Selection of sugarcane families and parents by Reml/Blup

Abstract

Were evaluated 113 full-sib families obtained by unbalanced diallel matings in three experiments in an augmented block design (ABD) planted side by side in the same area. The individual and joint analyses of the experiments were performed by the Reml/Blup method. The use of the ABD without replication did not prove adequate in experiments of family selection owing to the low heritability estimate at the level of family means in comparison to the joint analysis of the three experiments. The results presented predominance of the additive effects for all evaluated traits: number of stalks, tons of stalks per hectare and mean stalk weight. The Components of estimated means via BLUP allowed the selection of families and superior parents.