Thrombin converts fibrinogen to fibrin in two steps. First fibrinopeptide A and fibrin I are formed and then fibrinopeptide B (B beta 1-14) and fibrin II. Since it is postulated that fibrin II is important in the genesis of thrombosis, it is of interest to measure fibrinopeptide B in peripheral blood samples. Previous difficulties in interpreting fibrinopeptide B immunoreactivity in plasma resulted from crossreaction of fibrinogen and of plasmin digest peptides B beta 1-42 and B beta 1-21 and from rapid loss of fibrinopeptide B immunoreactivity resulting from cleavage of arginine 14 by blood carboxypeptidase B. We have obviated these difficulties by removing fibrinogen from plasma by precipitation with ethanol and peptides B beta 1-21 and B beta 1-42 by adsorption on bentonite. Fibrinopeptide B is then converted to a desarginine fibrinopeptide B, which is measured in a new specific assay. Studies of the kinetics of fibrinopeptide cleavage showed that when whole blood was allowed to clot in vitro, fibrinopeptide A was cleaved more rapidly than fibrinopeptide B. In 18 patients on an acute care medical ward, desarginine fibrinopeptide B levels were lower than fibrinopeptide A levels and did not correlate with the levels of fibrinopeptide A or B beta 1-42. Desarginine fibrinopeptide B levels were less than 1 pmol/ml in all but two patients. In six patients receiving intraamniotic infusions of hypertonic saline to induce abortion, desarginine fibrinopeptide B levels increased 10-fold from the preinfusion mean level of 0.4 pmol/ml and then decreased. The pattern of changes resembled that of the fibrinopeptide A levels rather than of the B beta 1-42 levels. On the basis of these data it is suggested that plasma desarginine fibrinopeptide B levels reflect fibrin II formation in vivo.