The chemical composition of cooperage oak wood is highly variable, depending upon the tree species (Quercus robur L. versus Quercus petraea Liebl.), its geographic location, and the single-tree effect. In the process of cask-making, natural seasoning and toasting contribute strongly to the modification of the oak wood chemical composition and therefore influence wine cooperaging. HPLC and GC quantification of ellagitannins and volatile compounds such as whiskey-lactones, eugenol, and vanillin over a sample set of 61 pedunculate oaks and 72 sessile oaks originating from six different forests showed that natural drying leads to a decrease of the ellagitannins and total extractives content level and a quasi constant level of the volatile compounds. Toasting (medium type) drastically enhanced the loss of ellagitannins and the gain in volatile compounds. Statistical treatment showed that the species effect remained significant throughout the process of drying and toasting, but not the provenance. The poor correlation with ring width of extractives levels measured on fresh timber remained unchanged as did the single-tree effect, with high variability found for all chemical parameters. These results provide further evidence that cooperage oak selection should not be based solely on the wood grain or the provenance but rather on a species-provenance combination.

Abstract. Dry extract, ellagitanins, free ellagic acid, whisky-lactone, eugenol and vanillin have been quantified for a sample set of 61 pedunculate oaks (Quercus robur L.) and 72 sessile oaks (Quercus petraea Liebl.) originating from 6 different forests. Despite a very high interindividual variability, pedunculate oak shows higher level of dry extract, ellagitanins and free ellagic acid but lower level in volatil compounds compared to sessile oak. Inside a same species, differences between forests are less marked. Extractible compounds level is poorly correlated to the grain (ring width). Regarding cambial age, it might have its significance but lack of scientific data could not permit to conclude in that sense. Our results suggest that cooperage oak selection should be based on a species-provenance combination rather than on the grain or the provenance solely.