Volatile Compound Evolution in Spanish Oak Wood (Quercus petraea and Quercus pyrenaica) during Natural Seasoning

Transcription

1 Volatile Compound Evolution in Spanish Oak Wood 163 Volatile Compound Evolution in Spanish Oak Wood (Quercus petraea and Quercus pyrenaica) during Natural Seasoning Estrella Cadahía, 1 Brígida Fernández de Simón, 1 * Raúl Vallejo, 1 Miriam Sanz, 1 and Miguel Broto 2 Abstract: The chemical composition of heartwood from the Spanish oaks Quercus petraea and Q. pyrenaica includes a wide range of volatile compound families (volatile phenols, phenolic aldehydes, furanic compounds, lactones, phenyl ketones, and others), which are potentially extracted into wine during aging. Changes in volatiles during the natural seasoning process followed in barrel cooperage were studied. Volatile compounds varied during seasoning, with some increasing in concentration while others decreased or did not show significant variations. The evolution of the wood was most significant during the first two years of seasoning, followed by stabilization during the third year, with variations depending on species, origin, and seasoning environmental conditions. Key words: volatile compounds, seasoning, oak wood, Quercus petraea, Quercus pyrenaica The oak wood quality used in wine aging has a great influence on wine s sensory properties and, in general, on its quality and economic value. During aging, physical, chemical, and physicochemical processes take place in which compounds are extracted from the wood into the wine. Phenolic compounds of wood are especially involved in wine oxidation and are directly related to color and sensory properties of wine such as astringency and bitterness (Monties 1992, Gómez-Cordovés and Gonzalez- SanJosé 1995, Vivas and Glories 1996a, Fernández de Simón et al. 2003a). In addition, numerous volatile compounds that migrate from wood into wine have been identified and affect flavor properties in wine, including the cis and trans isomers of β-methyl-γ-octalactone, furfural and its derivatives, phenolic aldehydes, and volatile phenols such as eugenol, guaiacol, and ethyl- and vinyl-phenols (Boidron et al. 1988, Cutzach et al. 1997, Fernandez de Simón et al. 2003b, Spillman et al. 2004). Chemical composition of barrel wood, including volatile substances, depends on two main factors: (1) oak species, geographical origin, and silvicultural treatment (Masson et al. 1995, Fernandez de Simón et al. 1996, Chatonnet and 1 Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria (INIA), Apdo. 8111, Madrid, Spain; 2 CESEFOR, Pol. Ind. Las Casas, Calle c, Parcela 4, Soria, Spain. *Corresponding author ( fdesimon@inia.es) Acknowledgments: This study was financed by project AGL C02-01 of the Ministry of Education and Science, Spain. The authors thank Mr. Antonio Sánchez for his help with the chemical analysis. Manuscript submitted May 2006; revised August 2006 Copyright 2007 by the American Society for Enology and Viticulture. All rights reserved. 163 Dubourdieu 1998, Doussot et al. 2002) and (2) processing of wood in cooperage, including method of seasoning (natural or artificial, length, and location) (Sefton et al. 1993, Chatonnet et al. 1994a,b, Fernandez de Simón et al. 1999, Masson et al. 2000, Cadahía et al. 2001a,b, Doussot et al. 2002) and method and degree of toasting (Chatonnet et al. 1989, Cadahía et al. 2001a,b, 2003). Barrel toasting likely has the most influence on volatile composition of wood. It is a necessary step in barrel manufacture and results in superficial degradation of wood biopolymers such as lignin, polyosides, polyphenols, and lipids, and it leads to the formation of new aromatic compounds, including volatile phenols, phenolic aldehydes, phenolic alcohols, phenyl ketones, furanic aldehydes, lactones, and other related compounds. The effect of seasoning on wood volatile composition cannot be ignored, although it is less dramatic than the effects of toasting. Seasoning causes wood to dehydrate until its humidity is in balance with the ambient humidity. A certain amount of fiber contraction occurs simultaneously. During this process, oak wood goes through cycles of dehydration, rehydration, and constant humidity, which slow the process, reducing the risk of fissures appearing in the staves. Natural seasoning results in wood maturation, decreasing bitterness and astringency and increasing aromatic properties by means of changes in chemical composition. Seasoning results primarily in loss of hydrosoluble polyphenolic substances such as ellagitannins, possibly because of different physical and chemical mechanisms, including rain leaching of staves, hydrolytic oxidative degradation (Chatonnet et al. 1994b, Vivas and Glories 1996b, Fernández de Simón et al. 1999, Cadahía et al. 2001b), and fungal enzymatic activity (phenol heterosidase, etherase, and depsidase) (Vivas 1997).

