The metabolism of human antithrombin III (heparin cofactor) was studied in four control subjects, in four subjects with peripheral obliterative arterial disease, in six patients with recent venous thrombosis and in one patient with clinically severe haemophilia A. The labelled antithrombin III has a high specific activity (5.75 units/mg) and displayed a single band on SDS-polyacrylamide gel electrophoresis. On Sephadex G-100 gel filtration the labelled material eluted in the same position as the antithrombin III activity in plasma. Crossed immunoelectrophoresis of a mixture of fresh plasma and labelled antithrombin III against a specific antiserum, revealed a single precipitin line in which radioactivity was concentrated. The changes in electrophoretic mobility of both the plasma antithrombin III and the labelled material following the addition of heparin to the mixture or following coagulation were identical. The purified antithrombin III behaved as a homogeneous protein in the turnover experiments. The plasma radioactivity data were approximated by a sum of two exponential terms and the metabolism of antithrombin III represented by a two compartment mammillary model. Results in the control subjects were as follows: plasma antithrombin III concentration 19.6 +/- 2.3 mg/100 ml; intravascular fraction 0.45 +/- 0.05; fractional catabolic rate 0.55 +/- 0.02 of the plasma pool per day; half-life of the plasma radioactivity 2.83 +/- 0.26 days. Circulating large molecular weight degradation products of labelled antithrombin III could not be detected by Sephadex G-100 gel filtration. No significant differences in these parameters were found in the patients with peripheral arterial insufficiency. The turnover rate of antithrombin III was normal in the patient with haemophilia A. In three patients with venous thrombosis not treated with heparin, the turnover of labelled antithrombin III was in the normal range. In three patients with venous thrombosis, treated with heparin, the plasma radioactivity half-life was significantly shortened (2.13 +/- 0.08 days) and the fractional catabolic rate increased (0.75 +/- 0.05) of the plasma pool per day). In one of these patients, the labelled antithrombin III had been incubated with an equimolar amount of heparin prior to injection. In this patient the plasma radioactivity half-life was in the same range as in the other two patients (2.15 days).