Sequential fractionation procedures that use an NaOH extraction for removal of P associated with iron and aluminum minerals, followed by an HCl extraction to dissolve P associated with calcium minerals, have been used extensively in investigating the chemical partitioning of phosphorus in soils. Due to the high pH conditions that occur during NaOH extraction, it is hypothesized that PO43− freed from association with Fe and Al minerals can resorb to, or precipitate as calcium minerals. This phosphate would subsequently be recovered in the acid extraction, resulting in overestimation of calcium bound P and underestimation of Fe and Al bound P. To test the hypothesis, Fe bound phosphate was subjected to sequential fractionation with and without CaCO3 present. Without CaCO3, 57% of the total recovered phosphorus was found in the 0.1N NaOH extraction, and 15% in the 0.5N HCl extraction. With CaCO3, 4% was found in the 0.1N NaOH extraction and 39% in the 0.5N HCl extraction. Fe-bound phosphate was added and then sequentially extracted from a typical Everglades Histosol soil. Mean recoveries of phosphate added as Fe-bound P (95% confidence intervals in parentheses) were estimated to be 33% (±2.4%) in HCl, 32% (±4.8%) as inorganic phosphate in NaOH, 24% (±13.6%) organic P in NaOH, 14% (±1.3%) in KCl, and 8% (±2.8%) in the residual fraction. These results show that the procedure underestimated the amount of iron- and aluminum-bound P and falsely attributed Fe-bound P to the Ca-bound P fraction.