Agricultural Sciences UNIVERSITY OF CALIFORNIA EFFECTS OF TEMPERATURE ON WINE MAKING C. SOUGH M.A.AMERINE AGRICULTURE LIBRARY NOV 7 1966 UNIVIRSITY OP CALIFORNIA BERKELEY CALIFORNIA AGRICULTURAL EXPERIMENT STATION BULLETIN 827 Effects of Temperatun H Controlled temperature is important for the production of fine wines. Its in- fluence begins when the grapes are ripening, and continues during fermentation and aging of the wine. Even high-quality grapes of the best varieties may not result in fine wines if the proper temperature is not maintained during the wine-making process. High-temperature control in wine making is important because: (1) At tempera- tures above 85°F (38°C), yeast fermentation becomes sluggish, and at about 100°F it stops. When this happens, the resulting wine will have residual sugar, and be biologically unstable. (2) At these high temperatures, problems arise with growth of Lactobacillus sp. and toxic products of the yeast. (3) Alcohol losses are also greater and yields are lowered. (4) The aromas and flavors of white wines decrease with in- creasing fermentation temperatures. (5) General instability of wines occurs at ex- cessively hot storage temperature. Low-temperature control is also necessary because: (1) At temperatures below 50°F (10°C) yeast growth and rate of fermentation are extremely slow. (2) Red wine fermentation should be above 65°F (18°C) for best color and flavor results. (3) Aging of wines at too low a temperature is very slow and hence uneconomical. In the traditional European wineries, the small-scale operations keep fermentation heat accumulation to a minimum, and the importance of temperature control is usu- ally not stressed. However, many problems of temperature control actually exist. In many countries, wine makers take advantage of natural winter cold to stabilize and even, in some instances, to freeze out water and concentrate the wines. In Germany, external heat is used to facilitate completion of fermentation. Warming the must to increase color extraction is practiced in Burgundy. In Bordeaux cellars, excess tem- perature during fermentation is not uncommon. In fact, as Sailer (1955) has per- suasively argued, the quality of many European wines could no doubt be improved if closer control of temperature were feasible. This bulletin brings together data resulting from laboratory experiments concerned with finding the best temperatures for the fermenting, stabilizing, storing, and aging of California wines. Practical recommendations are based on available domestic and foreign sources, published and unpublished, including work done at the California Agricultural Ex- periment Station. The recommendations are generally applicable to California wine making, but are by no means definitive. Experiments are continuing, but are limited by lack of ade- quate basic data. Furthermore, it is not yet known if, with proper temperature con- trol, the resulting increased quality of the wine justifies the extra equipment and maintenance costs for that control. The table (pp. 4-5) shows recommended temperature ranges for the various aspects of wine making. If the temperatures are kept within these ranges, and if other prac- tices are what would be considered standard, results should be near optimum. 1 Submitted for publication September 13, 1965. [2] >n Wine Making 1 GRAPE-GROWING REGIONS OF CALIFORNIA The regional classification is based on the amount of heat received during the growing season. The regions are defined as follows: Region I, the coolest in which grapes are grown. Alameda County, Mission San Jose; Napa County, Napa and Oakville; San Benito County, San Juan Bautista ; San Mateo County, Woodside ; Santa Clara County, Saratoga ; Santa Cruz County, Bonny Doon and Vinehill districts; Sonoma County, Guerne- ville, Santa Rosa, and Sonoma. Region II, the most important table wine district. Monterey County, Soledad ; Napa County, Rutherford, St. Helena, and Spring Mountain; San Benito County, Hollister; Santa Barbara County, Santa Barbara; Santa Clara County, Almaden district, Evergreen, Guadalupe district, and Los Gatos; Sonoma County, Glen Ellen. Region III, moderately warm. Alameda County, Livermore and Pleasanton; Mendocino County, Calpella, Ukiah, and Hopland; Monterey County, King City; Napa County, Calistoga; San Diego County, Alpine; San Luis Obispo County, Templeton; Santa Cruz County, Loma Prieta; Sonoma County, Alex- ander Valley, Asti, and Cloverdale. Region IV, warm. Merced County, Livingston; San Bernardino County, Guasti; San Diego County, Escondido; San Joaquin County, Acampo, Escalon, Lockeford, Lodi, and Manteca; Solano County, Cordelia; Stanislaus County, Ceres, Hughson, and Vernalis; Ventura County, Ojai; Yolo County, Davis. Region V, the warmest grape-growing area. Fresno County, Fresno and Sanger; Kern County, Bakersfield and Delano; Madera County, Madera; Merced County, Merced; Tulare County, Trocha. The Authors : C. S. Ough is Lecturer in Enology and Associate Specialist in the Experiment Station, Davis. M. A. Amerine is Professor of Enology and Enologist in the Experiment Station, Davis. OCTOBER, 1966 [3] Under California^ warm climatic con- ditions, control of temperature was early recognized as a critical factor. Bioletti (1895, 1906, and 1911) discussed the problems, recommended cooling musts during fermentation, and even devised an apparatus for that purpose. Hayne (1897) outlined the factors governing temperature control during fermentation, and explained the principles of operation of heat-exchange equipment. Jordan (1911) reported in detail on his follow- up commercial experiments which tested the earlier work of Bioletti. Jordan showed the advantages of cool fermenta- tion and adequate use of sulfur dioxide in the making of white table wines, and of warmer fermentations for red table wines. He also suggested other improve- ments in cellar practices. This early work indicated that for sound wine produc- tion, the fermentation temperatures must be maintained within certain limits, de- pending on the wine type. RECOMMENDED TEMPERATURE RANGES Type of wine Crushing Fermentation Stabilization Table: White dry. . . . White sweet*. Red dry Red sweet*. . . Dessert: Red White Sherry Sparkling: Red and white Concentrate: Red and white 55-65°F (13-18°C) 65-75°F (18-24°C) 70-90°F (21-32°C) 70-90°F (21-32°C) 70-90°F (21-32°C) 55-65°F (13-18°C) 50-65°F (10-18°C) 65-85°F (18-29°C) 70-85°F (21-29°C) 70-85°F (21-29°C) 70-85°F (21-29°C) 50-60°F (10-15. 5°C)H none 22-24°F (-5 to -4°C) 10 to 14 days 22-24°F (-5 to -4°C) 10 to 14 days 22-24°F (-5 to -4°C) 10 to 14 days 22-24°F (-5 to -4°C) 10 to 14 days 16-18°F (-9 to -7.5°C) 10 to 14 days 16-18°F (-9 to -7.5°C) 10 to 14 days 16-18°F (-9 to -7.5°C) 10 to 14 days r 22-24°F (-5 to -4°C) 10 to 14 days (prior to 2nd fermentation) none * A blend of dry table and dessert wines. t Close filtration and additive treatment may be substituted (used only at time of bottling). t Used to help remove colloids that are heat coaguable. (Bentonite treatments may be used instead.) § Usually made from dry table wines. K This is for the second fermentation— the sparkling wine production, whether by tank or bottle fermentation. [4] After the repeal of Prohibition, Joslyn and Cruess (1934) reviewed the earlier work done in California and elsewhere, and brought the discussion up to date. Berg (1953a) discussed stability factors affected by temperature, noting espe- cially that the rate of oxidative changes could be increased by two and one-half times with a temperature increase of 18°F (10°C) . Other stability factors con- cerning refrigeration were covered by Skofis (1953), who also briefly reviewed THIS TABLE summarizes the author's recommendations. These data are based on the current available information. As results of future research and economic conditions dictate, the recommendations will be adjusted. FOR SOME WINE-MAKING PRACTICES Heating Storage Pasteurization Cellar Market none none 55-65°F (13-18°C) 55-75°F (13-24°C) 180°F (82°C) 1 minutef or 130°F (54°C) and hot bottle none 55-65°F (13-18°C) 55-75°F (13-24°C) none none 55-65°F (13-18°C) 55-75°F (13-24°C) 180°F (82°C) 1 minutef (or 130°F (54°C) and hot bottle) none 55-65°F (13-18°C) 55-75°F (13-24°C) none 180°F (82°C) 1 minutef (for clarification and stability) 60-70°F (15.5-21°C) 55-75°F (13-24°C) none 180°F (82°C) 1 minutej (for clarification and stability) 60-70°F (15.5-21°C) 55-75°F (13-24°C) none 120-140°F (49-60 C) 2 to 6 months (baking) 60-70°F (15.5-21 °C) 55-75°F (13-24°C) none none 55-65°F (13-18°C) 55-75°F (13-24°C) none 70-100°F (21-38°C) (vacuum pan concen- tration temperature) 32-40°F (0-4°C) 32-40°F (0-4°C) [5 the history of cold stability of wines. Pilone (1953) considered the necessary equipment and the proper conditions for heat stabilizing by pasteurization, and recommended temperature and time val- ues. Other reviews, such as those of Ough and Amerine (19636) and Rosenstein (1964, 1965), have considered the prob- lem for warmer regions. A recent Con- gress of the Office International de la Vigne et du Vin considered the influence of temperature on wines in general. About 18 reports by representatives of 16 countries were summarized by Dimo- taki-Kourakou (1966). This summary mainly reflects the findings and the fu- ture needs of the European wine coun- tries. Singleton (1962) discussed tem- perature as it affects the aging of wines and related products. PREFERMENTATION TEMPERATURES Wine may be made directly from the grapes, or the extracted juice may be concentrated and either stored or shipped to non-grape-growing areas for fermen- tation at a later date. Grape handling The temperature of the air surround- ing the vine during fruit maturation has an important effect on the chemical com- position of the mature grapes. California has been divided into five grape-growing areas on the basis of the amount of heat during the growing season (see box on page 3). Certain varieties are best suited for the warmer regions and others, for the cooler. Recommended varieties for each region were given by Amerine and Winkler (1963a, 6). Vineyards must be closely supervised before harvest to ensure that the grapes are picked at the proper stage of matur- ity (Amerine and Roessler, 1958a, b; Roessler and Amerine, 1958, 1963; and Baker et at., 1965) . In these publications rational sampling techniques to ensure harvesting at the proper time are sum- marized. Cool weather can delay matur- ity; warm weather will hasten it. Grapes picked too early have low sugar content ; wines made from them are deficient in flavor and, particularly if from cooler areas, may be too high in acidity. On the other hand, excessive delay in har- vesting, under California conditions, may result in wines of very low acidity, resid- [6 ual sugar, and a flat and unbalanced taste. Such wines often oxidize easily and change in color and flavor (Berg, 19536). Alcohol yield is also influenced by pre- harvest temperature. Winkler and Amer- ine (1937) showed that greater alcohol per degree of must sugar occurs in warm years than in cool years. The data of Ough and Amerine (1963c) show that, for a given region, the per cent al- cohol by volume produced by each de- gree of Brix increases with the amount of accumulated heat to time of picking. On the other hand, when grown in the cooler regions, certain grape varieties at equivalent maturity produced a greater percentage of alcohol. The time of day for picking is impor- tant since the excess latent heat must be removed from the crushed grapes. The optimum time to harvest the grapes in California is normally as soon as the dew is off the fruit in the morning. The tem- perature of the fruit is then relatively low and the fruit, if moved