Additive and Non-Additive Genetic Variances for Tree Growth in Several Hybrid Poplar Populations and Implications Regarding Breeding Strategy

Populus species (P. deltoides, P maximowiczii, P. nigra) and their inter-specific hybrids were tested for growth rate over a five year period at four test locations in Minnesota, USA, to estimate genetic variance components. The breeding scheme incorporated recurrent selection of full-sib families of pure species parents, production of F, inter-specific hybrids from selected families, and selection of clones within the F,s. Improvement of yield through time using this scheme is predicated on the assumption that additive effects comprise a significant portion of the total genetic variance. The estimates of additive and non-additive variances reported are not traditional point estimates, because a fully balanced mating design was impossible due to parental incompatibilities which result in incomplete breeding matrices. Instead, bounded estimates, not previously used in tree genetics research, are derived from linear combinations of formulae of genetic expectations observed among-family, among-clone, and environmental variances. Our results suggest that combined family and mass selection would lead to increases in growth rate of 27 % and 47 % per generation in P. deltoides and P. nigra, respectively. Broad sense-based clonal selection within the F, could yield selection responses in excess of 90 % of the mean of such populations. Among-family variance comprised about 1/3 of total genetic variance while within-family variance was always about 2/3 of total genetic variance, regardless of pedigree.The results indicate that recurrent intraspecific selective breeding followed by interspecific hybridization and non-recurrent selection based on broad sense genetic variation would constitute an effective yield improvement strategy.