Fibrillization of tau protein is a hallmark lesion in Alzheimer's disease. To clarify the utility of electron microscopy as a quantitative assay for tau fibrillization in vitro, the interaction between synthetic tau filaments and carbon/formvar-coated grids was characterized in detail. Filament adsorption onto grids was hyperbolic when analyzed as a function of time or bulk protein concentration, with no evidence for competitive displacement or elution from other components in the reaction mixture. Filament length measurements were linear with filament concentration so long as the concentration of total tau protein in the sample was held constant, suggesting that measurement of filament lengths was accurate under these conditions. Furthermore, exponential filament length distributions were not significantly affected by adsorption time or filament concentration, suggesting that preferential binding among filaments of differing lengths was minimal. However, monomeric tau protein was found to be a strong competitor of filament adsorption, indicating that comparison of filament length measurements at different bulk tau concentrations should be interpreted with caution.