The paper demonstrates how a more profound theoretical understanding concerning the errors governing the accuracy of the flux in digital simulations of electrochemical experiments can be obtained by investigating how a particular discretized scheme performs if it is considered a box method, i.e. a discretized form of Flick's first law of diffusion. The behaviour of the flux is then directly reflected in the discretized equations and the accuracy of the flux can be estimated in a simple and straightforward way by investigating the underlying discretization errors. When discretising Fick's second law of diffusion the discretization errors reflect only the properties of the second-order space derivative and say only little about the accuracy of the simulated flux. (C) 2003 Elsevier B.V. All rights reserved.