Introduction

In Tunisia, viticulture is very ancient and the first historical record of grapevines dates back to 6000 BC (Zohary and Hopf, 2000). Autochthonous grapes are often grown by poor farmers in marginal, low-input and drought-stressed environments. These genetic resources, which may represent valuable reservoirs of interesting genes for crop improvement such as adaptation to biotic and abiotic stresses (Brush, 1995), represent a small population with a high risk of extinction due to the introduction of commercial high-yield foreign varieties (Hjalmarsson and Ortiz, 2000). All these factors contribute to the need for a detailed description and evaluation of the Tunisian grapevine genetic resources.

Initial efforts to identify genetic diversity in Tunisian grapevine cultivars were mainly based on molecular tools (Zoghlami et al., 2001; 2009). Despite the use of molecular markers, knowledge of the phenotype given by morphological and agronomical descriptors is still important for breeding programs, conservation and commercialization of new varieties (Franco et al., 2005; Gonçalves et al., 2008; Laurentin, 2009).

The description of the morphological characteristics is the usual methodology accepted from a legal point of view for patenting and registration of varieties (Badenes, 1991). In fact, ampelography is the first step in grapevine identification and selection and for resolving different classification problems (Martinez de Toda and Sancha, 1997). The complete characterization as well as the conservation of autochthonous cultivars is of great importance to prevent the loss of diversity (Rodrigues et al., 2008).

In the last years, morphological data have been used to resolve the complex problem of the definition and classification of crop accessions using multivariate statistical analyses such as principal component analysis (PCA) and hierarchical cluster analysis (HCA) (Manjunatha et al., 2007; Aghaei et al., 2008). These methods assist in the successful management of plant genetic resources and contribute to the determination of the agronomical value of the accessions in the germplasm collection (bank).

Morphological characterization continues to be the first step for the description and classification of germplasm accessions and statistical methods like PCA are useful tools for screening the accessions of a collection (Badenes et al., 2000). PCA transforms the original variables into a limited number of uncorrelated new variables. As reported by Martínez-Calvo et al. (2008), this method allows the visualization of differences among individuals, the identification of groups, and the identification of relationships among individuals and variables.

Keeping in mind the importance of morphological characteristics in varietal identification and registration, the characterization of the worldwide Vitis materials has been homogenized by the use of standardized OIV/IPGRI descriptors (Soylemezoglu et al., 2001; Santiago et al., 2007). These descriptors allow the inventory of the grapevine genetic resources with passport, primary and secondary descriptors, bibliography and photos.

In this study we provide, for the first time, an inventory of the Tunisian autochthonous grapevine genetic resources using 70 major OIV descriptors (OIV, 2007). By applying a multivariate analysis to these data, we characterized the autochthonous grapevine cultivars from the Tunisian national germplasm collection, estimated the phenotypic diversity among and within the cultivars, and identified the traits that contributed to cultivar diversity.

Materials and methods

1. Plant material

Sixty-one Tunisian autochthonous grapevine genotypes were included in this study (Table 1). They were collected from different parts of the country and are kept at the repository of the Centre of Biotechnology of Borj Cédria, Tunisia (CBBC).

2. Data collection

Seventy major morphological traits, selected from the OIV list (OIV, 2007) and comprising shoot (12 descriptors), leaf (39 descriptors) and fruit (19 descriptors) descriptors, were measured and used to design a numbered-data matrix (Table 2). Ten specimens per accession were individually evaluated for their morphological diversity across the 70 descriptors. Measurements were performed by the same two persons to avoid errors due to individual variation.

3. Statistical analysis based on morphological diversity using shoot, leaf and fruit descriptors

Morphological data were analyzed by multivariate analysis, clustering and ANOVA analysis using XLSTAT software (Addinsoft, www.xlstat.com); PCA was performed to identify accession groups and to determine the axes and the characters significantly contributing to the variation. In this procedure, the similarity matrix was used to generate eigenvalues and scores for the accessions. The first two principal components, which accounted for the highest variation, were then used to plot two-dimensional scatter plots. HCA was carried out using Ward’s minimum variance method as a clustering algorithm (Williams, 1976) and squared Euclidean distances as a measure of dissimilarity (Ward, 1963).

