POPULATION-DEPENDENT SELECTION-STRATEGIES NEEDED FOR 2-YEAR-OLD BLACKCOTTONWOOD CLONES

We conducted a study to determine whether genotypic and phenotypic cov ariances among important traits differed between collections of black cottonwood (Populus trichocarpa Torr. and Gray) from British Columbia and northern Idaho. We also determined how differences in estimated co variances could affect various multiple-trait clonal selection strateg ies. The experimental design consisted of 166 clones from British Colu mbia and northern Idaho intermixed at random in two replications of th ree-tree row plots. Measurements made during the second growing season included tree height, frequency and number of sylleptic branches, num ber of leaves, and the lengths and areas of leaf plastochron index 5, 10, and 15 leaves at the end of the growing season. The severity and e xtent of Melampsora leaf rust and Septoria leaf spot were also evaluat ed. We found high heritability for tree height phenology, leaf morphol ogy, and resistance to damaging agents. However, no advantage to using leaf morphology to improve genetic selection for tree height was demo nstrable. Restricted selection indices had high potential utility for controlling increases in damaging agents. However, the relation betwee n tree height and leaf diseases varied between the British Columbia an d northern Idaho collections of black cottonwood, and between black co ttonwood and an adjacent experimental population of balsam poplar (Pop ulus balsamifera L.). Such population-related variation in intertrait covariances suggests that selection strategies involving disease resis tance may need to be evaluated on an individual population basis.