Duplex ultrasonography of the lower extremity was used for routine non-invasive screening for asymptomatic proximal and distal deep venous thrombosis, for monitoring of potential propagation of deep venous thrombosis from distal to proximal, and for confirmation of the resolution of proximal deep venous thrombosis after treatment. In the first part of the study, to substantiate the accuracy of duplex ultrasonography, 130 lower limbs (seventy-nine patients) were studied with that modality as well as with venography after a total hip or total knee arthroplasty. Compared with venography, duplex ultrasonography demonstrated 100 per cent sensitivity, specificity, and accuracy for the detection of proximal deep venous thrombosis and 88 per cent sensitivity, 98 per cent specificity, and 98 per cent accuracy for the detection of distal deep venous thrombosis. In the second part of the study, 100 patients who had had a total knee arthroplasty and had been managed with pneumatic stockings and aspirin for prophylaxis against deep venous thrombosis had screening of both lower extremities with duplex ultrasonography on the fourth postoperative day. Duplex ultrasonography demonstrated proximal deep venous thrombosis in seven patients and distal deep venous thrombosis in twenty-two patients; all twenty-nine patients were asymptomatic. The patients who had distal deep venous thrombosis had surveillance with serial duplex ultrasonography on the seventh and fourteenth postoperative days; five of these patients were found to have had propagation of the thrombosis to the proximal deep veins.(ABSTRACT TRUNCATED AT 250 WORDS)