2 164 Cadahía et al. Seasoning can also have a significant effect on the aromatic profile of wood, but studies are limited and focused on few compounds with high aromatic potential, and often with contradictory results. The majority examined oak lactone, eugenol as the only volatile phenol, and phenolic aldehydes (vanillin, syringaldehyde, coniferaldehyde, and sinapaldehyde) (Sefton et al. 1993, Chatonnet et al. 1994a,b, Fernández de Simón et al. 1999, Sauvageot and Feuillat 1999, Masson et al. 2000, Cadahía et al. 2001a, Doussot et al. 2002). Phenolic aldehydes usually increase concentration in wood during natural seasoning, but there is conflicting data on eugenol and oak lactones. The latter of these, the β-methyl-γ-octalactones, are regarded as the most important oak volatiles contributing to the flavor of barrel-aged alcoholic beverages because of their low sensory threshold (Boidron et al. 1988, Chatonnet et al. 1992). Their concentrations vary by oak species, origin (Masson et al. 1995, Sauvageot and Feuillat 1999), and degree of toasting, during which their concentrations may decrease until they are undetectable (Chatonnet et al. 1989, 1999, Cadahía et al. 2003). Therefore, knowledge of volatile compound behavior, and especially of octalactones during the wood seasoning, is very important. There were two objectives of this work. The first was to understand evolution of oak wood volatiles during natural seasoning, including a wide range of compound families (volatile phenols, phenolic aldehydes, furanic compounds, lactones, phenyl ketones, and other related compounds) that are potentially extractable from oak wood into wine during aging, contributing to wine flavor. The second objective was to expand on studies of Spanish oak wood performance in barrel production, particularly changes in low molecular weight polyphenols and tannins during natural seasoning (Fernández de Simón et al. 1999, Cadahía et al. 2001a,b). Materials and Methods Wood sample collection. The oak heartwood used in this work was from Quercus petraea grown in Navarra (Garaioa) and Q. pyrenaica grown in Salamanca (Gata- Peña de Francia), Spain. Quercus petraea wood was harvested in January 2002 and placed in a cooperage in Álava province (average annual temperature = 13.6 C; total precipitation = 362 mm 3 /year; average humidity = 68.5%, from ). Quercus pyrenaica wood was harvested in March 2003 and placed in a cooperage in Navarra province (average annual temperature = 15 C; total precipitation = 420 mm 3 /year; average humidity = 66.3%, from ). The wood was split and the staves naturally seasoned in the open air for three years for Q. petraea and two years for Q. pyrenaica. During this process, staves were subjected to environmental conditions and standard watering at each cooperage. Wood samples were taken at different times during seasoning, consisting of a minimum three staves of each species and geographical origin. Three pieces of wood were cut from the center and the headboards of each stave. The wood pieces were ground and sieved, the sawdust ranging from 0.80 to 0.28 mm. Extraction. Volatile compounds were extracted by the method of Cadahía et al. (2003), which was based on Chatonnet et al. (1999). The sawdust samples (2 g) were soaked in 100 ml hydroalcoholic solution (12% ethanol, 0.7 g/l tartaric acid, 1.11 g/l potasium bitartrate) for 15 days at room temperature and in the dark. After filtration, an internal standard (hexalactone) and 15 g ammonium sulfate were added, and the solution was extracted with three washes of dichloromethane (a total of 45 ml). The organic fraction was dried with anhydrous sodium sulfate, concentrated to 0.5 ml under nitrogen flux in a Kuderna- Danish apparatus, and subjected to gas chromatographymass spectrometry (GC-MS) analysis. The analysis from each stave was carried out in duplicate, and coefficients of variation less than 5% were obtained. Standards. Reference compounds were purchased from several sources and are numbered in Table 1. Compounds 2, 3, 5, 6, 10, 11, 17, 18, 22, and 23 were purchased from Fluka Chimie AG (Buchs, Switzerland); 9, 12, 13, 14, 24, 26, 35, and 41 were from Aldrich Chimie (Neu-Ulm, Germany); 1, 4, 7, 15, 19, 25, and 27 were from Sigma Chemical (St. Louis, MO); 8 was from Chem Service (West Chester, PA); and 29, 37, 42, and γ-hexalactone were from Extrasynthèse (Genay, France). GC-MS analysis. GC-MS analysis of extracts was carried out in a HP 5890 gas chromatograph (Hewlett-Packard, Palo Alto, CA), equipped with a HP 5971A selective mass detector and a fused silica capillary column (Supelcowax- 10, 30 m x 0.25 mm id, and 0.25-μm film thickness), as modified by Cadahía et al. (2003) from Chatonnet et al. (1999). GC grade helium was the carrier gas at flow rate of 1.15 ml/min, 9.00 psi; the column temperature program was 45 C to 230 C at 3 C/min, held for 25 min, and then heated to 270 C at 10 C/min and held for 21 min. The injection temperature was 230 C. Detection was by electron impact mass (EI) in the full-scan mode, using an ionization energy of 70eV and an interphase detection temperature of 290 C. Quantitative determinations were calculated from total ion current (TIC) peak areas, using γ-hexalactone as internal standard, and calibration curves were made with commercial standards or with closely related chemical structures, analyzed under the same conditions. Statistical analysis. Univariate analysis was performed using ANOVA, applying the Student Newman-Keuls multiple range test. Multivariate canonical discriminant analysis was also carried out with all compounds evaluated, using the SAS statistical program (version 6; SAS Institute, Cary, NC). Results and Discussion Gas chromatography-mass spectrometry analysis of Spanish oak (Q. pyrenaica and Q. petraea) wood revealed most of the volatile compounds previously identified in other Spanish oak woods (Cadahía et al. 2003), and also in French (Cutzach et al. 1997, Chatonnet et al. 1989, Pérez-

