Seven fertilizer experiments were laid out in 1960-61 on soils of the «arenito Bauru» type, to study the effects of several treatments on the yields of dry beans. For determining the residual effect of the treatments, three of the experiments were repeated in 1961-62. In the average of the experiments of 1960-61, the responses to 60 and 120 kg/ha (kilograms per hectare) of P2O5 as ordinary superphosphate were +364 and +574 kg/ha of seeds and those due to 30 and 60 kg/ha of nitrogen as ammonium sulfate, +177 and +212 kg/ha, whereas those due to 45 and 90 kg/ha of K2O as potassium sulfate were -5 and -56 kg/ha. The average residual effect of phosphorus corresponded to +151 kg/ha and that of potassium was again negative. The addition of a mixture containing gypsum and micronutrients (Zn, Cu, B and Mo) to a NPK fertilizer practically sulfur-free increased the yields in all of the 1960-61 experiments, the average increase reaching 331 kg/ha. Although applied in the presence of a complete fertilizer, lime depressed the yields in many experiments and its average effect was practically null. Apparently, the rate of application, of a 4 metric tons per hectare, was excessive for the soils studied, which were low in clay and organic matter. While the 1960-61 average yield of the treatments without phosphorus and nitrogen corresponded to 444 kg/ha, that of the treatments which received 30 kg/ha of N as ammonium sulfate and 120 kg/ha of P2O5 as ordinary superphosphate increased to 1,307 kg/ha.