Abstract
Phenolic compounds play a major role in red wine quality, affecting color, taste, mouthfeel, and aging potential. The objective of this research was to compare the phenolic composition of Malbec wines from California and Mendoza, Argentina, produced in 150 L volumes. Sixteen vineyards in California and 26 blocks in Mendoza were selected based on their uniformity and regional representativeness. Anthocyanins and low molecular weight phenolics were analyzed by HPLC at approximately six to nine months of aging. Malbec wines were compared using chemometrics on 33 phenolic components comprising individual anthocyanins, low molecular weight phenolics, and total phenolics by HPLC. The results demonstrate that Malbec wines produced in Mendoza have different phenolic profiles than those produced in California.
Wine chemistry is influenced by grape chemistry and in particular, by phenolic compounds. Factors that are reputed to affect phenolic compound biosynthesis in vines include light, temperature, altitude, soil type, water, nutritional status, microbial interactions, and human interventions such as leaf thinning and irrigation practices (Downey et al. 2006). Few studies have addressed the phenolic composition of Malbec grapes and wines from Argentina. The first published HPLC analysis of Malbec wines was performed by Salagoïty-Auguste and Bertrand in France (1984). More recently, the effect of maceration time on total phenols, total condensed tannins, and polymeric color was evaluated using UV-visible spectrophotometric techniques (Vila et al. 2005). Others have evaluated the effect of cluster thinning on the phenolic composition of Malbec grapes and wines (Matus et al. 2006, Silva et al. 2008, Fanzone et al. 2011). Recently, the effect of oak aging in Malbec phenolic composition was assessed (Catania et al. 2011). In the United States, the only study that has been published on Malbec did not evaluate phenolic composition (Benz et al. 2007). The most comprehensive phenolic composition study surveyed wine samples from different wineries in Mendoza, Argentina (Fanzone et al. 2010).
In the present study, the phenolic composition of Malbec wines from Mendoza, Argentina, and California was compared. The study used controlled conditions for site selection and phenolic compound analysis. Monomeric anthocyanins, hydroxybenzoic acids, hydroxycinammic acids, flavan-3-ols, and flavonols were used to characterize the Malbec wines.
Materials and Methods
Malbec viticultural sites.
Mendoza.
Argentina is divided into provinces; the province of Mendoza is divided into departments, which are divided into districts, analogous to American AVAs. Departments in the Mendoza province that were included in the study were Luján, Maipú, Tupungato, and San Carlos. Within these departments, specific districts were chosen and 26 different Malbec blocks were selected during the second half of 2010. Each block consisted of the necessary number of vines to provide 0.5 to 1 metric tons fruit. Vineyard blocks were selected to provide a representative sample of what is usually referred to as Malbec in the Mendoza province. Selected blocks satisfied a uniformity condition that was verified by the viticulturist managing each evaluated vineyard. The following factors were the same for each vine: plant material (rootstock and scion), planting year, trellis system, vine spacing, hail netting protection (present or absent), pruning system, irrigation (system, operation, and schedule), crop reduction, leaf thinning, and green shoot pruning. A summary of the Mendoza vineyards is available (Supplemental Table 1).
California.
During the first half of 2011, 17 different Malbec blocks were selected in six counties in California (Supplemental Table 2). Vineyard blocks were selected with the intention of obtaining a representative sample of what is usually referred to as Malbec in California. Block characteristics and selection criteria were the same as those used for the Mendoza province blocks.
Malbec winemaking.
