During 1945 and 1947, five regional coffee trials were established at experimental stations of the Instituto Agronômico located, respectively, in Campinas, Ribeirão Prêto, Pindorama, Mococa and Jaú comprising 25 selected entries, among which varieties, strains (mixture of seeds of several plants of a single progeny) and progenies of selected coffee plants. These trials were planted in three replications of a balanced square lattice 5x5, without any shade and three other replications under shade of Inga edulis. Temporary shade was provided in shaded plots by other shade tree species. During the first years, missing plants occurred more frequently in Ribeirão Prêto and Pindorama. In 1956, at the end of the first phase of these trials, it was observed that the number of misses was higher in Jaú and lower in Campinas. A larger number of misses always occurred in the three replications under shade. Progenies and strains of Sumatra, Laurina and Caturra cultivars and of the Bourbon Vermelho n. 44, presented the highest number of misses in all localities. The average height of the plants, taken as an index of vegetative growth, was lower in the shaded replications in the trial of Ribeirão Prêto. In all other localities, the average height was practically the same under both environments. It was also observed that most plants continued to grow in height by the end of the first phase of the experiments (12 years old). In the shaded plots, most plants had a more reduced secondary and terciary branching, which probably contributed to the lower total yield. The yield, in kilograms of ripe fruits, was a determined for each plot during this period of time. However, in order to simplify the discussion, only the total yields of the 9 years for Campinas and Pindorama and 8 years of consecutive yields for the other localities, were analysed in detail (Tables 5, 10, 15, 20 and 25). In all five trials, the replications without shade gave higher yields. In Campinas they produced 82 per cent more than the shaded plots; in Pindorama 349%; in Ribeirão Prêto 1271%; in Mococa 69%; and in Jaú, 305% more. These differences were still more striking for certain annual yields. Differences in the amount of plant misses, of luminosity and of available water in the soil were suggested to account for these high yield differences. In general, it was observed, that the highest yielding progenies in the replications without shade, had the same behavior under shade. Ten entries in the trial of Campinas, two in Pindorama, six in Ribeirão Prêto, three in Mococa and two in -aú, of the unshaded plots, gave higher yields than the check (Bourbon Vermelho strain n. LC 370). In the shaded replications two entries in Campinas, five in Pindorama, seven in Ribeirão Prêto, two in Mococa and only one in Jaú, surpassed the check in yield. The outstanding yields of Bourbon Amarelo in Mococa and Jaú trial confirmed the risults previously obtained that this cultivar gives highery yields than the selected progenies of Bourbon Vermelho. The Bourbon Amarelo was not included in the trials established in Campinas and in Pindorama. Comparisons of annual yielding variances for replications with and without shade in the Campinas and Mococa trials, revealed not to be significant. The variances of replications without shade in Pindorama and Jaú were a little higher, a finding which probably is related to the higher yield of the unshaded plots in these localities. An analyses of total yields for periods of two, four, six and eight consecutive years indicated that most of the outstanding entries could have been selected, based only on the first four yielding years, in both environments (Tables 9, 14, 19, 24 and 29). Some progenies revealed to be early yielders, as Caturra and Bourbon Vermelho n. C 43, while others are late yielding as the Bourbon Amarelo from Mococa and Jaú, Sumatra n. C 408 from Campinas and the progenies of Bourbon Vermelho n. C 376-1. C 662 and LC 493 in various localities. The outturn (relations of cherry weight to green coffee, Table 37), revealed to be better tor replications grown under shade. The laurina variety had the lowest outturn whereas a good one was found in the maragogipe. The lower laurina outturn seems to be a dominant characteristic. Fruit setting (Table 30) was higher in the shaded replications of Campinas, Mococa and Jaú, and lower in Ribeirão Prêto and Mococa. No correlation was found between percentage of fruit setting and total yield. The laurina variety, which presented the highest fruit setting, revealed to be one of the poorest yielding entries in all five regional trials. The law fruit setting of the maragogipe variety seems also to be a dominant characteristic. The average cherry weight (Table 36) was practically the same for the entries in both environments, while the average weight of flat beans (Table 38) was a litle higher under shade, indicating a lower weight of the fruit pericarp. These results are in agreement with the better outturn observed in the shaded plots. The shade had only a very slight influence on the proportion of flat, peaberry and elephant beans, when compared with the observations made in unshaded plots The same occured with the amount of empty seeds, bean size and density of the flat green coffee beans. (Tables, 39. 40, 41, 42 and 43). Data on infestation of ripe fruits with Hypothenemus hampei from the Campinas trial have pointed out a higher incidence in the shaded plots (Table 39), confirming the results of previous investigations on this field. Analysis of the process of fruit ripening, made on fruits derived from flowers opened on a single day, failed to reveal any difference between shaded and unshaded plots. However, when the amount of green fruits derived from different blossomings was determined at various intervals, showed a tendency for earlier ripening in the unshaded plots. This may be the result of differences in the intensity of blossoming during the season, between both environments (Tables 31, 32, 33, 34 and 35). Cup quality was also considered. The results suggest that a higher acidity occurs in the samples derived from the shaded plots. No consistent differences were noted in relation to the other characteristics of the beverage. The results of these trialds revealed that mesides the drastic reduction in yield in the shaded plots, no other character was particularly affected by the shade. They also indicated that the most promising coffee strains in the unshaded plots give also high yields under shade, which is of particular interest for coffee improvement projects, carried on in areas where shading is a common procedure.