rapidly to the winery, will be closer to the desired fermentation temperature. Measurements at Davis (Kliewer and Schultz, 1964) have shown that there can be as much as 40°F (22°C) difference between day- time and nighttime cluster temperatures. Measurements made in Israel by Luden (personal communication) have shown that fruit directly exposed to the sun may reach temperatures in excess of 140°F (60°C). Dimotaki-Kourakou (1966) summar- izes the effect of the air temperature in countries of various climatic ranges, and reports that those with extremes, such as Germany, where the fruit is harvested at a temperature of 40°F (4°C) and Por- tugal, southern Russia, and other areas where the fruit is harvested at tempera- tures of 90° to 100°F (32° to 38°C), need to modify the temperature of the fruit before fermentation. In countries with a moderate climate, the temperature of the grapes is more satisfactory for immediate fermentation. If the harvested fruit is held in metal gondolas in the sun for appreciable lengths of time, the temperature may rise and removal of the excess latent heat may become a problem. Yeast growth may also occur, and result in a further increase in temperature. Amerine (1962) has shown that picked grapes held at 120°F (49°C) for four days or at 130°F (54°C) for one day gave evidence of hydroxymethylfurfural formation. At higher temperatures, the rate of oxida- tion is increased and there may be other undesirable enzymatic reactions which result in eventual darkening of the color of the wine. Grape concentrate The temperatures of concentration and of storage are quite critical if concentrate quality is to be maintained. Matalas et al. (1965a) indicate the damage that can occur to a concentrate produced at various temperatures and held at vari- ous storage temperatures. Their general recommendations were to concentrate at as high a vacuum and as low a temper- ature as is economically possible and to store at a temperature close to 32 °F (0°C). The main apparent damage to the concentrate is a more rapid darken- ing at higher storage temperatures, and increased production of hydroxymethyl- furfural. Amerine (1962) found no hy- droxymethylfurfural in fresh California grapes unless the temperature was held at a high level. Kern (1964) reviewed the effects of pasteurization and heating on hydroxy- methylfurfural formation. He pointed out that repeated heating of juice to 189°F (87°C) for two minutes caused the formation of very little hydroxymethyl- furfural (a few milligrams per liter). He also found that grape juice must be heated several hours at 203°F (95°C) to form about 100 mg of hydroxymethyl- furfural per liter. To achieve any sensory effects from addition of hydroxymethyl- furfural, it was necessary to add about 100 mg of the pure compound. Although heating the grape juice for 30 minutes at 203°F (95°C) produced only 4 mg of hydroxymethylfurfural per liter, it caused a perceptible change in the sen- sory characteristic of the juice. Kern concluded that hydroxymethylfurfural per se could not have been responsible for the flavor changes. But there is no doubt as to the harmful effects of heating grape juice at high temperatures. Matalas et al. (19656) found that the storage temperature of the concentrate was important to the quality of dry table wine. If the wine was made immediately after concentration, then concentration temperatures up to 132°F (56°C) caused little change in quality, except for a darkening in color, as compared with wines produced at lower concentration temperatures. However, Amerine and Ough (1960) have shown, with natural sweet wines, that when the concentrate was added to the wine after fermenta- tion, the highest quality concentrate was desirable. If the concentrate was added before fermentation, the fermentation tended to "clean up" the off-odors caused by higher concentration temperatures. The product from concentrate made at high vacuum and low temperature, and added after fermentation, was as good as that in which concentrate was added before fermentation, and much easier to [7] handle (Ough and Amerine, 1963a). Meffert (1964) has developed a tenta- tive test procedure for evaluation of the efficiency of concentrators, which relates the heat applied to the actual chemical conversion desired. The results are given as Damkohler number (roughly heat units per volume of water removed per time unit) . Only the conversion aspects are considered — sensory quality aspects of the product are neglected in the Dam- kohler value. FERMENTATION TEMPERATURES One effective method of influencing wine quality is to control the temperature of fermentation. Most wine makers are well aware of the necessity for adequate cool- ing capacity in their plant in order to maintain the proper temperatures during fermentation, stabilization, and aging. However, the philosophy that if a little cooling during fermentation is good, a lot should be better, must be avoided for several reasons. First, the fermentation season is short. The present trend in Cal- ifornia is toward an even shorter crush- ing season, so that the equipment is used for only a fraction of the year. The quan- tity of grapes that can be purchased and processed is thus limited. The fermenta- tion rate becomes very slow if too low a temperature is maintained; thus the fer- mentation room capacity must be unduly increased. Second, red table wine quality seems to be best if the must is fermented at higher temperatures. Dimotaki-Kourakou (1966) found the general consensus in most wine-produc- ing countries to be that white table wine fermentations should be conducted at 61° to 68°F (16° to 20°C), and those for red table wines at 77° to 86°F (25° to 30°C) . She also recommended that, if the fruit is too warm when harvested, sufficient sulfur dioxide be added to pre- vent yeast growth for 24 hours. This ad- dition permits the must to cool naturally and is helpful for settling and later rack- ing into cool tanks. Racking during fer- mentation was also suggested as standard practice to slow fermentations. Dimotaki- Kourakou gives the following fermenta- tion temperatures as optimum for the various countries. Country White Red TABLE WINES TABLE WINES °F °C °F °C Argentina 72-77 22-25 77-81 25-27 Austria 59-64 15-18 68-77 20-25 Bulgaria 64-68 18-20 79-82 26-28 France 77-86 25-30 Germany 61-64 16-18 68-75 20-24 Greece 77-86 25-30 Hungary 82-86 28-30 Israel 68 20 86 30 Luxembourg 68 20 Portugal 77-82 25-28 82-86 28-30 Russia 57-64 14-18 82-90 28-32 Yugoslavia 68 20 86 30 We have no explanation for the high temperatures recommended for white table wines in Argentina and Portugal. We suspect that these figures may be based on the commercial practices of wineries producing wines of ordinary quality. Rates of fermentation Early workers were aware that temper- ature affected rates of fermentation (Slator, 1906). Franz (1961) and Ough and Amerine (1961a) plotted log fer- mentation rates versus reciprocal of the absolute temperature, and obtained lin- ear relationships in the temperature ranee of 68°F (20°C) to 86°F (30°C). [8] O 70 F # 60 F — saturated with air before inoculation O 60 F — oxygen stripped out before inoculation 50 60 70 HOURS 110 120 Fig. 1. Yeast growth curves showing the effect of two temperatures and absence of air on starting time and growth rate. Counts Avere made in duplicate or triplicate at each point. Franz used synthetic media and Ough and Amerine used grape juice. Further work by Ough (1964) indicates that a better plot was obtained when fermenta- tion rate was plotted against temperature directly. Over the range 50°F (10°C) to 91°F (33°C), the relationship was es- sentially linear. There are interactions with temperature and other variables on the fermentation rates which need fur- ther study. (See Ough, 1966a, b.) The effect of temperature on growth of the yeast cells must also be considered. Figure 1 shows viable cell-count data and the reduction in cell growth rates that result from starting musts at too low a temperature. Also indicated is the effect of nitrogen stripping of the musts on delay in yeast multiplication. This particular must was low in nutrients, and the differences are perhaps larger than normal. For maximum and imme- diate growth, temperatures in the range of 70° to 80°F (21° to 27°C) are ap- parently satisfactory, and those below 70°F (21°C) will result in delays. Fer- reira (1949) gives a value of 86°F (30°C) for the maximum growth rate with selected California musts. It has been shown that between temperatures of 50° to 80°F (10° to 27°C), the follow- ing equation approximates effects of tern- 9] 70 80 TEMPERATURE (°F) Fig. 2. The linear relationship of the reciprocal of the starting time (hours) to the temperature of the medium, for two separate years. 1963 data are averages of 10 fermentations at each temper- ature ; 1964 are averages for 12. perature on delay in starting a fermen tation: 1 // kT 9 (1) where — is the reciprocal of the time H until visible fermentation starts, T is the temperature (°F) and k is a constant that will vary for each must. The con- stant also varies with the yeast type and with other conditions that affect yeast multiplication. Data in figure 2, for 10 wines from musts fermented at tempera- tures of 50° to 91°F (10° to 33°C) in 1963, and 12 in 1964, and with slightly different yeast inoculations each year, verify equation ( 1 ) . With white musts, the maximum yeast- cell count appears to be reached when the degree Brix drops to between 10° and 3° (about 6 to 9 per cent alcohol by volume produced) . During the initial cell multiplication period, the yeast pop- ulation increases exponentially, but after fermentation starts, the multiplication rate is slower, and either gradually stops, or reaches an equilibrium with dying cells. The rate of cell growth and the rate of fermentation are independent, but both react to temperature; therefore, the conversion of sugar to alcohol will de- pend to some degree on the temperature effects on both rates. As more yeast cells are produced, the overall fermentation rates increase. The overall fermentation rates (disappearance of sugar) are, how- ever, inhibited by sugar, alcohol, and pH, and are limited by the amount of thiamine and other micronutrients pres- ent in the juice. It can be seen that a very complicated picture results. Ough (1964a), in initial studies on 10 different musts fermented at five temperatures, was able to account for about 80 per cent of the rate varia- tions. He used multiple-regression tech- niques and three factors: ammonia, as an index of nutrients, initial Brix, and initial pH. The effect of these variables [10] 90 70 60- 50 z o 5 40 I— z 111 :> on LU "" 30 20 10 ntable effects \ pH \ .