Optimization of genetic gain in different variables for Eucalyptus grandis Hill ex Maiden breeding

This study aimed to optimize the genetic gain for initial growth, low rust infection, and wood quality in an open pollination progeny test of E. grandis. The progeny test was established in Anhembi-SP, in a randomized block design, with 176 progenies, single tree plots, 29 replications, and one external borderline. Three evaluations were carried out. For the variable related to growth, diameter at breast height (DBH); evaluation at 20 months of age; for the degree of rust infection (DRI) the evaluations were carried out at 9 months of age, and for wood quality the variable number of branches (NB) was evaluated at 14 months. The choice of these variables considered the estimates of heritability and genetic correlations. Eight selection methods were used to optimize genetic gains for each variable, such as indirect selection, independent levels, and combined selection indexes with equal and different weights for each variable. The breeding strategies were based on the formation of Seedling Seed Orchards (SSO) and Clonal Seed Orchards (CSO). Based on the results, the indirect selection was not recommended because genetic correlations between DBH, DRI, and NB were weak (< 0.50), and it was necessary to consider the three variables for the construction of selection indexes and independent levels. For both breeding strategies the independent selection levels were the best in optimizing simultaneous selection for growth, rust resistance, and wood quality; with gains for DBH, DRI, and NB of 5.83%, 13.88%, and 4.12%, respectively for the formation of SSOs; while gains of 7.23%, 18.76%, and 6.72% were calculated respectively for the formation of CSOs.