The error which results from calculation of a diffusion coefficient from desorption measurements which fail to account for the nonuniform concentration profile developed during prior sample storage is determined for the case of migration from a plane sheet into an infinite, well-stlrred bath. The incorrectly calculated diffusion coefficient, D', neglecting prior storage time, is smaller than the actual diffusion coefficient, D. For small storage times, the ratio D'JD is close to unity; however, D'JD approaches an asymptotic value of 0.70 for very large storage times if the well-known "half-time'' method is employed for evaluating the diffusion coefficients. I f the so-called "initial slope" method is used for determination of diffusion coefficients, the calculational error is large, with the ratio D'lD approaching zero for large storage time. A sigmoid response in plots of fractional loss vs. the square root of time is sometimes taken as evidence for the existence of non-Fickian transport processes. It is shown here that such sigmoid plots are also observed when elution occurs prior to the initiation of the formal desorption experiment, even when Fickian transport controls the desorption process.