The Effects of Different Cultural and Environmental Factors on Grapevine Growth, Winter Hardiness and Performance in Three Locations in Canada

Abstract
Twenty grapevines genotypes of diverse hybrid origin ('Chancellor', 'Delisle', 'ES-6-12-28', 'ES-4-7-25', 'GR-7', 'Hibernal', 'Sabrevois', 'Kay Gray', 'Lucie Kuhlmann', 'Michurinetz', 'Okanagan Riesling', 'Prairie Star', 'St. Croix', 'St. Pepin', 'Seyval noir', 'Seyval blanc', 'SV-18-307' 'Vandal-Cliche' and 'Vidal blanc') and one vinifera ('Siegerrebe') were evaluated under different winter protection methods where twelve hardy and semi-hardy genotypes received 3 treatments, seven tender genotypes received one treatment and only one tender genotype received 2 treatments (vines without protection: (M1); pinning half of the canes to the ground allowing for natural snow cover protection and the remaining canes left attached to the trellis wires without protection (M2); vine removed from the trellis, laid down and covered with geotextile fabric (M3) or soil (M4) at three locations in Quebec (Canada) differing in soil type and microclimate conditions. The effects of these methods on annual production, winter resistance and vegetative growth were measured. Total yield for all genotypes at Frelighsburg were 3 and 4 times higher than at L'Orpailleur (Dunham) and Dietrich-Jooss (Iberville) commercial vineyards, respectively. For protection system M1 and M2, the highest yields were recorded for 'Vandal-Cliche' and 'St. Croix' followed by 'ES-4-7-25' and 'St. Pepin'. The M3 treatment was the most effective protection for some semi-hardy genotypes. Under M3 protection average yield of semi-hardy genotypes remained higher (6 and 4 kg) than that of the hardy genotypes (3 and 1.7 kg) at L'Orpailleur and Dietrich-Jooss (2000), respectively. 'Seyval noir', 'Seyval blanc', 'Chancellor', 'Vandal-Cliche', and 'ES-4-7-25' exceeded 20 kg under M3 treatment. M4 protection for 'Seyval blanc' was not effective resulting in decreased yield, vigour and increased mortality. Higher bud survival levels were observed at Frelighsburg for hardy genotypes such as 'Sabrevois', 'St. Croix', 'Kay Gray', 'Vandal-Cliche', 'St. Pepin' and 'Michurinetz' under all protection treatments. In commercial vineyards, greater than 50 % bud mortality was observed for semi-hardy genotypes 'ES-6-12-28', 'GR-7' and Lucie Kuhlman', with even higher mortality in tender genotypes 'Siegerrebe', 'Vidal blanc' and 'SV-18-307'. The highest vigour was achieved in hardy and moderately hardy cultivars while the tender cultivars were less vigorous at all three sites and over both years. Site location was the most important factor affecting vine yield and mortality, with the best performances being recorded at Frelighsburg. This is assured due to the site in Frelighsburg having a slight south slope very well drained sandy soil, and excellent snow accumulation.
KEYWORDS: Vitis spp, winterhardiness, protection against cold, grape yield, juice composition.

