Mixtures of two redox-active compounds with dissimilar diffusion coefficients produce non-additive mass-transfer limited currents. Similarly, in the potential range where three redox-active species, decamethylferrocene (dMeFc), ferrocene (Fc) and N-methylphenothiazine (MePTZ), are oxidized simultaneously (E-MePTZ(o') > E-Fc(o') > E-dMeFc(o')) with rates controlled by linear diffusion, electrogenerated radicals diffusing outwards from the electrode react with original species diffusing towards the electrode from the bulk; thus, Fc(center dot+) reacts with dMeFc producing Fc and dMeFc(center dot+), while MePTZ(center dot+) reacts with both Fc and dMeFc producing MePTZ together with Fc(center dot+) and dMeFc(center dot+). These processes replace the flux of dMeFc with Fc at the second current plateau (referring to normal pulse voltammetry), and the fluxes of both dMeFc and Fc with MePTZ at the third plateau. Analogous results have been obtained and analyzed with two other multicomponent systems undergoing multiple sequential electron transfers, namely dMeFc/Fc/TPTA and dMeFc/TTF (TPTA: tri-N-p-tolylamine; TTF: tetrathiafulvalene). Since the diffusion coefficients of the three species are different, the mass-transfer limited currents of the second and third oxidation waves are not equal to the sum of the currents that each component would have produced if it were in the solution alone. Numerical simulations of the experimental voltammograms using diffusion coefficients measured independently support this mechanism. Multicomponent systems are encountered frequently in practice and our results identify one significant (similar to 10%) source of error in quantitative voltammetric analysis. Ways around the problem are summarized in the conclusions section. (c) 2005 Elsevier Ltd. All rights reserved.