Among the groups identified based on different classifications (shoot, leaf and fruit), we determined the cultivars corresponding to the barycentre, which is a geometrical measurement allowing the concentration of a set of cultivars onto one that best expresses the inherent morphological diversity per variation class.

4. Evidence for integrating shoot, leaf and fruit descriptors

Regression analysis was applied to set up evidence for integrating shoot, leaf and fruit descriptors in distinguishing between autochthonous grape cultivars. Thus, the correlation between Euclidean distances calculated from shoot-leaf, shoot-berry and leaf-berry descriptors was determined from the linear regression plot (Rousset, 1997). A positive correlation is expressed by a positive R² value. The significance of this correlation was tested by a Mantel test using 10000 permutations (Mantel and Valand, 1970).

5. Perspectives for the commercialization of Tunisian autochthonous cultivars: table or wine cultivars?

To assess the vocation of the Tunisian grapevine accessions as table or wine cultivars, the following characters were determined according to the OIV descriptor list (Table 2): bunch weight, 100-berry weight, and sugar content, pH and acidity of the must.

Results

1. Morphological diversity based on individual characteristics

1.1 Shoot characteristics

Twelve descriptors were used for the characterization of the shoots. The discrimination between the accessions under investigation revealed that 53.76% of the variation (Table 3) was explained by the first three axes of the PCA plot (Fig. 1a): the first axis was defined by the distribution and the intensity of the anthocyanin coloration on the bud scales (OIV 015 and OIV 015-2, respectively). These two descriptors were strongly correlated (0.940). The second axis was defined by both the anthocyanin coloration of the tip (OIV 003) and the form of the tip (OIV 001).

Based on shoot descriptors, the accessions clustered into three main groups (C1, C2 and C3) as revealed in the dendrogram (Fig. 1b). Group C1 was the largest and comprised the 31 accessions with the highest density of prostrate hairs on the tip (OIV 004), as inferred from the morphological data matrix. Group C2 contained the 16 accessions with the highest distribution and intensity of anthocyanin coloration on the bud scales (OIV 015 and OIV 15-2, respectively). Group C3 included 14 accessions that are grouped by the form of the tip (OIV 001), which varied from “half open to open”.

Save View full size Expand inline Collapse inline

Figure 1. Shoot descriptor analysis. (a) Principal component analysis (PCA) plotted along the first two axes and (b) UPGMA dendrogram of Euclidean distance illustrating the genetic relationships among the 61 studied grapevine cultivars based on shoot descriptors. Sd1, Sd2 and Sd3: barycentres of the groups.

The variance components within and between the individual groups (C1, C2 and C3) detected with ANOVA showed that most of the morphological variation was partitioned within (62.98%) rather than between (37.02%) the groups. These were both significant at p<0.01 (Table 4). Per variation class, the accessions Asli Hadab, Khalt Bouchemma Gabès and Khalt Abiadh (Sd1, Sd2 and Sd3, respectively) were identified as the barycentres of the groups C1, C2 and C3, respectively.

1.2 Leaf characteristics

Thirty-nine OIV descriptors were used for the characterization of the leaves (Table 2). According to the PCA plot (Fig. 2a), the first three principal components accounted for 28.61% of the total variation (Table 3). The variables with the greatest weight in the first principal component were the opening (OIV 079-1) and the general shape (OIV 079) of the petiolar sinus and the angle between veins N2-N3 at the first ramification (OIV 608). The second axis was defined by three characters: the length from the petiolar sinus to the upper leaf sinus (OIV 605), the length from the petiolar sinus to the lower leaf sinus (OIV 606) and the shape of the blade (OIV 067). The characters OIV 079-1 and OIV 079 were the most correlated descriptors (0.738%). Based on leaf descriptors, the 61 studied cultivars grouped into three major groups: C1’ (19 genotypes), C2’ (39 genotypes) and C3’ (3 genotypes), as illustrated in the dendrogram (Fig. 2b).