^ ->'-*\... \ x^ \ •o # \i y *V°Brix / X / \ K 50 60 70 80 FERMENTATION TEMPERATURE (°F) 90 Fig. 3. Effect of three must variables on rates of fermentation at different temperatures. on the fermentation rates at the various temperatures is indicated in figure 3. The causes of differences in the rates in re- sponse to temperature variations are ob- viously far from being fully understood. Great differences in rates of fermenta- tion between Pinot noir and Cabernet Sauvignon grapes from the University's [ Oakville vineyards have been reported by Ough and Amerine (1961a). Differ- ent responses to changes in the tempera- ture of fermentation in the Semillon variety are shown in figure 4. Data were collected for three different years, from two regions and, in one case, from an early-harvested and a late-harvested in sample from the same vines. These data illustrate the variations in fermentation rates that can occur under laboratory conditions. Many more data are needed before a full explanation of these vari- ables is possible, particularly under winery conditions. Unfinished fermentations Extremely high or low fermentation tem- peratures can cause premature stopping or "sticking" of the fermentation before complete conversion of the sugar. At un- usually high temperatures — above 90°F (32°C) — the fermentation rates are ex- cessive and the heat given off by the re- actions is sufficient to increase the temper- ature to the point at which the yeast cells are inhibited in their metabolic functions by both the temperature and the increas- ing alcohol toxicity. If bacteria are pres- ent their activity may result in toxic con- ditions due to formation of other material which inhibits or kills the yeast. If the thermal death-time data of Jacobs et al. (1964) are extended, it would take about 70 days of heating at 105°F (41°C) to kill the cells. The actual "heat sticking" of the fermentation takes place much more quickly. At temperatures below 50°F (10°C), enzymatic reactions pro- ceed very slowly. With certain musts, usually those low in micronutrients, the yeast counts are initially low. In such cases, the fermentation often stops with unfermented sugar still present. Appar- ently, the fermentation proceeds so slowly that the yeast cells eventually die off or are drastically reduced in number 0.70 * 0.60 o O O 0.50 0.40 0.30 0.20 0.10 A 1964 Region I o 1963 Region IV (late) A 1963 Region I X 1964 Region IV D 1963 Region IV (early • 1960 Region IV -Li. 50 60 70 80 FERMENTATION TEMPERATURE 90 (°F) Fig. 4. Variations in rates of fermentation over a temperature range as influenced by season, region, and fruit maturity of Semillon grapes. (Grape-growing regions are listed on page 3.) [12] before all the sugar can be fermented. In the case of "stuck" fermentations, the usual method of restarting is to blend a small amount of the stuck wine with a larger amount of actively fermenting must. As a general rule, a ratio of 1:3 or 1:4 is satisfactory. If no actively fer- menting must is available, the stuck wine should be brought to about 70°F (21°C), aerated slightly, and fortified with urea or ammonium phosphate. A small quantity of the stuck wine is then added to a pure yeast culture. To this is added daily about one-half its volume of stuck wine. Starting with 50 gallons (12.5 gallons stuck wine and 37.5 gal- lons culture), the volume would increase to 75 gallons on the second day and to nearly 2,000 gallons on the tenth day. The rate of addition may be increased if observation shows the fermentation is proceeding rapidly. These methods are a variation of that outlined in Amerine and Joslyn (1940, 1951) and Amerine and Cruess (1960). The usual method is preferable, and the alternative method should be used only if no other ferment- ing must of appropriate type is available. Cap temperatures In the musts for red wine fermentation, the skins remain in contact with the juice. Excessive temperatures often occur in the floating cap of skins. This prob- lem is not encountered with fermenta- tions of white musts, in which the skins are separated from the juice. Red musts usually are fermented on the skins until 30 to 60 per cent of the sugar is fer- mented. Almost immediately after crush- ing, the skins float above the must. There- after, they are occasionally mixed into the liquid by various methods. Heat ac- cumulates at the liquid-cap interface, and, depending on the rate of fermenta- tion, cap thickness, and other factors, may rise to very high temperatures. Table 1 shows the results of a typical pilot plant fermentation and indicates that maximum heat accumulation occurs when the initial sugar content is between TABLE 1 Cap-liquid temperature differences for a Pinot noir pilot plant fermentation Brix reading and temperature difference at cap-liquid interface, at fermentation temperatures of: Times cap punched down 60°F (15.5°C) 70°F (21°C) 77°F (25°C) ° Brix Temp, difference (°F) "Brix Temp, difference (°F) "Brix Temp, difference (°F) First 21.5 21.6 21.5 21 20 19.5 18.5 18 17 17 13 12.5 12 11 10.5 10 10 5 3 3 1 2 5 2 3 4 12 1 4 7 8 9 1 5 20 19 19 18 13 12 11 9.5 8.5 8.5 4 2 3 2 12 0.5 6 8 8 0.5 3 18 17 16.5 15 8 7.5 6 5 4 4 6 0.5 Third 9 3 Fourth 2 4 3 Fifth Sixth [13 18° and 10° Brix. The maximum differ- ences are reasonably independent of the temperature of fermentation. It should be noted that the temperature rise is very rapid. This rise is almost impossible to control completely by pumping the musts over the cap or by conducting the fer- mentation with a submerged cap. On a laboratory scale, the rise in some in- stances was one degree F (0.6°C) per hour. In commercial tanks, the rate of increase may be much faster, and reach higher maximums. Color extractions Data of Sudraud and Cassignard (1958) and Ough and Amerine (1961a) show that extraction of the tannins and the color pigments from the skins depends on and is proportional to the tempera- ture. The higher the temperature the more color is extracted, but above cer- tain temperatures — those in excess of about 85°F ( 29 °C)— although more color is extracted, it appears to be less stable. Whether the pigments are less stable or whether other factors influence color stability is not clear. Data on Ca- bernet Sauvignon from Region I are presented in table 2. Apparently about twice as much color is extracted (by the terms of these measurements) by raising the temperature of fermentation from 55° to 85°F (13° to 29°C) . If the percentage of the color retained after storage for the various time periods shown is calculated in 10-degree F (5.5°C) spans of fermentation tempera- ture, it indicates less color stability for the wines fermented above 80°F (27°C) . However, in final color density after storage, the warmer fermentations gen- erally are superior. Data in table 2 also demonstrate that tannin extraction in- creases with a rise in fermentation tem- perature. Singleton and Draper (1964), working with model systems, studied the effect of heat on extraction of grape-seed tannin over a temperature range of 52° to 86°F (11° to 30°C) and found that TABLE 2 Three series (three different years) of Cabernet Sauvignon grapes from Region l f fermented at various temperatures and all pressed at 2.5° to 5° Brix Duboscq* color at the following Tannin at the following Series Fermentation temperature months after press: months after press: 2 6 12 2 6 op °C gm/100 ml gm/100 ml 53 12 200 167 0.11 1 70 21 294 285 0.15 80 27 400 250 0.19 55 13 330 250 0.24 65 18 360 333 0.26 2 70 21 500 333 0.31 75 25 670 500 0.30 85 29 670 500 0.32 55f 13 364 238 0.13 55 13 377 238 0.15 3 68 20 540 312 0.19 68 20 541 312 0.18 85 29 870 384 0.25 85 29 1000 384 0.25 * Determined by means of color standards in a color comparator. The higher the figure the greater the color in- tensity. Red wines usually range from 100 for the lightly colored to as high as 1000 for the more intensely colored, t Fermentations duplicated at each temperature. [14] the amount of tannin extracted from the seeds approximately doubled in the ranges studied, with every rise in tem- perature of 18 degrees F (10°C). Alco- hol enhanced the rates of extraction. Color extraction by rapid heating and cooling is possible, but cannot as yet be recommended for high-quality wines. Recent work by Coffelt and Berg (1965) with color extraction of whole berries by heating indicated no effect on quality of port wines produced by this method as compared with those made by ferment- ing on the skins to extract color. Heating crushed grapes at higher temperatures (for times in excess of several seconds) is disadvantageous because of the off- flavor produced in the wine (Berg and Marsh, 1956) . Theoretically, color may be released by freezing the cells and thus rupturing them for color release, but this process has not been used commercially, and does not appear to be economically feasible. At present the usual method for color extraction in California is fermentation on the skins at some reasonable tempera- ture— 65° to 85°F (18° to 29°C)— and separation of the must from the skins when optimum color has been reached. Generally, at about 6 per cent alcohol, low-colored varieties from the warmer regions yield maximum color at these temperatures (Berg and Akiyoshi, 1962) . However, high-color varieties or varie- ties grown in the cooler regions do not yield maximum color until near dryness (Ough and Amerine, 1962). The opti- mum time of pressing also depends on the amount of tannin desired and how long the wine is to be aged before it reaches maturity. EFFECT OF TEMPERATURE ON WINE QUALITY Table wines The sensory quality of white table wine is generally greater when the juice is fermented at lower temperatures — 50° to 65°F (10° to 18°C) — (Stiihrk, 1960; Ough and Amerine, 1961a) . Yet to be determined is whether the change in the sensory quality is due to a greater flavor retention or to different metabolic reac- tions at lower temperatures. Ribereau-Gayon et al. (1963) found that heating of slightly moldy, white musts (stems removed) to 158° to 165°F (70° to 74°C) (up and down time, two hours) resulted in higher quality wines than those from the unheated controls. The use of higher fermentation tem- peratures to extract increased color and flavor from red must results in improved wine quality (Ough and Amerine, 1961a) . Generally, red musts should be fermented above 65°F (18°C), but not above 80° to 85°F (27° to 29°C). Prehoda (1963), using 9,240-gallon fer- menters, confirmed that the higher tem- peratures of fermentation for red must are superior, and recommended 82° to 86°F (28° to 30°C) as optimum tem- peratures. Observations of commercial fermentations (Ough, unpublished data), however, have shown that if tempera- tures above 85°F (29°C) are used, the product tends to have a "horsey" or pre- dominantly "barn-like" aroma which, in some instances, makes the wine most un- palatable, especially if the musts are of relatively high pH and low acidity. 15 Composition of the wine. The changes in the gross chemical composi- tion of wines fermented under various controlled temperatures would be ex- pected to be small. Amerine and Joslyn (1951) reviewed and discussed the available data. It is well known that alcohol is lost during fermentation as a result of evaporation and of entrap- ment by carbon dioxide. Stradelli (1951) reported small losses due to en- trapment of the alcohol, and stated that losses by evaporation from a hot cap or during transfer were largest. Flanzy and Boudet (1949) gave values of 0.1 per cent alcohol loss for a fermentation tem- perature of 86°F (30°C) and found es- sentially no loss at 41°F (5°C). These types of losses have been discussed fur- ther by Warkentin and Nury (1963) and by Zimmerman et al. 1964). Their find- ings and those of others can be sum- 55 60 65 70 75 80 85 FERMENTATION TEMPERATURE (°F) 100 105 Fig. 5. Expected loss of alcohol due to fermentation temperature, summarized from laboratory experiments and based on 22.5° Brix must. Values for several investigators' results are included within the shaded areas. [16] marized as follows: (1) The rate of loss of alcohol increases with temperature of fermentation, alcohol level of the wine fermented, amount of agitation of the fermentation, and presence of a pomace cap. (2) Maximum loss occurs during the most rapid period of fermentation. Expected alcohol losses at various temperatures and conditions are sum- marized for juice fermentations in figure 5. Probably an estimated 33 to 50 per cent more alcohol is lost in a fer- mentation with a cap than in one with- out. Changes in losses due to initial sugar concentration can be approximated by making percentage estimates from 15° Brix to 25° Brix from figure 5 (in- creasing the amounts above 22.5° and