dulcratin f White Wine with Xylitl and Seyval Blanc Juice Reserve C. L. DUITSCHAVR ~, CARL BUTAU 2, and G. C. ASHTN 3 Riesling and tw blends f white French hybrid grape varieties (Blend A and Blend B) were vinified t dryness at 13 C. The wines were edulcrated with either Seyval blanc juice reserve r with xylitl at varius cncentratins. Sensry evaluatin f the wines by experienced tasters shwed that juice reserve was the mst effective sweetener, but wines edulcrated with xylitl were highly acceptable. Wines, particularly white table wines and lw-alchl sparkling wines, are ften edulcrated t balance the acidity and astrigency and, therefre, t imprve the palatability (2,3,7,8). The mst cmmn edulcrant used fr this purpse is sucrse. This residual sugar is a substrate fr micrrganisms such as yeasts and bacteria that may ccur as cntaminants in wines and can cause refermentatin f the wine after bttling, resulting in cnsiderable ecnmic lsses fr the wineries. Additin f a sweetener that cannt be utilized by yeasts and bacteria wuld vercme this prblem. Althugh xylitl is at the present time nt apprved fr the sweetening f wines, the authrs have chsen it as an edulcrant because it has been allwed in sme fds and wuld appear t have pssibilities fr sweetening f beverages (6,9). The bjectives f this wrk were: 1) t vinify t dryness musts f several white grape varieties, Blend A, Blend B and Riesling using the lw temperature vinificatin prcess as described previusly (4); 2) t cntrast edulcratin f wines with Seyval blanc juice reserve and a nn-fermentable sweetener xylitl; and 3) t evaluate the relatin between certain wine characteristics and different cncentratins f the tw sweeteners. Materials and Methds Grape variety: Riesling and tw mixtures f varius experimental white French hybrid varieties (Blend A and Blend B) were prvided by Chateau des Charmes Wines, Ltd. (St. David's, ntari, Canada). Harvest date fr Riesling crrespnded with that f cmmercial practice fr 1981. Grapes fr Blends A and B were harvested at ptimal ripeness. Yeasts: Saccharmyces cerevisiae strain ST 61 frm the culture cllectin f the Department f Fd Science, University f Guelph, was used. Maintenance and preparatin f the inculum (3% vv) was as previusly described (3). Methd f vinificatin: The prcedure fr vinificatin has been utlined previusly (4). The grapes were crushed and pressed with a basket-type hand press; the yield was 0.65 L mustkg f grapes. Sulfur dixide was added in the frm f K2S205. The three musts had a sugar cncentratin f apprximately 20 Brix, a ph f 3.0 t 3.2 and a ttal acidity f 0.9 t 1.2 grams tartaric acid per 0 ml. They were, therefre, nt diluted r chaptalized. ~Assciate Prfessr and 2Research Assistant, Department f Fd Science, and 3Prfessr meritus, Department f Mathematics and Statistics, University f Guelph, Guelph, ntari, Canada N1G 2W1. This wrk was supprted in part by the ntari Ministry f Agriculture and Fd. Manuscript submitted fr publicatin 25 July 1983. 88 The vinificatin was carried ut in three replicatins f 20 L each. Sweeteners: Seyval blanc juice reserve (JR) was prepared as previusly described (3). Xylitl (Hffman- LaRche, 00 Blvd. Rche, Vaudreuil, Quebec, Canada J7V 6B3) was used in the granular frm. dulcratin: At time f bttling, the wines were edulcrated with either ne f the tw sweeteners at five different cncentratins fr each sweetener (Table 1). The actual amunt f sweetener added was nt the same fr each wine because the bjective was t btain wellbalanced wines rather than t edulcrate the wines t the same sugar cntent. Table 1. Amunt f 2 sweeteners added t each cncentratin per 1 liter f 3 different wines. Types f wines Sweeteners BlendA SeyvalblancJR, ml 0.0 Xylitl, g 0.00 Blend B SeyvalblancJR, ML 0.0 Xylitl, g 0.00 Riesling SeyvalblancJR, ml 0.0 Xylitl, g 0.00 50.0 0.0 150.0 200.0 6.75 13.50 20.25 27.00 37.5 75.0 112.5 150.0 