3 Volatile Compound Evolution in Spanish Oak Wood 165 Table 1 Volatile compounds evaluated by GC-MS in Spanish Quercus petraea and Q. pyrenaica oak woods. Compound (IUPAC) Common name RRT a Identif b 1 2-Furancarboxaldehyde Furfural L B S 2 1-(2-Furanyl)-ethanone Furanyl-1-ethanone L B S 3 Benzaldehyde L B S 4 5-Methyl-2-furancarboxaldehyde 5-Methylfurfural L B S 5 Dihydro-2(3H)-furanone Butyrolactone L B S 6 2-Furanmethanol Furfuryl alcohol L B S 7 2-Methoxyphenol Guaiacol L B S 8 Benzyl alcohol Phenylmethanol L B S 9 trans-4-methyl-5-butyldihydro-2(3h)-furanone trans-β-methyl-γ-octalactone L B S 10 2-Phenylethanol L B S 11 Benzothiazole L B S 12 cis-4-methyl-5-butyldihydro-2(3h)-furanone cis-β-methyl-γ-octalactone L B S 13 4-Methyl-2-methoxyphenol 4-Methylguaiacol L B S 14 3-Hydroxy-2-methyl-4H-pyran-4-one Maltol L B S 15 Phenol L B S 16 4-Ethyl-2-methoxyphenol 4-Ethylguaiacol L B 17 1H-Pyrrole-2-carboxaldehyde L B S 18 2-Phenoxyethanol L S 19 2-Methoxy-4-(2-propenyl)-phenol Eugenol L B S 20 4-Vinyl-2-methoxyphenol Vinylguaiacol L B 21 3-Hydroxy-2-furyl methyl ketone Isomaltol L B 22 2,6-Dimethoxyphenol Syringol L B S 23 2-Methoxy-4-(1-propenyl)-phenol Isoeugenol L B S 24 4-Methyl-2,6-dimethoxyphenol 4-Methylsyringol L B S 25 5-Hydroxymethyl-2-furancarboxaldehyde 5-Hydroxymethylfurfural L B S 26 4-Allyl-2,6-dimethoxyphenol 4-Allylsyringol L B S 27 4-Hydroxy-3-methoxybenzaldehyde Vanillin L B S 28 2-(4-Hydroxy-3-methoxyphenyl) acetaldehyde L 29 1-(4-Hydroxy-3-methoxyphenyl)-ethanone Acetovanillone L B S 30 1-(4-Hydroxy-3-methoxyphenyl)-2-propanone L 31 1-(4-Hydroxy-3-methoxyphenyl)-propanone Propiovanillone B 32 1-(4-Hydroxy-3-methoxyphenyl)-2-butanone L 33 1-(4-Hydroxy-3-methoxyphenyl)-butanone Butyrovanillone B 34 Methyl vanillyl ether L 35 4-Hydroxy-3,5-dimethoxybenzaldehyde Syringaldehyde L B S 36 2-(4-Hydroxy-3,5-dimethoxyphenyl) acetaldehyde L 37 Ethyl vanillyl ether L 38 1-(4-Hydroxy-3,5-dimethoxyphenyl)-ethanone Acetosyringone L B S 39 1-(4-Hydroxy-3,5-dimethoxyphenyl)-2-propanone L 40 1-(4-Hydroxy-3,5-dimethoxyphenyl)-propanone Propiosyringone B 41 3-Methoxy-4-hydroxycinnamaldehyde Coniferaldehyde L B S 42 3,5-Dimethoxy-4-hydroxycinnamaldehyde Sinapaldehyde L B S a Relative retention time in relation to eugenol. b Compound identification. L: by comparison of mass spectra obtained in the EI-mode with published mass spectra (Wiley Spectra Library, 1995). B: in wood by other authors. S: by comparison of mass spectra and retention time with that of standard compound. Coello et al. 1998) and American (Towey and Waterhouse 1996) oak of suitable quality for wine aging. Quantified compounds, IUPAC name, common name, and relative retention times are shown in Table 1. Volatile phenols. The average concentrations of volatile phenols in wood extracts from Q. pyrenaica and Q. petraea reveal three different behaviors during the first 24 months of natural seasoning (Table 2). Guaiacol and the related compounds 4-methylguaiacol, 4-ethylguaiacol, and 4-vinylguaiacol in addition to syringol, 4-methylsyringol, and 4-allylsyringol increased in concentration during seasoning. Phenol concentration decreased, while eugenol and isoeugenol fluctuated. Results were similar in both species, with wood evolution becoming more evident during the second year of seasoning. The evolution gap between the two species may have occurred because their

4 166 Cadahía et al. Table 2 Gas chromatography quantitative evaluation of volatile phenols in oak wood during natural seasoning. Volatile phenols in Q. pyrenaica (μg/g of wood ± SD) Seasoning (months) N o of samples a Guaiacol 0.18 ± 0.08b b 0.22 ± 0.10b 0.16 ± 0.12b 0.45 ± 0.15a 4-Methylguaiacol 0.51 ± 0.57b 0.39 ± 0.28b 1.09 ± 1.08a 0.33 ± 0.12b 4-Ethylguaiacol c 0.06 ± 0.03b 0.05 ± 0.03b 0.03 ± 0.01b 0.14 ± 0.06a 4-Vinylguaiacol c 0.61 ± 0.24c 1.11 ± 0.44b 1.21 ± 0.32b 2.19 ± 0.46a Phenol 0.25 ± 0.05a 0.24 ± 0.07a 0.23 ± 0.07a 0.08 ± 0.03b Eugenol 5.16 ± 3.51a 5.65 ± 2.71a 3.91 ± 2.19a 7.28 ± 5.43a Isoeugenol 0.61 ± 0.26a 0.49 ± 0.12ab 0.32 ± 0.10b 0.68 ± 0.19a Syringol 0.47 ± 0.27b 0.38 ± 0.20b 0.16 ± 0.06b 1.66 ± 0.78a 4-Methylsyringol 0.67 ± 0.25b 0.49 ± 0.20b 0.41 ± 0.14b 1.32 ± 0.52a 4-Allylsyringol 3.03 ± 2.23a 5.24 ± 2.86a 4.03 ± 2.32a 4.11 ± 2.22a Volatile phenols in Q. petraea (μg/g of wood ± SD) Seasoning (months) N o of samples a Guaiacol 0.09 ± 0.00c b 0.08 ± 0.01c 0.08 ± 0.01c 0.13 ± 0.04ab 0.11 ± 0.01bc 0.15 ± 0.02a 0.11 ± 0.01bc 4-Methylguaiacol 0.06 ± 0.02c 0.03 ± 0.01c 0.64 ± 0.14a 0.06 ± 0.02c 0.10 ± 0.07c 0.08 ± 0.03c 0.20 ± 0.02b 4-Ethylguaiacol c 0.01 ± 0.01bc nd c 0.01 ± 0.00bc 0.01 ± 0.01bc 0.02 ± 0.01bc 0.03 ± 0.02ab 0.05 ± 0.02a 4-Vinylguaiacol c 0.39 ± 0.11b 0.48 ± 0.19b 0.09 ± 0.01b 0.28 ± 0.04b 1.03 ± 0.36a 0.95 ± 0.41a 0.40 ± 0.08b Phenol 1.47 ± 0.16a 0.38 ± 0.04b 0.17 ± 0.02b 0.28 ± 0.02b 0.19 ± 0.01b 0.16 ± 0.05b 0.17 ± 0.19b Eugenol 2.05 ± 1.10a 0.40 ± 0.15a 1.47 ± 0.58a 1.34 ± 1.09a 1.31 ± 0.27a 3.05 ± 3.30a 1.46 ± 0.68a Isoeugenol 0.18 ± 0.05bc 0.09 ± 0.02c 0.21 ± 0.04bc 0.37 ± 0.13b 0.72 ± 0.18a 0.45 ± 0.03b 0.66 ± 0.16a Syringol 0.07 ± 0.03a 0.04 ± 0.01a 0.04 ± 0.00a 0.10 ± 0.04a 0.19 ± 0.06a 0.24 ± 0.06a 0.23 ± 0.16a 4-Methylsyringol 0.11 ± 0.02bc 0.03 ± 0.00c 0.07 ± 0.02c 0.12 ± 0.04bc 0.25 ± 0.09ab 0.24 ± 0.06a 0.30 ± 0.07a 4-Allylsyringol 0.57 ± 0.27a 0.23 ± 0.01a 0.53 ± 0.17a 0.86 ± 0.48a 0.65 ± 0.15a 0.57 ± 0.33a 0.99 ± 0.69a a Each sample analyzed in duplicate. b Average and standard deviation (x ± SD) were calculated for the number of samples taken at each step of seasoning. Different letters in a row denote a significant difference with 95% confidence level in the Student Newman-Keuls multiple range test. c Expressed as 4-methylguaiacol. seasoning began at different times of the year and was carried out under different environmental conditions in two cooperages. On the other hand, peaks observed for eugenol, isoeugenol, and other phenols could be