Chang and Slattery (1986) introduced a simplified model for dispersion that contains only two empirical parameters, both of which can be determined in one-dimensional experiments. The traditional model for dispersion (Nikolaevskii, 1959;Scheidegger, 1961;de Josselin de Jong and Bossen, 1961;Bear, 1961a;Peaceman, 1966;Bear, 1972) has three empirical parameters, two of which can be measured in one-dimensional experiments while the third, the transverse dispersivity, must be measured in experiments in which a two-dimensional concentration profile develops. For the common one-dimensional experiment in which the signs of the concentration gradient and of the velocity field are different, the simplified model and the traditional model give identical results. For a one-dimensional experiment in which the signs of the concentration gradient and of the velocity field are at least sometimes the same and for two-and three-dimensional flows, the simplified model of Chang and Slanery (1986) gives results that can differ from those predicted using the traditional model.Only the experimental data of Bear (1961b) and of Yule and Gardner (1978) are sufficiently complete to permit a comparison of the two models. Considering the quality of the experimental data, we can not distinguish between the predicted profiles based upon the simplified model and those based upon the traditional model.