decreasing the amounts below 22.5°). As an example, if the degree Brix was 19° for a must fermented at 70°F (21°C), the estimated loss would be the ratio of the degree Brix used over the degree Brix of figure 5 times the value read from figure 5 for the fermentation temperature : 19 22.5 x 0.55 = 0.46 per cent by volume A series of 10 juices from five white varieties of grapes grown in two regions (I and IV) in 1963 were fermented under carefully controlled conditions of temperature, agitation, sampling, etc. The same experiment was repeated in 1964, plus an additional variety from both regions. The temperature range covered was from 50° to 91°F (10° to 33°C), and a fermentation was done at approxi- mately every 10 degrees F (5.5°C) . Brix readings were made daily or oftener. In 1963, the wines were thoroughly shaken prior to sampling; in 1964, they were not shaken. The wines were held at 70°F (21°C) for 24 hours after the fermenta- tion to allow for gross yeast settling; they were then racked, put into a 32°F (0°C) room, and racked again after several months. The wines were left at this tem- perature for about eight to 10 months to ensure the attainment of tartrate equi- TABLE 3 Effect of fermentation temperature on chemical composition of the wine' Year Average values at fermentation temperatures of: Analyses 50°F 60°F 70°F 80°F 91°F (10°C) (15.5°C) (21°C) (27°C) (33°C) Total acid (gm tartaric acid per 100 ml) 1963 0.650 0.676 0.693 0.706 0.693 1964 0.745 0.752 0.785 0.791 0.773 Volatile acid (gm acetic acid per 100 ml) 1963 0.039 0.031 0.030 0.027 0.032 1964 0.022 0.022 0.020 0.021 0.028 pH 1963 1964 3.09 3.18 3.03 3.14 3.02 3.15 3.02 3.19 3.08 3.23 1963 0.81 0.83 0.73 0.87 1.72 1964 376 341 414 415 441 1963 2.87 2.74 2.80 2.81 2.98 1964 2.16 2.21 2.34 2.47 2.55 Reducing sugar (gm glucose per 100 ml) 1963 0.27 0.15 0.13 0.13 0.15 1964 0.15 0.18 0.13 0.20 0.12 1963 12.02 12.03 11.93 11.88 11.74 1964 13.14 12.94 12.74 12.56 12.42 Volatile esters (mg ethyl acetate per liter) 1963 53 56 61 60 55 1964 64 82 90 76 68 Acetaldehyde (mg acetaldehyde per liter) 1963 54.0 50.3 46.2 42.9 46.8 1964 30.1 29.2 28.5 33.1 41.1 * The average values of 10 fermentations in 1963 and 12 in 1964 (fermented in 1-gallon containers). Wine made rom five varieties of grapes grown in both Region I and Region IV in 1963, and the same five plus a sixth in 1964. [17] 50 60 70 80 90 FERMENTATION TEMPERATURE ( F) Fig. 6. Remaining effects of fermentation tem- perature on total acid and pH of 10 wines (1963) and 12 wines (1964) stabilized at 32° F for more than six months. ture span, and the citric (not shown) was also constant. The tartaric acid con- tent accounted for most of the variation. To determine if this was normal (it had been observed at other times) data for 17 wines were taken over a three-year period (table 4) . These wines were made from juices that had been pretreated with sulfur dioxide and yeast, mixed, and then divided into separate contain- ers to be fermented, at the two tempera- tures, in 12-gallon Pyrex containers. The results of the two experiments agree fairly well. Mean difference of the acid values of table 4 is 0.038 gm per 100 ml (as tartaric) . If the difference values for the two temperatures are estimated from figure 6 (1964 values), the difference is 0.040 gm per 100 ml (as tartaric). The pH differences are also in the same di- rection and of the same magnitude. Alco- hol differences were tabulated for both experiments because relatively large dif- ferences could cause equilibrium varia- tions. As expected, the alcohol differences are small. For the carefully controlled librium so that the true effect of the various fermentations on the total acid and pH could be evaluated. Before chemi- cal analysis, the wines were carefully de- canted off the sediment and then brought to room temperature. The average mean values for the total acid and other gross analyses are given in table 3. With increasing temperature of fermentation, the total titratable acid of the final wine rises to a maximum in wines fermented between 70° and 80°F (21° and 27°C) and decreases in those fermented above 80°F (27°C). The acid values as well as the pH values are plotted on figure 6 for both years. (Cor- rection for volatile acids does not alter the shape of the curves.) In an effort to determine which acids were reduced, we made paper chromatograms for the 1963 wines. Results are summarized in figure 7. The succinic plus lactic and the malic acids seemed stable over the tempera- 50 60 70 80 90 FERMENTATION TEMPERATURE ( F) Fig. 7. Estimated amounts of the major or- ganic acids as calculated from visual compari- sons of spot intensities of the 1963 wines to chromatogrammed standards. Wines stored at 32°F for about nine months before chromato- grams made. [18] TABLE 4 Effect of fermentation temperature* on total acid and pH of wines Year samples Total acid at fermentation temperatures of: pH of fermentation temperatures of: Alcohol at fermentation temperatures of: 53°F (12°C) 72°F (22°C) 53°F (12°C) 72°F (22°C) 53°F (12°C) 72°F (22°C) 1957 1957 gm/100 ml 0.69 0.79 0.60 0.54 0.75 0.75 0.63 0.72 0.63 0.63 0.67 0.68 0.62 0.68 0.72 0.75 0.65 gm/100 ml 0.79 0.76 0.63 0.62 0.83 0.80 0.63 0.68 0.65 0.67 0.72 0.74 0.68 0.72 0.79 0.76 0.68 3.41 3.40 3.52 3.38 3.17 3.10 3.52 3.58 3.29 3.28 3.08 3.11 3.10 3.11 3.08 2.97 3.03 3.35 3.42 3.48 3.38 3.22 3.12 3.49 3.43 3.23 3.20 3.06 3.11 3.09 3.13 3.03 2.97 3.01 per cent vol. 12.3 13.2 13.8 15.4 13.9 14.4 14.8 14.4 14.4 14.0 13.5 12 7 13.7 13.9 12.5 12 1 12.9 per cent vol. 11.5 13.0 1957 13.8 1957 15.2 1958 13.7 1958 14.3 1958 14.2 1959 14.1 1959 14.2 1959 14.1 1959 12.6 1959 12.8 1959 13 8 1959 13.6 1959 12 1 11 7 1959 12 5 Average 0.677 0.715 3.243 3.219 13.64 13.36 * Control was room temperature, so temperature variations during fermentation did occur. The average temper- ature in excess of the control tem D erature was 5.6°F (3.1°C) for the 72°F (22 e C) and 3.6°F (2.0°C) for the 53°F (12°C). experiment presented in table 3 (p. 17), for the temperature difference of 53° to 72°F (12° to 22°C), the alcohol differ- ence was estimated to be about 0.10 per cent by volume for 1963 and 0.27 for 1964. The average alcohol difference in the experiments presented in table 4 was about 0.30 per cent by volume. These alcohol-difference values agree quite well with the expected value obtained at these temperatures and shown in figure 5, namely, 0.35 per cent by volume. It is difficult to account for these small but consistent changes in acids. Rankine (1953), reporting on Australian wines fermented at 60°F (15°C) and 77°F (25°C), found that the fixed acids de- creased with the lower temperature. He attributed this to the reduced solubility of potassium bitartrate at the lower tem- perature and to the greater amounts of alcohol produced. The work of Berg and Keefer (1958) indicated that these small alcohol changes could not account for the acid variations. Also, if the changes were due to alcohol concentration effects on potassium bitartrate solubility, then the effects should increase at the highest temperatures. Since the design of the ex- periment allowed for bitartrate stability in each wine to be closely approached by lengthy storage at a low temperature, any changes in tartaric acid must have been associated with reactions that took place during fermentation. Work by Diemair and Maier (1962) and by Pilone and Berg (1965) has shown that the proteins in the wines may account for equilibrium shifts. Pilone and Berg (1965) theorized that since bentonite fining removed protein, and the wines after this treatment had lower solubility products for potassium bitartrate, the protein removed was thus involved in the tartaric acid equilibrium. If the limiting reactions fixing the equi- librium took place during the fermenta- tion and were not altered by the low [19] storage temperature afterward, it might be postulated that the low fermentation temperatures caused some proteins to be altered. At the higher temperatures, some may be denatured. However, the cause of the phenomenon is still not clear. Ough (1964a) found that the corre- lation of ammonia concentration to fer- mentation rate varied with fermentation temperature. When the per cent effect of ammonia on the fermentation rate was plotted against fermentation temperature, a curve resulted similar to that relating total acid to fermentation temperature (fig. 7). Possibly the relationship is merely coincidental. Since the tempera- ture for the maximum growth of yeast cells more or less agrees with the tem- perature for the maximum amount of total tartrates fixed in the wines, these two factors may be more than casually related. Also, the more rapid autolysis of the yeast at the warmer temperatures should cause greater release of protein into the wine before the first racking removes the bulk of the yeast. The same temperature effect on total acid in red wine fermentations is seen in the work of Ough and Amerine (1961a). The data in table 3 also show slightly increased extract values for both years, with increasing fermentation tempera- tures (if 1963 is corrected for sugar dif- ferences) . No explanation for this is ap- parent except that, at the higher fermen- tation temperatures, substances not usu- ally measured are formed (or extracted) . The differences in the volatile esters are significant and are plotted in figure 8 to show the relationship with fermentation temperature. A maximum was reached at 70°F (21°C) . The higher values for 1964 may possibly be attributed to the riper grapes used that year or to small differ- ences in fermentation and storage tech- niques or to slight differences in yeast. Nordstrom (1964) pointed out that dur- [ ing fermentation, volatile-ester formation is largely due to enzymatic reaction through the acyl-CoA system and is regu- lated through this reaction and the avail- ability of the alcohols. Ester formation during fermentation is apparently not due to esterification reactions per se. The dif- ferences between samples, within each series, were very consistent for the aver- age data of figure 8. These changes may have been due to slower rates of forma- tion at lower temperatures and faster rates of evaporation at the higher tem- peratures and to higher enzyme activity in the middle temperature range. The difference in the aldehydes seems to be real even though the range is small. A slight depression occurs at the middle temperatures and increases at the higher and lower temperatures. One explanation is that, with the longer fermentation times, which are a result of low tem- peratures, there is a longer time at the 50 60 70 80 90 FERMENTATION TEMPERATURE (°F) Fig. 8. Relation of volatile ester concentra- tions to fermentation temperature of wines stored at 32 °F for six to eight months. Each point for 1963 is an average of 10 different wines ; for 1964, an average of 12 wines. 