Abstract
A procedure to quantify volatile, organic sulfur compounds in wines was developed using solid-phase microextraction to preconcentrate the analytes followed by gas chromatography and detection with a sulfur chemiluminescence detector. The compounds studied (hydrogen sulfide, carbon disulfide, methanethiol, ethanethiol, dimethyl sulfide, diethyl sulfide, dimethyl disulfide, and diethyl disulfide) have low sensory thresholds and contribute aromas ranging from rotten egg to cabbagelike in wines. These compounds, particularly the thiols, are very sensitive to oxidation. In the developed method, several steps were taken to minimize effects of oxidation and artifact formation, including the use of internal standards and air exclusion during sample preparation. However, oxidation could not be completely prevented, making it difficult to obtain acceptable calibration curves in model wines. Therefore, to correct for these effects, the method of standard addition was used to measure the concentrations of sulfur compounds in wines. The developed method was shown to be reproducible (<10% relative standard deviation) with excellent recovery (>97%) and limits of detection similar to or below sensory threshold levels. Using the developed method, a range of sulfur compounds were identified in wines associated with sensory descriptors characteristic of the S-containing compounds.
- CS2, carbon disulfide
- DEDS, diethyl disulfide
- DES, diethyl sulfide
- DMDS, dimethyl disulfide
- DMS, dimethyl sulfide
- EMS, ethyl methyl sulfide
- EtSH, ethanethiol
- H2S, hydrogen sulfide
- MeSH, methanethiol
- PrSH, propanethiol
Low-molecular-weight organic sulfur compounds, including hydrogen sulfide (H2S) and the volatile organic thiols (e.g., methanethiol, ethanethiol), sulfides (e.g., dimethyl sulfide and diethyl sulfide), and disulfides (e.g., dimethyl disulfide and diethyl disulfide), contribute important aromas and flavors to wines (Mestres et al. 2000, Vermeulen et al. 2005). These compounds are generally formed during fermentation (Landaud et al. 2008), and they typically have low sensory thresholds and aroma descriptions ranging from rotten egg to cabbage to onion (Table 1⇓). Although often considered to contribute negative aroma characters to wines, at low concentrations the S-containing compounds may contribute to aroma complexity. In addition, dimethyl sulfide has been reported to contribute positively to the aroma of some wines (Spedding and Raut 1982) and to act as an “odor enhancer” of fruity aromas (Segurel et al. 2004, Escudero et al. 2007).
The analysis of sulfur volatiles is challenging. The range of volatilities makes chromatographic separation difficult, particularly for the most highly volatile compounds (e.g., H2S, CS2, SO2, and MeSH); therefore, few methods simultaneously measure these compounds together with the higher molecular weight thiols, sulfides, and disulfides. Previous strategies to accomplish improved separation, particularly for H2S, CS2, SO2, and MeSH, include serially connecting nonpolar and polar capillary columns (Park et al. 2000, Siebert et al. 2010) and use of polar columns with thick film stationary phases (Fang and Qian 2005, López et al. 2007). Additionally, the S-containing volatiles are usually present at low concentrations in wine and are prone to oxidation during the several stages of sample preparation and analysis, particularly H2S and the low-molecular-weight thiols (Wardencki 1998, Mestres et al. 2000, Lestremau et al. 2004, Lopez et al. 2007). Thus, preconcentration is often the first step in methods aimed at the quantification of these compounds in wine. However, because S-containing compounds are sensitive to aerial oxidation, it is also desirable to conduct sample preparation and analysis in as few steps as possible.
Solid-phase microextraction (SPME) is a safe, rapid, and straightforward procedure that is amenable to automation and has been successfully applied to the analyses of a wide variety of volatile compounds in wine and other food and beverage matrices (Polaskova et al. 2008). The ease of use and the short sampling times that can minimize artifact formation have resulted in application of SPME combined with gas chromatography (GC) for the analysis of sulfur compounds in wine. However, most previous methods have used carboxen-based SPME phases, which can be susceptible to competition among analytes for fiber sorption sites (Murray 2001). This competition has not been evaluated in the mixed carboxen-divinylbenzene (DVB)-polydimethylsiloxane (PDMS) phases that have become commercially available and that offer multiple sorption mechanisms that may minimize this effect.
