A unique data set of 3,380 British coal miners has been reanalyzed with major focus on nonpneumoconiotic respiratory conditions. The aim was to assess the independent contribution of smoking and exposure to respirable dust to clinically significant measures of respiratory dysfunction. Exposure to coal-mine dust was monitored over a 10-yr period. Medical surveys provided estimates of prior dust exposure and recorded respiratory symptoms. Each man's FEV1 was compared with the level predicted for his age and height by an internally derived prediction equation for FEV1. Four respiratory indices were considered at the end of the 10-yr period: FEV1 less than 80%, chronic bronchitis, chronic bronchitis with FEV1 less than 80%, and FEV1 less than 65%. Results were uniformly incorporated into logistic regression equations for each condition. The equations include coefficients for age, dust, and when indicated, an interaction term for age and dust. Dust-related increases in prevalence of each of the 4 conditions were statistically significant and were similar for smokers and nonsmokers at the mean age (47 yr). There was no evidence that smoking potentiates the effect of exposure to dust. Estimates of prevalences at the mean age of all 4 measures of respiratory dysfunction were greater in smokers. At intermediate and high dust exposure the prevalence of the 4 conditions in nonsmokers approached the prevalence in smokers at hypothetically zero dust exposure. Both smoking and dust exposure can cause clinically important respiratory dysfunction and their separate contributions to obstructive airway disease in coal miners appear to be additive.(ABSTRACT TRUNCATED AT 250 WORDS)