20] finish of the fermentation in which oxi- dative changes can produce aldehydes, whereas at the warmer temperatures the rates are much faster, and only slight in- creases occur. The work of Amerine and Ough (1964) did not show these fine differences, nor would it have been ex- pected to do so, due to lack of precision in the method which, at best, is ± 5 to 10 mg per liter. Thus the replications are necessary to delineate this effect. Rankine (1953) also was unable to show any cor- relation between fermentation tempera- ture and aldehyde production. Detailed analysis of higher alcohols in 12 wines fermented at 50°, 60°, 70°, 80°, and 91°F (10°, 15.5°, 21°, 26.5° and 33°C) showed the following results (Ough et ah, 1966) : maximum amounts of isoamyl, active amyl, and isobutyl alcohols were produced at about 75°F (24°C), with lesser amounts produced at lower or higher fermentation tem- peratures; a minimum amount of n- propyl alcohol was found at 75°F (24° C), and increasing amounts at higher or lower temperatures. The amounts of isoamyl formed at the different tempera- tures were significant, and might be a factor in wine or brandy quality. Otsuka et al. (1963) also found maximum amounts of isobutyl and isoamyl alcohols produced at a fermentation temperature of77°F(25°C). The amounts of 2,3-butanediol also were found to increase rapidly with tem- perature (Ough and Crowell, unpub- lished data), with significant increases noted at the higher temperatures of 80° to91°F (27°to33°C). Measurement of the combined diacetyl and acetoin indicated that, at the high temperature, these compounds are pro- duced in increased amounts. This in- crease, if mostly in diacetyl, could ac- count for some of the off-flavor produced at higher temperatures of fermentation. Castelli (1941) found increased acetoin at higher fermentation temperatures. Table 3 (p. 17) also shows the aver- age total nitrogen values for the wines (determined after racking and storage). The values are somewhat erratic, but tend to be lower at the lower tempera- tures. One explanation may be that the increased rates of yeast autolysis oc- curring at the elevated fermentation tem- peratures allow a measurable increase in the nitrogen freed from the yeast before the first racking and transfer to the cellar. This explanation is substantiated, in general, by the poor microbiological stability of wines fermented at higher temperatures. Stiihrk (1960) summarized the ad- vantages of cool fermentations — 55°F (13°C) — as follows: they result in more alcohol and less extract, lactic acid, and tartrate. Other possible advantages are the lower glycerin and the higher fixed sulfur dioxide (owing to reduced loss of aldehyde). Teodoresco (1965) reported on com- mercial fermentations made over three temperature ranges — 59° to 68°, 68° to 77°, and 77° to 86°F (15° to 20°, 20° to 25°, and 25° to 30°C)— and indicated significant differences in alcohol, total acidity, volatile acidity, glycerin, ex- tract, ash, alkalinity of the ash, and pH. Cantarelli (1965) made a series of fermentations with three different yeasts (Saccharomyces cerevisiae var. ellipsoi- deus, S. oviformus, and Torulospora rosei) at two fermentation temperatures —64° and 95°F (18° and 35°C). Dif- ferences in composition caused by the yeast were minor. The T. rosei at 95°F (35°C) fermentation temperature pro- duced greater amounts of glycerin than did the other yeasts. The total and vola- tile acids were increased at the higher temperature, as were the 2,3-butanediol and the amounts of glycerin. Acetoin content apparently did not depend on the fermentation temperature. Acetalde- hyde content decreased with higher fer- mentation temperature. The fusel oil con- tent did not vary significantly, but the [21] two temperatures chosen were inappro- priate to show the effects on this sub- stance. Otsuka et al. (1963) found lower volatile acids and aldehydes at warmer fermentation temperatures. Other effects of temperature on the composition of wines may be summa- rized : 1. Prolonged fermentation at low tem- perature with high sugar gives wines with higher volatile acids (Amerine and Ough, 1960) . 2. Higher temperatures of fermenta- tion increase amounts of materials ex- tracted from the skins and seeds (Sud- raud and Cassignard, 1958; Ough and Amerine, 1961a; and Singleton and Draper, 1964) . 3. High storage temperatures of con- centrate (Matalas et al., 19656) and of dessert wines (Amerine, 1948) cause in- creases in hydroxymethylfurfural. 4. Fermentation of red must at higher temperatures will encourage the malo- lactic bacteria to multiply and ferment during the regular fermentation (Ribe- reau-Gayon and Sudraud, 1958). 5. The rate of loss of aroma constitu- ents and alcohols by evaporation in- creases as the temperature is raised. Dessert wines Usually less care and control are main- tained with dessert wine fermentation temperatures. Because of early additions of alcohol to stop the fermentations, the wines seldom reach excessively high temperatures from heat accumulation during fermentation. Stopping the fer- mentation, however, will cause some variations in the composition other than in the sugar and alcohol. The acetoin will increase and the volatile acids will decrease. Acid concentration will be lower because of the effect of the in- creased alcohol on solubility of the acid salt (potassium bitartrate). In other re- spects, the effects of temperature on com- position of dessert wines should be more or less the same as on that of dry wines. Temperature at the time of racking off the initial lees after fortification also affects wine quality. If racking is de- layed, the yeast may autolyze rapidly, and off-flavors can result. For red dessert wines, which remain only a short time on the skins, fermenting at higher tem- peratures increases color extraction. Joslyn and Amerine (1964) recommend fermentation temperatures of less than 75°F (24°C) in white sweet wines and less than 85°F (29°C) for red sweet wines. In general, sweet wines are aged at warmer temperatures than are table wines. Special effects by heat treatments to produce "aged" flavors in dessert wines have been discussed by Joslyn (1935), Lipis et al. (1959), Singleton (1962), and Singleton et al. (1964). Joslyn and Amerine (1964) indicated that the commercial practice for some standard red sweet dessert wines is to heat to 110°F (43°C) for a week or so, occasionally with aeration to speed ma- turation. Rosenstein and Rosenthal (1964) described the following treat- ment for stabilization prior to bottling of sweet table wines in Israel: The wine is fined with bentonite and potassium ferrocyanide, filtered, and flash-pasteur- ized at 185°F (85°C). It is then cooled to 21°F (-6°C) until no further pre- cipitation occurs. Finally, it is filtered, the free sulfur dioxide is adjusted to 20 to 30 mg per liter, and the wine is "hot bottled" at 130° to 140°F (54° to 60°C) . These wines had only 10 to 12 per cent alcohol. California sherries are actually heated to temperatures around 120° to 140°F (49° to 60°C) for six weeks to six months, depending on the intensity of "baked" character desired. At lower tem- peratures, longer heating times are re- quired, but generally the flavor produced at the lower temperatures is superior. Hydroxymethylfurfural formation oc- curs with heating if any levulose is present. A small amount of alcohol is lost [22] by volatilization, and the aldehyde con- tent increases slightly. If the heated wines are aerated, the increase is greater, and so are the browning rates. Increase in rates of esterification was re- ported by Heitz et al. (1951). However, Guymon (1955) reported on the com- position changes of baked sherry with changes in heating time, and found that the aldehydes and the volatile esters de- creased with time. His laboratory ex- periments allowed considerable evapora- tion, and the results may not be typical of those found in commercially-produced, heated sherries. In a survey of the wine industry, Martini and Cruess (1956) re- ported slight losses in alcohol content and large decreases in sulfur dioxide during baking. Mattick and Robinson (1960a), work- ing with Concord-type sherry wines and using the Tressler process — oxygen bubbling through sherry material heated to 140°F (60°C)— found both volatile carbonyls and acids increased. The vola- tile esters also showed small increases. Mattick and Robinson (19606) also dis- cussed the presence and formation of the various volatile acids during the baking of sherry. These workers concluded that the major acids present are acetic, lactic, propionic, and formic. They suggested that the formic is possibly an oxidation product of the methyl alcohol produced either by hydrolysis of the pectins or by the degradation of furfural. During a six-week baking period, acetic and formic acids increased steadily; propi- onic acid remained constant; and lactic acid increased to a maximum at about three weeks, after which it decreased to about the starting concentration. Heating, along with the oxygen, speeds the browning rate of the sherries. With the trend toward paler sherries, preven- tion of excessive color changes is de- sirable. Two compounds, polyvinylpyr- rolidone (PVP) and nylon, have been suggested as possible materials for treat- ing white wines to protect against brown- ing and to decrease the color. They both act as fining agents by adsorbing brown- ing precursor compounds and polymeric browning products. Nylon can be used in a column, and be regenerated. Caputi and Peterson (1965) found that pre- treatment of wines with insoluble PVP did help prevent color changes in oxygen- saturated wines but that nylon treat- ments were of questionable value. They recommended 4 to 6 pounds per 1,000 gallons of PVP. Fuller and Berg (1965) found that nylon 66 could be used, in some instances, to prevent browning, but they also found that in other instances, nylon either had no effect or actually in- creased browning rates. They concluded that individual preliminary laboratory trials must be made on each wine lot before nylon is used on a wine. Other- wise, the resulting extra color has to be removed later by carbon treatment, which also removes some desirable flavor characteristics. The best way to prevent great color changes is to use tempera- tures in the lower ranges and to keep the sherry material from excessive contact with air during baking. There have been reports of failure of some sherry material to darken in color, even at high tempera- tures. These incidents appear to have been due to the use of low-alcohol sherry material produced from low-sugar, over- cropped Palomino and Thompson Seed- less grapes. Sparkling wines Details of sparkling wine production have been given by Amerine and Cruess (1960). The base material (cuvee) for spar- kling wine is a sound, dry wine that has been originally fermented below 60°F (15.5°C), clarified, and stabilized. The acid has been adjusted if necessary. Oc- casionally, if the base material is de- ficient in a nitrogen source, diammonium phosphate or urea may be added. [23] The yeast starter used should be ac- climatized to the temperature at which the sparkling wine fermentation will be made and to the alcohol level of the cuvee. Although sugar concentration is not generally considered important, it also may delay the start of yeast growth after transfer if that of the cuvee differs from that of the culture. Lewis (1964) showed the effect of temperature on the rate of amino acids released from the yeast into the media and readsorbed by the yeast (brewers') during the shock of transfer. At lower fermentation tempera- tures, a lag in yeast growth rates may be caused by the shock of transfer and the subsequent loss of amino acids from the cells into the media. For the best sparkling wine quality, the fermentations should be made below 60°F (15.5°C) (Tarantola, 1937; Am- erine and Monaghan, 1950). Fermenta- tions at the lower temperature have a more fruity aroma and a cleaner taste. For commercially-produced wines, tem- peratures as high as 70°F (21°C) may be used. In these cases, the quality of the base wine may not offer any oppor- tunity for improvement, or, for economic reasons, the more rapid production may justify the reduction in quality. Ough and Amerine (19616) showed clearly the advantage of cooler fermentation temperatures over warmer ones in so far as the sensory quality of the product is concerned. Datunashvili (1963) noted particular darkening of sparkling wines and over- oxidized taste at about pH 3.4 if excess amino acids or yeast