Health surveys were carried out in two factories, one in which only chrysotile had been used and one in which chrysotile, crocidolite and amosite had been used. The same health questionnaire, given by the same interviewer, and the same lung function tests recorded on the same apparatus were carried out as in previous studies with Quebec chrysotile asbestos miners and millers. Results were expressed as prevalence % abnormality. Despite shorter exposure to apparently lower levels of dust, the factory workers showed a greater prevalence of certain symptoms and of respiratory function abnormality than the miners and millers. The results suggest a greater effect of process than of fibre type on these health measurements, a finding with implications for control and one which deserves further study.