The objective of this work was to evaluate different sources, doses and methods of phosphorus application in the recovery of Brachiaria decumbens. An experiment was installed on the "Caminho das Pedras" Farm, in Uberlândia-MG, in Acroustox soil, medium texture, between November 2003 and April 2004. The treatments were three phosphorus sources (Triple Superphosphate, magnesium Thermophosphate and Gafsa hiperphosphate) in the doses of 0; 50; 100, and 150 kg ha-1 of P2O5, incorporating or not incorporating, in three repetitions, in a randomized blocks experimental design. The plots were of 2.0 x 5.0 m, that were fertilizer evenly with 330 kg ha-1 of N and 220 kg ha-1 K2O, using urea + ammonium sulphate and potassium chloride, respectively as fertilizer sources. The productivity of Brachiaria was evaluated in three cuts of the shoot, in an interval of 30 days between cuts. Phosphorus application with different sources and doses promoted increment in the dry matter production of Brachiaria, however, there was no significant difference in production between application methods. The greatest dry matter productivity was obtained with 150 kg ha-1 of P2O5 using triple superphosphate with an average productivity of 2997 kg ha-1 of dry matter per cut. The smallest productivity was obtained applying 50 kg ha-1 of P2O5 using the Gafsa hiperphosphate. With increased phosphorus doses, to all the sources, there was increased phosphorus in the leaf and in the soil, to both extractors. In the extractors evaluation, Mehlich-1 presented highest extraction for the source Gafsa hiperphosphate, not corroborating with the production results.
Phosphate fertilizer; degraded pasture; phosphates
