In soils with a high phosphate buffering capacity or in soils with large amounts of colloidal P in soil solution, E-values often overestimate P availability. These problems have been ascribed to a variety of causes including analytical difficulties in terms of measuring low concentrations of P, colorimetric interferences, and inadvertent measurement of non-isotopically exchangeable colloidal P. We investigated measurement of E-values in 8 soil types, of which 6 had properties likely to give overestimates of size of the available pool of P based on results of previous authors. The potential for overestimation of the E-value due to interferences described above was identified in several soil types. A simple anion exchange resin purification step was introduced to the E-value methodology. The resin adsorbs both P-31 and P-32 isotopes in proportion to their concentration in the solution and minimises transfer of colloids and Si into the analysed eluant, thereby providing a simple way to increase the eluant P concentration and to avoid analytical interferences.