Previous methods for sulfur volatiles in wines have used a variety of detectors such as mass spectrometry (MS) (Mestres et al. 2002, Fedrizzi et al. 2007), atomic emission detection (AED) (Campillo et al. 2009), and pulsed flame photometric detection (PFPD) (López et al. 2007, Fang and Qian 2005). An alternative to these detectors is the sulfur chemiluminescence detector (SCD). The SCD has been widely used in the petroleum industry (Yan 2006) and can be easier to operate, simpler to maintain, and less expensive than AED (Al-Attabi et al. 2009) with better sensitivity than is typically observed with MS in the total ion mode. The SCD is a specific and universal sulfur detector, resulting in simplified analysis of S-containing compounds in complex organic matrices such as wine. The SCD produces a linear and equimolar response over a wide range of analyte concentrations (contrasting with the nonlinear responses obtained with the PFPD). Moreover, the flameless SCD (used in this study) has significantly improved the detector performance by reducing or eliminating the effects of limited sensitivity, selectivity, and general stability that were typical of the flame SCD (Yan 2006). Gas chromatography with SCD has been used in combination with static headspace sampling, liquid-liquid extraction, and static headspace sampling combined with cool on-column injection for analysis of volatile, organic sulfur compounds in wine (Nedjma and Maujean 1995, Vidal et al. 1996, Rauhut et al. 1998, 2007, Siebert et al. 2010). However, to the best of our knowledge no method has been reported for the use of SPME-GC-SCD in wine.
The advantages of both SPME and the SCD as outlined above therefore provided the impetus for this work, which is focused on the development of a reproducible and sensitive SPME-GC-SCD method, with a minimum of artifact formation, for the improved analysis in wines of S-containing compounds with a range of volatilities.
Materials and Methods
Standards.
Na2S.9H2O (98%), sodium methanethiolate (95%), propanethiol (99%) (PrSH), diethyl disulfide (99%) (DEDS), and dimethyl sulfide (99%) (DMS) were purchased from Sigma Chemical Co. (St. Louis, MO); carbon disulfide (99.9%) and diethyl sulfide (98%) (DES) were from Fluka (St. Louis, MO). Dimethyl disulfide (99%) (DMDS), ethyl methyl sulfide (98%) (EMS), and ethanethiol (99%) (EtSH) were from Acros (Geel, Belgium). Standards purchased were of the highest purity available and were used without further purification. All pure standards were stored under nitrogen.
Na2S.9H2O was used as the source of H2S while sodium methanethiolate was used as the source for methanethiol (MeSH) (Siebert et al. 2010). Stock solutions of these two standards were prepared using degassed water to prevent their oxidation (Herszage and Afonso 2003, Herszage et al. 2003). All other stock solutions were prepared by diluting the compound of interest 1000 times in ethanol (1000 mg L−1). For preparation of final standards below 100 μg L−1, a second stock solution was used; this second stock was prepared by a further 1:1000 dilution (1000 μg L−1) of the original stock in the appropriate matrix (model wine or wine). Final standards were prepared from dilutions of both of these stock solutions in the matrix needed (Table 2⇓).
Several extra steps were taken during sample preparation to minimize oxidation of the compounds. All stock solutions, standards, and spiked wine samples were prepared fresh daily at room temperature, under nitrogen using a Captair Pyramid Glove Bag (Erlab, Rowley, MA). In addition, deactivated glass headspace sample vials (Restek, Bellefonte, PA) were used and were closed under nitrogen inside the glove bag and held until extraction. To avoid loss of analyte, sample vials were not sparged directly with nitrogen.
To minimize the amount of air inside the glove bag before use, the bag was filled with enough nitrogen to provide a positive internal pressure. With nitrogen still flowing, the bag was deflated slightly (to approximately atmospheric pressure). This fill cycle was repeated several times, after which sample preparation was done in the glove bag under positive nitrogen pressure.
Based on previous reports (Miracle et al. 2005), PrSH was used as internal standard for H2S, and the thiols and EMS was used as internal standard for sulfides and disulfides. These compounds were not observed in any of the wine samples. The internal standard solution was prepared following the same protocol as above, to a final concentration in wine of 0.28 μg L−1 for PrSH and 3.3 μg L−1 for EMS.
Model wine.
Standards were prepared in a model wine solution made by dissolving 2 g potassium hydrogen tartrate (Sigma) in a 12% ethanol aqueous solution and adjusting the pH to 3.2. Absolute ethanol for preparation of the 12% aqueous solution was obtained from Gold Shield Chemical (Hayward, CA).
