This paper reports the results obtained in 13 experiments conducted in the State of São Paulo, to study the effect of increasing rates of phosphorus application in the absence and presence of nitrogen plus potassium. The nutrients used per hectare were: 20 kilograms of N, 30 kilograms of K2O, and 40, 80 and 120 kilograms of P2O5. respeetívely in the forms of Chilean nitrate, potassium chloride and superphosphate. These materials were applied in the seed furrows just before sowing. Apparently due to the application in contact with the seed, the nitrogen plus potassium fertilization impaired the stands in about 50% of the experiments, and its effect on the yield, in the average of this group, was only 1/3 of that obtained in the group, whose stands were not damaged. The average response to phosphorus was very good with the dose of 40 kilograms of P2O5 per hectare and increased appreciably with that of 80 kilograms, but no further increase was obtained with the 120 kilograms rate of application. The effect of phosphorus was better in the "terra-roxa" soil, decreasing in the "massapê-salmourão" and especially in the "arenito Bauru". In the "terra-roxa" it still increased when the rate of application was increased up to 120 kilograms per hectare. The presence of nitrogen and potassium contributed to enhance the response to phosphorus in the "terra-roxa" and "arenito Bauru" soils, but not in the "massapê--salmourão", apparently because those nutrients damaged the stands in all of the experiments conducted on the latter type of soil. According to the prices of 1956-57 and the average of the 13 experiments the most profitable rates of P2O5 applications would be 50 and 60 kilograms per hectare, respectively n the absence and in the presence of nitrogen and potassium. These doses should be ncreased for the "terra-roxa" and reduced for the "arenito Bauru" soils. For the "massapê-salmourão" soil, in view of what happened to the stands the dose to be used in the presence of nitrogen plus potassium cannot be estimated; in the absence of these nutrients it should be of 50 kilograms. The results of limited number of experiments indicate the possibility of reducing the doses of phosphorus for soils fertilized with it in the previous crops.