5.00.00 15.00 20.00 50.0 0.0 150.0 200.0 6.75 13.50 20.25 27.00 Bttling: The wines were bttled by sterile filtratin as previusly described (3). Chemical analysis: Analyses fr ttal acidity, ph, vlatile acidity, residual sugar and alchl were perfrmed accrding t Amerine and ugh (1). A mdified enzymatic methd was used fr the analysis f xylitl (5). Sensry evaluatin: Sensry evaluatin f the wines was perfrmed six mnths after bttling by a panel f nine judges using the evaluatin frm shwn ~ Figure 1. The wine samples were served ne at a time, each with a separate evaluatin frm. The judges were prvided with a reference wine, Riesling 1980, Chateau des Charmes, which they had evaluated in a pre-experiment and assigned scres fr the characteristics - sweetness, acidity and balance (Fig. 1). The wines were served at a temperature f 13 t 15 C in 150 ml tulip-shaped glasses cded with randmized three digit numbers fr identificatin. xperimental design: The prcedure fr the prductin f the wines and their sensry evaluatins were carried ut accrding t a 3 2 5 factrial in a splitsplit-plt design in three replicatins. The three grape varieties (wines) frmed the main plt, the tw sweeteners the sub-plts, and the five sugar cncentratins the sub-sub-plts. n the assumptin that a judge culd
DULCRATIN F WHIT WIN -- 89 Name:... Date:... Cde:... Please taste the wine sample and answer by placing a vertical line (prperly identified) acrss the hrizntal line at the pint that best describes that prperty in the sample. A reference wine is prvided and identified as R; the sample shuld be judged by cmparisn with the reference. When tasting the wines, please swallw. Take sufficient time fr the evaluatin. If yu have any questins, ask the experimenter. Thank yu! very dry SWTNSS very sweet weak ACIDITY strng unbalanced BALANC well-balanced R dislike very much VRALL IMPRSSIN like very much Fig. 1. valuatin frm fr the sensry evaluatin. N.B.: The respnses f the panelists were recrded in cm measured frm left t right. Wine Blend A Table 2. Residual sugars, xylitl and alchl cncentratins, ph, vlatile and ttal acidity f 3 wines, Blend A, Blend B and Riesling; 5 cncentratins f either sweetener, juice reserve r xylitl; average f 2 repetitins. Average Blend B Average Riesling Average Sweetener Cde a Res. sugar (g0 ml) Xylitl (g0 ml) Alchl (% vv) ph V.A. b (g0 ml) T.A. c (g0 ml) A J0 1.06 A J1 1.60 A J2 1.99 A J3 2.41 A J4 2.79 AX0 -- 0.00 AXl ~ 0.66 AX2 ~ 1.55 AX3 ~ 2.08 AX4 m 2.48 11 84 11 35 11 03 53 991 11 80 11 36 11 39 11.36 11.31 11.18 B J0 0.62 m 11.60 B J1 1.07 ~ 11.29 B J2 1.43 --.89 B J3 1.72 --.78 B J4 2.16 ~.19 BX0 ~ 0.00 11.62 BXl ~ 0.53 11.57 BX2 m 0.97 11.44 BX3 ~ 1.44 11.81 BX4 ~ 2.13 11.50 ~ 11.60 3.33 0.048 0.79 3.20 0.033 0.80 3.31 0.033 0.80 3.28 0.032 0.80 3.28 0.033 0.84 3.33 0.048 0.78 3.33 0.038 0.75 3.34 0.042 0.76 3.34 0.036 0.72 3.33 0.036 0.73 3.31 0.038 0.78 3 34 3 33 3 33 3 32 3 35 R J0 0.79 m 11.76 3.55 R J1 1.09 --.89 3.54 R J2 1.60 m.50 3.51 R J3 2.02 --.03 3.49 R J4 2.39 m 9.63 3.46 RX0 ~ 0.00 11.76 3.55 RX1 ~ 0.70 11.76 3.56 RX2 ~ 1.28 11.37 3.57 RX3 ~ 2.06 11.29 3.57 RX4 m 2.65 11.20 3.57 -- 11.69 3.54 a J-4 = Seyval Blanc juice reserve cncentratin series; X-~X4 = Xylitl cncentratin series. b Vlatile acidity expressed as acetic acid. c Ttal acidity expressed as tartaric acid. 0.051 0.84 0.034 0.88 0.038 0.89 0.038 0.91 0.038 0.91 0.051 0.82 0,039 0.78 0.036 0.84 0.039 0.89 0.038 0.86 0.040 0.86 0.021 0.79 0.017 0.79 0.019 0.85 0.018 0.87 0.020 0.89 0.021 0.79 0.018 0.83 0.018 0.83 0.020 0.81 0.020 0.83 0.019 0.83