hydrolysate were present. His work indicates that lower pH, good prefiltration, and cool fermen- tation should be recommended. The time and temperature for com- plete fermentation will vary from about two to four weeks and from 60° to 50°F (15.5° to 10°C), depending on the yeast and the cuvee. If a temperature lower than 50°F (10°C) is used, the time may become excessively long. Bottles should be stored, for yeast au- tolysis, at a moderate temperature of about 55° to 60°F (13° to 15.5°C). Too-warm a temperature will allow un- desirable flavor to develop from the de- gradation of the yeast components. Jos- lyn and Vosti (1955) reported on tem- perature effects on yeast autolysis. Al- though their studies were confined to short-term, relatively high-temperature experiments at higher pH than those used for wine, and in an alcohol-free medium, the results indicated some autolysis at 90°F (32°C) in short periods of time (48 to 72 hours). In several months, at lower temperatures, a great deal of yeast autolysis should occur. Masuda and Muraki (1963) reported, for white table wine, optimum storage time (three to seven months) on the lees at 50° to 65°F (10° to 15°C) for the best autolysate flavors. During the process of disgorging or removing the yeast sediment packed in the bottle necks, it is necessary to chill the bottles below 32°F (0°C), and then freeze a plug of ice inside the necks by immersing the bottles, neck-down, in a refrigerated brine bath. The lowering of the sparkling wine temperature keeps much of the carbon dioxide in solution. As the bottle cap is removed, the ice plug is forced out, and pushes the yeast from the neck of the bottle. It is then possible to add the sweetening liquor and to fill the bottle from a second disgorged bottle with a minimum loss of pressure. The bottles are then corked or closed with polyethylene stoppers, warmed to pre- vent sweating, and labeled. In the tank fermentation, the fer- mented wine is usually cold-stabilized at 24° to 25°F (-4° to -5°C) to settle out yeast and to precipitate out excess bitar- trate if this has not been done previously. The wine is then filtered at near the freeze-point, under a counterpressure of either air or, preferably, nitrogen; sul- fur dioxide and sugar are added; and the wine is bottled. [24] Wines that fall in the category of drink industry— that is, to pasteurize slightly sweet carbonated table wines can and then carbonate at a temperature best be carbonated by using the tech- which permits the desired absorption of niques developed by the carbonated soft carbon dioxide. REFRIGERATION AND HEATING REQUIREMENTS Guymon and Marsh (1962) outlined clearly the heating and cooling meth- ods and requirements for normal winery operation during fermentation, tartrate removal, and chill proofing. Cooling of fermentations The enzymatic conversion of sugar to alcohol is exothermic. The accepted value for this reaction is about 23.5 kg cal per gm molecule. Calculations show that for each degree of decrease in the Brix, enough heat is given off to raise the tem- perature of the fermenting must by 2.34°F (1.3°C). However, some heat is lost by radiation and in the evolved carbon dioxide, the actual amount de- pending on differences in rates of fermen- tation, shape of tanks, and temperature of the surrounding air. The best estimates (Guymon and Marsh, 1962) indicate the following cooling requirements for dry table wines from start to dryness: Type of c Brix of COOLING MUSI CONTROL Intermittent 24 Intermittent >24 Constant 24 Constant >24 Field BTU/ TEMPS. 1,000 (°Fof GAL. must) req'd 70 150,000 >70 250,000 70 225,000 >70 400,000 These are rough estimates for dry table wine operations. For both dry wine and dessert wine operations, however, each plant situation must be thoroughly studied by an engineer. There are several methods for cooling. The fermentation tank can be equipped with coils, usually constructed of thin- wall stainless steel. Cool water may be circulated through the coils, or a re- frigerant gas may be expanded into them. How much surface needs to be in contact with the fermenting must de- pends mainly on the difference between the temperature of the fermenting must and that of the coolant, the transfer co- efficient of the tubes, the velocity of the coolant through the tubes, the circula- tion of the must about the tubes, and the heat capacity of the must and of the coolant. Plate coils are sometimes used. Each half of the plate is stamped out to allow a continuous passage of liquid. The two halves are welded together, the plate is put into the fermenting must, and coolant is circulated through. This type of coil is useful for small wineries with small fermenting tanks. Plate coils are portable and can be moved from tank to tank. Another method consists of pumping the wine from the fermenting tank through a counter-flow heat exchanger and back into the tank. Tube-in-tube (fig. 9) or tube-in-shell types are com- monly used, with cool water as the coolant, but a refrigeration coolant may be used directly. Use only a refrigerant such as Freon, that is harmless to hu- mans, in case of a break in the piping. Gervasi (1961) discusses refrigeration equipment and provides good schematic drawings of the general designs for re- [25] Fig. 9. Tube-in-tube heat exchanger in operation, cooling fermenting grape juice. frigeration, wine concentration, and sta- bilization. One type of equipment is shown in figure 10. If an adequate supply of water at a low enough temperature — 60°F (15.5° C) — is available, this will usually take care of most of the fermentation cooling needs for dessert wine operations. For dry table wine production, in which lower temperatures are required for the white wines, mechanical refrigeration is necessary. If water at 60°F (15.5°C) is not available, a cooling tower will usu- ally be able to lower the temperature. A [26] storage reservoir can accumulate low- temperature water for use at peak load times, and can increase the overall ca- pacity of the cooling tower or mechanical refrigeration equipment. Cooling for stabilization Wines are cooled to temperatures near the congealing point of the wine, and held there for some days for general sta- bilization purposes, and to prevent future precipitation and changes in flavor, aroma, and color. If the wines have not been ion-exchanged to remove excess po- tassium, they are usually held at those temperatures for two to four weeks to stabilize the wine against future deposi- tion of potassium bitartrate. Protein, color, and metal stability are increased by storage at near freezing-point tem- peratures. In chilling to effect tartrate stabiliza- tion, it is essential that the temperature be as close as possible to that of the freez- ing point of the wines, and that it be held there. A log relationship of time to temperature has been demonstrated by Berg and Keefer (1956). They found that an increase in temperature of about 11.7°F (6.5°C) will double the time re- quired to precipitate 50 per cent of the original bitartrate out of a red table, white table, or white dessert wine. A dessert wine required a rise of about 20.7°F (11.5°C) to double the time re- quirement. The minimum temperatures for chilling are 22° to 20°F (-5.5° to -6.5°C) for table wines and 18° to 16°F (-8° to -9°C) for dessert wines. Several methods may be used to chill the wine: (1) It may be pumped into tanks in a cold room, and allowed to come to temperature. This is the least efficient and most expensive method. (2) It may be pumped continually through a refrigeration unit and back into an insulated tank. With this method, the Fig. 10. A series of refrigeration compressors used to produce low-temperature conditions for winery operations. [27] agitation helps reduce the holding time. (3) It may be put into an insulated tank with cooling coils through which re- frigerant is pumped to maintain the tem- perature. This is probably the most eco- nomical method, but it requires proper arrangement of the coils so that the wine is chilled evenly and water is not frozen out on the coils. (4) A combination of methods may be used, in which the wine is circulated through a high-capacity ex- changer until the desired temperature is reached, then pumped either into a cold room at that temperature or into an in- sulated tank with cooling coils and with the necessary capacity to hold that tem- perature. Holstein (1960) gives a practical equation for estimating the tons of re- frigeration needed to cool a dry wine: qt 0.652 W(h-t 2 ) 1,000 (2) q t — cooling load, tons of refrigeration W — gallons of wine cooled/hour h = initial wine temperature, °F t 2 = final wine temperature, °F For sweet wine the factor 0.652 is re- placed by 0.625. As an example, suppose a winery is holding 30,000 gallons of dry table wine at 70°F (21°C), and the wine maker wishes to cool this wine to 21°F (-6°C) in a 48-hour period. How many tons of refrigeration would be required? For the purpose of these calculations, it will be assumed that once the wine reaches the desired temperature the operation is com- plete. In this case, the rate of cooling is 30,000 gallons per 48 hours, or 625 gallons per hour. Inserting the values into equation (2) : (0.652) (625) (70-21) q '=' 1^00 =20tonsof refrigeration. Heat gains from the surroundings must also be considered. Wines held in large (90,000-gal.), insulated concrete tanks at 15°F (-Q°C) may gain only a fraction of a degree F per day, but those held in uninsulated concrete tanks may gain over 1 degree F (0.55°C) per day. In 30,000-gallon, uninsulated redwood tanks, starting at 18°F (-8°C), the gain may be as much as 3 degrees F (1.7°C) per day. These figures (Guymon and Marsh, 1962) are for normal California wineries in the interior valley where the ambient winery temperature is about 65 o to70°F(18 o to21°C). Guymon and Marsh (1962) showed that, for efficient removal of bitartrates, the wines should be held within about 1 degree F (0.55°C) of their congealing point, and that little variation in this temperature should be allowed if maxi- mum precipitation rates are to be at- tained. As the temperature decreases toward the freezing point, all gases become more soluble. The concentration of oxygen in a wine being cold-stabilized may become very high. If this wine is removed to a warmer area, the wine will become super- saturated with oxygen, and may then rapidly become oxidized. Cant (1960) indicated the degree of saturation for several wines after two weeks at 25°F (-4°C). If care is taken to exclude air from the cold wine, no undue oxidation would take place then or later. Rossi (1960) reviewed the effects of sugar and alcohol concentrations on the solubilities of gases in liquid, and indicated that both depressed solubility of oxygen slightly, but were of minor importance. The effect of temperature on oxygen solubility may be approximated by the solubility of oxygen in water solutions (fig. 11). Reasonable estimations of the dissolved oxygen concentration can be made from figure 11 in the temperature range of 50° to 80°F (10° to 27°C). Almost any combination of sugar and alcohol nor- mally found in wines should give a satu- ration level between that of water and that of must. The higher the sugar con- tent and the higher the alcohol, the less oxygen will be needed for saturation. 