Indirect and Direct Selection for Yield in Coffea canephora Genotypes Under Drought-Stress and Non-stress Conditions

Abstract

It is a normal practice to select coffee in non-drought-stress environments for use in drought-stress and non-stress conditions. This study assesses the efficiencies of drought stress and non-stress selection environments to improve coffee for adaptability. In a nine-year experiment, 18 genotypes were tested simultaneously under drought-stress and non-stress conditions. Six traits related to Robusta coffee adaptation were evaluated to determine the best environment for improvement and traits amenable to selection for broad adaptation to drought stress. Drought-stress significantly reduced fruit set, fruit/node, number of primary branches, and yields. Span, girth, and diameter of primary branches were not reduced by stress. Genetic correlations between the environments were significant for all the traits (rG = 0.47–0.84) except fruit/node and number of primary branches. Selection of genotypes with broad adaptation to water-deficit stress should be possible through an index including span, girth, diameter of primary branches, fruit set, and yield itself. Lower genotypic and environmental variation coefficients but higher genetic variances and broad-sense heritabilities were observed under non-stress than under stress conditions. Correlated response to selection as against response to selection under stress (0.02–0.90) and non-stress (0.03–0.93) was lower than unity for all the traits. Hence, direct selection of genotypes is predicted to be more efficient in the environments in which they are intended to be used rather than in indirect selection. However, when resources are limited, it is better to select under non-stress conditions.