90 -- DULCRATIN F WHIT WIN Table 3. Analyses f variance in terms f mean squares fr the characteristics sweetness, acidity, balance and verall impressin. Mean square Surce f variatin Degrees f freedm Sweetness Acidity Balance verall impressin Replicates 2 1.214 0.647 1.767 1.258 Types f wine (A) 2 Blends A+B vs Riesling (A1) 1 13.890" 0.032 0.005 0.600 Blend A vs Blend B (A2) 1 34.948*.677" 0.013 2.600 rrr (a) 4 0.836 0.306 0.478 0.599 Types f sweetener (B) 1 4.134" 0.187 4.601 * 6.523* Types f wine x Sweeteners 2 A1 x B 1 0.206 0.050 1.023 0.897 A2 x B 1 3.015" 0.586 0.000 0.885 rrr (b) 6 0.446 0.119 0.232 1.038 Cncentratins (C) 4 Linear regressin 1 339.130" 24.642* 3.411 * 4.156" Quadratic regressin 1 0.001 0.002 25.690* 22.854* Remainder 2 0.335 0.024 0.943 0.597 Types f wine x Cncentratins 8 A1 x Linear regressin 1 1.293* 0.040 6.831 *.965* A1 x Quadratic regressin 1 0.083 0.037 1.482" 1.716" A2 x Linear regressin 1 0.736 0.619 3.457* 3.499* A2 x Quadratic regressin 1 0.548 0.18i 0.352 0.024 Remainder 4 0.6 0.233 0.360 0.173 Types f sweetener x Cncentratins 4 B x Linear regressin 1 1.245* 0.001 1.287* 2.128" B x Quadratic regressin 1 0.046 0.000 0.604 0.115 Remainder 2 0.075 0.168 0.077 0.045 Types f wine x Sweeteners x Cncentratins 8 A1 x B x Linear regressin 1 0.011 0.000 0.039 0.661 A1 B Quadratic regressin 1 0.002 0.049 0.073 0.368 A2 x B x Linear regressin 1 0.042 0.177 0.393 1.085" A2 x B x Quadratic regressin 1 0.052 0.044 0.025 0.013 Remainder 4 0.139 0.322 0.059 0.058 rrr (c) 48 0.232 0.195 0.302 0.256 Ttal 89 *Significant at the prbability level f ~ = 0.05 satisfactrily evaluate five samples (cncentratins) withut interruptin, and with a break f 20 minutes, culd then evaluate a secnd set f five cncentratins at ne sessin (ne main plt), it was arranged that the sensry evaluatin wuld be carried ut in nine subsessins (ne per week). These sessins crrespnded t the three main plts in each f the three replicatins. Since each f the nine judges evaluated all 90 (30 x 3) samples f wine, the averages f the nine judges were taken as the experimental units in the statistical analyses f the data. Because the levels f added sugar were different fr the tw sweeteners and the three types f wines, statistical analyses cncerning cdncentratins were dne n the basis f cded levels f additins as C, C 1, C2, C3 and C4 (Table 1). The analysis f variance plan is indicated in Table 3. Results and Discussin Wine cmpsitin: A summary f the grss chemical cmpsitin f the three wines edulcrated with tw sweeteners is given in Table 2. The ph f the Blends A and B was slightly lwer than that f the Riesling wine. The ttal acidity, hwever, was similar fr all wines. N attempts were made t deacidify the musts in rder t accentuate the pssible effect f added sweetener. The mean values fr vlatile acidity were 0.019, 0.038 and 0.040 g0 ml fr Riesling, Blend A and Blend B, respectively, indicating a nrmal fermentatin. The de- crease in alchl cntent ver the cncentratin series was mre prnunced when juice reserve was added as indicated by the slpes f the regressin lines given belw: Blend A Blend B Riesling Juice reserve -0.47-0.33-0.51 Xylitl -0. 0.00-0.16 This decrease reflects the dilutin effect brught abut by the additin f juice reserve. Blend A needed slightly mre juice reserve than Blend B. Sensry evaluatin: Sweetness: In a previus publicatin (4) dealing with the effects f sweeteners n the sensry quality f wines, it was shwn that grape juice reserve was the preferred sweetener. Fr this reasn, Seyval blanc juice reserve was included in this study. There was a linear increase in the sweetness scres with the increase in sugar cncentratin irrespective f the type f wine (Fig. 2) and the type f sweetener (Fig. 3). The apparent nn-parallelism in bth figures was statistically significant (Table 3). Analysis f the interactin f types f wine x sweetener was significant. The assciated means are tabulated belw: Juice reserve Xylitl Blend A 7.5 8.3 Blend B 6.4 6.3 ach f the abve means was based n 15 scres and had a standard errr f 0.17 cm with six degrees f freedm (Table 3). The difference in scres between juice reserve and xylitl was 0.8 fr Blend A and -0.1 fr Blend B.