Grapes in both regions were harvested at 24.4 ± 1.7 Brix. Malbec grapes were handpicked and transported to the winery in 0.5 tonne plastic macro bins. Grapes were destemmed and crushed, with the resulting must transferred into fermentation vessels. In Mendoza, vinifications were carried out at Catena Zapata winery using 500 L plastic macro bins in duplicate. In California, fermentations were performed at the University of California, Davis Winery in 250 L stainless steel fermentors. Sulfur dioxide (SO2, 150 mg/L) was added at the crush pad as a potassium metabisulfite solution. After 24 hr, 200 mg/L of Lalvin EC-1118 yeast was inoculated into the fermentors (Scott Laboratories, Inc., Petaluma, CA, and Lallemand América Latina, Mendoza, Argentina). One day after the inoculation, 100 mg/L diammonium phosphate was added as a nitrogen source for the yeast. In Mendoza, fermentations began at 22°C with daily punch downs and reached 25°C at peak fermentation. In California, fermentations were pumped-over daily and held at 25°C throughout. Fermentation progress was monitored by measuring density (Brix) and temperature at least twice a day. Wines were pressed ~11 days after crush, when the residual sugar content was below 4 g/L. Only the free run fraction was preserved and inoculated with 100 mg/L of Lalvin VP41 malolactic bacteria culture (Scott Laboratories, Inc. and Lallemand América Latina). Malolactic fermentation (ML) took place at ~25ºC and was considered complete when the malic acid content dropped below 200 mg/L. Once ML was finished, SO2 was added as a potassium metabisulfite solution to achieve a concentration of 35 mg/L free SO2. Wines were stored in 26 L stainless steel drums at 10ºC for at least two weeks, then racked off gross lees into an empty, clean new drum. Finally, twelve 750 mL bottles were filled for each Malbec block replicate. Tin screw caps were used instead of natural cork as stoppers to prevent potential cork taint or variable oxygen incorporation. Wines were stored at 15ºC until analysis.
HPLC analyses of anthocyanins and of low molecular weight phenolic compounds.
Instrumentation.
The chromatographic system used was an Agilent 1100 series HPLC consisting of a photodiode array detector, a single quadrupole mass spectrometer with an API-electrospray source, a quaternary pump, a thermostated column compartment, and an autosampler (HPLC-DAD/API-ES-MS; Agilent, Palo Alto, CA). The instrument was coupled to an Agilent Chem Station (version A.09.03) data-processing station. A reversed-phase Chromolith Performance C18 column (100 mm × 4.6 mm i.d., 2 μm; Merck, Darmstadt, Germany) was used for individual anthocyanin analysis. A reversed-phase Nova-Pak C18 column (300 mm × 3.9 mm i.d., 4 μm; Waters Corp., Milford, MA) was used for low molecular weight phenolic compound analysis.
Standards and reagents.
Standards of gallic acid, caffeic acid, caftaric acid, (+)-catechin, (−)-epicatechin, and quercetin-3-glucoside, were purchased from Sigma-Aldrich (St. Louis, MO), while malvidin-3-glucoside chloride was supplied by Carl Roth (Karlsruhe, Germany). Acetic acid (99%, glacial, Certified ACS), formic acid (88%, Certified ACS), and methanol (HPLC grade) were purchased from Fisher Scientific (Fair Lawn, NJ), while acetonitrile (HPLC grade) was supplied by Sigma-Aldrich.
HPLC analysis of anthocyanins.
HPLC analysis was adapted from Fanzone et al. with minor modifications (2010 (2011). Briefly, bottled wine was thoroughly mixed by inverting the bottle ten times. One mL wine per vineyard block replicate was withdrawn from the bottle, filtered through a 0.45 μm pore size PTFE membrane, and a 25 μL sample was injected into the HPLC-DAD system. Separation was performed at 25ºC. A gradient consisting of solvent A (water/formic acid, 90:10, v/v) and solvent B (acetonitrile) was applied at a flow rate of 1.1 mL/min from 0 to 22 min and at an increasing flow rate 1.1 to 1.5 mL/min from 22 to 35 min as follows: 96 to 85% A and 4 to 15% B from 0 to 12 min, 85 to 85% A and 15 to 15% B from 12 to 22 min, then 85 to 70% A and 15 to 30% B from 22 to 35 min. This was followed by a final wash with 100% methanol for 4 min at 1.0 mL/min and re-equilibration of the column to the starting conditions for 4 min. Photodiode array data acquisition was performed from 250 to 600 nm and an individual DAD signal was stored for 520 nm. The analyses were performed in duplicate. For identification, API-ES-MS was used in positive polarity mode with a mass range of 100 to 1000 m/z. The spray chamber parameters were set as follows: drying gas flow (N2) 12.0 L/min, drying gas temperature (N2) 350ºC, nebulizer pressure 60 psig, and capillary voltage +3000 V.
Anthocyanins were identified using their UV-vis spectra and retention times. Identification of malvidin-3-glucoside was performed by comparing its spectrum and retention time to those of the standard (Buscema 2012). Identification of the remaining anthocyanins was conducted by comparing their spectra and retention times to the ones published by Fanzone et al. (2010) and others provided by Fanzone (personal communication). Confirmation of the identity of anthocyanin pigments was performed by HPLC-DAD/API-ES-MS as described (Monagas et al. 2003).