Save View full size Expand inline Collapse inline

Figure 2. Leaf descriptor analysis. (a) Principal component analysis (PCA) plotted along the first two axes and (b) UPGMA dendrogram of Euclidean distance illustrating the genetic relationships among the 61 studied grapevine cultivars based on leaf descriptors. Ld1, Ld2 and Ld3: barycentres of the groups.

The variance components within and between the individual groups (C1’, C2’ and C3’) detected with ANOVA showed that most of the morphological variation was partitioned within (85.58%) rather than between (14.42%) the identified groups (Table 4). Per variation class, the accessions Khalt Bouchemma Gabès, Arich Djerba and Chaouche Djerba (Ld1, Ld2 and Ld3, respectively) were identified as the barycentres of the groups C1’, C2’ and C3’, respectively.

Further observations of leaf descriptors showed that the shape of the blade was either wedge-shaped or pentagonal (OIV 067) with open petiole sinus (OIV 079-1). From a disease sensitivity point views, the leaves of the accessions Châaraoui, Khamri Tozeur and Arich Dressé exhibited high resistance to both Oidium (OIV 455) and Plasmopara (OIV 452), while the accessions Khédhiri 1, Sakasly Baddar, Muscat Rafraf, Akhal Mguergueb, Bidh el Hamem Rafraf and Bezzoul el Khadem Rafraf displayed lower resistance to Plasmopara but high resistance to Oidium.

1.3 Fruit characteristics

Nineteen ampelographic descriptors were used for the description of the fruits. The discrimination between all cultivars revealed that the first three axes of the PCA plot explained 37.32% of the variation (Table 3). The highest loadings on the first PCA axis corresponded to berry length (OIV 220) and bunch density (OIV 204) (Fig. 3a). The variables with the highest loadings on the second PCA axis were color of the berry flesh (OIV 230) and total acid content of the must (OIV 506). The color of the berry skin (OIV 225) and color of the berry flesh (OIV 230) were the most correlated characters. Three major groups were determined using cluster analysis: group C1” (29 genotypes), C2” (14 genotypes) and C3” (18 genotypes) (Fig. 3b).

Save View full size Expand inline Collapse inline

Figure 3. Fruit descriptor analysis. (a) Principal component analysis (PCA) plotted along the first two axes and (b) UPGMA dendrogram of Euclidean distance illustrating the genetic relationships among the 61 studied grapevine cultivars based on fruit descriptors. Fd1, Fd2 and Fd3: barycentres of the groups.

The variance components within and between the individual groups (C1”, C2” and C3”) detected with ANOVA showed that most of the morphological variation was partitioned within (79.12%) rather than between (20.88%) the groups (Table 4). Per variation class, the accessions Blanc 3, Arbia and Balta 2 (Fd1, Fd2 and Fd3, respectively) were identified as the barycentres of the groups C1”, C2” and C3”, respectively.

The bunch density character (OIV 204) varied from loose to dense and the cluster analysis showed that the accessions with the lowest bunch density were grouped together. Across all accessions, whatever the size or the shape of the bunch or the berry, the weight of a bunch did not exceed 500 g and the weight of a single berry ranged between 2 and 5 g. The sugar content of the must (OIV 505) varied from low (Saouadi, Amokrane, Arbia, Djebbi) to high (Asli Hadab, Asli Dar Slimane, Beldi Rafraf, Beldi Sayeb, Khédhiri 1, Tounsi Djerba, Khédhiri 2, Turky). A low total acid content of the must (OIV 506) characterized all the studied accessions and most of the accessions displayed low ph (OIV 508). Considerable variation was observed for the fruit shape (OIV 223). The intensity of the skin color (OIV 225) was quite diverse, ranging from very light green-yellow to dark red-violet, with medium to firm flesh (OIV 235).

