Relations entre caractéristiques de retrait et densité du bois chez le pin sylvestre, le sapin pectiné et l'épicéa commun

Abstract

— Des relations entre les caractéristiques de retrait du bois et son infradensité sont re-cherchées pour le sapin pectiné, l'épicéa commun et le pin sylvestre. Les échantillons utilisés sont des carottes de sondage de 5 mm de diamètre, prélevées sur un ensemble de 928 arbres. Les relations obtenues sont significatives mais restent cependant faibles. Le retrait radial est toujours mieux lié à l'infradensité que le retrait tangentiel. L'introduction d'autres variables telles que la largeur moyenne de cerne, l'âge et la circonférence de l'arbre, n'améliore pas de manière significative les relations mises en évidence. Pinus sylvestris / Abies alba / Picea abies / retrait du bois / infradensité Summary — Relationships between wood shrinkage properties and wood density for Scots pine, silver fir and Norway spruce. Relationships between basic wood density, tree circumference, age and wood shrinkage properties were investigated. Increment cores were used from 928 trees; the number of trees sampled for each species is given in table I. The location of the stands in the "Vosges" department is given in figure 1. The shrinkage coefficients, representing the dimensional variation of wood when its moisture content decreases by 1%, were calculated. The compression wood was taken into account as it influences the shrinkage coefficient values (table II). The results of the 3 species were compared (table III): pine wood had the highest values; the values for fir and spruce were identical. The relationship between the different shrinkage coefficients was also investigated (table V). In order to study shrinkage variability, a step-by-step multiple regression was used, with basic density, ring width, age and tree circumference as the parameters. The results are given in table VI and can be summarized as follows: 1) For all species, basic density is the variable which correlates best with all the shrinkage coefficients. Therefore basic density is always the 1st variable in the regression equation; 2) The radial shrinkage coefficient is better correlated with basic density than with volumetric or tangential shrinkage; 3) The other parameters (ring width, age and tree circumference) contribute little to the explanation of shrinkage variability. The correlations between the shrinkage coefficients and basic density are low, especially for fir wood. Figures 3, 4 and 5 show these correlations (each shrinkage coefficient as a function of basic density). All the results are discussed with reference to the literature. Pinus sylvestris /Abies alba / Picea abies / wood shrinkage / wood density * Correspondance et tirés à part.