In order to compare the effect of ammonium sulfate and urea, when sidedressed and incorporated just after thinning, on cotton cv IAC 20, sixteen field experiments were conducted in different soils of the State of Silo Paulo, Brazil. The experiments were carried out during the growing seasons of 1984/85 to 1986/87. Rates of 0, 25, 50, 75, and 100 kg/ha of N were used in a factorial design (2 x 5), with four replications in randomized blocks. Two extra treatments were added in some experiments, in which the N sources were sidedressed without incorporation at the rate of 75 kg/ha of N. The highest yields and the smallest responses of cotton to N application occurred in soils previously covered with old pastures or grown with legumes in rotation. The most profitable rate of nitrogen fertilization, calculated for groups of experiments based on expected yields, varied from 17 to 88 kg/ha of N, the highest rates for soils more intensively cultivated. Yield responses to N could be reasonably predicted in soils grouped according to P extracted by ion exchange resin and, in some way, with the inorganic-N content of the 0-60cm soil layer. Soils with high available P, long under cultivation, showed high responses to N application than soils with median P levels. Nitrogen sources or methods of fertilizer application did not affect cotton yield and total N content of leaves; however, total S concentration in leaves decreased linearly with the increase of urea application. Total N around 3.8% is suggested as the critical limit for N in cotton, determined in blades of the most recently matured leaf, collected from the main stem, 80 to 90 days after planting.
cotton; nitrogen fertilizers; method of fertilizer application; soil and foliar analysis