Quantification was carried out by peak area measurements at 520 nm and comparison with an external malvidin-3-glucoside standard (Fanzone et al. 2010).
HPLC analysis of low molecular weight phenolic compounds.
The HPLC analysis was adapted from Fanzone et al. (2010). Briefly, bottled wine was thoroughly mixed by inverting the bottle ten times. One mL wine per vineyard block replicate was filtered through a 0.45 μm pore size PTFE membrane and then 30 μL was injected in the HPLC-DAD system. Separation was performed at 25ºC. Two mobile phases were employed for elution: A (water/acetic acid, 98:2, v/v) and B (water/acetonitrile/acetic acid, 78:20:2, v/v/v). The gradient profile was 0 to 55 min, 100 to 20% A and 0 to 80% B; 55 to 57 min, 20 to 10% A and 80 to 90% B; 57 to 90 min, 10% A and 90% B isocratic, followed by a 100% methanol wash for five min and re-equilibration of the column for five min. The flow rate was 1.0 mL/min from 0 to 55 min and 1.2 mL/min from 55 to 90 min. Methanol wash was performed at a flow rate of 1.0 mL/min. Detection was performed by scanning from 190 to 400 nm and individual DAD signals were stored for 280 nm, 316 nm, and 365 nm. For identification, API-ES-MS was used in negative polarity mode with a mass range of 100 to 1000 m/z. The spray chamber parameters were as follows: drying gas flow (N2) 12.0 L/min, drying gas temperature (N2) 350ºC, nebulizer pressure 60 psig, and capillary voltage −3000 V.
Specific compounds were identified by comparison of their UV-vis spectra and retention times with those of standards for gallic acid, trans-caffeic acid, trans-caftaric acid, catechin, and epicatechin. Compounds for which standards were not available were identified by comparing their UV-vis spectra and retention times to those published by Monagas et al. (2005). All individual phenolic compounds were confirmed by HPLC-DAD/API-ES-MS as described by Monagas et al. (2005). Quantification was carried out by peak area measurements at 280 nm, 316 nm, or 365 nm according to the evaluated compounds. Quantitative determinations were made by using external standards of commercially available gallic acid, trans-caffeic acid, catechin, epicatechin, trans-caftaric acid, and quercetin-3-glucoside. Compounds for which no standards were available were quantified using the calibration curves of catechin (syringic acid), trans-caftaric acid (trans-p-coumaric acid and unk 1) and quercetin-3-glucoside (unk 2, unk 3, quercetin-3-glucuronide, myricetin, and unk 4). The calibration curves were obtained by injection of standard solutions, under the same conditions as for the samples analyzed, over the range of concentrations observed. For further details of the HPLC method and the corresponding concentrations, refer to Buscema (2012).
The total phenolic content was measured as the total peak area of the HPLC-DAD analysis at 280 nm.
Time of HPLC analysis.
Mendoza Malbec wines were analyzed for monomeric anthocyanins at seven to nine months and for low molecular weight phenolics (LMWP) at eight to nine months after pressing. California wines were analyzed for monomeric anthocyanins and LMWP at five to six months after pressing.
Statistical Analysis.
Analyses of variance (ANOVA) were performed for all phenolic compounds evaluated by HPLC (α = 0.05). Discriminant analyses (DA) was conducted for combined phenolics (monomeric anthocyanin, LMWP, and total phenols (α = 0.1). Canonical variate analysis (CVA) with 90% confidence ellipses was used to illustrate the relationships among different Malbec-producing regions. Principal component analysis (PCA) was used to characterize individual Malbec-producing regions using significant variables (p < 0.05). Statistical analysis was performed with XLSTAT Version 2012.3.01 (Addinsoft, New York, NY).
Results
Total phenols.
Total phenols, measured as the total peak area at 280 nm during the HPLC analysis, ranged from 836.7 mg/L to 1764.4 mg/L gallic acid equivalents (GAE), with an average of 1166.8 mg/L and a coefficient of variation of 15%. Malbec wines from Mendoza and California were significantly different in total phenol concentration (p = 0.004, α = 0.05). Sonoma County wines had the most total phenols (1308.8 mg/L GAE), while Lake County wines had the least (962.2 mg/L). The Mendoza wines had intermediate values, with Tupungato having the highest concentration (1238.7 mg/L GAE), which was not significantly different from San Carlos or Lujan, the lowest, at (1084.0 mg/L).
