To determine whether exposure to crocidolite is associated with excess mortality, the authors calculated standardized mortality ratios based on deaths in South African crocidolite mining districts from 1968 to 1980 for selected causes of death. Contiguous districts were used as controls. To take account of background geographic variability, they divided the control districts into groups of population size similar to those of the crocidolite mining districts. Standardized mortality ratios in crocidolite mining districts were elevated for asbestosis and/or mesothelioma, and cancer of the lung and stomach. These findings could not be explained by background geographic variability in mortality and are likely to be due to exposure to South African crocidolite during mining and milling or to environmental contamination. The increased standardized mortality ratios for stomach cancer are of particular interest, since excess deaths have not previously been shown in individuals exposed to crocidolite alone.