Recently, several studies have proposed models describing the mechanisms of Alzheimer's beta-amyloid fibril formation in vitro. However, these models are somewhat controversial and no exact kinetic analyses measuring the polymerization velocity as an indicator of the reaction, have thus far been available. We first formed beta-amyloid fibrils from a synthetic peptide, beta-amyloid (1-40), and determined the optimum conditions for quantitative fluorometry of these beta-amyloid fibrils with thioflavine T. Optimum fluorescence measurements of beta-amyloid fibrils were obtained at the excitation and emission wavelengths of 446 and 490 nm, respectively, with the reaction mixture containing 5 microM thioflavine T and 50 mM of glycine-NaOH buffer, pH 8.5. We then focused our study on the extension phase of beta-amyloid fibril formation in vitro. When beta-amyloid fibrils were incubated with monomeric beta-amyloid (1-40) in conditions where de novo seed formation does not occur, the extension of beta-amyloid fibrils was observed with electron microscopy. Quantitative fluorometry revealed that:(a) extension of amyloid fibrils proceeded by a pseudo-first-order exponential increase as measured by the fluorescence of thioflavine T;(b) the rate of extension was maximum around pH 7.5, and was dependent on the incubation temperature. Between 20 and 37 degrees C, good linearity was observed between the common logarithm of the initial rate and the reciprocal of the absolute temperature;(c) the rate of polymerization was found to be proportional to the product of beta-amyloid fibrils number concentration and the beta-amyloid (1-40) concentration;(d) the net rate of extension was the sum of the rates of polymerization and depolymerization. These results show that beta-amyloid fibril formation can be explained by a first-order kinetic model: ie, the extension of beta-amyloid fibrils proceeds via the consecutive association of beta-amyloid (1-40) onto the ends of existing fibrils.