In a prospective study all positive phlebographies within the well-defined population of the city of Malmö, Sweden, during 1987 were studied in order to determine the incidence of deep venous thrombosis (DVT). Epidemiological data were analysed for the detection of patient groups at increased risk of DVT. The incidence was found to be equal for both sexes, i.e. 1.6 per 1000 inhabitants a year. Risk factors were found to be in accordance with earlier studies. The median age for men was 66 years, compared to 72 years for women. At diagnosis of DVT, 19% of subjects had a known malignancy and within 1 year 5% (19 cases) developed a new malignancy. Of the men, 29% had postoperative or post-traumatic (fracture) DVT, compared to 46% of the women. Fewer patients with DVT than expected (39%) belonged to blood group 0 (31%) (P less than 0.005). Pulmonary embolism (PE) was clinically suspected in only 5% of cases, and diagnosis was verified scintigraphically in 2% of cases. None of these died of PE, but of 6 patients who were found to have PE at autopsy, four died about 4 weeks after the DVT was diagnosed.