Effects of Different Data Transformation Methods on the Reduction of the Genetic, Environmental and Phenotypic Variance in the Progeny Trial of Eucalyptus grandis W. Hill ex Maiden

Abstract

This research work was developed in order to evaluate progeny trials of Eucalyptus grandis W. Hill ex Maiden using the software SELEGEN-REML/BLUP. The best trees were identified in order to be used in seedling and clonal orchards. Fifty three half-sib progenies of three Australian provenances were tested in the municipalities of Mogi Guaçu, Boa Esperança do Sul and Caçapava, all located in the State of São Paulo. A compacted families block experimental design was used with variable number of replicates, linear plots of six trees each, and a 3.00 x 2.00 m spacing. Two methods of data standardization or transformation were used in order to evaluate their efficiency in the reduction of the genetic, environmental and phenotypic variances. The transformation or  orrection of the data, performed with the ratio (hi/him) between the square root of heritability in locality i and the mean of square roots of heritability in each locality, presented higher efficiency than the non-correction as well as the data correction obtained with the phenotypic standard deviation ( f sˆ ) usually used. Additionally, results of joint analysis of data from different locations have confirmed that high magnitude gains can be obtained for diameter growth though the selection of individuals based on their genetic values. This can be achieved either by thinning of the trial, resulting in a Seedling Seed Orchard or by vegetative propagation of the best individuals for the establishment of a clonal seed orchard. From the results obtained for genetic correlation among locations, one single Selected Population and only one Clonal Seed Orchard can be established which represent all the three experimental locations, which in turn result investment savings. Genetic variability was achieved with the moderates values obtained from individual heritability, in the narrow sense, for growth in DBH in the three locations studied. The adoption of strategies and criteria proposed for the selection provides the composition of a selected population with two hundred individuals of high genetic value and an adequate effective number of progenies, producing gains for DBH between 12.89% and 24.33% in relation to the overall experimental average, for the establishment of a Seedling Seed Orchard. The selection of twenty individuals with the highest breeding values, for the establishment of a clonal seed orchard may provide gains for DBH between 17.18% and 50.95% in relation to the experimental average. On the other hand, the selection of the best twenty individuals, with the highest genotypic values for the establishment of a Seedling Seed Orchard, may provide gains for DBH between 22.40% and 82.16% in relation to the experimental average, for clonal plantations established with the selected material.