Introduction

Low winter temperature is the major environmental factor limiting vineyard productivity in Quebec (Dubois and Deshaies, 1997). The majority of Quebec's commercial vineyards are concentrated between 45° and 47° north latitude where winter minima regularly reach -30 °C and occasionally near -35 to -45 °C (Jolivet et al., 1999). Under these extreme conditions, cold injury is expected mid-winter, but also occurs in late fall before the vines have fully acclimated or in late spring after sap flow is re-initiated. Autumn frosts can cause premature defoliation, limit the normal vegetative cycle and make a normal harvest season difficult (Galet, 1993). Severe yield losses have also been attributed to late spring frosts, when both primary and secondary buds can suffer irreversible damage at temperatures from -2 °C to -4 °C (Dereudre et al., 1993). These constraints force the grape grower to use cold adapted short season cultivars, with reasonable yields and fruit composition for commercial production. Cold tolerance of many plant species has been extensively reviewed and studied (Weiser, 1970; Stergio and Howell, 1977; Levitt, 1980; Fowler et al., 1981; Gusta et al., 1982; Sakai and Larcher, 1987; Khanizadeh et al., 1989a; Khanizadeh et al., 1989b; Khanizadeh et al., 1992; Reisch et al., 1993; Khanizadeh et al., 1994; Rioux et al., 2000; Richer and Rioux, 2001). Winter injury can occur in all parts of the vine such as buds, canes, trunks and even roots. Differences in cold hardiness among diverse grape vine genotypes and the effect of site on cold acclimation and deacclimation (Stergio and Howell, 1977), grape bud survival (Clore et al., 1974; Pierquet and Stushnoff, 1980; Wolf and Cook, 1994; Clark and Watson, 1998) and productivity (Wolf and Warren, 2000; Wolf and Miller, 2001) have been studied extensively. Several research programs are also in progress to improve hardiness and wine quality (Reisch et al., 1993; Hemstad and Luby, 2000; Fisher and Jamieson, 2000; Gal, 2000) or to adapt techniques to minimize winter injury using proper location, cultural practices like irrigation application, microclimate modification, rootstocks and chemical application for increasing hardiness (Ahmedullah, 1985; Frances et al., 1974; Jolivet and Dubois, 2000; Stushnoff and Hamman, 2002). The dormant bud is usually considered the most cold susceptible part of the mature grapevine and frequently exhibits injury even though other vine tissues survive the same conditions (Ahmedullah, 1985; Quamme, 1986; Clark and Watson, 1998; Jolivet et al., 1999). Several works have shown that buds and cane tissue of grapevine supercool as a mechanism of freezing tolerance (Pierquet and Stushnoff, 1980; Quamme, 1986). Cultivars normally resistant to the cold withdraw interstitial cellular water and modify its molecular structure to prevent ice crystal formation (Pierquet et al., 1977; Audran et al., 1993; Wolf and Cook, 1994). Other studies have reported that grapevine bud and shoot winter survival are linked to different water regimes from January to March (Skorokhod, 1975). Frost resistance has also been correlated positively with high levels of water retention and in particular the ability to retain water in a liquid form despite extremely low temperatures (Pogosyan, 1975). Significant differences in cold hardiness were observed among the Vitis species and cultivars within each species. The bud tolerance of many hybrids varies from -15 °C to -35 °C (Vandal, 1986; Galet, 1988) while almost all varieties of Vitis vinifera L. are somewhat damaged between -15 °C and -20 °C (Galet, 1993). The rustic hybrids derived from Vitis riparia Michx and Vitis amurensis Rupr species can tolerate temperatures as low as -35 °C and -40 °C, respectively, and still produce fruit of reasonable quality (Vandal, 1986; Ahmedullah, 1985). To prevent the annual losses due to winter damage, a common viticultural practice is to bury the vines with at least 20 to 50 cm of soil in late fall depending on snow cover (Skorokhod, 1975; Vandal, 1986; Dubois and Dehaies, 1997). This method is commonly used in cold climates like Eastern Europe, Northern China, Minnesota and Quebec. However, the effectiveness of the vine burial process varies considerably. Skorokhod (1975) found that bud survival was higher when vines were covered with 40 cm of soil. Whereas at Stavropol area (Russia) Prostitova (1977) found that primary bud survival of unprotected vines was limited to 6-15 % when submitted to hard winters. In Moldavia, Kondo et al. (1972) observed that during a normal winter, the covering process reduced winter damage to frost-sensitive cultivars but increased damage to frost-resistant varieties. This practice is very expensive and time consuming, especially when one removes the soil in early spring. Moreover, the burial process causes cane damage, increasing the probability of cane diseases, delaying bud break and increasing soil erosion when there is no cover crop between the rows (Pierquet et al., 1977; Vandal, 1986; Jolivet and Dubois, 2000). In Russia, to replace the burial method, Stetsenko (1978) found that winter survival was good and yields were increased up to 200 percent even in the colder areas, using polyethylene covers in combination with 1-3 kg straw. Bordelon et al. (1997a) also compared closed-cell polyethylene foam sheeting (PE) with straw grapevine covering at Lafayette Indiana (central USA) and reported that with PE, primary bud survival averaged 80-92%, while with straw it was 53-55%. In studying the efficiency of two covering agents for the protection of vines against late spring frosts in Quebec, Jolivet et al. (1999) have shown that polystyrene cones are the best and they maintain the mean cane temperature 1.7 °C above those protected by geotextile fabric alone. Moreover, the temperature of canes under geotextile fabric was colder by 1.5 °C in May than those without any protection. Snow cover can also lessen the amount of cold damage by insulating the vines from cold air temperatures. Lavoie (1971) showed that the yields of cultivated blueberries were 1.2 to 4.3 times the unprotected control when the snow cover was 15 and 30 cm thick, respectively. Jolivet et al. (1998) observed that the temperature at the graft union of a 'Muscadet Melon' vine was -1 °C under 40 cm of snow and -26 °C at the ambient air. Thus, the utilisation of cold resistant cultivars and protection of semi-hardy and non-hardy cultivars by snow, soil or fabrics is essential to assure a stable annual production in Quebec and other northern commercial vineyards. In spite of increasing numbers of vineyards in Quebec, no study has been yet conducted to evaluate commonly grown vine cultivars in Quebec and their responses to different cultural practices.

The aim of this study was to:
 1) evaluate the winter survival of 20 genotypes thought to be of value to the eastern Canadian viticultural industry and 
2) determine the effects of environmental factors and cultural practices on winter hardiness, vegetative and reproductive growth.

 

Materials and Methods

Plant Materials and sites
This study was conducted using 20 nongrafted grape genotypes including diverse interspecific hybrids and one vinifera type (Table 1), and planted in 1998. These genotypes had previously been classified as hardy, semi-hardy and tender (Reisch et al., 1979; Odneal, 1983; Bordelon et al., 1997b; Dubois and Deshaies, 1997; Plocher and Parke, 2001). Three locations with different soil and microclimate conditions were chosen: 1) Agriculture and Agri-Food Canada experiment station site at Frelighsburg, Quebec (45o N-72o W and elevation 205 m) with a sandy loam soil containing pebbles and gravel, having a pH of 6.0. This site is sloped, allowing excellent cold air drainage and has a wind break to the south. 2) A commercial vineyard l'Orpailleur in Dunham, Quebec (45o N-72o W and elevation 125 m) with a schistous podzol soil with natural pH of 4.9-5.3 (improved to pH 6.2 on the top 15 cm with lime application). This site in level land surrounded by woods increase cold air accumulation. 3) A commercial vineyard Dietrich-Jooss in Iberville, Quebec (45o N-73o W and elevation 43 m) with a sandy loam, stony soil, having a pH of 6.6. This site is also on level land, but has no wind break.