Wines.
Three wines produced in the University of California (UC), Davis winery from grapes grown in the UC Davis vineyard (2005 Merlot, 2006 Syrah rosé, and 2008 Torrontes) were analyzed by the method presented here. These wines were identified by winemakers and laboratory personnel as having characteristic sulfur-containing aromas associated with rotten egg, rubber, or cabbagelike aromas. Free SO2, at typical concentrations present in wines, has a broad chromatographic peak that masks the signal for MeSH (Table 2⇑). As its presence in wine samples can be deleterious for the GC column, 5 μL acetaldehyde (Sigma) was added to wine samples to bind the SO2 present in the sample.
SPME method.
Wine samples or standards (15 mL) were placed in 20-mL glass crimp-top deactivated vials (Restek) containing 3 g NaCl for SPME extraction. To decrease the chance of oxidation, the headspace was minimized as much as possible while leaving enough room to expose the SPME fiber. Crimp-top vials provided better closure than screwcap vials. Immediately after crimping, samples were extracted manually for 30 min at 24 ± 1°C. Three different SPME fibers were evaluated as potential candidates for the method: 75 μm carboxen/polydimethylsiloxane (CAR/PDMS, black hub), 65 μm polydimethylsiloxane/divinylbenzene (PDMS/DVB, blue hub), and 50/30 μm divinylbenzene/carboxen/polydimethylsiloxane (DVB/CAR/PDMS, grey hub).
GC analysis.
The method was adapted from a previous one for sulfur-containing volatiles in beer (Miracle et al. 2005). A 5890 HP GC (Hewlett Packard/Agilent, Little Falls, DE), equipped with a J&W GS-Gaspro PLOT (60 m x 0.32 mm) column (Agilent, Folsom, CA), connected to a Sievers 355 sulfur chemiluminescence detector (Agilent) was used for all analyses. The GC was set for splitless injection with an inlet temperature of 250°C and using a 0.7 mm i.d., deactivated glass SPME injection liner (Supelco, Saint Louis, MO). The split flow was opened at 3 min following the injection and closed again at 13 min after injection. A volumetric flow of 3 mL min−1 using helium as a carrier gas was used with a constant column head pressure of 150 kPa. The oven temperature program started with an initial setting of 40°C followed by an immediate ramp of 10°C min−1 to 260°C, followed by a hold of 3 min. The SCD burner temperature was 800°C with a hydrogen flow rate of 100 mL min−1 and an air flow rate of 40 mL min−1. The SCD pressure was 6 Torr with the controller at 200 Torr.
Samples were manually injected; the fiber was exposed in the injector for 2 min, retrieved, and exposed to the headspace of the following sample. A fiber blank was run at the beginning of each day to ensure the fiber was clean. The detector was periodically conditioned, as recommended by the manufacturer, to avoid loss of sensitivity. A SPME fiber was exposed for 10 to 15 min in the headspace of a 1:1000 (v/v) ethanolic solution of CS2 and subsequently placed in the injector. The oven was then ramped as normally for a sample injection with the fiber in the inlet.
Calibration and quantitation.
Peaks corresponding to the different analytes were integrated using the data analysis package included in ChemStation version A.03.34 (Agilent, Little Falls, DE). Calibration curves in model solutions or wines were created by plotting the peak area ratio of the analyte peak area to the internal standard peak area for a range of analyte concentrations (Table 2⇑) and linear regression equations were calculated. In real wine samples, the method of standard addition was used for quantitation. Standards and internal standards were added to the wine matrix as described above. Peak area ratios of the analytes to the area of internal standards were plotted against the analyte concentrations, and the concentrations in the original sample were determined at the point where the linear regression line intersected the x axis (y = 0). In all cases, linear regression analyses were done using Excel (Microsoft, Redmond, WA).
Precision and recovery studies.
Wines were spiked with known concentrations of analytes and analyzed in triplicate. The relative standard deviation (%RSD) was used as a measurement of precision. To calculate recovery factors, a standard of known concentration was spiked into the wine samples and standard addition calibrations were performed as described above using standards with concentrations above that of the recovery spike. The calculated concentration of the recovery spike was then determined from the standard addition calibration curve. Percent recovery was determined from the calculated analyte concentration divided by the known (spiked) concentration (times 100%).
