Relative growth trend as an early selection parameter in a Douglas-fir provenance test

Abstract

Height data from 89 Douglas-fir provenances planted at seven sites in northern Spain were used to explore and discuss the utility of the relative height growth trend as an early selection parameter. Total height was measured at each site at different ages between 2 and 18 years after planting. A modification of the joint regression analysis was used to analyze and interpret the provenance x age interaction at each site. The analyses of variance showed a significant provenance x age interaction in five out of the seven sites. Most of this interaction at these sites was explained by the linear relation over years between the provenance mean height at each site and the overall site mean height. The relative growth trend was defined as the slope of these linear relations for each provenance and was considered as a temporal stability parameter. Both the high genetic variability and the biological significance of this parameter suggest its applicability in early selection. A strong linear relationship was observed between this growth parameter and the initial provenance mean height at all sites where provenance x age interaction was significant except one. This relationship indicates that the differences between provenances at an early stage will hold up and be amplified in future measurements, and then there is a relative security of early selection. One site showed no significant relation between the relative growth trend and the initial performance. At this site, rank changes among measurements are likely to be more frequent, and early selection becomes harder. The use of the relative growth trend as an early selection parameter in these sites improves the early selection efficiency. The results suggest that the relative growth trend can help in both the early selection and the evaluation of the early selection efficiency.