The clinical and epidemiological features of 80 cases of malignant mesothelioma (as proved by examination of biopsy specimens) referred to the clinic at the National Centre for Occupational Health between January 1977 and June 1983 are reviewed. There was a positive history of asbestos exposure in 89% of cases. The mean survival time from diagnosis to death was 8,6 months and from the onset of symptoms to death 13,6 months. Survival time was unaffected by stage of the tumour, treatment, histological features, smoking status, presenting symptoms, presence or absence of effusion and asbestosis, side of the lesion, source of exposure and lag period from first exposure to diagnosis. The duration of survival was significantly affected by age at diagnosis, duration of asbestos exposure and the number, rather than the type, of treatment regimens used. Caution is advocated in interpreting these data since the number of cases was small and the study design was retrospective. A reference group of 546 cases notified over the same period was drawn from the records of the South African Asbestos Tumour Reference Panel. The incompleteness of national mesothelioma incidence data was noted, and an incidence figure of 7,2 per million per year was calculated from the best-available data for South Africa. This figure is an underestimate because not all diagnosed cases are reflected and, more important, significant numbers of cases are never diagnosed. The extent to which the compensation machinery functions is mentioned.