This collaborative study, organized by the Hemostasis Subcommittee of C.I.S.M.E.L., evaluated the accuracy, precision, and comparability of the following six widely used fibrinogen assays: total clottable fibrinogen (Blombäck and Blombäck), clotting time (Von Clauss), turbidimetry (Ellis and Stransky), Chromotime System, prothrombin time (PT)-derived, and radial immunodiffusion (RID). The same frozen samples, with normal and high contents of fibrinogen, were examined in four laboratories. The methods were calibrated with an internal standard whose fibrinogen content was determined gravimetrically. Both the Von Clauss and the RID methods were reliable, accurate, and precise, if adequate calibration was used. The PT-derived method was highly reproducible, but had some problems with accuracy. We demonstrate that an adequate calibration procedure is indispensable for reliable fibrinogen measurements whatever method is used. Because neither the calibration procedures proposed by the manufacturers nor the use of lyophilized commercial plasmas is adequate for this purpose, we urge that an international standard for fibrinogen measurement be promptly established.