We performed the analysis of variance and covariance on height, branch length, height to branch length ratio, and needle length measurements from thirty open-pollinated families of Pinus sylvestris L. from Siberia, Russia. This progeny trial was replicated on three sites in central Alberta, Canada. At six years from seeds, there was statistically significant variation for height, branch length to height ratio and needle length. On individual test sites, values of individual-tree heritability for height ranged from 0.20 to 0.35, whereas values of heritability for family means ranged from 0.41 to 0.59. These low to moderate heritabilities suggest that a combination of family and within-family selection would be effective in improving height growth for this population on individual sites. Across sites, values, of individual-tree. heritability for height ranged from 0.03 to 0.06, whereas values for heritability of family means ranged from 0.17 to 0.29. These low heritabilities across sites were due to high genotype by environment (GE) interaction. Analysis showed that 87% to 99% of the GE interaction was due to lack of genetic correlation among sites. Heterogeneity of the genetic variance among test sites contributed 1% to 13% of the GE interaction. This shows that families in this population of Scots pine are not broadly adapted and are therefore suitable only for a site-specific breeding programme. The paper also presents and discusses results of other traits with emphasis to breeding Scots pine for production of Christmas trees, which is the main use of this species in North America.