related to the time of year in which sample collection took place. Information in the literature on volatile phenol concentration during wood seasoning is scarce and limited to eugenol because of its high aromatic potential. This compound had highly variable concentrations in wood (0 to 10 μg/g), depending on the species and geographical origin of the wood (Chatonnet and Dubourdieu 1998, Doussot et al. 2002, Cadahía et al. 2003). Studies have examined the relationship among eugenol concentration and seasoning time and method, but have had contradictory results. One study noted either no change or a regular and constant decrease in eugenol during natural seasoning (Sefton et al. 1993), and another described the latter effect during kiln drying (Masson et al. 2000). Yet another study suggested the opposite effect: a slight increase during seasoning duration (Chatonnet 1994a). A comparison of all results indicates no general tendency for all wood types studied, but there were some commonalities. For example, eugenol evolution during seasoning is highly variable, changes generally are not significant, and there is frequent fluctuation. Volatile phenols may be eliminated from wood during seasoning by physical or chemical mechanisms such as solubilization in water, evaporation, or oxidative degradation process, and at the same time are formed from precursors or lignin degradation (Philips and Goss 1932). For each compound group, one or several of these mechanisms will dominate or they will compensate for each other. It is possible that eugenol losses during seasoning compete with eugenol formation from lignin degradation, with one or another effect prevailing depending on environmental conditions. Toasting produces only small variations in eugenol concentration (Chatonnet et al. 1989, Cadahía et al. 2003), so concentrations in wood can be optimized by adjusting seasoning conditions. A canonical discriminant analysis of compounds listed in Table 2 was conducted for Q. pyrenaica and Q. petraea to determine the overall evolution of volatile phenols in wood during natural seasoning (Figure 1A and 1B, respectively). The first axis leads to a mathematical model that distinguishes three Q. pyrenaica wood groups by season-

5 Volatile Compound Evolution in Spanish Oak Wood 167 Figure 1 Canonical discriminant analysis of volatile phenols from wood after different lengths of natural seasoning (see Table 2). (1A) Q. pyrenaica. Projections of the points of each oak wood sample on a plane defined by the two principal canonical (PC) axes: (A) 0, (B) 12, (C) 18, and (D) 24 months of seasoning. Can1 accounted for 79.56% and can2 for 19.07% of total variance. Eigenvalues for can1 and 2 were and , and canonical correlations were and , respectively. Total canonical structure coefficients for the main discriminant variables in can1: 4-vinylguaiacol , phenol , syringol , guaiacol , eugenol , 4-methylsyringol ; in can2: isoeugenol (1B) Q. petraea. Projections of the points of each oak wood sample on a plane defined by the two PC axes: (a) 0, (b) 6, (c) 11, (d) 15, (e) 24, (f) 30, and (g) 36 months of seasoning. Can1 accounted for 49.26% and can2 for 32.21% of total variance. Eigenvalues for can1 and 2 were and and canonical correlations were and , respectively. Total canonical structure coefficients for the main discriminant variables in can1: isoeugenol , 4-methylsyringol , 4-ethylguaiacol , syringol ; in can2: 4-methylguaiacol , phenol ing stage. Group 1 contains wood seasoned for 0 months (unseasoned) (A); group 2 contains wood seasoned for 12 (B) and 18 (C) months; and group 3 contains woods seasoned for 24 months (D) (Figure 1A). Statistical distance among the groups was particularly large between 24 months and the other three treatments. This distance resulted from canonical function 1, which was derived from concentrations of 4-vinylguaiacol, phenol, syringol, guaiacol, eugenol, and 4-methylsyringol, all of which had positive coefficients except for phenol. All these compounds have positive sensory characteristics, except for phenol, which gives wine an undesirable aroma (Boidron et al. 1988, Chatonnet et al. 1992). Discriminant analysis of Q. petraea data also separates wood by length of seasoning (Figure 1B). There was considerable variability at 0, 6, 11, and 15 months, followed by stabilization of volatile compounds starting at 24 months, indicating there is no significant evolution of volatile phenols during the third year. Canonical function 1 had principal discriminant variables isoeugenol, 4- methylsyringol, 4-ethylguaiacol, and syringol all with positive coefficients in total canonical structure and positive sensory characteristics. β-methyl-γ-octalactones, furanic compounds, cyclic ketones, pyrrole structures, and related compounds. Evolution of cis- and trans-β-methyl-γ-octalactone during seasoning was highly variable in both Q. pyrenaica and Q. petraea (Table 3). The concentration of these compounds varied considerable in green wood samples as well (4.84 ± 8.37 μg/g and 18.4 ± 20.5 μg/g for the trans isomer, 35.9 ± 60.9 μg/g and 36.9 ± 38.5 μg/g for the cis isomer), which suggests that the differences are not significant, as can be confirmed by the variance analysis. Similar variability was described in French oak Q. robur as Q. petraea and in American oak Q. alba (Masson et al. 1995, Doussot et al. 2002), with reported variation coefficients of 36 to 162%. However, the mean concentrations of cis isomer, which has greater aromatic potential in wine than the trans isomer (threshold value in synthetic solution of mg/l