This paper reports and discusses results refering to a field trial set up to study irrigation and fertilization problems of old coffee groves. The seven year study, started 1953, was carried at a representative well coffee grove of Bourbon variety, about 35 years old, located in a typical sandy soil of the Bauru superior formation of the Pindorama Experiment Station, considered to well represent on extensive port of that formation where the bulk of coffee is at present produced in the State of São Paulo. Eight randomized blocks were used with a 2x2x2 factorial for the small treatments in each block, the following treatments being compared: complete chemical fertilizer in two dosages, with and without organic manure, with and without green manure as an annual intercrop. The irrigation treatments, the first two years, included application of 1 and 2 inches of water every three weeks, starting July, 2 inches storting August and no irrigation. The remaining 5 years the treatments compared were simply with irrigation and no irrigation. Irrigation application time was then mostly determined by soil moisture depletion, when about 50 millimeters were consumed from the 0-80 centimeter soil layer. Irrigation effect was significant only one year, with substantial increase in yield. The first two years irrigation apparently decreased yields. However, this result can be attributed to the effect of insufficient number of replications and to the fact that the blocks with no irrigation were benefited by their particular location in the field, where soil fertility was higher. The remaining years, although a balanced re-sampling was made to double the replications by reducing the irrigation treatments, the climate was generally more favorable to coffee production, thus minimizing the effect of irrigation. The use of green manure as an annual intercrop was shown to significantly decrease yields, in spite of the fact that seeds were inoculated and there was a basic substract of chemical fertilizer in all plots. The use of organic manure in the dosage of about 12 kg every year per coffee tree (which means o group of 3-6 trees originally planted in the some hole) did not increase yields. In no year the effect of double dosage of fertilizer was significant. The average basic dosage used was the following per tree per year: 311 g of ammonium sulphate (20,5% N), 162 g of calcium superphosphate (20% P2O5) and 145 g of potassium chloride (60% K2O). When yields of the treated plots were compared with yields of untreated border plot trees, it was shown an increase of about 38% due to the effect of chemical fertilizer alone. However, the present coffee economical situation in Brasil does not favor the use of fertilizers on o profit basis on the old coffee plantations of the type reported. Considering the advantages of forming new coffee groves according to modern techniques, where there was before pasture lond, a practice which is sure to grow in importance, it would be outdated to recommend expenditures to try to improve old coffee plantations. The irrigation equipmente used for the experiment and part of the laboratory equipment needed was acquired through a donation of the Rockefeller Foundation,, to which we owe due recognition.