We have used activation peptide release assays to compare factor VII and activated factor VII (VIIa) activation of factor X, normal factor IX (IXN), and a variant factor IX (IXBmLE), which, after activation, is unable to back-activate factor VII. In purified systems, factor VII and VIIa each rapidly activated factor X, but after a one minute lag for factor VII. VIIa also readily activated both IXN and IXBmLE. Factor VII initially failed to activate substantial amounts of either IXN or IXBmLE; on further incubation factor VII activated IXN but not IXBmLE. Activation of IXN began when approximately 10% of factor VII had been converted to VIIa, as measured by 125I-factor VII radioactivity profiles. Adding factor VII to VIIa slowed its activation of IXBmLE. However, in the presence of factor X, factor VII alone rapidly activated IXBmLE. Unlike purified systems, 1 nmol/L VIIa added to factor VII-deficient plasma failed to activate factor IX. Increasing factor VII to 10 nmol/L (plasma concentration) either as native VII or VIIa yielded similar activation curves for factor IX and similar activation curves for factor X. Adding 5% VIIa to factor X-deficient plasma and to factor XII-deficient plasma substantially shortened the dilute tissue factor clotting time of only the former. These data support the hypothesis that factor VII/tissue factor complex initiates tissue factor-dependent clotting through a minimal generation of Xa. This Xa then rapidly back-activates a small amount of factor VII, following which the rates of activation of both factors IX and X increase dramatically.