Experimental design and field layout
The experimental design for the three trials was an unbalanced, randomized, complete block, split plot with winter protection as the main plot and genotype as the sub plot with four replications (blocks). Each replication had 5 vines spaced 1.5 m in the row and 3 m between the rows. The three middle vines were used for data collections and the rest kept as guard vines. The training system was composed of four horizontal wires attached to 1.8 m cedar posts at 45 cm, 80 cm, 125 cm and 175 cm from soil level. Pruning was done based on the need of the winter protection methods used in each vineyard. Hardy and semi-hardy genotypes were trained to a Mini-J-Style training system, while tender genotypes were trained to a Fan training system. All genotypes were cane pruned and had multiple trunks. The vine genotypes replications were randomized within each winter protections methods. Four different methods winter protection were use as follow: Method 1 (M1): The canes were left attached to the trellis with no winter protection of any sort. Only hardy and semi-hardy genotypes were assigned to this treatment. Method 2 (M2): Half the canes were pinned to the ground so in the winter they would be covered by snowfall. The remaining canes were left attached to the trellis without protection. Only hardy and semi-hardy genotypes were assigned to this treatment. Method 3 (M3): All the canes were removed from the trellis, laid down on the ground and covered with geotextile fabric (Polyester, Arbotex Plast 0100 2.00X50M, Canada). This method was used for hardy, semi-hardy and tender genotypes. Method 4 (M4): All the canes were buried by 0.35 m of soil taken between the rows. This method was used only for 'Seyval blanc' the most common genotype in Quebec. Among the 20 grapevine genotypes, 12 genotypes (hardy and semi-hardy) received three of these treatments, 7 genotypes (tender) received one treatment and only one genotype received two treatments. All the winter protection methods were applied in late fall when the vines had lignified and the day and night temperature was approximately 0 C to avoid etiolation and continuous growth.

Temperature measurements
Temperature measurements were carried out during the 1999-2000 and 2000-2001 winters in all three sites. Data probe measurements were taken every six hours with a temperature probe plugged into a Hobo external temperature sensor (H08-002-02, Onset computer corporation, Bourne MA). The temperature sensors were placed at several positions depending on the method of vine protection. For all 4 protections treatments, a temperature sensor was attached to the 2nd trellis wire at ~80 cm from the soil level to measure the ambient air temperature. In addition, for M2 a second probe was placed ~5 cm above the soil and attached to a cane; for M3 a second probe was placed on a cane under the geotextile at ~2.5 cm from the soil and for M4, a second probe was placed on a cane ~25-30 cm below the soil surface.

Assessments
Winter hardiness was evaluated before pruning, at bud-break, the following spring. Mortality was visually judged in the spring and rated as percentage of shootless nodes. Vine vigour was determined by scoring vines in July of each growing season and rated as length and shoots density using the following scoring: 1= not vigorous, 2= low vigour, 3= intermediate, 4= vigorous and 5= very vigorous. Yield was determined by measuring the total fruit weight of the three middle vines occupying 4.5 m of row. Cluster and berry weight were also calculated using 5 clusters and 50 berries randomly sampled from each of three vines for each replicate, but only at the Frelighsburg site. Soluble solids concentration (SSC) was determined using a digital refractometer (Abbe Mark II, Baxter Division Canlab, Canada) and titratable acidity (TA) was determined by titrating 5 ml juice sample with 0.1 N NaOH to a pH of 8.1 (665 Dosimat, Metrohm Ltd. Switzerland). SSC and TA were measured only at the Frelighsburg site and only for treatment M3. Fruits were harvested when the juice composition reached the desirable levels for commercial wine making.

Statistical analysis
Data were analysed by SAS (SAS, 1989) using an unbalanced randomized complete block, split plot model. The means separation and comparison were done by LSD. The arc-sin transformed data on percentage bud mortality were used to perform the analysis of variance. Under M3 winter protection, specific groups of genotypes (hardy, Semi-hardy and tender) were compared through the use of orthogonal contrasts.

 

Results
Mean maximum and minimum temperatures for the November-April period of the study are presented in Table 2. According to the temperatures recorded at all three locations, 2000-2001 was colder than 1999-2000 winter season. Means maximum and minimum temperatures varies with winter protection methods. In both 1999-2000 and 2000-2001 seasons, M3 followed by M2 recorded higher mean maximum temperatures than M4. In contrast, the lower mean minimum temperatures were registered in M2 method (Table 2). Winter absolute minimum temperatures were generally much lower than 0 °C and varied according to both the type of winter protection and the site location. As showed in Table 2, the lowest minimum was recorded at L'Orpailleur on the second wire for M1 (-33.7 and -32.5 °C) during the two study periods. Due to the large size of each experiment and short harvest period to avoid frost, it was not possible to collect berry weight and cluster weight for all three experimental sites. Therefore berry and cluster weights were measured only at Frelighsburg for all four treatments methods and all 20 genotypes. Total yield was completely evaluated for all three sites for 2000 but only at Frelighsburg and L'Orpailleur for 2001.The two commercial sites were not compared to Frelighsburg site since they were completely different in terms of soil and microclimate.

