Incorporating wood density in breeding programs for softwoods in Europe : A strategy and associated methods

In this paper, we introduce a strategy for incorporating wood density into breeding programs in Europe. We develop some related methods, and we present the preliminary results. We used covariance analysis to study the negative relationship between volume growth and wood density and its genetic variation in two important softwood species in Europe: Norway spruce and Douglas-fir. For both species, the same proportion of the observed variability for ring density can be explained by growth traits and genetic origin (respectively 54% and 57%). But the most explanatory factors are very different among the species: in Norway spruce, the ring width variable explains most part of the density variation, while in Douglas-fir the weight of genetic variability of ring density itself is much bigger. To explore such a relationship, we used a more refined description of the ring density profile than the usual earlywood-latewood model. This method was based on wavelets decomposition. Preliminary results suggest that a part of the within-ring density variation not accounted by the earlywood-latewood model is of genetic origin. Wavelets decomposition was also used to quantify a trait often quoted, wood homogeneity. High coefficients of variation at the between-clone level for both species let expect a possible high relative genetic gain for wood homogeneity.