Three hundred and twenty autopsy cases of sarcoidosis during a 32 year period were collected from the Annuals of the Pathological Autopsy Cases in Japan, published yearly since 1958, and from a literature survey. A statistical analysis of these reviewed autopsy cases was carried out on the epidemiological features of the disease and on the causes of death. The proportion of sarcoidosis autopsy cases relative to the total autopsy cases had increased during this 32 year period. The increase of sarcoidosis autopsies during this period was chiefly due to the increase in aged females; the total number of female cases was approximately two times more than that of males. As over half of the total cases had only a pathological diagnosis and not a clinical diagnosis for sarcoidosis, the actual morbidity from sarcoidosis that was estimated from the autopsy data and corrected by autopsy rate was over five times higher than that of the clinically recognized cases. Age and sex distribution of these cases peaked in the thirties for both sexes, while another very high peak was noted in females over 50 years of age. In approximately 60% of the sarcoidosis autopsies, the cause of death related to sarcoid lesions in the heart, lung or nervous system, the majority of which involved cardiac sarcoidosis. In the remaining 40% of the cases, the cause of death was from non-sarcoidosis diseases.