Finalizing the selection of parents for the Sitka spruce (Picea sitchensis (Bong.) Carr) breeding population in Britain using Mixed Model Analysis

Abstract

The Best Linear Unbiased Predictor method was used to derive breeding values of over 1000 Sitka spruce (Picea sitchensis (Bong.) Carr) plus trees based on data from 69 half-sibling progeny tests. Roughly half the progeny was derived from open pollination of the tree standing in the forest. The balance followed controlled pollination of female flowers on grafted copies of plus trees growing in clonal achives with a polymix (PO) of selected pollens. It was necessary to adjust for the selected nature of the PO pollens during analysis. The breeding objective was to maximize gain for stem diameter and stem form while avoiding a fall in wood density. The study investigated numerous economic weightings attached to the three selection traits and used a multi-trait selection index method to rank parents. Sub-populations of the best 120 parents required to complete the breeding population, and the best 40 parents required for new tested clonal orchards were reselected and tested against the breeding objective. The strong inverse relationship between stem diameter and wood density remains the main challenge to meeting the breeding objective. The study demonstrates, however, that provided sufficient plus trees are initially selected, it is possible to meet the breeding objective despite this difficult inter-trait relationship.