compared with 0.11 mg/l for the trans isomer) (Chatonnet et al. 1992), tended to increase with seasoning duration, even during the third year in Q. petraea wood. The minima reached at 15 and 30 months of seasoning in this last species seemed to be transitory and may be due to environmental conditions previous to sampling, which could have favored the leaching of this compound or evaporation from wood over its formation from precursors (Sefton et al. 1993, Masson et al. 2000). However, our seasoning conditions contributed to the long-term formation and slight accumulation of cis-methyl-octalactone, in agreement with that found in French oak wood (Chatonnet 1994a). Apparent contradictions in published results could be due to a combination of special conditions during seasoning and different physical and chemical characteristics of various oak woods. Although these compounds are sensitive to cooperage toasting, their concentrations in wines have reflected those observed for those in oak prior to toasting (Spillman et al. 2004, Fernandez de Simón et al. 2003a). Similar to eugenol, their concentrations in green and seasoned woods can determine the aromatic potential of wood during wood-wine interaction. For furanic compounds, cyclic ketones (furanones and pyranones), pyrrole structures, and related structures, the only compound that showed a change in concentration during wood seasoning was phenylmethanol, which increased significantly starting at 12 or 15 months depending on species (Table 3). Multivariate discriminant analysis carried out on all the volatile compounds evaluated in Table 3 yielded a pattern of sample distribution in which groups corresponding to different wood seasoning stages

6 168 Cadahía et al. Table 3 Gas chromatography quantitative evaluation of lactones, furanic compounds, pyranones, and other volatile compounds (μg/g of wood) in oak wood during natural seasoning. Compound in Q. pyrenaica (μg/g of wood ± SD) Seasoning (months) N o of samples a Furanyl-1-ethanone 0.34 ± 0.30a b 0.28 ± 0.08a 0.30 ± 0.17a 0.33 ± 0.10a Butyrolactone 0.44 ± 0.44a 0.45 ± 0.27a 0.51 ± 0.16a 0.67 ± 0.44a Benzaldehyde 0.17 ± 0.15c 0.32 ± 0.06bc 0.45 ± 0.09b 0.89 ± 0.26a Benzothiazole 0.09 ± 0.06a 0.07 ± 0.01ab 0.04 ± 0.01b 0.05 ± 0.01ab 1H-Pyrrole-2-carboxaldehyde 0.02 ± 0.01c 0.07 ± 0.21b 0.12 ± 0.03a 0.14 ± 0.04a Phenylmethanol 0.95 ± 0.55d 7.82 ± 1.30c 9.89 ± 1.14b 15.7 ± 2.11a 2-Phenylethanol 0.28 ± 0.15a 0.44 ± 0.20a 0.39 ± 0.11a 0.28 ± 0.28a 2-Phenoxyethanol 0.13 ± 0.11b 0.84 ± 0.46a 0.28 ± 0.21b 0.19 ± 0.06b Maltol 0.28 ± 0.04c 0.37 ± 0.11ab 0.42 ± 0.11a 0.33 ± 0.07bc Isomaltol c 0.14 ± 0.01b 0.14 ± 0.02b 0.15 ± 0.01a 0.16 ± 0.01a Furfural 2.56 ± 1.43b 3.46 ± 1.04b 4.92 ± 2.79a 3.18 ± 1.08b 5-Methylfurfural 0.22 ± 0.05c 0.91 ± 0.19b 1.61 ± 0.56a 0.98 ± 0.26b Furfuryl alcohol 0.24 ± 0.11bc 0.15 ± 0.07c 0.54 ± 0.43a 0.42 ± 0.31ab 5-Hydroxymethylfurfural 3.71 ± 2.03a 4.15 ± 2.35a 3.43 ± 1.58a 4.89 ± 2.06a trans-β-methyl-γ-octalactone 4.84 ± 8.37a 3.52 ± 5.31a 4.58 ± 6.16a 5.25 ± 6.51a cis-β-methyl-γ-octalactone 35.9 ± 60.9a 50.4 ± 72.0a 72.7 ± 110a 68.2 ± 91.2a cis/trans ratio 10.1 ± 8.86a 27.0 ± 29.2a 23.9 ± 15.4a 19.9 ± 22.2a Compound in Q. petraea (μg/g of wood ± SD) Seasoning (months) N o of samples a Furanyl-1-ethanone 0.02 ± 0.00bc b 0.01 ± 0.00c 0.04 ± 0.01ab 0.04 ± 0.02ab 0.05 ± 0.01a 0.05 ± 0.01a 0.04 ± 0.02ab Butyrolactone 0.20 ± 0.03ab 0.09 ± 0.02b 0.08 ± 0.01b 0.31 ± 0.09a 0.24 ± 0.12ab 0.25 ± 0.12ab 0.07 ± 0.01b Benzaldehyde 0.06 ± 0.02b 0.02 ± 0.01b 0.04 ± 0.00b 0.07 ± 0.01b 0.35 ± 0.06a 0.34 ± 0.10a 0.10 ± 0.01b Benzothiazole 0.03 ± 0.00c 0.04 ± 0.00b 0.04 ± 0.00bc 0.08 ± 0.01a 0.04 ± 0.01bc 0.03 ± 0.01c 0.02 ± 0.00d 1H-Pyrrole-2-carboxaldehyde 0.05 ± 0.02a 0.04 ± 0.01a 0.03 ± 0.00a 0.07 ± 0.02a 0.07 ± 0.02a 0.05 ± 0.03a 0.05 ± 0.01a Phenylmethanol 0.34 ± 0.06c 0.28 ± 0.09c 0.30 ± 0.08c 0.70 ± 0.15c ± 1.87a ± 1.65a 5.98 ± 0.57b 2-Phenylethanol 0.36 ± 0.21a 0.36 ± 0.06a 0.30 ± 0.17a 0.37 ± 0.08a 0.47 ± 0.15a 0.29 ± 0.15a 0.28 ± 0.10a 2-Phenoxyethanol 0.07 ± 0.01abc 0.06 ± 0.01bc 0.06 ± 0.00bc 0.06 ± 0.01bc 0.09 ± 0.02ab 0.10 ± 0.02a 0.04 ± 0.01c Maltol 0.24 ± 0.02a 0.25 ± 0.02a 0.27 ± 0.03a 0.42 ± 0.07a 0.34 ± 0.06a 0.48 ± 0.25a 0.29 ± 0.03a Isomaltol c 0.12 ± 0.00a 0.12 ± 0.00a 0.12 ± 0.00a 0.13 ± 0.02a 0.13 ± 0.00a 0.22 ± 0.14a 0.15 ± 0.01a Furfural 1.19 ± 0.45bc 0.42 ± 0.04c 2.13 ± 0.13bc 2.02 ± 0.85bc 3.54 ± 2.10ab 4.70 ± 1.49a 1.99 ± 0.82bc 5-Methylfurfural 0.22 ± 0.04ab 0.11 ± 0.01b 0.14 ± 0.01b 0.25 ± 0.10ab 0.23 ± 0.05ab 0.50 ± 0.32a 0.20 ± 0.04ab Furfuryl alcohol 0.22 ± 0.02a 0.05 ± 0.01c 0.07 ± 0.01bc 0.23 ± 0.01a 0.20 ± 0.08a 0.15 ± 0.07ab 0.07 ± 0.02bc 5-Hydroxymethylfurfural 0.63 ± 0.19c 1.01 ± 0.25c 0.73 ± 0.12c 2.07 ± 0.73b 2.78 ± 0.28ab 2.83 ± 0.21ab 3.32 ± 0.83a trans-β-methyl-γ-octalactone 18.4 ± 20.5a 19.5 ± 29.2a 34.9 ± 33.7a 16.0 ± 11.4a 14.7 ± 18.2a 4.42 ± 7.72a 4.12 ± 0.64a cis-β-methyl-γ-octalactone 36.9 ± 10.2a 38.5 ± 28.6a 43.7 ± 15.5a 18.6 ± 12.2a 55.9 ± 36.1a ± 11.5a 65.8 ± 26.8a cis/trans ratio 3.90 ± 2.57a 20.1 ± 23.2a 1.94 ± 1.43a 2.74 ± 3.03a 8.95 ± 7.29a 9.75 ± 8.53a 16.2 ± 6.96a a Each sample analyzed in duplicate. b Average and standard deviation (x ± SD) were calculated for the number of samples taken at each step of seasoning. Different letters in a row denote a significant difference with 95% confidence level in the Student Newman-Keuls multiple range test. c Expressed as maltol. can be distinguished (Figure 2). Quercus pyrenaica woods presented four easily distinguished groups along canonical axis 1, separating according to concentrations of phenylmethanol, 1H-pyrrole-2-carboxaldehyde, benzaldehyde, and 5-methylfurfural, as deduced from their canonical structure coefficients (Figure 2A). This proves progressive chemical evolution, resulting at the end of seasoning (group D) in woods with different aromatic profiles than green woods (group A). For Q. petraea, separation of wood by seasoning stages can also be observed, but the behavior is more complex, showing recession phases during the evolution. In examining the statistical distance between groups, it is necessary to consider both axis 1 and 2 (Figure 2B). The third year of Q. petraea

