Repeatability estimates for fruit quality characters of sweet orange

The objective of this work was to determine the repeatability coefficient of fruit quality characteristics of sweet orange (Citrus sinensis) and the minimum number of evaluations that can provide levels of prediction certainty of the true genotype value. Thirty nine genotypes of sweet orange, collected in nine municipalities of the state of Acre, Brazil, were evaluated in five crop cycles. A randomized complete block design was used, with three replicates. The following were evaluated: average fruit weight, juice yield, peel thickness, soluble solids (SS), total acidity (TA), SS/TA ratio, and technological index. To estimate the repeatability coefficients, the methods of analysis of variance, principal component analysis, and structural analysis were used. All evaluated characters showed variability, except for juice yield. The evaluated characters showed a cyclic pattern, which was best captured by the multivariate methodologies for estimating the repeatability coefficient. Fifteen evaluations are needed to determine, with 90% certainty, the characters peel thickness and soluble solids, and 11, 6, 3, 2, and 1 evaluations, respectively for average fruit weight, total acidity, technological index, juice yield, and SS/TA ratio, respectively.