Total anthocyanins.
Total anthocyanins, expressed in mg/L malvidin-3-glucoside equivalents (ME), ranged from 84.6 mg/L to 708.5 mg/L, with an average of 358.8 mg/L and a coefficient of variation of 44%. Malbec wines from Mendoza and California were significantly different in total anthocyanin concentration (p = <0.0001, α = 0.05). Wine from all Mendoza departments had total anthocyanin concentrations significantly lower than all California wines except those from Yolo County. Napa County wines reached the highest concentration of all the evaluated regions (554.8 mg/L ME), Tupungato had the highest total anthocyanin concentration among the Mendoza set (321.4 mg/L ME), and San Carlos wines had the least combined anthocyanins of the combined set (182.7 mg/L ME).
Combined phenols.
Total phenols, monomeric anthocyanins, and LMWP were combined together and used to compare Malbec wines from the four departments in Mendoza and six counties in California. Total anthocyanins was not included, as this measure is the sum of the individual monomeric anthocyanins quantified; therefore, the total anthocyanin value does not provide any extra information. Thirty-one of the 33 phenolic compounds evaluated were significantly different among the Malbec wines from the four departments in Mendoza and the six counties in California (α = 0.05). Among the LMWP, only quercetin-3-O-glucuronide (p = 0.376) was not significantly different among the Malbec wines from the ten evaluated regions. Within the monomeric anthocyanins, only A-Unk 4 (p = 0.149) was not significantly different between the Mendoza and California Malbec wines. Total phenol concentration was significantly different (p = 0.004). A PCA for the significant phenolic compounds evaluated in the Malbec wines from Mendoza and California generated two components, PC1 and PC2, which together explained 67.7% of the variance (Figure 1). PC1 explained 39.9% of the variation among the wine samples. It was heavily loaded in the positive direction by the vectors malvidin-3-glucoside and trans-malvidin-3-(6″-p-coumaroylglucoside), and in the negative direction by syringic acid and gallic acid. PC2 explained 27.8% of the variation among the wines; it was positively loaded by vectors for vitisin B and delphinidin-3-(6″-acetylglucoside) and negatively by Unk 2 and Unk 3, two of the LMWP. On the first principal component, the California wines showed higher values for malvidin-3-glucoside and trans-malvidin-3-(6″-p-coumaroylglucoside), while the Mendoza set had higher concentrations of syringic and gallic acids.
Factor loadings and scores for a principal component analysis (PCA) performed on the phenolic profile of Malbec wines from four department in Mendoza and six counties in California. Only variables that differed significantly among departments are included (α = 0.05). Abbreviations: GA, gallic acid; t-CaftA, trans-caftaric acid; Cat, catechin; t-CaffA, trans-caffeic acid; SA, syringic acid; Epicat, epicatechin; t-p-cmA, trans-p-coumaric acid; U1, unknown 1; U2, unknown 2; U3, unknown 3; Myr, myricetin; U4, unknown 4; A-U1, A-unknown 1; D-3-g, delphinidin-3-glucoside; C-3-g, cyanidin-3-glucoside; P-3-g, petunidin-3-glucoside; Peo-3-g, peonidin-3-glucoside; M-3-g, malvidin-3-glucoside; D-3-6-ac g, delphinidin-3-(6-acetylglucoside); Vit A, malvidin-3-glucoside pyruvate (vitisin A); Vit B, malvidin-3-glucoside acetate (vitisin B); M-3-6ac g pyr, malvidin-3-(6″-acetylglucoside) pyruvate; C-3-6ac g, cyanidin-3-(6″-acetylglucoside); P-3-6ac g, petunidin-3-(6-acetylglucoside); Peo-3-6ac g, peonidin-3-(6-acetylglucoside); M-3-6ac g, malvidin-3-(6-acetylglucoside); M-3-6caf f, (malvidin-3-(6-caffeoylglucoside); P-3-6pcm g, petunidin-3-(6-p-coumaroylglucoside); Peo-3-6pcm g, peonidin-3-(6-p-coumaroylglucoside); M-3-6pcm g trans, trans-malvidin-3-(6-p-coumaroylglucoside); and TP, total phenols.