Results and Discussion
A procedure using HS-SPME analysis combined with gas chromatography and sulfur chemiluminescence detection was developed to analyze volatile, organic sulfur compounds in wines. Our approach involved use of a porous layer open tubular (PLOT) GC column specifically designed for analysis of volatile gaseous analytes combined with careful sample handling to minimize oxidation and, importantly, the use of internal standards and standard addition calibrations to control for matrix effects and artifact formation, which are common problems in the analysis of S-containing analytes.
GC analysis.
The PLOT capillary columns used in this study gave excellent separation in model solutions for all of the analytes studied (Table 2⇑). The PLOT columns have a solid stationary phase instead of a liquid film, as is present in most capillary GC columns. Originally developed as an improvement over packed columns to separate gaseous mixtures, PLOT columns often display a unique selectivity based on the molecular shape and size of the solid phase (Ji et al. 1999). The PLOT column used here consisted of a silica material chemically bonded to the silica column, enabling it to withstand high carrier gas flows and gas head pressures without the risk of particle dislodging. The column also has excellent inertness (Ji et al. 1999). This inertness and the ability to separate low molecular weight gases makes the column ideally suited to analysis of volatile sulfur compounds using a sulfur-specific detector like the SCD; however, no applications of this column for wine analysis have been previously reported. This PLOT column has been designed for the analysis of gases or volatile liquids, with poor separation for analytes of boiling points >150–200°C; therefore, other less volatile sulfur compounds usually reported in wine (such as 2-mercaptoethanol, 3-methylthio-1-propanol (methionol), 4-methylthio-1-butanol, and 3-mercaptohexanol) could not be analyzed using this column. The injection of single standard solutions for these compounds showed no signal, suggesting that they are retained in the column.
In wines with high free SO2 concentrations, the SO2 and MeSH peaks can be poorly resolved due to the large and tailing SO2 peak, which masks the smaller, later-eluting MeSH peak (Table 2⇑). The addition of acetaldehyde or other aldehydes binds the SO2, making it nonvolatile without affecting the peaks of other compounds (Fang and Qian 2005, López et al. 2007).
SPME fiber, matrix effects, and extraction conditions.
Three SPME fiber types (CAR/PDMS, PDMS/DVB, DVB/CAR/PDMS) were evaluated based on previous reports of volatile sulfur compound analysis in wine (Mestres et al. 2002, Miracle et al. 2005). Fibers containing CAR or DVB were reportedly the best choices because they had higher sensitivities (Mestres et al. 2002). Of the three fibers evaluated in our study, the DVB/CAR/PDMS had the best reproducibility in a model wine matrix for all compounds tested (expressed as %RSD, n = 7), ranging from 5 to 9% over the concentration ranges (Table 2⇑). The other two fibers had poorer reproducibility (>15% RSD) over the same concentration range (data not shown). The sensitivity was 5 to 20% higher for the DVB/CAR/PDMS fiber compared to the PDMS/DVB fiber and was 40 to 50% higher compared to the CAR/PDMS fiber (data not shown).
Using a carboxen-based SPME fiber, sample matrix effects in the analysis of sulfur compounds have been previously reported where matrix components appeared to compete with the analytes for the SPME fiber sorption sites, limiting the sensitivity and dynamic range of the analysis (Murray 2001). Similar results were also observed using a combination CAR/PDMS SPME fiber (López et al. 2007), and GC responses of S-containing analytes did not decrease linearly following serial dilutions of wines spiked with the analytes. Consequently, matrix dilution was suggested as one way to eliminate this effect in SPME extractions with a CAR/PDMS fiber (López et al. 2007). However, a significant disadvantage of this method is that greater oxidation occurs with increased sample handling. Here we used a mixed-phase SPME fiber, DVB/CAR/PDMS, which showed excellent sensitivity and reproducibility for the S-compounds studied. However, competition for fiber sorption sites with this fiber has not been directly evaluated. Therefore, we performed an experiment similar to one reported elsewhere (López et al. 2007) to establish if the same effects were observed in the mixed phase we used. Serial dilutions of red wine (Merlot) showed a nonlinear decrease in signal response with increasing dilution for H2S, CS2, DMS, and DES, indicating that competition for fiber sorption sites may also influence quantitation of S-containing analytes with this SPME fiber (Figure 1⇓).
