Repetibilidade e seleção em genótipos do gênero Urochloa spp.

Abstract

The objective was to verify the feasibility of direct, indirect, simultaneous selection and to estimate the repeatability coefficients and genetic parameters of genotypes of the genus Urochloa ssp. under semiarid conditions. The design was in randomized blocks with three replications and twenty seven treatments. Twenty-five genotypes of the genus Urochloa and two controls (Urochloa brizantha cv. Marandu and Convert HD-364) were evaluated in seven harvests. Repeatability was estimated by analysis of variance (ANOVA), covariance matrix-based principal components (CPCOV), correlation matrix-based principal components (CPCOR), AECOR structural correlation analysis and Mixed Models. The genotypes were selected by the direct and indirect selection methods and by the simultaneous selection index of Mulamba and Mock. The characteristic total dry mass MST presented repeatability values between 0.1370 and 0.3421 by the Mixed Models and CPCOR methodologies. The% FOL leaf percentage characteristic presented values between 0.0950 and 0.2993, also for the Mixed Models and CPCOR methodologies. The CPCOV methodology presented higher repeatability values for most of the evaluated characteristics except for plant height, where the principal component methodology based on correlation matrix provided higher values (0.4422). Genotypic variance components were significant for all evaluated agronomic characteristics. Mean heritability values ranging from 0.47 for leaf: stem ratio (RFC) to 0.78 for height (ALT) and regrowth (REB) were observed. Simultaneous selection resulted in more uniform gains in all evaluated characteristics, ranging from 3.76% for leaf percentage (% FOL) to 11.70 for REB. Urochloa spp genotypes have low repeatability estimates for agronomic characteristics by different methodologies. Direct selection results in high gains in the main trait and discrete or negative gains in indirect selection of the other traits. Simultaneous selection results in more uniform gains and allowed the identification of BARG171, BARG155, T1510, T1709 and BARG153 genotypes as being the most adapted to the study conditions.