Resumo:
The purpose of this study was to verify the possibility of using direct selection in nine traits underlying indirect selection for yield and determine which traits should participate in the selection process. Data of 40 Conilon coffee genotypes were analyzed in two experiments in the growing seasons of 1996, 1998, 1999, 2000 and 2001 in random blocks with four and six replications. The significance of phenotypic associations was evaluated by the t test and the genotypic and environmental associations by bootstrap resampling. The genotypic associations were higher than the phenotypic, indicating a prevailing influence of the genotypic over the environmental effects in the relationship between significant traits; equal signs indicated a lack of contrary action among the effects. The traits related to cycle; yield; ratio of fresh ripe cherries to clean coffee; empty or flat grains; and sieve 17 should be maintained in the selection, evaluation and study of genetic divergence. The estimated gains in grain yield by indirect selection for any trait studied are not satisfactory.