2. Statistical evidence for integrating all morphological data in discriminating between Tunisian autochthonous grapevines

Different dendrograms were obtained and different relationships between the accesssions were observed when based upon shoot, leaf or berry criteria (Fig. 1b; 2b; 3b). Therefore, we performed statistical analysis to see whether or not all morphological data ought to be integrated in the characterization of Tunisian grapes. Thus, Euclidean distances calculated for shoot/berry, shoot/leaf and leaf/berry were compared by Mantel tests (Rousset, 1997) using 10000 permutations (XLSTAT software). As shown in Figure 4, low correlation values were obtained between shoot and berry descriptors (r=0.023) and between leaf and berry descriptors (r=0.027). However, a higher significant correlation value was obtained between shoot and leaf descriptors (r=0.125). Based on these results (Table 5), all descriptors ought to be integrated in clustering Tunisian autochthonous grapevines.

Save View full size Expand inline Collapse inline

Figure 4. Linear regression between shoot/leaf, shoot/fruit and leaf/fruit Euclidean distances in 61 grapevine genotypes.

3. Morphological diversity based on total characters

The use of all the 70 ampelographic descriptors (Table 2) yielded a high number of morphotypes and permitted the discrimination of all cultivars. High morphological variation was recorded among the studied accessions. The majority of the descriptors were significantly correlated, though with very heterogeneous values for the coefficients of correlation.

The first three PCA axes accounted for 21.78% of the total variation (Table 3; Fig. 5a). Besides, 12 descriptors out of 70 were identified as the most useful morphological descriptors for the classification of the accessions. These were the following: the shape of petiolar sinus (OIV 079), the opening/overlapping of the petiolar sinus (OIV 079-1), the density of prostrate hairs between the main veins (lower side) (OIV 084), the angle between N2 and N3 measured at the first ramification (OIV 608), the intensity and distribution of anthocyanin coloration on the bud scales (OIV 015-2 and OIV 015, respectively), the intensity of berry skin color (OIV 225), the width of teeth N2 of mature leaf (OIV 613), the sugar content of must (OIV 505), the anthocyanin coloration of shoot tip (OIV 003), the density of prostrate hairs on shoot tip (OIV 004) and the sex of flower (OIV 151) (Table 2).

Save View full size Expand inline Collapse inline

Figure 5. Total descriptor analysis. (a) Principal component analysis (PCA) plotted along the first two axes and (b) UPGMA dendrogram of Euclidean distance illustrating the genetic relationships among the 61 studied grapevine cultivars based on total descriptors. Td1, Td2 and Td3: barycentres of the groups.

Using all descriptors, three major groups were identified by the cluster analysis: groups C1”’ (14 accessions), C2”’ (36 accessions) and C3”’ (11 accessions) (Fig. 5b). The variance components within and between the individual groups (C1”’, C2”’ and C3”’) detected with ANOVA showed that most of the morphological variation was partitioned within (89.31%) rather than between (10.69%) the groups (Table 4). These were both significant at p<0.01. Per variation class, the accessions Khalt Bouchemma Gabès, Blanc 3 and Blanc 2 (Td1, Td2 and Td3, respectively) were identified as the barycentres of the groups C1”’, C2”’ and C3”’, respectively. Consequently, these accessions are representative of all the morphological variation within autochthonous grapes in Tunisia.

4. Perspectives for the commercialization of Tunisian autochthonous cultivars: table or wine cultivars?

The question that was dealt with in this section was how these accessions ought to be introduced into the market: are they wine or table cultivars?