28 10 - 5 7 6 - ▲ HO -Ifc A 12%ethanol + H 2 20%ethanol + H 2 O \ l- O 20% sugar + H 2 - ^V D Average of several saturated dry wines - \ ^ 50 60 70 TEMPERATURE ( : F) Fig. 11. Solubility of oxygen in various solu- tions saturated with air at 760 mm pressure. (Calculated from data in Hodgman, 1943; Rossi, 1960; Ough and Amerine, 1959.) Pasteurization Sterilization by heating has long been used with table wines containing re- sidual sugar. Although it is very impor- tant to be sure that the yeast in the bottled wine is dead, too much heat may cause off-flavors. A time-temperature re- lationship is established in the thermal- death time (the time required, at a given temperature, to kill all but 0.01 per cent of the yeast) . As the temperature in- creases linearly, the thermal-death time decreases logarithmically. An increase in temperature of approximately 9 degrees F (5°C)— in the range of 120° to 180°F (49° to 82°C)— will decrease thermal- death time about tenfold. Some yeasts are also more heat sensi- tive than others, but work by Jacobs et al. (1964) indicated that, under ideal conditions, the appropriate pasteuriza- tion time at 168°F (76°C) would be about 0.1 minute. Commercially, it is difficult to reach desired high tempera- tures quickly and to recool on the basis of a holding time of 0.1 minute. Actually, in practice, higher temperatures and longer holding times are used. Radio- frequency heating is very fast and could quickly raise the temperatures to de- sired levels, but the problem of recooling would still exist. Furthermore, radio- frequency heating is more expensive, on either an equipment or a maintenance basis, than are standard heat-exchange methods. Plate heat exchangers are preferable to tubular types because of efficiency and ease of cleaning. The wine being flash- pasteurized (fig. 12) is usually preheated by exchange with already heated wine, and then is finally cooled by refrigera- tion. Temperatures of 180° to 185°F (82° to 85°C) for about one minute are the usual practice, and provide as quick cooling as is economically possible. Wines may also be pasteurized by being heated to 130°F (54°C) or higher, put into the bottle, and capped at that temperature. The bottles are then usu- ally cooled by water sprays. With this method, the effective heating time is, of course, longer than when heat exchangers are used for cooling. As wines are heated, the gas satura- tion concentrations go down. At the boil- ing point, the amount of oxygen (or other gases) in solution would be zero. If wines heated to 177°F (81°C) were saturated with air at 68°F (20°C), they would lose only about two-thirds of the dissolved oxygen; of course the gas would redissolve if not removed from contact with the wine before cooling. If wine originally saturated with air at 68° F (20°C) is bottled hot, at 145°F (63° C), it will lose about one-half its dis- solved oxygen. Wines may be heated to coagulate heat-sensitive proteins. The usual pro- cedure, after adding sulfur dioxide at 100 mg per liter, is: Raise the tempera- ture to 140° to 145°F (60° to 63°C) for a short time; cool to 120° to 130°F (49° to 54°C) ; add 3 to 6 pounds of bentonite per 1,000 gallons; hold at this temperature for several days; and then [29] Fig. 12. A flash-pasteurizing heat-exchanger installation. filter. This procedure is not recom- mended (Amerine and Cruess, 1960) for table wines or dessert wines of high quality. Bentonite fining to remove pro- teins is the preferred treatment. The method recommended by Berg (private communication) for bentonite pretreat- ment is to add 83.4 pounds of bentonite to 100 gallons of water, with stirring, to give a 10 per cent bentonite slurry. This slurry is mixed by steam, allowed to age 24 hours, and mixed again by steam. After 24 hours more of aging, it is ready to be used in appropriate amounts. AGING AND MARKETING Even when the finest raw material is used, and necessary precautions are taken, wine can easily be degraded in quality by improper storage conditions. Proper maturation of the wine depends largely on correct storage temperature. After the wine is ready for sale, unless it is kept within certain general ranges of temperatures it may also decrease in quality even though the best stabiliza- tion practices have been used. Aging in the cellar Significant changes occur in the odor and flavor of wines during their storage in either glass or oak. The rate of these changes is in proportion to the storage temperature. A standard estimate is that the rates of reactions double with every rise in temperature of 18 degrees F (10°C). Therefore, aging should be oc- curring twice as fast at 68°F (20°C) as 30 at 50°F (10°C) . This is only an approxi- mation, of course, but it indicates what to expect. However, certain reaction rates may not follow the predicted pattern. Furthermore, the reaction products may be more volatile at elevated tempera- tures, and hence more rapidly lost. Storage temperatures for aging of wines are usually within the 55° to 65°F (13° to 18°C) range. In this range, development of off-flavors from yeast autolysis is not too rapid. Brown- ing reactions can also be controlled in white wines (Berg and Akiyoshi, 1956) without preventing the desired flavor changes. At storage temperatures lower than 55°F (13°C), settling and clarifica- tion are slower as are the flavor changes associated with aging (Ough et ah, 1960). At temperatures higher than 65° F (18°C) , oxidation and off-flavors due to yeast autolysis occur rapidly, and brown- ing reactions cannot easily be controlled. Fruit flies (Drosophila sp.) are not a problem at temperatures below about 52°F (12°C). At higher temperatures, bacteria can quickly affect the wine quality, partly because of the more favorable oxidative and nutritional conditions, and because of the difficulty in maintaining sulfur di- oxide levels. In many instances, we have observed that the malo-lactic fermenta- tion can be delayed for years by low temperatures. The same wines, stored at somewhat higher temperatures, may rapidly undergo this secondary fermen- tation. Wines stored in oak cooperage also would be expected to acquire barrel age more rapidly if stored at warmer tem- peratures— 70° F (21°C) versus 55°F (13°C). Again, however, adverse ef- fects of storage at elevated temperatures (more rapid oxidation and loss of vol- ume) suggest that the best range is 55° to 65°F (13° to 18°C) for table wines and possibly slightly higher, 60° to 70°F (15.5° to 21°C), for dessert wines. In Spain, sherry is usually aged in butts, at warmer temperatures — 59° to 68°F (15° to 20°C). However, similar wine types of the Jura region in France are kept at cooler temperatures — 54°F (12°C) — because of the lower alcohol content (12 per cent by volume). Cali- fornia sherry is usually aged in oak casks in relatively warm areas — 60° to 80°F (15.5° to 27°C). For a more complete review see Singleton (1962). Aging in bottles Rapid aroma and flavor changes can be induced in bottled wines by heating under reducing conditions (Singleton et al., 1964) . These treatments, if prop- erly controlled, yield aromas and flavors resembling those in wines aged for sev- eral years in the bottle. Such treatments might be useful in blending for addi- tional complexity of flavor and aroma in certain types such as red table and des- sert wines. The present preference in varietal white table wines seems to be for young fruity wines, for which this treatment is inappropriate. However, the experimental work should be supple- mented with industry and consumer tests, especially with the nonvarietal wines that have minimum fruitiness and aroma. Dimoutaki-Kourakou (1966) states that bottled white table wines can be binned for up to 15 years at 48° to 54°F (9° to 12°C) storage temperatures. The practice appears impracticable. Dimou- taki-Kourakou indicated that warmer storage temperatures for a short time would increase esterification of the vola- tile acids and increase bouquet, but that extended storage at above the recom- mended temperature was undesirable. Red wines held anaerobically, for ex- tended periods of time, at 77°F (25°C) or higher will develop off-flavors. [31] Marketing storage Once wines have been bottled, tempera- ture variations should be minimized, es- pecially if the wines are to be binned for more than a few months. Excessive tem- peratures, over 65°F (18°C), may cause the same general problems that occur during storage in tanks. If wines are shipped in bulk, care should be taken to avoid great temperature changes. High temperature — over 75°F (24°C) — dur- ing transport can cause the rapid loss of sulfur dioxide. When the bulk shipment is unloaded, the wine may become rap- idly oxidized or brown, and may be sub- ject to bacterial spoilage. Direct sunlight on wine in bottles will cause unwanted changes in some wines in relatively short periods. Storage of wine bottles in an upright position may allow air to penetrate the wine even though the cork or the metal closure would not allow liquid to leak out. This penetration of air is increased with fluc- tuations in temperature. Corked bottles stored with the liquid in contact with the closure are much less subject to air leak- age than are bottles stored upright, in which the cork is not in contact with the wine. Stabilization practices are good only for limited temperature ranges and times. If a normally stable wine is subjected to extreme cold or heat for unusual lengths of time, it may become unstable. ACKNOWLEDGMENTS The authors wish to thank Professor H. W. Berg and Dr. V. L. Singleton for their careful evaluation and criticism of the manuscript. They also wish to thank the following people for their contribu- tions to the research reported here: Mr. Harry Brenner and Mr. Walter Winton for their technical assistance in making the wines and operating the tasting room; and Miss Elizabeth Shapkin, Mrs. Claire Bailey, Mr. George Root, Mr. Mark Morris, Mr. Charles Haugen and Mr. Sam Balakian for the chemical anal- ysis of the wines. LITERATURE CITED Amerine, M. A. 1948. Hydroxymethylfurfural in California wines. Food Research 13:264-69. 1962. Physical and chemical changes in grapes during maturation and after full maturity. Internatl. Hort. Cong. Proc. XVI : 479-83. Amerine, M. A., and W. V. Cruess 1960. The technology of wine making. Westport, Conn.: The Avi Publishing Co. Inc. 709 pp. Amerine, M. A., and M. A. Joslyn 1940. Commercial production of table wines. California Agr. Exp. Sta. Bui. 639:1-143. 1951. Table wines. The technology of their production in California. Berkeley and Los An- geles: University of California Press. 397 pp. Amerine, M. A., and M. W. Monaghan 1950. California sparkling wines: bulk and bottle fermented. Wines & Vines 31(8):25-27; 31(9) :52-54. Amerine, M. A., and C. S. Ough 1960. Methods of producing sweet table wines. Wines & Vines 41 (12) :23-29. 1964. Studies with controlled fermentation. VII. Factors affecting aldehyde accumulation. Amer. Jour. Enol. and Vitic. 15:23-33. Amerine, M. A., and E. B. Roessler 1958a. Methods of determining field maturity of grapes. Amer. Jour. Enol. 9:37-40. 19586. Field testing of grape maturity. Hilgardia 28(4) :93-114. 32 Amerine, M. A., and A. J. Winkler 1963a. California wine grapes: Composition and quality of their musts and wines. California Agr. Exp. Sta. Bui. 794:1-83. 19636. Grape varieties for wine production. California Agr. Exp. Sta. Leaf. 154. Baker, G. A., M. A. Amerine, and E. B. Roessler 1965. Characteristics of sequential measurements on grape juice and must. Amer. Jour. Enol. and Vitic. 16:21-28. Berg, H. W. 1953a. Wine stabilization factors. Proc. Amer. Soc. Enol. 1953:91-111. 19536. Varietal susceptibility of white wines to browning. II. Accelerated storage tests. Food Research 18:407-10. Berg, H. W., and M. Akiyoshi 1956. Some factors involved in browning of white wines. Amer. Jour. Enol. 7:1-8. 1962. Color behavior during fermentation and aging of wines. Amer. Jour. Enol. and Vitic. 13:126-32. Berg, H. W., and R. M. Keefer 1958. Analytical determination of tartrate stability in wine. I. Potassium bitartrate. Amer. Jour. Enol. 9:180-93. Berg, H. W., and G. L. Marsh 1956. Effects of heat treatment of must on color characteristics of red wine. Food Technol. 