The aim of the study was to increase knowledge on differences in mechanical properties of Norway sawn timber from selected log procurement regions in Finland and north-western Russia and to study the prediction of bending strength and modulus of elasticity, which are important properties in the construction uses of sawn timber. Support to the optimal sourcing of logs for structural products of sawmills as well as for the development of manufacturing value added products by improving sorting and grading of logs and sawn timber was targeted for, the sorting being based on log measurements, sawn timber measurements or both together.

Spruce logs and centre yield sawn pieces (N=1.162) were collected from five geographic growing regions, three from Finland and two from Russia. The levels and statistically significant differences in local longitudinal modulus of elasticity (E₁₂) and bending strength (f_m,12) of sawn timber were determined, along with several physical characteristics of sawn timber and logs. Multiple linear regression models were calculated to predict E₁₂ and f_m,12 using sawn timber and log properties as predictors, and it was analysed if any regional differences remained thereafter.

Generally, the levels of f_m,12 were higher in Finland than in Russia, being highest in northern Finland, albeit the large variation in that region. For E₁₂, the variation between countries was not so clear, but more fertile growing conditions seemed to provide sawn timber with lower E₁₂. Air-dry density (12% MC), knot area ratio (KAR) and annual ring width (RW) were the best predictors for E₁₂, and E₁₂, KAR and RW for f_m,12. If only log properties, only sawn timber properties or log and sawn timber properties together were considered, 40%, 70% or 57% of the variation in f_m,12 could be accounted. The results of this study were mainly in line with the previous studies concerning the levels of bending properties for spruce sawn timber, but the geographic variation was more obvious, systematic and larger.

Key words: Bending strength, model, modulus of elasticity, multiple linear regression, Picea abies, sawn timber