The three genotypes that were identified as the barycentres of the groups (Fig. 5b) were considered to determine the use and analyzed for the following characters: weight of the bunch (g), weight of 100 berries (g), must sugar content (brix), must organic acid composition (titratable acidity; TA) and must pH (Table 6). The brix and TA parameters were the most compulsory criteria that ought to be considered when distinguishing between table and wine grapes. As shown in Table 6, the sugar content of the analyzed accessions (between 14.2 and 17.1 brix) was lower than that usually reported for wine grapes (between 24 and 26 brix) and closer to values reported for table grapes (between 17 and 19 brix)".

Discussion

The present study revealed the morphological diversity within a collection of autochthonous grapevine germplasms grown in different geographical regions of Tunisia using 70 morphological descriptors. These were used to estimate the phenotypic diversity among and within the cultivars, to identify the traits contributing to the heterogeneity, to classify them using PCA and HCA and to see how these accessions ought to be commercialized.

The use of ampelographic descriptors comprising shoot, leaf and fruit traits yielded a high number of morphotypes and permitted the discrimination of all cultivars. This discrimination was found to be higher than those reported previously using isozyme (Ben Abdallah et al., 1998) and molecular markers (Zoghlami et al., 2001; 2009) for the same varieties. In fact, morphological markers assume higher degree of genomic coverage (Vetelӓinen et al., 2005) and most individual phenotypic markers are multigenic. Therefore, variation at more than one locus is being analyzed.

Conversely, the set of OIV descriptors used here allowed for the first time to easily split the Tunisian autochthonous grapevine accessions into wine and table grapes. These findings were not clearly attained when using molecular tools.

Several authors have analyzed morphological diversity in crop plants (Bozokalfa et al., 2009; Jesus et al., 2009; Aghaee et al., 2010; Sarıkamış et al., 2010). In grapevine, morphological investigations have been carried out on Brazilian (Leão et al., 2011), Croatian (Sladonja et al., 2007), Egyptian (Hassan et al., 2011), Georgian (Ekhvaia and Akhalkatsi, 2010), Italian (Muganu et al., 2009; Alba et al., 2011), Portuguese (Cunha et al., 2009), Spanish (Santiago et al., 2007) and Turkish grapevines (Ates et al., 2011).

Multivariate analyses based on morphological characters provide information allowing the breeder to improve populations by selecting from specific geographical regions (Souza and Sorrells, 1991). The multivariate techniques have been applied in viticulture with several objectives: morphological and agronomical descriptions (Coelho et al., 2004; Borges et al., 2008), management (Intrieri et al., 2001) and disease resistance (Nascimento et al., 2006).

Because grapevine shoot is an important character for the description of cultivars, many studies were conducted using such descriptors (Santiago et al., 2007; Zdunic et al., 2008; Barth et al., 2009; Sabir et al., 2009). A thorough knowledge of the quantitative aspects of shoot development is therefore necessary to understand the determination of crop quality and productivity as a function of environment. In our case, the highest amount of variation was attained using shoot descriptors (PCA; 53.76%) (Table 3; Fig.1a). The number of shoot descriptors (n=12) was higher than those reported previously: 4 descriptors (Santiago et al., 2007; Alba et al., 2011), 5 descriptors (Sabir et al., 2009), and 9 descriptors (Muganu et al., 2009).

The tip of the young shoot was open for all the cultivars. This character allows the differentiation among the Vinifera and other Vitis species. Additionally, the coloration of the shoot tip, which defines the second axis of the PCA plot, seems to be an important character in distinguishing between the grapevine cultivars (Morton, 1979). Moreover, this character may vary in relation to exposure to light (Kara, 1990). The accessions Asli Hadab, Khalt Bouchemma Gabès and Khalt Abiadh were established as being the barycentres of the identified groups (Fig. 1b). This means that using shoot descriptors all the morphological variation could be explained by these accessions.