10:4-9. BlOLETTI, F. T. 1895. Fermentation. Rept. of the Prof, of Agr. and Dir. of the Exp. Sta. 1894-1895. University of California, pp. 379-421. 1906. A new wine-cooling machine. California Agr. Exp. Sta. Bui. 174:1-27. 1911. The principles of wine-making. California Agr. Exp. Sta. Bui. 213:395-442. Cant, R. R. 1960. The effect of nitrogen and carbon dioxide treatment of wines on dissolved oxygen levels. Amer. Jour. Enol. and Vitic. 11 : 164-69. Cantarelli, C. 1965. Influence du temperatures de fermentation et de conservation du vin ets vins speciaux sur leur caracteres chimiques, microbiologiques et organoleptiques. (Report of Italy.) Office Intern. Vin. IX Cong. Paris, 13 pp. See Bui. Office Internatl. Vin. 38(416): 1174-77. Caputi, A., Jr., and R. G. Peterson 1965. The browning problem in wines. Amer. Jour. Enol. and Vitic. 16:9-13. Castelli, T. 1941. Temperature e chimismo dei blastomiceti. Ann Microb. 2(1) :8-22. Coffelt, R. J., and H. W. Berg 1965. Color extraction by heating whole grapes. Amer. Jour. Enol. and Vitic. 16:117-28. Datunashvili, E. N. 1963. O pereokislennosti vin. Biokhim. Vinodeliya 1963(7) :91-101. Diemair, W., and G. Maier 1962. Beitrage zur Wein Analytik. III. Zur Ermittlung der Wein steinloslichkeit in Weinen. Z. Lebensm.-Untersuch. u.-Forsch. 118:348-413. DlMOTAKI-KoURAKOU, V. 1966. Influence de temperatures de fermentation et de conservation du vin et vins speciaux sur leur caracteres chimiques, microbiologiques et organoleptiques. (General report.) Bui. Office Internatl. Vin. 39:73-102; 191-222; 327-55; 482-97. Ferreira, J. D. 1959. The growth and fermentation characteristics of six yeast strains at several temperatures. Amer. Jour. Enol. and Vitic. 10:1-7. Flanzy, M., and V. Boudet 1949. Les pertes d'alcool. Vitic. Aboric. 95:104-07. Franz, B. 1961. Untersuchungen iiber die Kinetik der alkoholischen Garung unter den Bedingungen der Backhefezuchtung. Nahrung 5:457-81. Fuller, W. L., and H. W. Berg 1965. Treatment of white wines with nylon 66. Amer. Jour. Enol. and Vitic. 16:212-18. [33] Gervasi, E. 1961. Die Kalte in der Kellerwirtschaft. Genio Rurale 24(12) :1-16. Guymon, J. F. 1955. Effect of brandy on the changes in composition of California-type sherry during baking. Amer. Jour. Enol. 6:36-41. Guymon, J. F., and G. L. Marsh 1962. Beverage process, Part II: Wine making. New York: Amer. Soc. Heating, Refrigeration and Air-Conditioning Engineers, Guide and Data Book, pp. 396-401. Hayne, A. P. 1897. The control of the temperature in wine fermentation. California Agr. Exp. Sta. Bui. 117:1-19. Heitz, J. E., E. B. Roessler, M. A. Amerine, and G. A. Baker 1951. A study of certain factors influencing the composition of California type sherry during baking. Food Research 16:192-200. Hodgman, C. D. (ed.) 1943. Handbook of chemistry and physics. 27th ed. Cleveland, Ohio: The Chemical Rubber Publishing Co., p. 1329. HOLSTEIN, J. H. 1960. Refrigeration, evaluation and application. Rept. to Tech. Adv. Comm., Wine Institute, San Francisco. May 23, 1960. 4 pp. Jacob, F. C, T. E. Archer, and J. G. B. Castor 1964. Thermal death time of yeasts. Amer. Jour. Enol. and Vitic. 15:69-74. Jordan, R., Jr. 1911. Quality in dry wines through adequate fermentations. San Francisco: Perman Pub- lishing Co. 146 pp. Joslyn, M. A. 1935. Preliminary observations on the mellowing and stabilization of wine. Fruit Products Jour. 15:10-12,24. Joslyn, M. A., and M. A. Amerine 1964. Dessert, appetizer and related flavored wines. The technology of their production. Berkeley: University of California Division of Agricultural Sciences. 483 pp. Joslyn, M. A., and W. V. Cruess 1934. Elements of wine making. California Agr. Ext. Serv. Circ. 88:1-64. Joslyn, M. A., and D. C. Vosti 1955. Yeast autolysis. II. Factors influencing the rate and extent of autolysis. Wallerstein Laboratory Communications 18:191-205. Kern, A. 1964. Die Bedeutung des Hydroxymethylfurfurols als Qualitatsmerkmal fur Fruchsafte und Konzentrate. Internal!. Fruchtsaft-Union. Fruchtsaftforschung und -technol. Vienna, pp. 203-18. Kliewer, W. M., and H. B. Schultz 1964. Influence of environment on metabolism of organic acids and carbohydrates in Vitis vinifera. II. Light. Amer. Jour. Enol. and Vitic. 15:119-29. Lewis, M. J. 1964. Aspects of nitrogen metabolism of brewers' yeast. Wallerstein Laboratory Communica- tions 27 :29-39. Lipis, B. V., V. N. Nykandrova, and L. A. Kalguskina 1959. Thermal treatment of ordinary fortified wines. Sadovodstvo, Vinogradarstvo i Vinodelie Moldavii 14(4) :48-50. (Chem. Abst. 54 : 7059c.) Martini, L. P., and V. W. Cruess 1956. Sherry production practices survey. Rept. to Tech. Adv. Comm. Wine Institute, San Francisco. Feb. 20, 1956. 16 pp. Masuda, H., and H. Muraki 1963. Utilization of yeast cells in wine-making. I. Prolonged keeping of white wine on sedi- mented yeast. Bui. Res. Inst. Ferm. (Japan) 38(10) :15-22. Matalas, L., G. L. Marsh, and C. S. Ough 1965a. The effect of concentration conditions and storage temperatures on grape juice concen- trate. Amer. Jour. Enol. and Vitic. 16:129-35. 19656. The use of reconstituted grape concentrate for dry table wine production. Amer. Jour. Enol. and Vitic. 16:136-43. [34] Mattick, L. R., and W. B. Robinson 1960a. Changes in the volatile constituents during the baking of sherry wine by the Tressler process. Food Technol. 14:30-33. 19606. Changes in the volatile acids during the baking of sherry wine by the Tressler baking process. Amer. Jour. Enol. and Vitic. 11 : 113-16. Meffert, H. F. Th. 1964. Zum Problem der Verweilzeiten und Umsetzungen beim Eindampfen von Fruchtsaften Ein Versuch zur Bewertung von Eindampfanlagen. Internatl. Fruchtsaft-Union. Fruchts- aftforschung und -technol. Vienna, pp. 219-37. Nordstrom, K. 1964. Studies on the formation of volatile esters in fermentation with brewers' yeast. Svensk Kemish Tidskift 76:1-34. Otsuka, K., S. Hara, and S. Imai 1963. Effect of temperature on the fermentation of wine. Nippon Jozo Kyokai Zasshi 58(7) : 631-35. OUGH, C. S. 1964. Fermentation rates of grape juice. I. Effects of temperature and composition on white juice fermentation rates. Amer. Jour. Enol. and Vitic. 15:167-77. 1966a. Fermentation rates of grape juice. II. Effect of initial °Brix, pH and fermentation tem- perature. Amer. Jour. Enol. and Vitic. 17:20-26. 19666. Fermentation rates of grape juice. III. Effect of initial ethyl alcohol, pH and fermentation temperature. Amer. Jour. Enol. and Vitic. 17(2) :74-81. Ough, C. S., and M. A. Amerine 1959. Dissolved oxygen determination in wine. Food Research 24:744-48. 1961a. Studies with controlled fermentation. VI. Effects of temperature on rates, composition and quality of wines. Amer. Jour. Enol. Vitic. 12:117-28. 19616. Polyethylene and cork closures and the fermentation temperature for sparkling wines. Wines & Vines 42(10) : 27-28. 1962. Studies with controlled fermentation. VII. Effect of antefermentation blending of red musts and white juice on color, tannins and quality of Cabernet Sauvignon wine. Amer. Jour. Enol. and Vitic. 13:181-88. 1963a. Use of grape concentrate to produce sweet table wines. Amer. Jour. Enol. and Vitic. 14:194-204. 19636. The production of table wines in regions IV and V. Wines & Vines 55(6) :56-58, 60-62. 1963c. Regional, varietal, and type influences on the degree Brix and alcohol relationship of grape musts and wines. Hilgardia 34(14) :585-600. Ough, C. S., J. F. Guymon, and E. A. Crowell 1966. Formation of higher alcohols during grape juice fermentations at various temperatures. Jour. Food Sci. 31:620-25. Ough, C S., E. B. Roessler, and M. A. Amerine 1960. Effects of sulfur dioxide, temperature, time, and closures on the quality of bottled dry white table wine. Food Technol. 14:352-56. Pilone, B. F., and H. W. Berg 1965. Some factors affecting tartrate stability in wine. Amer. Jour. Enol. and Vitic. 16:195-211. Pilone, F. J. 1953. The role of pasteurization in stabilization and clarification of wines. Proc. Amer. Soc. Enol. 1953:77-83. Prehoda, J. 1963. Homersekletszabalyzas a voros borok ergesztesenel. Borgazdasay H : 15-23. Rankine, B. C. 1953. Quantitative differences in products of fermentation by different strains of wine yeasts. Australian Jour. Appl. Sci. 4:590-602. Ribereau-Gayon, J., R. Cassignard, P. Sudraud, J. Blouin, and J. C. Barthe 1963. Sur la vinification en blanc sec. Compt. Rend. Acad. Agr. France. 49:509-12. Ribereau-Gayon, J., and P. Sudraud 1958. Etude experimentale de la vinification en rouge V. Experimentation 1955. Chimie et In- dustrie 79:774-81. Roessler, E. B., and M. A. Amerine 1958. Studies on grape sampling. Amer. Jour. Enol. 9:139-45. 1963. Further studies on field sampling of wine grapes. Amer. Jour. Enol. and Vitic. 14:144-47. [35] ROSENSTEIN, E. 1964. Wine production in countries with hot climates. Wines & Vines 45(8) : 29-32. 1965. Probleme des vinification en pays chauds. Bui. Office Internatl. Vin. 38:164-84. Rosenstein, E., and P. S. Rosenthal 1964. Emploi des antiseptiques dans les vins en association avec d'anhydride sulfureux en vue du replacement de ce dernier. Bui. Office Internatl. Vin. 37 : 620-22. Rossi, E. A., Jr. 1960. Observations on oxygen absorption in wine. Rept. to Tech. Adv. Comm. Wine Institute. San Francisco. May 23, 1960. 9 pp. Saller, W. 1955. Die Qualitatsverbesserung der Weine und Siissmoste durch Kalte. Frankfurt/Main: S. Horn, 98 pp. Singleton, V. L. 1962. Aging of wines and other spiritous products, acceleration by physical treatments. Hil- gardia32(7):319-92. Singleton, V. L., and Diana E. Draper 1964. The transfer of polyphenolic compounds from grape seeds into wine. Amer. Jour. Enol. andVitic. 15:34-40. Singleton, V. L., C. S. Ough, and M. A. Amerine 1964. Chemical and sensory effects of heating wines under different gases. Amer. Jour. Enol. andVitic. 15:135-45. Skofis, E. 1953. The role of refrigeration on stabilization and clarification of wines. Proc. Amer. Soc. Enol. 1953:69-76. Slator, A. 1906. Studies in fermentation. I. The chemical dynamics of alcohol fermentation by yeast. Jour. Chem. Soc. (London) 89:128-42. Stradelli, A. 1951. Evaporazione di alcol durante la fermentazione dei mosti.Riv.Viticolt.Enol. (Conegliano) 4:50-53. Stuhrk, A. 1960. Der Einfluss niedriger Temperaturen auf den Garverlauf in Traubenmost und seine Auswirkung auf Zusammensetzung und Qualitat des Weine. Qualitas Plant, et Mater. Veg. 7:147-58. Sudraud, P., and R. Cassignard 1958. Influence de certaines conditions dans la vinification en rouge. Ann. de Technol. Agr. 2:209-16. Tarantola, C. 1937. La preparazione dell "Asti Spumante" con fermentazione a bassa temperature. Annuario della Region Stazione Enologica Sperimentale di Asti 2(11) :315-21. Teodoresco, S. C. 1965. Influence du temperatures de fermentation et de conservation du vin et vins speciaux sur leur caracteres chimiques, microbiologiques et organoleptiques. (Report of Rumania.) Office Intern. Vin. IX Cong., Paris. 8 pp. See Bui. Office Internatl. Vin. 38(416) : 1174-76. Warkentin, H., and M. S. Nury 1963. Alcohol losses during fermentation of grape juice in closed fermenters. Amer. Jour. Enol. andVitic. 14:68-74. Winkler, A. J., and M. A. Amerine 1937. What climate does. The relation of weather to the composition of grapes and wine. The Wine Review 5(6) :9-ll; 5(7) :9-ll, 16. Zimmermann, H. W., E. A. Rossi, Jr., and E. Wick 1964. Alcohol losses from entrainment in carbon dioxide evolved during fermentation. Amer. Jour. Enol. and Vitic. 15:63-68. To simplify the information, it is sometimes necessary to use trade names of products or equip- ment. No endorsement of named products is intended nor is criticism implied of similar products not mentioned. 15ra-10,'66(G4464)L.L.