OBJECTIVE: To evaluate the potential for confounding from asbestos exposure, primarily chrysotile, on the relation between crystalline silica and mortality from lung cancer among diatomaceous earth (diatomite) workers. METHODS: A reanalysis of a cohort mortality study of diatomite workers was performed to take into account quantitative estimates of asbestos exposure. The reanalysis was limited to a subset of the original cohort, comprising 2266 white men for whom asbestos exposure could be reconstructed with greatest confidence. Comparisons between mortality from lung cancer (standardised mortality ratios (SMR)) were made between rates for 1942-87 for United States white men, and workers cross classified according to cumulative exposures to crystalline silica and asbestos. Comparisons of internal rates, involving Poisson regression modeling, were conducted for exposure to crystalline silica, with and without adjustment for asbestos exposure. Exposures were lagged by 15 years to take into account disease latency. RESULTS: There was an overall excess of lung cancer (SMR 1.41; 52 observed). The SMRs for four categories of increasing crystalline silica among the workers not exposed to asbestos were 1.13, 0.87, 2.14, 2.00. An SMR of 8.31 (three observed) was found for workers with the highest cumulative exposure to both dusts. Internal analysis, after adjustment for asbestos exposure, yielded rate ratios for categories of exposure to crystalline silica: 1.00 (reference), 1.37, 1.80, and 1.79. CONCLUSIONS: Asbestos exposure was not an important confounder of the association between crystalline silica and mortality from lung cancer in this cohort. Although based on a small number of deaths from lung cancer, the data suggest possible synergy between these exposures. An extended follow up of this cohort is in progress and should enable better assessments of independent and combined effects on risk of lung cancer.