FAMILY-SITE INTERACTIONS IN RADIATA PINE FAMILIES IN NEW-SOUTH-WALES, AUSTRALIA

The magnitude and practical importance of family x site interactions for growth and form traits in radiata pine (Pinus radiata D. DON) breeding in New South Wales were analysed using several different techniques. Data were from 73 half-sib families in progeny tests. Considerable interactions for all traits were indicated by combined analyses of variance and by rank changes between sites. The ratio of interaction to family variance component showed the interactions to be of little practical importance, except for stem straightness. Calculated losses in potential genetic gain in a breeding population were found to be of concern for all traits under individual selection but only for stem straightness under family selection. Deletion of the eight most interactive families reduced these losses considerably, indicating that this would be an effective strategy in selection for stability in performance across sites. In practice families should preferably be selected for average performance across the range of trial sites, with some interactive but overall superior families retained. Genetic correlations between the same trait at different sites were moderate to high (0.62 to 1.00), and generally lowest for stem straightness. The greatest genetic gains were generally expected from selecting and planting families on the same site, but relative efficiencies of selecting and planting on different sites were mostly high. These and the other analyses of the practical importance of the interactions do not give substantial support to a strategy of regionalising breeding operations in New South Wales, particularly where diameter is the main selection trait.