Liming of coffee crop is based on soil analysis of samples from the projection of the crown, which is the most acidic part of the soil due to nitrogen fertilization. This procedure is doubt ful since there are less acidic areas of the soil, mainly in the mid rows, due to the accumulation of bases promoted by squaring. Another problem of coffee management is the routine use of NPK formulas with high N and K contents, independently of the actual need of the crop. In this paper the effects of liming, nitrogen and potassium fertilization were studied in two plantations that had received normal application of these insumes. A fractional factorial design 1/2 (4 x 4 x 4) was used, with the following amounts of applied treatments: limestone - 400, 1,600, 3,600 and 6,400 kg/ha; nitrogen - 64, 121, 196 and 289 g/hole; potassium (K2O) - 36, 81, 144 and 225 g/hole. The experiments were performed in Garça, State of São Paulo, Brazil, on a lins-and-marília-podzolized soil. Five crops were recorded from 1987 to 1992. Nitrogen produced the largest yield increases, although not very high, followed by liming; potassium did not affect yields. In the case of liming, low responses were obtained in spite of the fact that the observed base saturation values were much lower than the recommended 70%. Leaf nitrogen contents were rather high, which explained the fair response of the coffee tree to nitrogen fertilization. Medium to high levels of potassium were found in the 0-20 cm soil layer and medium contents in the 20-40 and 40-60 layers, which explained the lack of response to this nutrient. It can be concluded that it is necessary to review the actual base saturation needs for liming coffee. Nitrogen fertilization can be based on leaf and potassium on soil analyses.
coffee; liming; fertilization; nitrogen; potassium; soil analysis; leave analysis