A retrospective cohort study was performed to determine the cause of death among 5923 farmers in Iceland. Information on deaths occurring between 1977 and 1985 was obtained through the Statistical Bureau of Iceland. The vital status could be ascertained for all subjects in the study. Expected death rates were calculated, based on the national rates for males in the corresponding age groups and calendar years. The number of deaths from all causes, malignant neoplasms, lung cancer, ischaemic heart disease, respiratory diseases and accidents was less than expected in the total cohort and in nearly all subcohorts. There was no statistically significant excess risk, however: SMR for skin cancer was 2.30, SMR for Hodgkin's disease was 1.71, for leukaemia SMR was 1.60, and for brain cancer SMR was 1.23 in the total cohort. The results are in agreement with those of most previous studies of farmers, but because of the short follow-up time, the excess risk found for deaths from skin and haematological malignancies did not reach statistical significance. Further follow-up is planned in the future.