The effects of the presence of supporting electrolyte and the beam diameter on probe beam deflection experiments were investigated both theoretically and experimentally. A mass-transfer : limited potential step experiment with the ferricyanide/ferrocyanide redox couple is used to illustrate these effects. Although the supporting electrolyte does not react, the transference number of the supporting electrolyte is close to unity, which results in concentration gradients forming in the supporting electrolyte. We show here that these concentration gradients should not be neglected. Analytic solutions for a potential step and fast kinetics are developed both with and without considering the concentration gradients in the supporting electrolyte. The model agrees quantitatively with experiments performed in K2SO4, KCl, and KOH. Typical probe beams are too large in diameter to be approximated as an infinitely thin ray. The models include consideration of the finite diameter of the probe beam. The results show that a 200 mu m diameter probe beam is too large to detect unique features of the potential step experiment and that a 100 mu m diameter probe beam should be able to detect these features. The difficulties of determining diffusion coefficients from the probe beam deflection are discussed. It is concluded that it is impossible to determine the diffusion coefficients of the reactant and product from the extremum of the probe beam deflection. A non-linear regression of the model to experimental data is possible in theory, but is susceptible to error.