Leaf descriptors have been generally used as powerful tools for identifying grapevine genotypes (Santiago et al., 2007; Celik et al., 2008; Sabir et al., 2009; Gago et al., 2009; Harbi-Ben Slimane et al., 2010). Thirty-nine leaf descriptors were used in a set of 61 autochthonous grapes. This number was higher than those reported previously by Sabir et al. (2009), Santiago et al. (2007) and Gago et al. (2009) using 12, 17 and 19 descriptors, respectively. According to the PCA plot (Fig. 2a), although using 39 descriptors, the first three principal components accounted for only 28.61% of the total variation.

The discrimination between all grapevines, based on fruit descriptors, revealed that the first three axes of the PCA plot explained 37.32% of the variation using only 19 ampelographic descriptors (Fig. 3a). The number of descriptors was higher than those used by Ates et al., 2011 (6 descriptors), Alba et al., 2011 (11 descriptors) and Leão et al., 2011 (12 descriptors). It was observed that all accessions had low acidity values, which is one characteristic of cultivated grapevine (Navarro et al., 2001; Liu et al., 2006).

Based on the differences in clustering that occurred with shoot, leaf or berry descriptors (Fig. 1b; 2b; 3b) and the correlation values registred between shoot and berry (r=0.023), leaf and berry (r=0.027) and shoot and leaf descriptors (r=0.125) (Table 5), all descriptors ought to be integrated in clustering for the discrimination of the Tunisian autochthonous cultivars. Significant correlations were detected between all descriptors (Fig. 4), confirming the results of Ocampo et al. (2006), who found positive correlations between all morphological traits.

Therefore, a total number of 70 major OIV descriptors has been used to discriminate among autochthonous grapes. This number appears to be high if compared to previous studies in grapevine cultivars: Leão et al., 2011 (12 descriptors); Sabir et al., 2009 (17 descriptors); Cunha et al., 2009 (22 descriptors); Alba et al., 2011 (30 descriptors); Muganu et al., 2009 (34 descriptors); Ekhvaia and Akhalkatsi, 2010 (43 descriptors); Sladonja et al., 2007 (50 descriptors); and Celik et al., 2008 (61 descriptors).

As inferred from the PCA plot (Fig. 5a), when using the 70 descriptors to draw a final cluster scheme, the eigenvalues of the first, second and third axis of the principal components accounted for 8.65, 15.37 and 21.78% of the total variation, respectively (Table 3).

The relative magnitude of the first three PCA eigenvectors showed that 12 descriptors out of 70 were identified as the most important morphological descriptors for the classification of the accessions. These were OIV 079, OIV 079-1, OIV 084, OIV 608, OIV 015-2, OIV 015, OIV 225, OIV 613, OIV 505, OIV 003, OIV 004, and OIV 151 (Table 2). These descriptors ranked within the primary descriptor list established by the OIV for the characterization of cultivars (OIV, 2012). Iezzoni and Pritts (1991) mentioned that associations between traits uncovered by PCA may correspond to a genetic linkage between loci controlling traits or a pleiotropic effect.

To access the grapevine fresh fruit market, we must characterize the cultivar as table or wine grapes. The overall flavor is one of the most important qualities for establishing a continuous consumer preference. Flavor composition has been defined as a complex quality attribute, in which the mix of sugars, acids, and volatiles plays a primary role (Baldwin, 2002). Among the flavor metabolites already mentioned, sugar and organic acid compositions, which are measured through total soluble solids (TSS) and titratable acidity (TA), are most commonly associated with the taste of fruits, including table grapes (Shiraishi et al., 2010). In our case, the accessions Khalt Bouchemma Gabès, Blanc 2 and Blanc 3 were identified as table grapes (Table 6), according to their flavor composition and bunch weight, which meet the market requirements for table grapes previously set up by the United Nations Economic Commission for Europe (UNECE) (Norme CEE-ONU FEV-19, 2010). These indicate that table grapes must have a refractive index of at least 16 to 19 brix and a minimum bunch weight of 75 g. Based on the aforementioned statements, Khalt Bouchemma Gabès possess the required characteristics to be introduced into the market as table grapes. Their overall flavor composition that is associated with their taste (Baldwin, 2002; Shiraishi et al., 2010) may therefore strongly satisfy consumer preference.