All SPME extractions were performed at room temperature (25 ± 1°C). Although increasing the temperature increases the volatility of analytes, it also increases the rate of oxidation (Mestres et al. 2002). An unidentified peak appeared immediately after the signal for EtSH in chromatograms from samples that were extracted at 30°C, possibly because of oxidation and/or the formation of mixed disulfides. Therefore, extraction temperature was maintained at 25 ± 1°C for all further analyses.
Initially a 15-min extraction time was used, but the time was increased to 30 min to improve sensitivity (data not shown). Although longer times were not tested, the 30-min extraction was the best compromise between obtaining the greatest extraction from the headspace while keeping analysis time under reasonable limits and minimizing the time samples sit on the bench, which limits oxidation.
Calibration, reproducibility, accuracy: Oxidation and artifact formation.
In a model wine matrix, sulfides and disulfides showed excellent response linearity over the concentration ranges studied with correlation coefficients (r) >0.99 (Table 2⇑). Spiked recoveries of these analytes were also greater than 97% (Table 3⇓). However, poor linearity and recovery of the volatile thiols (H2S, MeSH, and EtSH) was observed in model wine (Table 2⇑, Table 3⇓). Oxidation of volatile sulfur compounds is a common problem complicating their quantification, especially at the low levels typical of wines and other fermented beverages. Preparing samples immediately before analysis may minimize oxidation (López et al. 2007); however, this technique limits the use of autosamplers for running many samples over extended time. We attempted to avoid oxidation of the sulfur compounds in model wine solutions by preparing and capping the samples under nitrogen (including preparations in a glove box saturated with nitrogen gas) and by using deactivated vials and GC inlet liners. However, significant oxidation of H2S, MeSH, and EtSH to the corresponding disulfides was still observed even with these precautions. Although one recent report suggested the oxidation can be avoided (Fang and Qian 2005), according to other reports (López et al. 2007) and our own experience, oxidation of thiols in particular seems very difficult to avoid. In a recent study, all standards and samples were prepared and held at 4°C in nitrogen sparged containers and samples were injected using cool on-column injection to minimize oxidation (Siebert et al. 2010). Although maintaining the samples at low temperatures appears beneficial, it was not evaluated for the current study. One study examined the causes for artifact formation in the analysis of sulfur compounds in gaseous samples by SPME GC (Lestremau et al. 2004). The authors concluded that although carboxen (which is part of the fiber sorption material used in this study) may catalyze the oxidation of sulfur compounds, the main factor contributing to the oxidation was the high temperature of the injector that was required for fiber desorption. Nevertheless other factors also seem play a role in our analyses because we do not see extensive oxidation and poor recoveries and linearities for the thiols when analyzed in a wine matrix (Table 2⇑, Table 3⇓). If the SPME fiber and/or injector temperature were responsible for the oxidation of thiols during analysis of these wine samples, then there would be similar amounts of oxidation in wine matrices and in the model wine matrix.
In a study of the oxidation of 3-mercaptohexanol by oxygen in a model wine matrix, the authors concluded that while some polyphenols (e.g., catechin) can catalyze this oxidation, SO2 and anthocyanins are effective antioxidants in wine (Blanchard et al. 2004). Similarly SO2 effectively limited oxidation of 3-mercaptohexanol, 2-furanmethanethiol, and 4-methyl-4-sulfanylpentan-2-one (Nikolantonaki et al. 2010), with the amount of thiol oxidation dependent on the thiol structure (primary>secondary>tertiary). These findings are consistent with our studies indicating that during analysis, oxidation of sulfur compounds is less likely to occur in real wine matrices (i.e., containing SO2 and polyphenols) than in a model wine matrix.
The degree of oxidation of EtSH in model wine was reduced to only 5 to 8% of the initial EtSH concentration when 96 μg L−1 total SO2 was added to the solution compared to oxidation of 20 to 25% of the EtSH without SO2 addition (based on the peak areas for both EtSH and its oxidation product DEDS). Despite the apparent beneficial effects of SO2 on thiol stabilization, it was not added to the model wines because its highly acidic nature in the GC-SCD system could have deleterious effects over time.
