SummaryDiffusion coefficients of soil phosphate calculated from flux to chloride form of anion‐exchange resin‐paper, assuming total depletion of the labile pool at the boundary, were too small and unrealistic. Thus not all the phosphate in the labile pool contributed to the diffusion process while being desorbed at constant pH in the presence of an indifferent anion—Cl—from the resin‐paper. Desorption relationships under these conditions, using CaCl2 at atmospheric CO2 pressure, were markedly different from the relationship between the long‐term 32P‐exchangeable phosphate and the corresponding pore‐solution concentrations. In the same soil containing different amounts of labile phosphate, a different desorption relationship was found for each phosphate level.The constant proportionality between amount of phosphate diffusing into resin‐paper and square root of time was indicative of a constant concentration at the boundary. The resin‐paper: soil system was therefore considered as an infinite composite medium in which diffusion through both resin‐paper and soil were rate‐limiting. The constant boundary concentrations were estimated by the use of the diffusion coefficients in the resin‐paper, the phosphate adsorption isotherm for the resin‐paper and the desorption relationship for each phosphate level in the soil.Diffusion coefficients, calculated using the boundary concentration appropriate to each phosphate level, were related to the slopes of the corresponding desorption relationships, resulting in values of the impedance factor similar to those found for K diffusion under similar conditions.The resin‐paper method, however, does not provide an accurate enough measure of the diffusion coefficient of soil phosphate to be of any practical use. Until better and simpler methods are found, the diffusion coefficient may be calculated using the slopes of the desorption relationship and the separately determined value of the impedance factor.