Conclusion

The detailed ampelographic description presented in this study highlights clear morphological differentiation between 61 Tunisian autochthonous grapevines characterized for the first time using 70 OIV descriptors.

Morphological traits were shown to enable cultivar comparison and classification in germplasm collection. Numerical analyses showed that the number of morphological traits that are effectively contributing to the characterization of the cultivars could be reduced to 12. Furthermore, three cultivars out of 61 were found to represent the morphological variability observed in Tunisian grapes, representing the barycentres of the individual groups in cluster analysis.

Finally, the autochthonous Tunisian grapes ought to be classified as table grapes according to the standards of the UNECE (Normes CEE-ONU FEV-19, 2010).

Acknowledgements: The authors are thankful to the Ministry of Higher Education and Scientific Research (Tunisia) for financial support.

References

• Aghaee M., Mohammadi R., Nabovati S., 2010. Agro-morphological characterization of durum wheat accessions using pattern analysis. Aust. J. Crop Sci. 4, 505-514.
• Aghaei M.J., Mozafari J., Taleei A.R., Naghavi M.R., Omidi M., 2008. Distribution and diversity of Aegilops tauschii in Iran. Genet. Resour. Crop Evol. 55, 341-349. doi:10.1007/s10722-007-9239-0
• Alba V., Anaclerio A., Gasparro M., Caputo A.R., Montemurro C., Blanco A., Antonacci D., 2011. Ampelographic and molecular characterisation of Aglianico accessions (Vitis vinifera L.) collected in Southern Italy. S. Afr. J. Enol. Vitic. 32, 164-173.
• Ates F., Coban H., Kara Z., Sabir A., 2011. Ampelographic characterization of some grape cultivars (Vitis vinifera L.) grown in south-western region of Turkey. Bulg. J. Agric. Sci. 17, 314-324.
• Badenes M.L., 1991. Caracterización e identificación de variedades de albaricoquero por métodos pomológicos y bioquímicos. Ph.D. Thesis, Universidad Politécnica de Valencia, Valencia (Spain).
• Badenes M.L., Martínez-Calvo J., Llácer G., 2000. Analysis of a germplasm collection of loquat (Eriobotrya japonica Lindl.). Euphytica 114, 187-194. doi:10.1023/A:1003950215426
• Baldwin E.A., 2002. Fruit flavor, volatile metabolism and consumer perceptions. In: Fruit Quality and its Biological Basis, Knee M. (ed.). Sheffield Academic Press, Sheffield, UK., pp. 89-106.
• Barth S., Forneck A., Verzeletti F., Blaich R., Schumann F., 2009. Genotypes and phenotypes of an ex situ Vitis vinifera ssp. sylvestris (Gmel.) Beger germplasm collection from the Upper Rhine Valley. Genet. Resour. Crop Evol. 56, 1171-1181. doi:10.1007/s10722-009-9443-1
• Ben Abdallah F., Chibani F., Fnayou A., Ghorbel A., Boursiquot J.M., 1998. Caractérisation biochimique des variétés tunisiennes de vigne. J. Int. Sci. Vigne Vin 32, 17-25.
• Borges R.M.E., Gonçalves N.P.S., Gomes A.P.O., Alves E.O.S., 2008. Divergência fenotípica entre acessos de uvas de mesa no semi-árido brasileiro. Pesq. Agropec. Bras. 43, 1025-1030. doi:10.1590/S0100-204X2008000800011
• Bozokalfa M.K., Eşiyok D., Turhan K., 2009. Patterns of phenotypic variation in a germplasm collection of pepper (Capsicum annuum L.) from Turkey. Span. J. Agric. Res. 7, 83-95. doi:10.5424/sjar/2009071-401
