Determining the optimal age for selection by modelling the age-related trends in genetic parameters in Eucalyptus hybrid populations

Ten factorial mating designs using a combined total of 88 females, 107 males, 684 families and 37,206 individual trees were used to model the age-related trends in genetic parameters and genetic gain between four and 65 months in the Eucalyptus urophylla x grandis breeding population in Republic of Congo. Selection was either of pure species (as parents for continued breeding) or individual hybrids for commercial plantations based on clonal varieties. The variance components were significantly different from zero for female, male and female-by-male interaction effects for volume. The age-related trends in additive, dominance and environmental variances, modelled by nonlinear functions, showed three phases corresponding to different stages of competition and growth. Male and female narrow sense heritabilities were high (h2A. = 0.70 and h2M = 0.90, respectively for highest estimates) compared with individual broad sense heritability (h2. d= 0.45). They were modelled by polynomial func- tions teat did not display specific trends with age. The age-age correlations, modelled by a response surface, were higher than 0.8 after 36 months. A similar trend with age was observed for additive and total genetic effects. Genetic gain was calculated by combining the different models. The genetic gain was higher for female than for male. Considerable gains can be achieved by clone selection. The efficiency of selection indicated an optimal age of 54 months for juvenile selection of males and females and a mature age for ortets. The trend in efficiency of selection per time unit showed that juvenile selection for volume is much more efficient than adult selection whatever the age. (Résumé d'auteur