in primary school children living in Lake Munmorah (LM), a coastal town near two power stations, and in Nelson Bay (NB), a coastal town free from any possible major sources of outdoor air pollution. A prevalence survey and longitudinal follow-up study were performed 1 year apart. In both studies, the prevalence of ever wheezed, current wheezing, breathlessness, wheezing with exercise, diagnosed asthma, and use of drugs for asthma at LM were all approximately double the prevalence at NB (all P values less than 0.01). The prevalence of bronchial reactivity was significantly greater at LM than NB (P less than 0.01) at the first but not the second survey. By contrast, no significant differences were found between the two areas for skin test atopy or for parental history of allergic disease. Multivariate analysis supported the conclusion from the univariate analysis that there was more wheezing at LM compared to NB at both studies, when adjusted for atopy, smoking in the home, age, and sex. As expected, a positive skin test reaction to house dust mite was the predominant explanatory variable. Asthma was more common in the community near power stations (LM) than in the NB area. The absence of significant differences in skin test atopy and parental history of allergic disease argued against major genetic differences between the two groups. By contrast, the more common reporting of siblings' chest disease and asthma in Lake Munmorah supported an environmental cause.