Performance of 12 hardy and semi-hardy genotypes under M1, M2 and M3 treatments Frelighsburg
There were significant differences in total yield among genotypes within treatment and significant cultivar x treatment interaction for total yield and vigour which suggests that grapevine genotypes reacted differently to winter protection methods (Table 3). Table 4 shows the effect of winter protection methods on yield of grapevine genotypes. In 2000 and 2001, the total yield for all genotypes at Frelighsburg was 3 and 4 times higher than at L'Orpailleur and Dietrich-Jooss commercial vineyards, respectively. 'Vandal-Cliche' (30.4, 27.8 and 21.1 kg for M1, M2 and M3 respectively) and 'St. Croix' (17.4, 14.1 and 14.4 kg for M1, M2 and M3, respectively) were among the highest yielding genotypes in 2000. The lowest yielding was for 'Prairie Star' (5.4, 3.9 kg and 3.6, 4 kg for M1 and M2, respectively). One of the interesting observations of the 2000 data, is that the semi-hardy genotypes 'ES-6-12-28' and 'Lucie Kuhlmann' yielded much more when covered with geotextile fabric than 'St. Pepin' and 'Kay Gray' hardy genotypes which had a decrease in total yield with M2 and M3 treatments compared to M1 (Tables 4-6). In contrast, 2001 was characterized by different relative yield patterns. With M1, yield was greater in 'Vandal-Cliche', 'ES-4-7-25', 'Sabrevois' and 'ES-6-12-28' (12.2, 12.1, 10.2 and 10.0 kg, respectively) than in 'Prairie Star' and 'GR-7' (3.6 kg). In 2001, total yield obtained in 'Vandal-Cliche' and 'ES-6-12-28' is 2x higher and in 'GR-7' and 'Lucie Kuhlmann' is 3x higher in M3 than M1 (Tables 4-6). At Frelighsburg, the highest overall bud survival levels were seen for those selections already exhibiting high levels of cold resistance such as 'Sabrevois', 'St. Croix' and 'Kay Gray' (mortality 15%) followed by 'Vandal-Cliche', 'St. Pepin', 'Michurinetz' and 'GR-7' (between 15 and 25%) for M1, M2 and M3 protection (Table 7). In contrast, 'Lucie Kuhlmann' showed a higher bud mortality rate (>30%) under M1 method in 2000, whereas in 2001 it was 'Delisle' who recorded the higher rate ( 30%), under M1 and M3 than M1 method (Tables 6-7). For 2000, the highest vigour was achieved in hardy 'St. Croix' and semi-hardy 'GR-7' and 'Lucie Kuhlmann' genotypes followed by 'Kay Gray' in all three treatments. For 2001, almost all genotypes record a higher vigour under M3 than under M1 or M2 treatment (Tables 3-8).

Commercial vineyards
Significant cultivar and treatment effects were observed for all measured parameters and also significant cultivar x treatment interaction for total yield, bud mortality and vigour, except at Dietrich-Jooss vineyard where no treatment effect and no interaction occurred for vigour parameter. Depending on year, the results in commercial vineyard (L'Orpailleur) showed that all semi-hardy and some hardy genotypes such as 'Michurinetz', 'St. Croix', 'St. Pepin' and 'Delisle' had higher total yield using geotextile fabric (M3) compare with cane training to the ground (M2) or no protection at all (M1) (Tables 3-4-5). On the other hand in 2000, 'ES-4-7-25' recorded a higher yielding under M2 than under M1 or M3 method. In 2000, 'Vandal-Cliche' had the highest yields (7.2 kg) without any protection whereas, yield of the remaining genotypes varied between 0.25 and 3.40 kg. Under M2 treatment, the highest yielding was for 'ES-4-7-25' with 5.75 kg. In 2001, 'ES-4-7-25' had once again the highest yielding for both M1 and M2 treatments. Under treatment M3, there was a range of 2.40-10.93 kg among genotypes from lowest ('Prairie Star') to highest ('ES-4-7-25') yield. In Dietrich-Jooss vineyard, yield records were only available for 2000. The semi-hardy genotypes 'GR-7', 'Lucie Kuhlmann' yielded more than all the hardy genotypes using M3 (Table 4). Interestingly, treatment M2 was more effective at Dietrish-Jooss location for semi-hardy 'GR-7' and 'ES-6-12-28', yielding higher than M1 or M3 method (Tables 5-6). For the remaining genotypes no significant difference was observed between treatments.
Large variation was found in mortality among genotypes in each site. 'ES-6-12-28', 'GR-7' and 'Lucie Kuhlmann' semi-hardy genotypes suffered the most winter injury with mortality rates 70% when compared to the hardy genotypes with M1 protection. As expected, the hardy genotypes were relatively little affected but showed a lower mortality rate with M3 compared to M1 and M2 protections while at Dietrich-Jooss this behaviour was observed for 'Sabrevois', Mitchurinetz' and 'St. Croix' compared to M1 treatment (Tables 5-6). Despite the fact that Frelighsburg location differs from L'Orpailleur commercial vineyards, the highest vigour was also found in 'St. Croix' under the three treatments, followed by 'ES-4-7-25', 'Sabrevois' under M2 method and semi-hardy genotypes under M3 method. At Dietrich Jooss, all genotypes were less vigourous to intermediate, mean vigour ranged from 2.2 to 3.6, 2.5 to 4.2 and 2 to 3.8 in M1, M2 and M3 treatment, respectively (Table 7).

Performance of all twenty genotypes under M3 treatment
Significant differences in total yield, mortality and vigour were observed among cultivars (P<0.0001 to P=0.0014). In 2000, yield ranged from 0.4-21.1 kg among genotypes with the highest being 'Vandal-Cliche' and the lowest 'SV-18-307' at Frelighsburg. However in 2001, 'Seyval noir', 'Seyval blanc', 'Chancellor' followed closely by 'Vandal-Cliche' and 'ES-4-7-25' could exceed 20 kg (Table 4). The lowest yielding genotypes were 'Delisle', 'Prairie Star' and 'Michurinetz'. These yielded 9 kg or less and all the remaining genotypes were intermediate with a total yield varying between 10.7 and 18.8 kg (Table 4). At commercial vineyard L'Orpailleur highest yielding were recorded by 'ES-6-12-28', 'Lucie Kuhlman' and 'Vandal-Cliche' in 2000 and by 'SV-18-307' and 'ES-4-7-25' in 2001. The least productive were 'Siegerrebe' and 'Prairie Star' with yield less than 2.5 kg. At Dietrich-Jooss, 'GR-7' and 'Lucie Kuhlman' and 'Seyval blanc' had the highest yields ( 5 kg) and the lowest were 'Siegerrebe' and 'Prairie Star' (=0.5 kg).
Orthogonal contrast showed significant differences between groups of genotypes hardy vs semi-hardy (H vs SH) for yield in both commercial vineyard. This indicates that M3 method was more effective for semi-hardy group than for hardy group associated with low bud damage at L'Orpailleur site (Tables 4-8). At Frelighsburg, no significant contrast was observed for yield and bud mortality, but was significant for vigour indicating that semi-hardy group was most vigourous than the hardy group (Tables 7-8). In the case of hardy and semi-hardy vs tender group (H&SH vs T) significant contrasts were noted on yield, mortality and vigour at Frelighsburg, but were not significant on yield, neither at L'Orpailleur in 2001 nor at Dietrich-Jooss vineyards (Table 8). Indeed, year 2000 was associated with high production, low bud damage and high vigour for hardy and semi-hardy group, in both sites of Frelighsburg and L'Orpailleur. In contrast, 2001 growing season was more productive for tender genotypes although they recorded relatively greater bud mortality and a lower vigour than the hardy and semi-hardy group.

Performance of 'Seyval blanc' under M4 treatment
The only significant differences between the M3 and M4 winter protections were found in 2001 for yield, mortality and vigour at Frelighsburg and L'Orpailleur sites (Table 9). The M4 treatment appears not effective and yields of 'Seyval blanc' were low when compared with M3 winter protection. 2001 harvest 'Seyval blanc decreased half to more half from M3 to M4 (7.01 to 3.5 kg and 26.6 to 7.4 kg) at L'Orpailleur and Frelighsburg, respectively. Under geotextile fabric, 'Seyval blanc' achieved the lowest vigour and suffered the most injury with mortality rates of 90%.

Berry, cluster and fruit composition
No significant effect of winter protection was noted on berry and cluster weights but significant differences were found between genotypes within each protection method (Table 3). The hardy genotype 'Kay Gray' had the largest berry weights over the two seasons, whereas the semi-hardy genotypes 'GR-7' and 'Lucie Kuhlmann' recorded the lowest values of all genotypes evaluated in this trial (Table 10). 'Seyval blanc' had the largest cluster (>240 g) in the M3 treatment followed by clusters of 200 g or less ('Seyval noir', 'SV-18-307', 'Vidan blanc' and 'Chancellor'. The hardy genotypes 'ES-4-7-25', 'St. Pepin', 'Vandal-Cliche', 'St. Croix' and 'St. Pepin' had medium size clusters between 100 and 150 g in M1 and M2 (Table 10). In contrast, the lowest value was found for 'Delisle' (< 70 g) in the three treatments over the two years. Significant differences in juice composition (SSC and TA) were observed among genotypes within each growing season (P 0.001). SSC ranged from 15 to 22.4 °Brix and from 13.0 to 20.5 °Brix, TA values ranged also from 9.20 to 18.9 g/l and from 5.9 to 16.1 g/l in 2000 and 2001 season, respectively (Table 10). In 2000, the highest SSC value was found for 'GR-7' and 'Michurinetz' (22.4 and 20.4 °Brix, respectively) while 'Kay Gray' and 'Vandal-Cliche' had the lowest SSC (15 °Brix). Similar to SSC, the highest TA was for 'Michurinetz' (18.9 g/l) whereas the lowest acidity was for 'ES-4-7-25' and 'Delisle' (9.2 and 9.9 g/l, respectively). In year 2001, the American hybrids 'Prairie Star' and 'St. Pepin' showed higher SSC (20 and 20.5 °Brix, respectively) than 'Hibernal' and 'Seyval noir' (13.0 and 15.0 °Brix, respectively). The lowest TA was for 'Kay Gray' (5.9 g/l) followed by 'Siegerrebe' (6.8 g/l) and the highest for 'Michurinetz' and 'Lucie Kuhlmann' and (16.1 and 15.9 g/l, respectively).

Discussion
The 20 grapevine genotypes reacted differently to winter protection methods and to trial locations. Average berry and cluster weights did not vary significantly as a function of winter protection, indicating that lower yields were as a result of reduced cluster number not berry size (data not shown). On a total yield basis, 'Sabrevois', 'Prairie Star', 'Delisle' followed relatively by 'Michurinetz', St. Croix' and 'St. Pepin' seem to be relatively indifferent to winter protection method, their total yield was more or less stable, thus showing their genetic potential for these cold areas. However, at Frelighsburg some of the semi-hardy genotypes yielded much more when covered with snow (M2) or geotextile fabric (M3) than some of the hardy genotypes, which had a decrease in total yield with these same winter protections compared to the M1, a finding similar to Kondo et al. (1972). In a Sherbrooke (Quebec) vineyard, Jolivet et al. (1998) reported the temperature variation of 20 °C between a covered vine plant with snow and the ambient air temperature. In this study, total yields obtained at Frelighsburg were far higher than those obtained in the two commercial vineyards. In addition to the various winter protection methods, site location was the most important factor, more important than protection treatments affecting vine mortality and total yield. According to Sayed (1992), site location remains the most important factor in minimizing the effect of the climate and maximizing the moderating effects of mesoclimates. Similar observations were noted on apple tree mortality in different locations in Quebec (Khanizadeh et al., 1989b; Khanizadeh et al., 1992). L'Orpailleur and Dietrich-Jooss vineyards are characterized as low, flat sites with little snow accumulation. Frelighsburg is a higher site with vines planted on a slight south slope that results in good cold air drainage away from vineyard. Moreover, a nearby wind break to the mount decreases the effect of cold winds during the winter, enhancing the warmth of the south slope. This study shows that of the three winter protections (M1, M2 and M3), M3 protection has been the most efficient for some of the semi-hardy genotypes over the two years, especially at Frelighsburg but less at L'Orpailleur and Dietrich-Jooss vineyards. Total yield of these semi-hardy genotypes such as 'ES-6-12-28', 'GR-7' and 'Lucie Kuhlman' was greater than of some hardy genotypes with or without protection. Although it is classified hardy, the total yield obtained with M3 for 'ES-4-7-25' was greater than 'Prairie Star' or 'Michurinetz'. Unexpectedly, buried 'Seyval blanc' (M4) did not give good results and yields were low in both Frelighsburg and L'Orpailleur sites compared to M3 protection. Although the temperatures were warmer under the soil, probably other problems connected to the vine burial process (rot, fungus and moisture) could explain the poor results obtained here. However, Stetsenko (1978) reported that polyethylene covers with straw improved winter survival even in the colder areas and yields increased 200 percent over buried vines. The percent mortality was higher for L'Orpailleur and Dietrich-Jooss than for Frelighsburg over the two seasons (P=0.01). Our results also showed that hardy genotypes like 'Sabrevois', 'St. Croix', 'Kay Gray', 'Vandal-Cliche', St. Pepin', 'Michurinetz' and semi-hardy genotype 'GR-7' were little affected by cold winter in Frelighsburg with survival of primary buds remaining relatively high (percent mortality between 15 and 25%). Previous findings based on laboratory cold tests showed some agreement with the above observations (Rekika et al., 2003). Our results with controlled freeze tests, on the same twenty genotypes, showed that hybrids such as 'Sabrevois', Prairie Star', 'St. Pepin', 'St. Croix' and 'Kay Gray' derived from hardy American species, and 'Michurinetz' with V. amurensis in its genealogy were the most hardy, having a higher survival of primary buds at -30 °C. French-American hybrids were less hardy, being crosses with vinifera in their pedigree and the vinifera genotype 'Siegerrebe' was the most tender (Rekika et al., 2003). Hemstad and Luby (2000) evaluated 15 genotypes for winter hardiness in Minnesota after experiencing -38 °C, and found 'St. Croix', 'Kay Gray', and 'Michurinetz' among the hardiest genotypes, while 'St. Pepin' was the least hardy. Bordelon et al. (1997b) also evaluated percent survival of primary buds following -32 °C in January 1994 at 2 locations in Indiana and 6 locations in Ohio. These authors rated 'St. Pepin' as very hardy, 'Chancellor' as moderately hardy and 'Vidal blanc' as the most winter tender. However, great differences were observed within the same genotype grown in different locations. Again, our observations closely matched those reported in the literature, which postulated that vineyard conditions that predispose vines to good acclimation or susceptibility to cold injury (location/aspect, vine health, soil drainage, soil fertility, and particularity crop load) may account for bud survival differences within a cultivar. Moreover, climatic patterns prior to and the timing of the cold incident that also predispose the vines to good acclimation or injury could also explain this difference. With regard to fruit juice composition, some genotypes produced good quality fruit with good balance in SSC and TA for wine, namely 'Prairie Star', 'GR-7', 'Delisle', 'ES-6-12-28', 'Vandal-Cliche', 'Kay Gray', 'Okanagan Riesling' and 'SV-18-307'. The other genotypes showed lower SSC and higher TA than was reported in the literature (Wolf and Warren, 2000; Kaps and Odneal, 2001; Plocher and Parke, 2001; Reisch and Luce 2002) because of the short season and premature harvest to control bird damage. In fact, some genotypes did not completely reach the desirable SSC and TA levels because of the truncated season. According to Spayd et al. (1989) for vinifera and French-American hybrids genotypes, fruit is harvested between 19-21 °Brix for white wine and between 20 and 24 °Brix for red wine. Harvest criteria of the American hybrids are similar to those described above. In another words, the absolute level of sugar and acid that determines ripeness will vary between genotypes and the achievement of mature fruit composition values may not be directly related to desired varietal character (Howell et al., 1998; Plocher and Parke, 2001). In the case of 'Seyval blanc', sugar content of 19 °Brix is not enough for full development of its varietal flavor. In contrast, genotypes such as 'Kay Gray' and 'Vandal-Cliche' produce excellent wine when harvested at low sugar content to avoid an undesirable flavor in the fully matured fruit (Plocher and Parke, 2001). In summary, the selection of a proper site, a proper slope, soil drainage and wind breaks increased yield, decreased mortality and increased vigour and productivity of grapevines. Geotextile fabric winter protection was more effective than burying vines. It was costly but the material can be used for several years. Geotextile covering or pruning the vines close to the soil level could constitute an alternative to increase yield and decrease mortality.

	Ahmedullah M. 1985. An analysis of winter injury to grapevines as a result of two severe winters in Washington. Fruit Varieties Journal 39 (4): 29-34.
	Audran J.C., Leddet C., Dereudre J., Ait Barka E. and O. Brun. 1993. Réponse de la vigne (Vitis vinifera L.) aux températures inférieures à 0 °C. I. Effets d'un refroidissement contrôlé sur les sarments aoûtés. Agronomie 13 (6): 491-498.
	Bordelon B.P., Henick-Kling, T., Wolf, T.E. and E.M. Harkness. 1997a. Winter protection of cold-tender grapevines with insulating materials. Proceeding of the fourth International Symposium on cold climate viticulture & enology, Rochester, New York, USA, 16-20 July 1996. I:32-35.
	Bordelon B.P., Ferree D.C. and T.J. Zabadal. 1997b. Grape bud survival in the Midwest following the winter of 1993-1994. Fruit Varieties Journal 51 (1): 53-59.
	Clark J.R. and P. Watson. 1998. Evaluation of dormant primary bud hardiness of Muscadine grape cultivars. Fruit Varieties Journal 52 (1): 47-50.
	Clore W.J., Wallace M.A. and R.D. Fay. 1974. Bud survival of grape varieties at sub-zero temperature in Washington. Amer. J. Enol. Vitic. 25 (1): 24-29.
	Dubois J.M.M. and L. Deshaies. 1997. Guide des vignobles du Québec. Les presses de l'Université Laval, Sainte-Foy, 297 p.
	Dereudre J., Audran J.C., Leddet C,. Ait Barka E. and O. Brun. 1993. Réponse de la vigne (Vitis vinifera L.) aux températures inférieures à 0 °C ; III : Effets d'un refroidissement contrôlé sur des bourgeons en cours de débourrement. Agronomie 13 (6) : 509-514.
	Fisher K.H. and A.R. Jamieson. 2000. L'Acadie, a cold hardy, white wine grape cultivar for low heat unit regions. Proceeding of the Seventh International Symposium on Grapevine Genetics and Breeding. Ed A. Bouquet and J.M. Boursicot. Acta horticulturae 528 (2): 563-567.
	Fowler D.B., Gusta L.V. and N.J. Tyler. 1981. Selection for winter hardiness in wheat. III. Screening methods. Crop Sci. 21:896-901.
	Frances V., Chinchetru G. and P. Esteran. 1974. The use of morphactins for the control of spring frosts in viticulture. Bulletin de l' O.I.V. 47: 758-764.
	Gal L. 2000. Wine quality of new fungus disease resistance grapevine varieties. Proceeding of the Seventh International Symposium on Grapevine Genetics and Breeding. Ed A. Bouquet and J.M. Boursicot. Acta horticulturae 528 (2): 559-562.
	Galet P. 1988. Précis de viticulture. 4e éd., Déhan, Montpellier, 586 p.
	Galet P.1993. Précis de viticulture. 6e éd, Déhan, Montpellier, 575 p.
	Gusta A.L., Fowler D.B. and N. J. Tyler. 1982. Factors influencing hardening and survival in winter wheat in Plant cold hardiness and freezing stress, (P. H.
	Li & A. Sakai, eds.), pp. 23-40. Academic Press, New York, NY.
	Hemstad P.R. and J.J. Luby 2000. Utilization of Vitis riparia for the development of new wine varieties with resistance to disease and extreme cold. Proceeding of the Seventh International Symposium on Grapevine Genetics and Breeding. Ed A. Bouquet and J.M. Boursicot. Acta horticulturae 528 (2): 487-490.
	Howell G.S., Miller D.P. and T.J. Zabadal. 1998. Wine grape varieties for Michigan. MSU Extension Fruit Bulletins. http://www.msue.msu.edu/msue/imp/modfr/26439701.html
	Jolivet Y., Dubois J.M.M. and H. Granberg. 1998. Évaluation du régime thermique du cépage Vitis vinifera L. var. Melon durant la saison froide au Québec. J. Int. Sci. Vigne Vin. 32 (2): 51-58.
	Jolivet Y., Dubois J.M.M. and H. Granberg. 1999. Évaluation de l'efficacité du cône de polyester et de la toile géotextile comme méthodes de protection de la vigne contre les gels tardifs printaniers au Québec pour les petites exploitations et les pépinières. J. Int. Sci. Vigne Vin. 33 (3): 99-104.
	Jolivet Y. and J.M.M. Dubois. 2000. Évaluation de l'efficacité du buttage de la vigne comme méthode de protection contre le froid hivernal au Québec. J. Int. Sci. Vigne Vin. 34 (3): 83-92.
	Kaps M. and M.B. Odneal. 2001. Grape cultivar performance in the Missouri Ozark region. J. Amer. Pom. Soc. 55 (1): 34-44.
	Khanizadeh S., Buszard D. and C.G. Zarkadas. 1989a. Effect of crop load on seasonal variation in chemical composition and spring frost hardiness of apple flower buds. Can. J. Plant Sci. 69: 1277-1284.
	Khanizadeh S., Buszard D. and C.G. Zarkadas. 1989b. Seasonal variation of protein and amino acids in apple flower buds (Malus pumila Mill. cv. McIntosh/M7). Journal of Agricultural and Food Chemistry 37:1246-1252.
	Khanizadeh S., Buszard D. and C.G. Zarkadas. 1992. Effect of crop load on hardiness, protein and amino acid content of apple flower buds at several stages of development. J. Plant Nutrition 15(11): 2441-2455.
	Khanizadeh S., Buszard D. and C.G. Zarkadas. 1994. Seasonal variation of hydrophilic, hydrophobic, and charged amino acids in developing apple flower buds. Journal of Plant Nutrition 17(11):2025-2030.
	Kondo I.N., Pudricova L.P. and S. Ya Ginzburg. 1972. A study on the frost resistance of grape varieties in the conditions of Moldavia. Plant Breeding Abstracts 45: 386.
	Lavoie V. 1971. Importance de la couverture de neige. In La recherche sur le Bleuet. Université Laval, rapport de travail 1970-1971 : 154-156.
	Levitt J. 1980. Responses of plants to environmental stress. 2nd edition, Volume I. Chilling, freezing, and high temperature stresses. Academic Press, New York.
	Odneal B.M. 1983. Winter bud injury of grapevine 1981-1982. Fruit Varieties Journal 31 (2): 45-51.
	Pierquet P., Stushnoff C. and M.J. Burke. 1977. Low temperature exotherms in stem and bud tissues of Vitis riparia Michx. J. Amer. Soc. Hort. Sci. 102: 54-55.
	Pierquet P. and C. Stushnoff. 1980. Relationship of low temperature exotherms to cold injury in Vitis riparia Michx. Amer. J. Enol. Vitic. 31:1-6.
	Plocher T.A. and B. Parke. 2001. Northern Winework: Growing grapes and making wine in cold climates. In: Plocher T.A and R.J. Parke (eds). Winework Inc. Minnesota, USA, 178 p.
	Pogosyan K.S. 1975. Physiological characteristics of cold resistance in different varieties and forms of vine. Vitis 13: 287-291.
	Prostitova V.S. 1977. Frost resistance varieties in the Stavrol area. Horticultural Abstracts 47: 461.
	Quamme H.A. 1986. Use of thermal analysis to measure freezing resistance of grape buds. Can. J. Plant Sci. 66:945-952.
	Reisch B.I., Pool R.M., Peterson D.V. and M.H. Martens. 1979. Grape varieties for New York State. New York Food and Life Science. Bulletin N° 80. A Cornell Cooperative Extension Publication.
	Reisch B.I., Goodman R.N., Martens M.H. and N.F. Weeden. 1993. The relationship between Norton and Cynthiana, red wine cultivars derived from Vitis aestivalis. Am. J. Enol. Vitic. 44: 441-444.
	Reisch B.I. and R. Luce 2002. Recent releases and numbered selections from the Geneva grape breeding program. http://www.nysaes.cornell.edu/hort/faculty/reisch/cultivar.html.
	Rekika D., Cousineau J., Levasseur A., Richer C., Fisher H. and S. Khanizadeh. 2003. The use of a freezing bud technique to determine the hardiness of 20 grape genotypes. Small Fruits Review (in press).
	Richer C. and J.A. Rioux. 2001. Evaluation de la tolerance du Thuja occidentalis L. et de cinq cultivars aux conditions climatiques du Nord-Est Canadien. Partie II. Can. J. Plant Sci. 82: 169-175.
	Rioux J.A., Marquis P., Richer C. and M.P. Lamy. 2000. Evaluation de la tolérance du Thuja occidentalis et de huit de ses cultivars aux conditions climatiques du Nord-Est Canadien. Can. J. Plant Sci. 80: 631-637.
	Sakai A. and W. Larcher. 1987. Frost survival of plants. Responses and adaptation to freezing stress. SpringerVerlag, New York, NY.
	Sayed H. 1992. Vineyard site suitability in Ontario. Ontario Grape and Wine Adjustment Program. OMAFRA and Agriculture Canada Publication N.10.92.
	Skorokhod V.A. 1975. The reasons for different frost resistance of vines under soil cover. Sadovodstvo, Vinogradarstvo I vinodelie Moldavii No 8:21-24.
	Spayd S.E., Morris J.R., Ballinger W.E. and D.G. Himelrick. 1989. Maturity standards, harvesting, post harvest handling, and storage. In: Galletta J. and D.V. Himelrick (eds). Prentice Hall, Englewood Cliffs, New Jersey 07632. 504-530.
	Stergio B.G. and G.S. Howell. 1977. Effect of site on cold acclimation and deacclimation of concord grapevines. Am. J. Enol. Vitic. Vol. 28 (1): 43-48.
	Stetsenko V.M. 1978. Winter protection for grapevines. Sadovodstvo No 8:41. Abstracted in Horticultural Abstracts 46 (1976): 388.
	Stushnoff C. and R.A. Hamman. 2002. Freeze Damage and Protection of Horticultural Plants. Colorado AES Projects 2002-2003. http://www.colostate.edu/Depts/AES/projs/278.htm. Vignes du Québec. http://vignesduquebec.com
	Weiser C.J. 1970. Cold resistance and acclimation in woody plants (A review). HortScience 5: 403-408.
	Wolf T.K. and M.K. Cook. 1994. Cold hardiness of dormant buds of grape cultivars: Comparison of thermal analysis and field survival. HortScience 29 (12): 1453-1455.
	Wolf T.K. and M.K. Miller. 2001. Crop yield, quality, and winter injury of 12 red-fruited wine grape cultivars in Northern Virginia. J. Amer. Pom. Soc. 55 (4): 241-250.
	Wolf T.K. and M.K. Warren. 2000. Crop yield, quality, and winter injury of eight wine grape cultivars in Northern Virginia. J. Amer. Pom. Soc. 54 (1): 34-43.
	Vandal J.O. 1986. La culture de la vigne au Québec. À compte d'auteur, Sainte-Foy. 143 p.