Plasma antithrombin activity was measured using an amidolytic method (substrate Chromozym TH) and a clotting method. The mean antithrombin values found in 76 hospital outpatients were 9.4 micronmol/min/ml with the amidolytic procedure and 100.1% of antithrombin activity with the clotting procedure. The two methods correlate fairly well (r = 0.85, P less than 0.01) and show satisfactory reproducibility. Coefficients of variation of 5.9% and 8.8% were obtained respectively with the amidolytic and the clotting procedures. In the presence of very high levels of fibrinogen degradation products, falsely elevated antithrombin activity levels were observed with the clotting procedure but the amidolytic method is essentially unaffected. It is concluded that both methods are suitable for determining antithrombin activity but a well-standardised amidolytic procedure has advantages.