DULCRATIN F WHIT WIN -- 91,& " m,= t 0 r,3 v r.3 J ~ t'- t-- 8 6 5,,,?" (3, 4 Fig. 2. Linear regressins f sweetness (S) n cded cncentratins (C) f sweeteners fr Blends A + B and fr Riesling. BlendsA+B (A) S=4.48+ 1.313Cr 2=0.998n= 12 Riesling, (k) S=3.28+ 1.493Cr 2=0.998n = 6 rrr mean square = 0.232 rrr degrees f freedm = 48 n = number f bservatins fr each pltted pint. The cded cncentratins ~n the regressin equatin represent the assciated values in Table 1. Cded cncentrat ins Fig. 3. Linear regressins f sweetness (S) n cded cncentratins (C) f sweeteners fr juice reserve and fr xylitl. Juice reserve() S=4.03+ 1.290Cr 2=0.996n = 9 Xylitl () S = 4.12 + 1.458C r 2 = 0.999 n = 9 rrr mean square = 0.232 rrr degrees f freedm = 48 n = number f bservatins fr each pltted pint. The cded cncentratins in the regressin equatin represent the assciated values in Table 1. Acidity: There were n significant interactins between the varius factrs invlved. An increase in the cncentratin f the sweeteners caused a decrease in the perceptin f acidity by the tasters as reflected by the significant linear effect f cncentratin in the analysis f variance (Table 3). The scres fr acidity ver the cncentratin series C t C4 were 9.8, 9.4, 9.0, 8.6 and 8.3, respectively. The ther significant main effect invlved the types f wine, Blend A versus Blend B (Table 3). Blend A was judged less acidic than Blend B (8.6 vs 9.5, respectively). Riesling received a scre fr acidity f 9.0 and was, therefre, nt different frm the average f Blends A and B. Balance and verall impressin: The analysis f variance indicates a significant three-factr interactin f type f wine x sweetener x cncentratin fr the characteristic verall impressin (Table 3). Since this is ne ut f eight significant F values fr this variable, and als since the calculated value f F.0~(4.24) and the tabulated value f F.0~(4.04) were s nearly equal, the authrs chse t prceed as if it were the result f chance variatin. This actin was supprted by the fact that such a situatin did nt ccur with the variable balance, which is t sme extent lgically related t the variable verall impressin. These tw characteristics each exhibited tw significant tw-factr interactins, namely, Blends A + B versus Riesling and Blend A versus Blend B x sweetener cncentratin illustrated in Figures 4 and 5, respectively. The scres fr Blends A + B and Riesling shwed a cnvex respnse with the maximum near the C2 cncentratin fr bth characteristics. In bth cases, hwever, the curves fr the tw types were nt parallel but crssed
92 -- DULCRATN F WHIT WIN J kj t- t~ (a) \ J z~,m ut~ v 9 8 Y t- "~, 6 L_ c~ 5 > A j f-- 7~. (b) Fig. 4. Quadratic regressins f balance (B) and verall impressin (I) n cded cncentratins (C) f sweeteners fr Blends A + B fr Riesling. (a) Balance Blends A+B (A) 13 = 7.84 + 0.000C - 0.264C 2 r 2 = 0.875 n = 12 Riesling (%) [3 = 7.83 + 0.413C - 0.428C 2 r 2 = 0.977 n = 6 rrr mean square = 0.302 rrr degrees f freedm = 48 (b) verall impressin Blends A+B (A) I = 7.50-0.023C - 0.242C 2 r 2 = 0.931 n = 12 Riesling (k) I = 7.68 + 0.501C - 0.419C 2 r 2 = 0.980 n = 6 rrr mean square = 0.256 rrr degrees f freedm = 48 n = number f bservatins fr each pltted pint. The cded cncentratins in the regressin equatin represent the assciated values in Table 1. Fig. 5. Quadratic regressins f balance (B) and verall impressin (I) n cded cncentratins (C) f sweeteners fr Blend A and fr Blend B. (a) Balance Blend A (11) 13 = 7.83-0.170C - 0.3C 2 r 2 = 0.812 n = 6 Blend B (IZ]) [3 = 7.86-0.170C - 0.220C 2 r 2 = 0.932 n = 6 rrr mean square = 0.302 rrr degrees f freedm = 48 (b) verall impressin Blend A (11) I = 7.71-0.194C - 0.230C 2 r 2 = 0.954 n = 6 Blend B (C]) I = 7.30 + 0.147C - 0.255C 2 r 2 = 0.869 n = 6 rrr mean square = 0.256 rrr degrees f freedm = 48 n = number f bservatins fr each pltted pint The cded cncentratins in the regressin equatin represent the