The purpose of this work was to investigate the effect of solute size, pore water velocity, and intraparticle porosity on dispersion and to delineate conditions under which it is appropriate to use a tracer-derived dispersivity to represent the dispersivity of other, dissimilar-sized solutes. This was accomplished by evaluating the physical properties of four porous media and the results of several column experiments performed with nonreactive tracers and these media. Intraparticle porosities of three sandy media were shown to constitute negligible fractions of the total porosities. As a result, intraparticle diffusion did not contribute to dispersive flux. The contribution of intraparticle diffusion, as well as film diffusion, to solute dispersion in fine-grained media having small intraparticle porosities will be negligible for most conditions. In addition, the importance of axial diffusion decreases as pore water velocity increases. Hydrodynamic dispersion is the predominant source of dispersion under these conditions, and, as such, dispersivity is essentially independent of solute size. Dispersivities determined with a tracer may therefore be used to represent the dispersivity of other, dissimilar-sized solutes. Under these conditions, the practice of using a nonreactive tracer such as 3H2O to characterize the dispersive properties of a soil column is valid. Solute dispersion in an aggregated soil was shown to be caused by a combination of hydrodynamic dispersion, film diffusion, and intraparticle diffusion. In addition, axial diffusion was shown to be important at low pore water velocities. Hence, the apparent dispersivities obtained by applying the single-component dispersion equation to transport in structured soils will generally be a function of solute size. The use of a tracer-derived dispersivity for solutes of different sizes would not be valid in this case. ¿ American Geophysical Union 1993