7 Volatile Compound Evolution in Spanish Oak Wood 169 Figure 2 Canonical discriminant analysis of lactones, furanic compounds, pyranones, and other volatile compounds from wood after different lengths of natural seasoning (see Table 3). (2A) Q. pyrenaica. Projections of the points of each oak wood sample on a plane defined by the two PC axes: (A) 0, (B) 12, (C) 18, and (D) 24 months of seasoning. Can1 accounted for 87.25% and can2 for 2.55% of total variance. Eigenvalues for can1 and can2 were and and canonical correlations were and , respectively. Total canonical structure coefficients for the main discriminant variables in can1: phenylmethanol , 1H-pyrrole-2-carboxaldehyde , benzaldehyde , 5-methylfurfural ; in can2: 2-phenylethanol , 5-methylfurfural , and maltol (2B) Q. petraea. Projections of the points of each oak wood sample on a plane defined by the two PC axes: (a) 0, (b) 6, (c) 11, (d) 15, (e) 24, (f) 30, and (g) 36 months of seasoning. Can1 accounted for 77.27% and can2 for 16.13% of total variance. Eigenvalues for can1 and 2 were and and canonical correlations were and , respectively. Total canonical structure coefficients for the main discriminant variables in can1: phenylmethanol , benzaldehyde , furfural ; in can2: benzothiazole , butyrolactone seasoning resulted in a recession or stabilization phase, ending in a final wood chemical composition (group g) that was statistically closer to green wood (group a) than to the intermediate phases (groups d, e, f). Canonical function 1 derived mainly from phenylmethanol and benzaldehyde, as before, plus furfural; function 2 derived from benzothiazole and butyrolactone. Most of these compounds are produced during the toasting stage of cooperage (Chatonnet et al. 1999, Cadahía et al. 2003), and changes in their concentrations during seasoning may seem of little interest enologically, but such changes are indicative of the natural evolution of wood. Phenolic aldehydes and related compounds. Concentrations of phenolic aldehydes, vanillin, syringaldehyde, coniferaldehyde, and sinapaldehyde increased significantly in both species during the first two years of seasoning (Table 4), as shown in previous studies of Spanish, French, and American oak woods (Chatonnet et al. 1994b, Cadahía et al. 2001a). However, their concentrations decreased during the third year, as shown for Q. petraea (Table 4). This increase can be explained by lignin degradation, depolymerization, and the posterior hydrolytic and oxidative degradation of monomers, either by enzymatic or chemical processes. In addition, the phenolic aldehydes generated are subject to leaching, photodegradation, and biochemical degradation, which decrease their concentrations, primarily in the superficial layers of wood (Chatonnet et al. 1994b). The prevalence of different mechanisms under different environmental conditions could explain contradictory results reported by other authors, who found no significant changes in vanillin during airdrying of the wood (Sefton et al. 1993). Concentrations of related compounds in wood also increased during the first two years of seasoning, with this increase being greater for phenyl ketones, acetovanillone, propiovanillone, acetosyringone, and propiosyringone than for acetaldehyde and 2-one-isomers (HMPA, HMPP, HDMPA, and HDMPP). Concentrations of ethyl vanillyl ether in both species and methyl vanillyl ether in Q. petraea also increased slightly during the same period. During the third year, concentrations decreased or remained constant compared to the second seasoning year. The origin of phenyl ketones in wood is not well known, but it is accepted that they come from thermodegradation of polyphenolic structures in wood, more likely the lignans in oak wood (Seikel et al. 1971) rather than lignin (Chatonnet et al. 1989). Most of these compounds also increase in concentration during toasting (Cadahía et al. 2003); and although some, such as the phenolic aldehydes, play a major role in wood aromatic potential because of their low sensory thresholds (0.065 mg/l for vanillin, in synthetic solution; Chatonnet et al. 1992), they are not especially useful for optimizing seasoning conditions. Discriminant analysis of these phenolic aldehydes and related compounds proves that their concentrations in Q. pyrenaica and Q. petraea woods evolved similarly to the preceding compound groups (Figure 3). During the two first years of seasoning (groups A, B, C, D and groups a, b, c, d, e), both species presented a progressive evolution, corresponding to canonical function 1, resulting in the seasoned woods (groups D and e) with different aromatic profiles than green woods (groups A and a). However, during the third year (groups f and g) the woods evolved toward compositions statistically closer to that of green wood (group a). The principal discriminant variables for axis 1 were HDMPP, vanillin, syringaldehyde, acetosyringone, and HDMPA, in this order and with positive coefficients of total canonical structure.

