Plasma antithrombin levels were measured by clotting, immunological, and amidolytic methods on two groups of subjects: 20 normal individuals and nine patients studied serially post-operatively (hip replacement). The postoperative patients were observed for the emergence of deep-vein thrombosis using 125I-fibrinogen uptake measurements (FUT). The three methods gave similar ranges for the normal subjects, were reproducible (cv less than 5%), and detected early postoperative reduction of antithrombin levels. All three methods failed to show any significant differences in preoperative antithrombin levels between the positive and negative FUT groups. Correlation studies were performed on the pooled data from the normal and postoperative group (range 60-130% of normal; 100 samples). The best correlation (r = 0.75; P less than 0.01) was achieved with the chromogenic kit assay method versus the Mancini immunoassay technique. The thrombin agarose (total antithrombin) gel diffusion technique correlated less well with the chromogenic (r = 0.65; P less than 0.01) and Mancini immunoassay (r = 0.45; P less than 0.01) methods. It is concluded that the chromogenic kit method gives a rapid, reproducible, and specific measurement of antithrombin III. The thrombin agarose diffusion method, though not specific for antithrombin III, is a cheap and simple method to perform. The potential of the three methods for detecting the prethrombotic stage and early thrombosis is discussed.