Results indicate that effects of matrix composition on competition for SPME sorption sites with carboxen-based fibers can be minimized by matrix/sample dilution (as recently also proposed by López et al. 2007), while oxidation appears to be minimized in the presence of the matrix components. To minimize these counteracting effects of the matrix, standard addition calibrations using the response ratios from internal standard additions were used for all subsequent quantitative analyses (with PrSH the internal standard for H2S and thiols; EMS the internal standard for sulfides and disulfides). Although time-consuming because of the many standards run for each sample, standard addition calibrations can be ideal when the matrix is complex and affects the analytical response (Harris 1999). Using our optimized method, all standard addition calibrations gave reproducible results (% RSD < 10%; Table 4⇓) with linear correlation coefficients (r) ≥ 0.99 (Table 5⇓). Overall recovery (accuracy) was greater than 97%, indicating a high degree of accuracy for the developed method (Table 3⇑).
Stable isotope internal standards are frequently used to counteract matrix effects, but they are not compatible with analysis by SCD. It has been proposed that when using a non-isotopically labeled internal standard for quantitation of H2S, as used here, a statistical variance component model can satisfactorily account for matrix-induced deviations (Lavagnini et al. 2009). This variance component model was not evaluated in the present study, but may provide an alternative to the more time-consuming standard addition approach.
Limits of quantitation and detection.
Method detection limits were estimated as the concentrations where the signal to noise ratio equals 3, while quantitation limits correspond to a signal to noise ratio of 10 (Christian 2004) (Table 2⇑). Comparison of the limits of detection (Table 2⇑) to the sensory thresholds (Table 1⇑) indicates that, with the exception of MeSH, the method presented here provides the necessary sensitivity for the determination of these important aroma compounds in wines with a minimal sample preparation and in a reasonable amount of time (<10 min sample preparation followed by 30-min extraction).
When liquid injection is the method of choice, the SCD has been reported to have a higher sensitivity (~1 pg sulfurs−1) than the flame photometric detector (FPD) (3 pg sulfurs−1) and is comparable to the sensitivity of PFPD (0.7 pg sulfur −1) and AED (1 pg sulfur s−1) (Yan 2006, Al-Attabi et al. 2009). Although significant losses in sensitivity after a few analyses were reported elsewhere (Hill and Smith 2000), such sensitivity losses with repeated analyses were not observed, in agreement with previous reports (Miracle et al. 2005, Siebert et al. 2010).
Analysis of wine samples.
Three wine samples identified by winemakers and laboratory personnel as having characteristic sulfur related off-aromas were analyzed using the method described here (Table 5⇑). All compounds except DMDS were quantified in at least one of the samples; however, the predominant sulfur compound was different for each wine. MeSH was the predominant compound in the 2008 Torrontes; H2S was predominant in the 2006 Syrah rosé; and DMS was predominant in the 2005 Merlot. Only CS2 was quantified in all three wines. Further sensory analysis is needed to relate specific sensory attributes to the individual volatile sulfur compounds. When combined with such sensory studies, this method now makes it possible to understand more fully the role of these sulfur compounds in the sensory properties of wines.
Conclusions
A GC method using PLOT columns for separation of highly volatile compounds, and combining the ease of sample preparation provided by SPME with the specificity and sensitivity of the SCD, provided the opportunity to quantify a range of volatile, organic sulfur-containing compounds in wine with >97% accuracy and relative standard deviations of <10%. The detection limits for the compounds studied were generally below their sensory thresholds.
Oxidation of S-containing compounds during their analysis remains one of the main obstacles for their quantification. Although this issue was minimized by taking several precautions during the sample preparation, analyte oxidation could not be completely avoided, particularly for the thiols studied. The optimal strategy for compound analysis was to use the method of standard addition for quantification of these analytes in the wine samples. This method provides the highest accuracy and precision and was successfully used to measure eight low-molecular-weight sulfur compounds in wine samples characterized as having undesirable sulfur-related aromas. The optimized method offers the opportunity to further understand the effects of fermentation conditions on the production of sulfur aromas and to relate wine composition information to sensory properties.
Footnotes
-
Acknowledgments: Funding for this work was provided in part by the American Vineyard Foundation and the California Competitive Grants Program for Research in Viticulture and Enology.
- Received September 2010.
- Revision received November 2010.
- Accepted November 2010.
- Published online December 1969
- Copyright © 2011 by the American Society for Enology and Viticulture
Literature Cited
Sign in for ASEV members
ASEV Members, please sign in at ASEV to access the journal online.
Sign in for Institutional and Non-member Subscribers
Log in using your username and password
Pay Per Article - You may access this article (from the computer you are currently using) for 2 day for US$10.00
Regain Access - You can regain access to a recent Pay per Article purchase if your access period has not yet expired.