8 170 Cadahía et al. Table 4 Gas chromatography quantitative evaluation of phenolic aldehydes and related compounds (μg/g of wood) in oak wood during natural seasoning. Compound in Q. pyrenaica (μg/g of wood ± SD) Seasoning (months) N o of samples a Vanillin 3.42 ± 0.95b b 5.29 ± 1.40a 4.47 ± 0.96a 4.91 ± 0.74a HMPA c 0.59 ± 0.42a 0.17 ± 0.15b 0.46 ± 0.26ab 0.49 ± 0.35ab Acetovanillone 0.38 ± 0.22b 0.47 ± 0.12b 0.34 ± 0.07b 0.72 ± 0.20a HMPP c 0.95 ± 0.89a 0.80 ± 0.16a 0.41 ± 0.11a 0.93 ± 0.33a Propiovanillone 3.88 ± 2.29ab 4.44 ± 1.32ab 3.24 ± 1.47b 5.61 ± 1.40a HMPB c 0.06 ± 0.09a 0.06 ± 0.04a 0.06 ± 0.03a 0.09 ± 0.04a Butyrovanillone d 0.06 ± 0.05a 0.05 ± 0.02a 0.07 ± 0.03a 0.09 ± 0.04a Methyl vanillyl ether e 0.70 ± 0.41b 1.04 ± 0.55ab 0.74 ± 0.33b 1.20 ± 0.53a Syringaldehyde 4.35 ± 1.41b 6.45 ± 1.04b 4.97 ± 1.10b 9.73 ± 3.86a HDMPA c 0.47 ± 0.28b 0.47 ± 0.11b 0.59 ± 0.20b 1.32 ± 0.45a Ethyl vanillyl ether e 1.83 ± 0.88b 3.13 ± 0.88a 2.09 ± 0.71b 2.93 ± 0.73a Acetosyringone 0.45 ± 0.22b 0.72 ± 0.15b 0.46 ± 0.20b 1.18 ± 0.54a HDMPP c 1.32 ± 0.81ab 1.35 ± 0.33ab 0.91 ± 0.32b 1.80 ± 0.55a Propiosyringone f 1.37 ± 1.03a 1.49 ± 0.74a 1.17 ± 0.67a 1.79 ± 0.90a Coniferaldehyde 2.16 ± 0.89b 3.66 ± 0.87a 2.26 ± 0.69b 4.23 ± 0.83a Sinapaldehyde 2.49 ± 1.30b 2.50 ± 0.86b 2.08 ± 0.67b 4.18 ± 1.61a Compound in Q. petraea (μg/g of wood ± SD) Seasoning (months) N o of samples a Vanillin 2.19 ± 0.16c b 3.56 ± 0.35bc 3.25 ± 0.95bc 4.85 ± 0.91b 6.99 ± 1.45a 4.79 ± 0.43b 4.81 ± 0.84b HMPA c 0.13 ± 0.06b 0.15 ± 0.01b 0.16 ± 0.05b 0.17 ± 0.06b 0.28 ± 0.12ab 0.41 ± 0.09a 0.24 ± 0.14ab Acetovanillone 0.16 ± 0.03c 0.21 ± 0.01c 0.19 ± 0.01c 0.32 ± 0.10b 0.39 ± 0.03ab 0.40 ± 0.09ab 0.48 ± 0.05a HMPP c 0.17 ± 0.02b 0.11 ± 0.02b 0.13 ± 0.01b 0.25 ± 0.02b 0.63 ± 0.15a 0.48 ± 0.09a 0.51 ± 0.12a Propiovanillone 0.64 ± 0.47b 1.26 ± 0.95ab 1.23 ± 0.16b 2.29 ± 1.61ab 3.60 ± 1.65a 2.21 ± 0.57ab 3.38 ± 0.75a HMPB c 0.03 ± 0.01a 0.10 ± 0.07a 0.04 ± 0.02a 0.06 ± 0.02a 0.13 ± 0.10a 0.14 ± 0.13a 0.07 ± 0.01a Butyrovanillone d 0.56 ± 0.34a 0.51 ± 0.02a 0.49 ± 0.23a 0.82 ± 0.31a 1.85 ± 0.14a 3.09 ± 3.40a 1.00 ± 0.34a Methyl vanillyl ether e 0.31 ± 0.08a 0.73 ± 0.20a 0.52 ± 0.25a 0.73 ± 0.62a 1.28 ± 0.68a 0.77 ± 0.32a 1.20 ± 0.49a Syringaldehyde 2.71 ± 0.80d 5.65 ± 0.63bc 4.22 ± 0.06cd 6.95 ± 1.94ab 8.42 ± 0.44a 6.89 ± 0.86ab 6.86 ± 0.80ab HDMPA c 0.12 ± 0.02b 0.16 ± 0.02b nd b 0.16 ± 0.03b 0.28 ± 0.16b 0.58 ± 0.19a 0.13 ± 0.10b Ethyl vanillyl ether e 1.18 ± 0.52c 1.45 ± 0.45bc 1.26 ± 0.56bc 2.05 ± 0.23abc 2.44 ± 0.30ab 2.51 ± 0.34ab 2.97 ± 0.76a Acetosyringone 0.12 ± 0.02d 0.19 ± 0.00cd 0.27 ± 0.16bcd 0.36 ± 0.18abc 0.55 ± 0.07a 0.45 ± 0.13ab 0.44 ± 0.07ab HDMPP c 0.14 ± 0.05c 0.10 ± 0.02c 0.21 ± 0.17c 0.41 ± 0.09bc 1.19 ± 0.19a 0.84 ± 0.47ab 0.52 ± 0.17bc Propiosyringone f 0.75 ± 0.15b 2.80 ± 0.96a 1.08 ± 0.19b 4.36 ± 1.62a 3.92 ± 0.52a 3.67 ± 1.23a 2.66 ± 0.28a Coniferaldehyde 0.87 ± 0.13d 1.18 ± 0.11cd 0.98 ± 0.62d 2.69 ± 0.62b 4.19 ± 1.04a 2.30 ± 0.75b 2.04 ± 0.23bc Sinapaldehyde 1.64 ± 0.60b 1.87 ± 0.14b 1.63 ± 0.08b 2.50 ± 0.32b 4.27 ± 1.21a 1.61 ± 0.04b 2.07 ± 0.29b a Each sample analyzed in duplicate. b Average and standard deviation (x ± SD) were calculated for the number of samples taken at each step of seasoning. Different letters in a row denote a significant difference with 95% confidence level in the Student Newman-Keuls multiple range test. c HMPA: 2-(4-hydroxy-3-methoxyphenyl) acetaldehyde; HMPP: 1-(4-hydroxy-3-methoxyphenyl)-2-propanone; HMPB: 1-(4-hydroxy-3- methoxyphenyl)-2-butanone; HDMPA: 2-(4-hydroxy-3,5-dimethoxyphenyl) acetaldehyde; HDMPP: 1-(4-hydroxy-3,5-dimethoxyphenyl)- 2-propanone. d Expressed as acetovanillone. e Expressed as vanillin. f Expressed as acetosyringone. Conclusions We can conclude that during natural seasoning in the open air, the volatile compound composition of Spanish Q. petraea and Q. pyrenaica oak wood undergoes significant evolution, which may appear limited when compared to that of toasting during cooperage, but which should not be underrated. The volatile compounds showed different behaviors: some increased in concentration, while others decreased or did not show significant variation. Over-

9 Volatile Compound Evolution in Spanish Oak Wood 171 Figure 3 Canonical discriminant analysis of phenolic aldehydes and related compounds from wood after different lengths of natural seasoning (see Table 4). (3A) Q. pyrenaica. Projections of the points of each oak wood sample on a plane defined by the two PC axes: (A) 0, (B) 12, (C) 18, and (D) 24 months of seasoning. Can1 accounted for 71.15% and can2 for 23.90% of total variance. Eigenvalues for can1 and can2 were and and canonical correlations were and , respectively. Total canonical structure coefficients for the main discriminant variables in can1: HDMPA , coniferaldehyde , syringaldehyde , acetosyringone ; in can2: vanillin , HMPA (3B) Q. petraea. Projections of the points of each oak wood sample on a plane defined by the two PC axes: (a) 0, (b) 6, (c) 11, (d) 15, (e) 24, (f) 30, and (g) 36 months of seasoning. Can1 accounted for 64.42% and can2 for 19.14% of total variance. Eigenvalues for can1 and 2 were and and canonical correlations were and , respectively. Total canonical structure coefficients for the main discriminant variables in can1: HDMPP , vanillin , syringaldehyde , acetosyringone , HDMPA ; in can2: HDMPA , butyrovanillone , propiosyringone , vanillin all, the variations have a predominantly positive effect on the volatile composition of oak wood. Wood evolution occurred primarily during the first two years of seasoning, followed by stabilization during the third year, with variations depending on species, origin of wood, and seasoning environmental conditions. Literature Cited Boidron, J.N., P. Chatonnet, and M. Pons Influence du bois sur certaines substances odorants des vins. Conn. Vigne Vin 22: Cadahía, E., B. Fernández de Simón, and J. Jalocha Volatile compounds in Spanish, French, and American oak wood after natural seasoning and toasting. J. Agric. Food Chem. 51: Cadahía, E., L. Muñoz, B. Fernández de Simón, and M.C. García- Vallejo. 