In order to study the effects of fertilizers on sweet potato, in the conditions of the State of São Paulo, a series of experiments was undertaken to cover the different aspects of the problem. In this paper, results are presented only for 31 experiments planned to study the effects of N, P and K on root development, yield, number of potatoes and mean weight, in the principal types of soil in the State. The results of the other experiments will be considered in further papers. Various plans and experimental designs were used. As a source of mineral elements, the following commercial fertilisers were employed : Chile saltpetre, sulphate of ammonia, superphosphate, degelatinised bone meal, potassium chloride and potasssium sulphate. The formulas for each plot were made up in advance and the mixtures broadcast over the soil before trenching. It appeared that the natural fertility of the soil and other agricultural factors had much more influence on the production of sweet potato than had the application of fertilizers. Substancial responses to fertilizer were generally observed only in soils of very low fertility. In these cases, even although the percentage increases in yield were considerable, the absolute increases (often less than 1.1 tons per acre) were insufficient to pay for the application of fertilizar. Among those trials producing satisfactory yields (over 4 tons per acre), only one showed a response to nitrogen. This was in Mococa on "massapê-salmourão" soil. Responses to phosphorus or potassium were noted only on soils of "glacial" origen in the "Sorocaba" region. A single case of nitrogen-phosphorus interaction was observed on "Arenito de Bauru" in Pindorama. Whenever the development of vines showed a response, this was always due to nitrogen. This was generally found on soils of "glacial" origim. It must be pointed out however that no correlation was found between vine development and yield. Other data, such as percentage of failures, mean weight of potato, etc., failed to show significant response to manuring. One factor which showed a strong influence on yield and growth of sweet potato, was crop-rotation. It was found that, in those trials where the plant was cultivated continuously, the yield fell drastically in subsequent years, often to a third or quarter. The results obtained show that, for conditions prevailing in the State of São Paulo, the direct application of mineral fertilizers to sweet potato is not usually an economic proposition. This plant should be grown in rotation with more exhausting crops, which demarnd heavy dressings. The sweet potato can then benefit from the residual effects of these dressings in the following year.