Possible associations between the incidence of progressive massive fibrosis (PMF) over periods of average length 11 years during 1963-80 and a range of explanatory variables, both environmental and medical, were examined in a study group of4772 miners and ex-miners using casecontrol methods. Cases were members of the study group who developed PMF during the study period; controls were men who remained free of the disease. Cases and controls were matched on cumulative dust exposure to the start of the study period, colliery of employment, and whether or not employment before and during the study period had been at a single colliery only. Of 257 cases of PMF, 142 were matched to four controls each, 39 to three controls, 33 to two controls, and 31 to a single control only. Twelve cases could not be matched. Age and category of simple pneumoconiosis were each found to be associated with the incidence of PMF, confirming the results of many previous studies. Quetelet's index of body mass (a measure of weight, normalised for height) was found to differ considerably between cases and controls, cases tending to be lighter for their height. The difference did not vary significantly between three groups of collieries, defined by coal rank. Regression analysis confirmed the high level of statistical significance of the difference (p < 0 001), allowing for the effects of age and prior category of simple pneumoconiosis. Of the environmental factors considered, an index of residence time of dust in the lungs was found to have the most statistically significant association with the incidence of PMF. The crudeness of the measure used, however, together with unreliability in estimates of early exposure to dust for some of the men studied, means that further work is necessary to investigate the magnitude and significance of this apparent association. In low and medium rank collieries there were large differences between the proportion of cases and controls showing a range of respiratory symptoms, cases consistently displaying a higher prevalence. In the high rank collieries of south Wales, however, symptoms occurred with equal frequency in cases and controls, the prevalence at these collieries being the same as among cases generally. Regression analysis (low and medium rank collieries only) showed that the presence of breathlessness was the symptom most strongly associated with the risk of attack of PMF, allowing for the effects of age and Quetelet's index.A recently reported statistical analysis' of data gath-their colliery of employment were all shown to be ered as part of the British National Coal Board's associated with their risk of developing PMF over five Pneumoconiosis Field Research (PFR) between 1953 year periods after allowing for the effects of exposure and 1980, and pertaining to more than 30 000 British to dust. coalminers, confirmed that cumulative exposure to This large scale study of working miners involved mixed coalmine dust plays a central part in the relatively few explanatory variables' The same limi...