The change in extractable P due to fertilization should be quantified for reliable P recommendations, but this effect is dependent on the P buffering capacity of the soil. To determine this change, rates of fertilizer P were applied in six field experiments on Ultisols that varied in clay content. Mehlich-3 extractable P (M3P) was measured annually, and an initial increase in M3P due to fertilization followed by an exponential decrease in M3P due to sorption were modeled. The change in extractable P after 1 yr at each site was calculated from the model. These data were combined with those from previous studies on two Ultisols in North Carolina and five Oxisols in Brazil to broaden the range in clay content observed. The increase in M3P per unit of applied P after 1 yr was 0.7 for soils with only 10% clay, and it decreased exponentially to <0.2 for soils with >50% clay. The results of this study were similar to conclusions from a P incubation study extracted with Bray-1 in Natal, South Africa, on 54 soils from five U.S. taxonomic orders, including Ultisols and Oxisols. It appears there is a broad range of cultivated soils for which knowledge of the clay content would allow more reliable predictions of fertilizer P to keep extractable P at a level sufficient for optimum crop yields.