A prerequisite of most AT III assays is the defibrination of the samples. Defibrination of plasma, however, is generally associated with a loss of AT III. This especially true of heat defibrination. Furthermore, according to our experience, most of the functional determinations don’t correlate well with the immunological ones, and don’t provide reproducible results. Therefore, it was the aim of our investigations to establish a method which omits the defibrination and yields reproducible results.The method reported is based on the identity of AT III and heparin cofactor antithrombin II (AT II) respectively. AT III is converted by heparin from a progressive into an immediate inhibitor, allowing a plasma dilution of 1:50. Due to this high dilution, the defibrination step can be omitted and AT III determined in native plasma. To transform AT III completely into its heparin complex, 5 USP units heparin are used. α2-macro-globulin up to 400% of normal plasma concentration does not influence the assay. When determinations are performed at a certain pH value, good reproducibility is obtained. The mean error of determinations of the same sample does not exeed 4%. Maximal deviations were in the range of 5 to 10%. The results of functional determinations are compared with those obtained by two immunological methods. Deviations scarcely exeed the limits of methodical errors.