A multi-way ANOVA (MANOVA) test indicated that at least one of the ten regions was significantly different from the rest in the combined phenolic profile (p < 0.0001). The Mendoza departments grouped together and were significantly different from the California counties (F1 + F2 = 85.5%; Figure 2). All California counties separated from each other by combined phenolic composition except Sonoma and Monterey. The phenolic compounds with the highest loading on the F1 axis were malvidin-3-(6-acetylglucoside) (4.131) and malvidin-3-glucoside (−3,690), while on the F2 axis cyanidin-3-glucoside (5.390) and A-Unk 4 (−5.254) predominated (Table 1). It is important to note that all four of these variables are monomeric anthocyanins.
Canonical variate analysis (CVA) plot of combined phenolics measured in individual fermentation replicates of Malbec wines from four departments in Mendoza and six counties in California. Ellipses that overlap are not significantly different from one another at the 90% level.
Standardized canonical discriminant function coefficients for F1 and F2 for combined phenolics measured in Malbec wines from four departments in Mendoza and six counties in California.
When evaluated alone, all four Mendoza departments separated from each other; however, San Carlos and Luján were closely associated (Figure 3). The phenolic compounds with the highest loading on the F1 axis were malvidin-3-glucoside (8.816) and malvidin-3-(6-acetylglucoside) (−14.191), while on the F2 axis petudinidin-3-glucoside (26.606) and delphinidin-3-glucoside (−17.321) predominated. It is important to note that all four of these variables are monomeric anthocyanins.
Canonical variate analysis (CVA) plot of combined phenolics measured in individual fermentation replicates of Malbec wines from four departments in Mendoza. Ellipses that overlap are not significantly different from one another at the 90% level.
Discussion
Total phenols.
Malbec wines from Mendoza and California evaluated in this study showed similar mean concentrations for total phenols: 1171.8 mg/L GAE and 1161.5 mg/L GAE, respectively. These concentrations are less than the reported values for commercial Malbec wines from Mendoza, which used the acidified-butanol assay (Fanzone et al. 2010). However, the concentrations found in this study are comparable to those observed in pilot-scale Cabernet Sauvignon wines from California that received no oak aging (Oldfield 2001). Apparently, the small scale used in that study and the current one led to a lower total phenols concentration in the resulting wines.
Total anthocyanins.
Malbec wines from California had more total anthocyanins than Malbec wines from Mendoza: 475.2 mg/L ME and 249.2 mg/L ME, respectively. These concentrations are comparable to those reported for commercial Malbec wines from Mendoza (Fanzone et al. 2010). The concentration of free monomeric anthocyanins decreases considerably during the first year of aging, by forming polymeric pigment or as a result of oxidative reactions (Nagel and Wulf 1979, Boulton et al. 1996, Monagas et al. 2003, Zimman and Waterhouse 2004). The lower average concentration found in the Mendoza Malbecs than in the California Malbecs may be partially explained by the fact that the first group was analyzed seven to nine months after pressing and was shipped from Argentina to the USA, while the second group was analyzed five to six months after pressing and kept in a cool cellar at 15ºC until analysis. Additionally, slightly cooler fermentation temperatures in the Mendoza set might have also influenced the lower total anthocyanin concentrations found in these wines.
Combined phenols.
Chemometrics has been employed effectively in wine differentiation and classification by geographic origin (Arvanitoyannis et al. 1999, Makris et al. 2006, Kallithraka et al. 2007, Rastija et al. 2009, Saurina 2010). In the first principal component of the PCA analysis, the California wines had more malvidin-3-glucoside and malvidin-3-(6″-p-coumaroylglucoside) transisomer, while the Mendoza wines had higher concentrations of syringic and gallic acids (Figure 1). This finding might suggest that the California wines would exhibit a more intense color, as anthocyanins are responsible for this property. However gallic acid and myricetin, as copigmentation cofactors, were more abundant in the Mendoza wines than in the California wines and may result in more purple-colored wines. This finding might also impact the perceived astringency and bitterness of Mendoza wines, as gallic acid can elicit these mouthfeel properties (Robichaud and Noble 1990, Peleg and Noble 1995, Brossaud et al. 2001). Syringic acid may be a by-product of malvidin degradation (Douglas Adams, personal communication). Higher concentrations of syringic acid in Mendoza wines might be due to the fact that these wines were three months older than the California set at the time of analysis, allowing more time for malvidin to degrade and form syringic acid. Additionally, it is important to highlight that total phenol concentration was not a main discriminant factor among the Malbec wines from Mendoza and California. This discovery suggests that wines from both regions might have similar overall antioxidant capacity because of their total phenols concentration (Frankel et al. 1995). Others have proposed that aging potential is related to total phenols concentration and antioxidant capacity (Jaffre et al. 2009). Nevertheless, as the study only evaluated one season, additional vintages should be evaluated to support these suggestions.
