For patients on oral anticoagulation controlled with Quick's prothrombin time test using rabbit brain thromboplastin American clinicians proposed in the early 1940s that the lower limit of prolongation be taken at 1.5 and the upper limit at 3, corresponding to 10-30% prothrombin activity on a saline dilution curve. Thromboplastins derived from other tissues and species were later introduced, methods were modified, and adsorbed plasma was used instead of saline in the construction of the dilution curve to obtain percentage prothrombin activity; ratios and percentages lost their initial significance, but too often without the clinicians' awareness. With the detection of PIVKAs, finally, it became clear that transformation of prothrombin times into percentage activity, as obtained from dilution curves, could never be a valid means of standardization. It also became evident that only direct comparison of thromboplastins with (fresh) plasma from patients on stabilized oral anticoagulants could be used to determine equivalent therapeutic ranges for different types of thromboplastins. Reference thromboplastins were established and calibration procedures developed. A series of well-controlled clinical trials has provided sufficient information to define, in terms of these reference thromboplastins, therapeutic ranges for the prophylaxis and treatment of venous as well as arterial thrombosis. Definition of the ranges in terms of the biochemical defect induced by coumarin congeners remains, however, to be established.