assciated values in Table 1. c~ t~ i (a) C.m u~ (u 9 8 7 6 5 L_ 4 > 0 3 1 J (b) Cded cncentrat ins in the vicinity f the C2 cncentratin. Figure 5a indicates a similar situatin fr the characteristic balance. Hwever, in Figure 5b, the data fr verall impressin fr Blends A and B shw nn-parallelism but a slight crssing ver frm C3 t C4. All the wines were acceptable as demnstrated by ratings which were abve seven ut f a maximum f 15, except fr Riesling at C and Blend B at C, which were judged as being t acid. The significant tw-factr interactins between the sweeteners and the cncentratins fr the characteristics balance and verall impressin are presented in Figures 6a and 6b. The scres fr the tw sweeteners were identical at C after which they digress frm each ther as the cncentratins increase. At cncentratin C, the wines were nt edulcrated and, therefre, received a similar rating. Wines edulcrated with juice reserve were given higher scres fr balance and verall impressin than wines sweetened with xylitl. Maximum ratings fr balance and verall impressin were btained at cncen-
DULCRATIN F WHIT WIN -- 93. D, kj t~ t:z3 3 t (a) C. m 9 ~- 8 7 C "~ 6 5 > 3 (b) Fig. 6. Quadratic regressins f balance (B) and verall impressin (I) n cded cncentratins (C) f sweeteners fr juice reserve and fr xylitl. (a) Balance Juice reserve () I~ = 8.06 + 0.221C - 0.368C 2 r 2 = 0.931 n = 9 Xylitl ( ) 13 = 7.61 + 0.053C - 0.269C 2 r 2 = 0.953 n = 9 rrr mean square = 0.302 rrr degrees f freedm = 48 (b) verall impressin Juice reserve () I = 7.83 + 0.261C - 0.324C 2 r 2 = 0.976 n = 9 Xylitl ( ) I = 7.29 + 0.043C - 0.279C 2 r 2 = 0.922 n = 9 rrr mean square = 0.256 rrr degrees f freedm = 48 n = number f bservatins fr each pltted pint The cded cncentratins in the regressin equatin represent the assciated values in Table 1. tratin C2 fr bth sweeteners. Althugh juice reserve was judged the mre effective sweetener, wines edulcrated with xylitl were fund highly acceptable having all scres equal t r greater than 7.0 (Figs. 6a and 6b). Cnclusins The sensry qualities f three white wines, f which ne was a vinifera, were nt adversely affected when edulcrated with xylitl. Scres fr sweetness increased linearly, and the scres fr acidity decreased linearly with increasing cncentratins f sweetener irrespective f type f wine r type f sweetener. The slpe fr sweetness was greater fr xylitl than fr juice reserve and als greater fr Riesling than fr Blend A + B. Fr the three types f wine and irrespective f the type f sweetener, cncentratin C2 received, in general, the highest rating fr the characteristics balance and verall impressin. The wines were cnsidered t acid at cncentratin C and t flat at cncentratin C4. Riesling received the highest ratings fr balance and verall impressin, clsely fllwed by Blend A and then by Blend B. Literature Cited 1. Amerine, M. A., and C. S. ugh. Wine and Must Analysis. J. Wiley and Sns, New Yrk (1974). 2. Amerine, M. A., and. B. Resler. Wines: Their Sensry valuatin. W. H. Freeman and C., San Francisc (1976). 3. Buteau, C., C. L. Duitschaever, and G. C. Ashtn. Vinificatin f three white grape varieties by three different methds. I. Fermentatin prcess and wine cmpsitin. Am. J. nl. Vitic. 30:139-45 (1979). 4. Duitschaever, C. L., C. Buteau, and G. C. Ashtn. dulcratin f white wine with fur different sweeteners. Am. J. nl. Vitic. 31:-43 (1980). 5. Duitschaever, C. L., C. Buteau, G. C. Ashtn, and A. Smiles. Mdificatin f the enzymatic methd fr the determinatin f xylitl. Analyst (In press)(1983). 6. mdi, A. Xylitl, its prperties and fd applicatins. Fd Technl. Jan:28-32 (1978). 7. Rib6reau-Gayn, J.,. Peynaud, P. Rib6reau- Gayn, and P. Sudraud. Sciences et Techniques du Vin. Tme 2. Dund, Paris (1975). 8. Trst, G. Technlgie des Weines. Verlag ugen Ulmer, Stuttgart (1972). 9. Ylikahri, R. Metablic and nutritinal aspects f xylitl. Adv. Fd Res. 25:159-80 (1979).