Implication of genotype × site interaction on Pinus radiata breeding in Galicia

The magnitude and practical importance of family × site interactions for growth and form traits considered in radiata pine (Pinus radiata D.Don) breeding were analysed by four different methods: type B genetic correlation, regression coefficient, mean rank deviation and ecovalence. The material analysed consisted in growth, form and frost resistance assessments of 58 open pollinated half-sib families at 3 to 4 years old across three sites in Galicia (NW Spain). Analysis of variance revealed that family × site interactions were quantitatively important for all traits (diameter, volume, branch angle, straightness and frost resistance). The losses in potential genetic gain for a breeding population were important for all traits under individual selection but only for diameter and frost resistance under family selection. Omission of the ten most interactive families from the analysis considerably reduced the losses for all traits, although losses in potential genetic gain remained important for frost resistance under family selection. The results indicate that elimination of these families from breeding programmes would be an effective strategy for selecting for stability in performance across sites for all traits except frost resistance. In order to overcome this problem, selecting varieties for frost resistance specifically adapted to various regional contexts would be an option. However further studies under controlled conditions are required before making final decisions for radiata pine breeding programmes.