In the current study, the application of DA to the phenolic profile of Malbec wines from Mendoza and California allowed successful discrimination between the Argentinian and the American wines with 90% confidence (Figure 2). This finding involved the characterization of the phenolic profile of Malbec wines from California for the first time. A similar approach was followed in two other studies. In the first, wines from four Spanish appellations of origin were successfully discriminated, applying chemometrics to the wines’ phenolic compositions (Peña-Neira et al. 2000). In the second, young red wines from five major viticultural areas of Greece were successfully discriminated by region based on their phenolic profiles (Makris et al. 2006). Unfortunately, both studies included multiple varieties in each region and the regional effect on a specific variety was not evaluated in detail.
When DA was jointly applied to the Mendoza and California Malbec wines, the Mendoza departments did not discriminate from each other significantly (Figure 2). However, when only the Mendoza set was evaluated, all four departments separated from each other with 90% confidence (Figure 3). Malbec wines from three provinces in Argentina were successfully distinguished by applying chemometrics to the phenolic profile and multi-element composition of the wines (Di Paola-Naranjo et al. 2011). A previous group successfully discriminated among commercial Malbec wines from three areas of the Mendoza province, using only the phenolic profile of the wines (Fanzone et al. 2010). The current study is the first to successfully discriminate Malbec wines among Mendoza departments (Figure 3) and California counties (Figure 1) based on their phenolic profiles. Furthermore, the present study is the first to distinguish wines made from a single cultivar in Mendoza departments and California counties based on their phenolic profile.
When DA was jointly applied to Mendoza and California wines, the phenolic compounds with the greatest contribution to the F1 axis were malvidin-3-(6-acetylglucoside) (4.131) and malvidin-3-glucoside (−3.690), while on the F2 axis, cyanidin-3-glucoside (5.390) and A-Unk 4 (−5.254) predominated (Table 1). This finding is in agreement with previous studies that showed that sunlight properties, particularly UV-B radiation, vary in different departments in Mendoza, causing noticeable variation in the phenolic composition of the resulting grapes and wines (Berli et al. 2008, Berli et al. 2011).
The ability to discriminate among Malbec wines, and varietal wines in general, by origin in a department/county or even vineyard has tremendous potential for screening new vineyards or regions for suitability for producing specific wines. The results of this research could also be applied to the verification of origin of a given wine. Finally, the results of this study suggest that Malbec wines with different phenolic profiles could be produced in Mendoza and California, and even within these two regions.
Conclusion
In this study, the phenolic patterns of young Malbec wines from Mendoza and California were evaluated and compared. The Malbec wines were distinguished using chemometrics on 33 phenolic components comprising anthocyanins, LMWP, and total phenolics. These differences suggest that Malbec wines with different phenolic profiles can be produced in Mendoza and California.
Acknowledgments
Acknowledgments: The authors would like to recognize the generosity of the Catena Zapata family, the Stephen Sinclair Scott Endowment, and the vineyard owners and winemakers in the United States who contributed to this project. The authors also acknowledge the support of the Catena Institute of Wine, Cary Doyle (UC Davis Advanced Instrumentation Lab), and Chik Brenneman (UC Davis Pilot Winery). Special appreciation is included for Doug Adams and Jean-Jacques Lambert for their valuable comments throughout the project. Finally, the authors are grateful for funding support from the Wine Spectator Scholarship, the Horace O. Lanza Scholarship, the David E. Gallo Educational Enhancement Fund, and an ASEV Scholarship.
Footnotes
-
Supplemental data is freely available with the online version of this article at www.ajevonline.org.
- Received January 2014.
- Revision received June 2014.
- Revision received August 2014.
- Accepted October 2014.
- Published online January 2015
- ©2015 by the American Society for Enology and Viticulture
Literature Cited
Vol 66 Issue 1
Jump to section
Related Articles
Cited By...
More from this TOC section
Similar Articles
