Results from genetic tests of selected parents of Douglas- fir (Pseudotsuga menziesii [Mirb.] Franco) in an applied tree improvement program

Abstract

Results from genetic tests, and genotype by environment interaction studies in six low-elevation breeding zones of Weyerhaeuser Company's Western Washington and Oregon Douglas-fir tree improvement program are summarized. Phenotypic selection in natural stands resulted in a 5% improvement in juvenile height over nonselect seed lots. Comparisons with nonselect sources, of offspring from parents in the top 50% performers in tests, indicated a 10% increase for the same trait. Seed produced in a 50% rogued seed orchard is thus expected to provide improved planting stock with a gain of 10% in juvenile height growth. Several select parents are producing offspring that are consistently performing in excess of 10% over nonselects. Estimates of breeding zone, breeding zone by location, and family by location interaction effects are small relative to family and planting location effects. Tests of families established on environmentally diverse sites indicate a striking lack of large family by planting location interaction. In tests showing statistically significant interactions, such interactions are caused by a relatively small number of families. We did observe earlier budbreak and a higher spring frost susceptibility of Oregon sources established on Washington sites. We conclude that allocation and utilization of select families within Weyerhaeuser's Oregon and Washington ownership should not be constrained by the currently defined breeding zone boundaries, but based on parental performance and stability for growth and adaptive traits in general. We further suggest that the necessity for maintaining separate breeding zones, within Washington and Oregon, in subsequent cycles of recurrent selection is questionable. Average individual tree heritability, from 65 6- parent disconnected diallels, is 0.13 for age 6 and age 8 height. Dominance genetic variance is estimated to be one-half that of additive genetic variance for the same traits. The environmental designs and test establishment methods used resulted in acceptable levels of statistical precision for juvenile growth traits. Individual test least significant range values (LSR) resulted in significant family mean percent height differences of 8 to 26, and averaged 13% for diallels, and full-sib and half- sib tests.