Four fertilizer experiments were planted at the Usina Tamôio, Araraquara, São Paulo. Two of the trials were located in the "terra roxa misturada" type of soil (Sections Serra D'Água and Sede) ; one in sandy soil (Mantuana Section), and one in "terra roxa legítima" (Bela Vista Section). A 2x4x4 factorial design was used in the four tests, and the fertilizer levels per hectare were as follows : Nitrogen, 0 and 30 kg of N ; Phosphorus and Potassium, 0, 30, 60 and 90 kg of P2O5 and K2O, respectively. Nitrogen was used as cotton seed meal, and phosphorus and potassium were employed as superphosphate and potassium chloride. The fertilizers were added to the furrows before planting, and the variety used was Co. 290. Harvesting was done after 18 months. The statistical analysis of the data from the four trials gave the following indications : a) Nitrogen applied as cotton seed meal at, the rate of 30 kg of N per hectare prior to planting gave a significant increase in yield in the three types of soil; b) yield increases due to phosphorus application were recorded in three'of the tests; the rate of 60 kg/ha of P2O5 is indicated for the Bela Vista Section, whereas for the Serra D'Água and Mantuana Sections the rate of 90 kg/ha was better; c) Potassium should be employed at the rate of 60 kg/ha in the Mantuana Section, and at the rate of 90 kg/ha in the Sede Section.