Repetibilidade e número de colheitas para seleção de clones de cana-de-açúcar

This work aimed at the establishment of the repeatability coefficients (r), determination of the predictability degree (R2) and the needed measurement numbers for tons of cane per hectare, percentage of pol (sucrose) in the juice of cane and tons of pol in the juice of cane per hectare in sugarcane genotypes. A randomized block experimental design, with three replicates in four experiments and four replicates in the other experiments was used. The repeatability estimates were obtained by the statistical methods: anova; main components based on correlation matrix; and structural analysis (correlation, average r). The estimates of the repeatability coefficients showed values with very similar magnitude. The general average repeatability for those three traits was above 0.60, therefore showing regularity in the genotype performance at several measurements (cuts) and reliability in the genotype discrimination that was higher than 87%. The results showed that for those three traits there is a need for the use of, at least, three cuts so that the selection can be accomplished with predictability of a real value above 80% for the genotype.