This paper reports the results obtained in three experiments conducted in the Mococa, Tietê and Tatui experiment stations comparing the usual method of application of nitrogen with its application as top-dressing. The former consists in the addition of the fertilizers (nitrogen, phosphorus and potassium) to the furrows just before planting, slightly mixing them with the soil. Nitrogen was used as Chilean nitrate and Calnitro IG. The experiments were carried out from 1937-38 to 1941-42, those at Mococa and Tietê during three and that at Tatui during two consecutive years. Calnitro and Chilean nitrate gave similar results throughout the three experiments. The two methods of application gave also equal results at Mococa, but at Tietê and Tatui nitrogen when applied by the usual method very little increased or even depressed the yield, whereas increased it appreciably when top-dressed. In Mococa the after sowing weather was humid, and no leaching rain occurred during the growing seasons. The poor results obtained with the usual method in Tietê were ascribed principally to the injury caused by the nitrogenous fertilizers to the sprouting seed; in one year of the Tatui experiment the fertilizers were leached before the plants had reached the stage of active nitrogen uptake, while the delayed application as top-dressing escaped the leaching rains. Leaching of nitrogen, which is more frequent when applied at planting time, was not of great influence in the reported experiments, probably because two of them (Mococa and Tietê) were conducted on heavy soils and the rains were relatively moderate. In lighter soils, as those used for cotton in large areas of the State of São Paulo, leaching as well as injury due to excessive salt concentration would have been of much greater importance and therefore the advantage of top-dressing over the usual method of application would have been still more pronounced.