2001a. Changes in low molecular weight phenolic compounds in Spanish, French, and American oak woods during natural seasoning and toasting. J. Agric. Food Chem. 49: Cadahía, E., S. Varea, L. Muñoz, B. Fernández de Simón, and M.C. García-Vallejo. 2001b. Evolution of ellagitannins in Spanish, French and American oak wood during natural seasoning and toasting. J. Agric. Food Chem. 49: Chatonnet, P., and D. Dubourdieu Comparative study of the characteristics of American white oak (Quercus alba) and European oak (Quercus petraea and Q. robur) for production of barrels used in barrel aging of wines. Am. J. Enol. Vitic. 49: Chatonnet, P., I. Cutzach, M. Pons, and D. Dubourdieu Monitoring toasting intensity of barrels by chromatographic analysis of volatile compounds from toasted oak wood. J. Agric. Food Chem. 47: Chatonnet, P., J.N. Boidron, and M. Pons Incidence du traitement thermique du bois de chêne sur la composition chimique. 2e Partie: Évolution de certains composés en fonction de l intensité de brûlage. J. Int. Sci. Vigne Vin 23: Chatonnet, P., D. Dubourdieu, and J.N. Boidron Incidence des conditions de fermentation et d élevage des vins blancs secs en barriques sur leur composition en substances cédées par le bois de chêne. Sci. Aliments 12: Chatonnet, P., J.N. Boidron, D. Dubourdieu, and M. Pons. 1994a. Evolution of oak wood volatile compounds during seasoning. First results. J. Int. Sci. Vigne Vin 28: Chatonnet, P., J.N. Boidron, D. Dubourdieu, and M. Pons. 1994b. Evolution of oak wood polyphenolic compounds during seasoning. First results. J. Int. Sci. Vigne Vin 28: Cutzach, I., P. Chatonnet, R. Henry, and D. Dubourdieu Identification of volatile compounds with a toasty aroma in heated oak used in barrelmaking. J. Agric. Food Chem. 45: Doussot, F., B. De Jéso, S. Quideau, and P. Pardon Extractives content in cooperage oak wood during natural seasoning and toasting; Influence of tree species, geographic location, and singletree effects. J. Agric. Food Chem. 50: Fernández de Simón, B., E. Cadahía, E. Conde, and M.C. García- Vallejo Low molecular weight phenolic compounds in Spanish oak woods. J. Agric. Food Chem. 44: Fernández de Simón, B., E. Cadahía, E. Conde, and M.C. García- Vallejo Evolution of phenolic compounds in Spanish oak wood during natural seasoning. First results. J. Agric. Food Chem. 47: Fernández de Simón, B., E. Cadahía, and J. Jalocha. 2003b. Volatile compounds in a Spanish red wine aged in barrels made of Spanish, French and American oak wood. J. Agric. Food Chem. 51: Fernández de Simón, B., T. Hernández, E. Cadahía, M. Dueñas, and I. Estrella. 2003a. Phenolic compounds in a Spanish red wine aged in barrels made of Spanish, French and American oak wood. Eur. Food Res. Technol. 216: Gómez-Cordovés, M.C., and M.L. González-SanJosé Interpretation of color variables during the aging of red wines: Relationship with families of phenolic compounds. J. Agric. Food Chem. 43:

10 172 Cadahía et al. Masson, E., R. Baumes, M. Moutounet, and J.L. Puech The effect of kiln-drying on the levels of ellagitannins and volatile compounds of European oak (Quercus petraea Liebl.) stave wood. Am. J. Enol. Vitic. 51: Masson, G., E. Guichard, N. Fournier, and J.L. Puech Stereoisomers of β-methyl-γ-octalactone. II. Contents in the wood of French (Quercus robur and Quercus petraea) and American (Quercus alba) oaks. Am. J. Enol. Vitic. 46: Monties, B Chemical composition of oak wood: Phenolic compounds and their relation with physical and chemical properties related to quality of wines and spirits. In The Wood and the Quality of Wines and Spirits. J. Int. Sci. Vigne Vin (special issue): Pérez-Coello, M.S., J. Sanz, and M.D. Cabezudo Gas chromatographic-mass spectrometric analysis of volatile compounds in oak wood used for ageing of wines and spirits. Chromatographia 47: Philips, M., and J.M. Goss Chemistry of lignin VII. Distillation of alkali lignin in reduced atmosphere of carbon dioxide. Ind. Eng. Chem. 24: Sauvageot, F., and F. Feuillat The influence of oak wood (Quercus robur L., Q. petraea Liebl.) on the flavor of Burgundy Pinot noir. An examination of variation among individual trees. Am. J. Enol. Vitic. 50: Sefton, M.A., I.L. Francis, K.F. Pocock, and P.J. Williams The influence of natural seasoning on the concentration of eugenol, vanillin and cis and trans-β-methyl-γ-octalactone extracted from French and American oak wood. Sci. Aliments 13: Seikel, M.K., F.D. Hostettler, and G.J. Nieman Phenolics of Quercus rubra wood. Phytochemistry 10: Spillman, P.J., M.A. Sefton, and R. Gawel The effect of oak wood source, location of seasoning and coopering on the composition of volatile compounds in oak-matured wines. Aust. J. Grape Wine Res. 10: Towey, J.P., and A.L. Waterhouse The extraction of volatile compounds from French and American oak barrels in Chardonnay during three successive vintages. Am. J. Enol. Vitic. 47: Vivas, N., and Y. Glories. 1996a. Role of oak wood ellagitannins in the oxidation process of red wines during aging. Am. J. Enol. Vitic. 47: Vivas, N., and Y. Glories. 1996b. Study and optimization of phenomena involved in natural seasoning of oak wood. Rev. Fr. Oenol. 158: Vivas, N., K. Amrani-Joutei, Y. Glories, B. Doneche, and C. Brechenmacher Development of microorganisms in oak heartwood (Quercus petraea Liebl.) during natural drying in open air conditions. Ann. Sci. For. 54: