The 19 inter-specific red and white wine grapes varieties, listed in Table 1 and present in the field collection of the CRA-VIT (Research Centre for Viticulture) in Conegliano (Treviso, Italy 45˚51'8.92''N, 12˚15'31.53''E)

were replanted in an experimental vineyard located near Motta di Livenza (Treviso, Italy 45˚47'5.91''N, 12˚35'38.90''E).

The hybrids were monitored for 3 years (2004 to 2006) and compared for agronomic parameters and disease resistance with “Pinot gris”, as this is an international cultivar widely cultivated in the area. The site is characterrised by a heavy soil (35% - 40% of clay). The climatic data accumulated for 3 years of study, are reported in Table 2. Approximately 11 - 13 chemical treatments per year against powdery and downy mildew are usually applied because the climate and soil conditions make this site habitat particularly suitable for high-severity fungal diseases. Anyway, to assess the resistance against fungal diseases, the varieties under investigation were treated only four times during the vegetative period, using formulations without copper before flowering and with copper hydroxide after flowering. Canopy and soil management were conducted according to practices commonly adopted in the vineyards of the area. The vine spacing was 3 × 1.7 m (1960 vines/ha), and the vines were trained using the Sylvoz system, with 3 canes of 10 - 12 buds each. The cultivars were grafted on the rootstock Kober 5 BB. The experimental design was replicated using three 15-plant blocks for each hybrid.

The main phenological stages, such as budbreak, bloom and veraison (defined as 50% of plants upon the

Table 1. Inter-specific varieties investigated in the study: colour of berry, origin of varieties and their pedigree are reported.

Table 2. Average values of mean temperature (˚C) and rainfall (mm) in the three years under investigation.

stage) were evaluated and the variation quantified using standard deviation. Sugar accumulation and acid degradation were monitored upon the initiation of veraison (data not shown) to determine the ripening level and the harvest time. The shooting percentage (shoots/buds), potential fruitfulness (bunches/shoots) and real fruitfulness (bunches/buds) were determined for all the varieties. The real fruitfulness of the first four buds of each cane was also determined in 2005. All these parameters were compared among the hybrids and to “Pinot gris”. At harvest, the average cluster weight and yield per vine were determined. The grape quality was analysed by measuring the soluble solids, titratable acids and pH using a digital refractometer (ATAGO PR-101), a manual titrator (Digitrate Pro 500-Jencons) and a Crison pH metre, respectively.

Based on the agronomic and grape qualitative data, 0.3 tons of grapes from only the most promising varieties were harvested yearly and microvinified (Table 3). Single vinifications were performed with each treated grape. The winemaking process was identical for all vinification experiments. The assessment of the overall wine quality was performed using blind tasting.

Taste panelists evaluated the general organoleptic characteristics, such as aroma and flavour intensity, complexity, balance, structure and finesse.

The wines from two of the most widespread varieties cultivated in the area were used as standards for the sensory analysis. “Pinot blanc” was used as a standard for white wines, as it presents more neutral aroma than “Pinot gris”, and “Franconia” was used as a standard for the red wines. Both standard wines were obtained using

Table 3. List of varieties microvinified in the three years under investigation.

• marks up the year of microvinification.

grape from the experimental site and by using the same winemaking process adopted for the hybrids.

The damage caused by the most common diseases associated with North-Eastern Italy climate (Downy Mildew, Botrytis and Black Rot) was assessed at harvest in 2004 and 2005 using visual inspections and evaluating leaves and bunches separately. A total of 150 clusters and 200 leaves sampled from different parts of the grapevines for each replicate were used in the assessment. Infected leaves or bunches were ranked according to a scale: 0 = no symptoms; 1 = 1% - 10%; 2 = 10% - 25%; 3 = 25% - 50%; 4 = 50% - 75%; and 5 = 75% - 100% of infected leaf areas or infected berries per bunch, respectively. The data regarding disease severity were processed according to Townsend-Heuberger formula [26] to calculate the percentage of infection (I%):

where I% = percentage of infection; n = number of leaves or bunches in each class; v = class value; z = highest class value; N = total amount of assessed leaves or bunches. The data were statistically analysed using ANOVA and Duncan’s multiple range test, using the statistical package “Statistica 8.1”, (StatSoft Inc., 2007). The average, standard deviation and P were calculated.

The full phenological expressions of the hybrids demonstrated their adequate adaptation to the North Italian area. The mean dates for the three years, relating to the phenological stages, compared with “Pinot gris”, are reported in Figure 1. In the North-East Italian areas,

“Pinot gris” is considered as an early wine grape variety, registering budbreak in the first/second decade of April, flowering in the first 10 days of June, and veraison in the last decade of July. The phenological data showed a high variability among genotypes in the budbreak phase (Figure 1(a)), reflecting a higher and different sensitivity to spring weather compared with “Pinot gris”. The budding mean date of “SV 39639”, “GA 52-42”, “GF 64- 170-1”, “Ambror” and “Seibel 5178” occurred in the second decade of April. “Ambror” and “Seibel 5178” were the most stable varieties observed during budbreak. The earliest cultivars were “A × GM 64-94-5”, “Saphira”, “GA 48-12” and “Sirius”. On average, the blooming, veraison and harvest dates for these varieties all occurred later than observed with “Pinot gris”. At flowering, the delay was particularly evident for “Seyve-Villard 12390” and “GM 723-4”, which flowered 9 - 10 days after “Pinot gris”. Regarding the veraison date (Figure 1(b)), most varieties registered a 2-week delay compared with Pinot gris, occurring at the end of July and the beginning of August. Almost all hybrids exhibited a harvest time within the first two weeks of September (approximately 7 days after “Pinot gris”). “Staufer”, “Phoenix”, “Regent”, and “Seibel 7052” exhibited the earliest, completing the growing season at the end of August, together with “Pinot gris”, “Seyve-Villard 12390” was the only late variety, completing berry ripening approximately one month after “Pinot gris”. Thus, this hybrid was not suitable for the investigated area, as long phenological cycles might represent a limiting factor due to autumnal frosts and rainfall that could inhibit berry maturation. The hybrids showed a longer budbreak-veraison period (107 days compared with 97 for “Pinot gris”) and a quite short veraison-harvest phase (35 days compared with the 39 days observed for “Pinot gris”). The only exceptions were “Regent”, which exhibited late budding and precocious harvesting, and “Seibel 7052”, which was consistently precocious for all parameters tested. Shooting percentage, potential and real fruitfulness values are reported in Table 4. The shooting percentage was lower for the hybrids compared with “Pinot gris” (84.7% vs. 92.5%, respectively). In contrast, the values of fruitfulness were similar to “Pinot gris”. It was evident that no fruitfulness decrease was directly associated with the use of hybrids. However, “Seibel 7052”, “GM 7743-8” (Figure 2(a)) and “SV 39639” (Figure 2(b)) exhibited high fruitfulness potential, which frequently presented three bunches per shoot. Considering real and potential fruitfulness, “Seibel 7052”, “GM 7743-8” and “Seyve-Villard 39639” exhibited better performances than the V. vinifera cultivar, while “Seyve-Villard 12390” and “GA 52-42” exhibited the worst performance. Alaa Al-Joumayly [27] reported that fertility coefficients are genetically determined and only slightly conditioned according to season. Notably, the fruitfulness of the first basal buds is a relevant aspect to consider the manual and mechanical spur pruning [28, 29]. “Seibel 7052”, “Seyve-Villard 39639” and “GA 52- 42” were unsuitable for these methods of pruning because of their low fruitfulness in the first 4 buds (Table 4). The standard deviation showed a high variation among different years, particularly for budbreak percentage.

Cluster weight (g) and yield values (kg/vine) of these cultivars showed an opposite trend compared with “Pinot gris”, as cluster weight was, on average, lower (137 g) than that of “Pinot gris” (167 g; Table 5). “Seyve-Villard 12390” and “Villard Blanc” showed the highest cluster weight; however, “Seibel 5178”, “Phoenix”, “GF 138-3”, “GA 52-42” and “Seyve-Villard 39639” recorded the lowest values at nearly 100 g. The yield (tons per hectare)

Table 4. Budbrake (%), real fruitfulness, real fruitfulness of first 4 basal buds and potential fruitfulness of the 19 hybrids compared with “Pinot gris”. The data reported are average of the data registered in 2004, 2005 and 2006, and their variability was measured as standard deviations.

Note: Letters represents the significance of variability among the hybrids with p < 0.05 (Duncan test).

Table 5. Cluster weight and yield per vine of the 19 hybrids and “Pinot gris”. The data reported are average of the data registered in the three years, and their variability was measured as standard deviations.

Note: Letters represents the significance of variability among the hybrids with p < 0.05 (Duncan test).

was significantly high, with over 20 t/h obtained for the most productive hybrids and 4 - 5 t/ha obtained for the least productive varieties. On average, the hybrids produced more (8.1 kg/vine) than “Pinot gris” (7.3 kg/vine), suggesting fruitfulness, as reported by previous studies [30,31]. The most productive varieties were “Sirius”, “Regent”, “Seibel 7052”, “Orion”, “GA 48-12” and “GM 7743-8”, all presenting a yield greater 9 Kg/vine. Overall, we observed high variability among the different genotypes. Over the three-year period, the “Regent” and “Ambror” registered the highest variation. The soluble solids, titratable acids and pH values were measured to assess the grape quality [32]. A high variability was observed among the hybrids (Figure 3), which reflect differences in their genotypes and the annual climate conditions, consistent with the observations of Pavloušek and Kumšta [33], showing the strong effect of the year and cultivar on the pH values and titratable acid contents. “Ambror” showed the highest soluble solid values (20.1˚ Brix), followed by “Regent”, “Seibel 7052” and “GF 138- 3”, which all exhibited values superior to “Pinot gris”, “Seibel 5178” and “Saphira” were similar to “Pinot gris”

(18.1˚ Brix), while “Staufer” recorded the lowest value (14.8˚ Brix). A comparison of the yield per vine (Table 5) to the sugar richness (Figure 3) revealed that “Regent” and “Seibel 7052” had good production and high-soluble solids; no negative regression was observed among the hybrids between the two parameters. The estimated titratable acids values were primarily high for most hybrids (Figure 3(b)), which also reflected moderately low pH values (Figure 3(c)). “Seibel 7052”, “Phoenix”, “Sirius” and “GF 138-3” exhibited acids levels similar to “Pinot gris” (6 - 7 g/L). High acid values indicate a positive trait in the present global warming scenario and also represent desirable characteristics, particularly for sparkling wines. The increase of air temperature corre-

sponded with a similar increase in pH values in must. As Keller [34] suggested, the pH values of musts should not exceed 3.6 to avoid a decrease of quality of wine produced. As shown in Figure 3(c), all hybrids presented pH values below this limit.

Regarding the wine characteristics and quality, in 2004 no white wines obtained from these hybrids performed as well as the reference commercial variety (“Pinot blanc”; Figure 4(a)). The quality of “Phoenix”, “GA 52-42” and “Ambror” were among the best reported, mostly for the olfactory intensity with a pronounced aroma and an adequate complexity. These qualities indicated low finesse, which is a characteristic that was confirmed in the 2005 and 2006 tastings (Figures 4(c) and (e)). Overall, “GA 48-12” demonstrated the best performance (not included in the 2004 tasting), exhibiting scores similar to “Pinot blanc”. Among the remaining white varieties, “Ambror” achieved interesting results, particularly for olfactory intensity and taste persistence.

Concerning red wines obtained from hybrids, the quality was lower than the reference variety wine (“Franconia”) in all three years (Figures 4(b), (d) and (f)). In 2004 (Figure 4(b)), the black wines did not exhibit high parameters associated with the bouquet or taste, except for “GF138-3”, which showed appreciated values for the olfactory aromatic intensity. “Regent” exhibited increased values from the first to the third year. In 2006 (Figure 4(f)), the lowest yield (8.1 Kg/vine) and sensory characteristics for “Regent” were similar to “Franconia”, but the olfactory value was low.

The percentage of Downy Mildew, Botrytis and Black rot infection is reported in Table 6. During 2004 the rainfall and temperatures were standard for the area; hence, the antifungal treatments sufficiently limited disease damage. In 2005, high rainfall occurred during the blooming and berry-set periods, when the vines are more susceptible to Downey Mildew infections, and at the beginning of September, when more severe Botrytis infections occur. In the second year of the trial, “Saphira” and “Seibel 7052” were more susceptible to Downy Mildew than “Pinot gris”, while “SV 39639”, “SV 12390” and “Villard Blanc” were the most tolerant for leaf and cluster, exhibiting negligible symptoms of infection (Table 6). However, “Villard Blanc” showed the worst results for black rot, and “Phoenix” exhibited the highest percentage of infection for Botrytis (Table 6). Although each of the hybrids showed a high resistance to at least one of the diseases evaluated, it was not possible to identify a hybrid with a low percentage of infection for all three diseases. Even though “GF 138-3” showed a low disease incidence, “SV 12-390” resulted the most tolerant among the red wine cultivars. Whilst, three white wine cultivars, “GA 48-12”, “Ambror” and “Regent”, resulted adequately tolerant to Downy Mildew, consistent with the data reported in Germany [35] and in Poland [3].