Using stopped flow light scattering, we show that assembly of fibrin following activation with non-rate-limiting amounts of thrombin or reptilase occurs in two steps, of which the first is end-to-end polymerization of fibrin monomers to protofibrils and the second is lateral association of protofibrils to fibers, in agreement with Ferry's original proposal. Polymerization is found to proceed as a bimolecular association of bifunctional monomers; the overall rate varies as the inverse first power of the concentration; end-to-end association of two monomers, of a monomer and an oligomer, and of two oligomers occurs with the same rate constant. The value of the rate constant is 8.2 C 10(5) M-1 s-1 in 0.5 M NaCl, is three times larger in 0.1 M NaCl (0.05 M Tris, pH 7.4), and is the same following activation by reptilase and by thrombin. The onset of growth of fibers from protofibrils takes 12 times longer in 0.5 than in 0.1 M salt, i.e. thick fibers ("coarse" gels) form from short protofibrils, and thin fibers ("fine" gels) form from longer protofibrils. Jumps of salt concentration at times when protofibrils, but not fibers, have formed result in immediate growth of thick fibers at low salt from long protofibrils formed at high salt. The rate of fiber growth in these experiments varies as the inverse first power of the concentration. 3the instant of gelation (formation of a network of fibers) falls in the later half of the time during which the scattering rises due to fiber growth; the rise of gel rigidity after gelation is found to continue beyond the end of this period. Jumps from low to high salt result in retention of whatever fibers have formed at low salt and a very small additional increase of the scattering due to further fiber growth at high salt. From a variety of evidence, we conclude that the properties of fibrin are determined by kinetics and not equilibria of assembly steps. Results obtained here agree with the following scheme of fibrin assembly: monomers polymerize to protofibrils; long protofibrils associate laterally to fibers; occasionally a long protofibril associates with two different fibers to form an interfiber connection; fiber growth does not reverse to yield stabler, more compact, structures and terminates in formation of a network of fibers. The typical delay of fiber growth is the time during which protofibrils form from monomers. Measurements at rate-limiting concentrations of thrombin have allowed estimation of turnover rates of fibrinopeptides that agree with kinetic parameters obtained with direct assay of fibrinopeptide. Release of fibrinopeptide B causes more rapid fiber formation. Addition of thrombin after activation by reptilase, at a time when protofibrils, but not fibers, have formed, is followed rapidly by fiber formation; this proves that thrombin readily removes fibrinopeptide B from protofibrils. On the basis of these new results and earlier work (in particular, Blombäck, B., Hessel, B., Hogg, D., and Therkildsen, L...