Grain dust exposure has been associated with both acute and chronic respiratory abnormalities. We sought to answer two questions with regard to this exposure: (1) is there a dose-response relationship between grain dust and respiratory abnormalities, and (2) is there evidence to suggest that the allowable exposure (currently 10 mg/m3 in Canada) be lowered? We compared respiratory symptoms and lung function among 454 grain elevator workers and 55 civic workers to estimates of lifetime average grain dust exposure. A total of 781 personal air samples representing 20 different job titles over a 15-yr period were used to construct a job title-time period matrix for average dust exposure levels. The matrix was applied to each worker's detailed job history to obtain a value for estimated cumulative and average dust exposure. Significant dose-response relationships were seen for chronic phlegm production, breathlessness on exertion, FEV1, and FVC; and dose-response trends were evident for the longitudinal change in both FEV1 and FVC. Workers with estimated average exposure between 4 and 9 mg/m3 were found to have significantly lower values for FEV1 and FVC compared to both grain workers exposed to less than 4 mg/m3 on average and civic workers, despite no difference in duration of employment among these groups. The results indicate a strong dose-response relationship between grain dust exposure and both respiratory symptoms and lung function and also suggest that the current Canadian allowable exposure level of 10 mg/m3 is too high.