Most natural pastures in the Campos biome in Southern Brazil grow in soils with low phosphorus (P) availability, but with high contents of total and organic P. The present study aimed to evaluate the changes in forms of soil P resulting from the application of phosphate sources over a growth cycle of natural pastures. In three experiments under natural pasture, P was applied in the form of Gafsa rock phosphate, triple superphosphate, and a control treatment in a randomized block design with three replications. In the trials carried out in natural pastures in the municipality of Candiota (RS) in an Alfisol and an Entisol, 100 kg ha-1 of P2O5 was applied in September 2010. In natural pasture in an Ultisol in the municipality of Santa Maria (RS), 180, 90, 100, and 100 kg ha-1 of P2O5 were applied in the years 1997, 1998, 2002, and 2010, respectively. Soil samples were collected (0-10 cm) throughout the natural grassland growing season (0, 55, 116, 171, and 232 days after P application at Candiota and 0, 50, 83, 129, 159, and 186 days after application of phosphate at Santa Maria). The levels of available P were analyzed by anion exchange resin, as well as P immobilized in the soil microbial biomass, P extracted by 0.1 mol L-1 NaOH, total organic P, and total P. The available P content increased rapidly with the application of soluble phosphate, but at the end of the evaluation period it was equivalent to the Gafsa rock phosphate, which was similar to the control. The application of phosphate fertilizers on natural pastures with low P availability increased the importance of the labile inorganic P forms to the plants, making them less dependent on the mineralization of organic P fractions. The organic fractions, including microbial P, are not good indicators of the bioavailability of P in natural pastures under Alfisols, Entisols, and Ultisols in Southern Brazil fertilized with phosphates.
microbial phosphorus; phosphorus availability; grassland ecosystems; rock and soluble phosphate
