This paper reports the results obtained in six preliminary experiments conducted between 1941-42 and 1950-51 in the State of São Paulo, to study the effect of moderate rates of nitrogen, phosphorus and potassium on the yield of castor beans. The response to nitrogen was satisfactory in three tests, but only in one it was statistically significant. Phosphorus also promoted satisfactory responses, in three experiments, but they did not reach the level of significance. The effect of potassium was negative or very poor in all the trials. Some of the experiments were located on soils apparently well provided with one or more of the nutrients tested. Other cases of poor results are partly attributed to the application of the fertilizers in the traditional method, which consists in adding them (inclusive the whole dose of nitrogen) to the furrows just before planting and mixing them slightly with the soil. Contributed also for the poor results obtained: a) two of the experiments were planted too late, and b) the plant density used in all of them was much lower than the optimum as learned in later investigations.