Experiments were carried out in 1957, 1958, and 1959 to determine the influence of the ripening stage a:harvest on the beverage quality of coffee from the Paraiba Valley area. The coffee samples tested had the following origins: (1) pulped cherries picked from the tree; (2)same, unpulped; (3) dried fruits from the tree; (4) green fruits from the tree; (5) dried fruits picked from the ground. Two additional samples, (6) soft and (7) riada were introduced in the tests as standards. A well trained panel with ten members served as tasters. The results for the three years indicated differences between the samples that were statistically significant and consistent, except for minor discrepancies in 1957. The samples (1) pulped and (2) unpulped cherries gave always the best beverage and did not differ from the soft standard. The samples (3) dried fruits from the tree and (4) green fruits did not differ from one another in 1958 and 1959 and gave an average hard beverage. Fruits picked from the ground gave a beverage that differed from all treatments (except in 1957) that could be classed as the «rio» type.