• Brush S.B., 1995. In situ conservation of landraces in centers of crop diversity. Crop Sci. 35, 346-354. doi:10.2135/cropsci1995.0011183X003500020009x
• Celik H., Kose B., Cangi R., 2008. Determination of Fox grape genotypes (Vitis labrusca L.) grown in Northeastern Anatolia. Hort. Sci. 35, 162-170.
• Coelho I., Cunha J., Cunha J.P., Carneiro L.C., Castro R., Eira Dias J.E., 2004. Ampelometric comparison of wild vine Vitis vinifera L. populations and old grapevine cultivars of the south of Portugal. Ciência Téc. Vitiv. 19, 1-12.
• Cunha J., Teixeira-Santos M., Carneiro L.C., Fevereiro P., Eiras-Dias J.E., 2009. Portuguese traditional grapevine cultivars and wild vines (Vitis vinifera L.) share morphological and genetic traits. Genet. Resour. Crop Evol. 56, 975-989. doi:10.1007/s10722-009-9416-4
• Ekhvaia J., Akhalkatsi M., 2010. Morphological variation and relationships of Georgian populations of Vitis vinifera L. subsp. sylvestris (C.C. Gmel.) Hegi. Flora 205, 608-617. doi:10.1016/j.flora.2009.08.002
• Franco J., Crossa J., Taba S., Shands H., 2005. A sampling strategy for conserving genetic diversity when forming core subsets. Crop Sci. 45, 1035-1044. doi:10.2135/cropsci2004.0292
• Gago P., Santiago J.L., Boso S., Alonso-Villaverde V., Grando M.S., Martinez M.C., 2009. Biodiversity and characterization of twenty-two Vitis vinifera L. cultivars in the Northwestern Iberian Peninsula. Am. J. Enol. Vitic. 60, 293-301.
• GENRES-081, 2001. Primary and Secondary Descriptors List for Grapevine Cultivars and Species (Vitis L.). Institut für Rebenzüchtung Geilweilerhof, Siebeldingen.
• Gonçalves L.S., Rodrigues R., Amaral A.T. Jr., Karasawa M., Sudré C.P., 2008. Comparison of multivariate statistical algorithms to cluster tomato heirloom accessions. Genet. Mol. Res. 7, 1289-1297. doi:10.4238/vol7-4gmr526
• Harbi-Ben Slimane M., snoussi h., bouhlal r., nahdi h., 2010. ampelometry to test for genetic diversity in Tunisian Vitis sylvestris. Afr. J. Plant Sci. Biotechnol. 4, 17-22.
• Hassan N.A., El-Homosany Gomma A.H., Shaheen M.A., 2011. Morphological and Issr polymorphisms in some Egyptian grapes (Vitis vinefera L.) collection [sic]. World Appl. Sci. J. 15, 1369-1375.
• Hjalmarsson I., Ortiz R., 2000. In situ and ex situ assessment of morphological and fruit variation in Scandinavian sweet cherry. Sci. Hortic. 85, 37-49. doi:10.1016/S0304-4238(99)00123-5
• Iezzoni A.F., Pritts M.P., 1991. Applications of principal component analysis to horticultural research. HortScience 26, 334-338.
• Intrieri C., Silvestroni O., Filippetti I., Bucchetti B., 2001. Assessment of a cv. Sangiovese clone’s morphological variability in differing environmental and management conditions. Adv. Hortic. Sci. 15, 79-84.
• Jesus O.N., Ferreira C.F., Silva S.O., Câmara T.R., Soares T.L., Pestana K.N., 2009. Characterization of recommended banana cultivars using morphological and molecular descriptors. Crop Breed. Appl. Biotechnol. 9, 164-173. doi:10.12702/1984-7033.v09n02a09
• Kara Z., 1990. Determination of the ampelographic characters of grape varieties grown in Tokat. Ph.D. Thesis, Ankara University, Ankara (Turkey).
• Laurentin H., 2009. Data analysis for molecular characterization of plant genetic resources. Genet. Resour. Crop Evol. 56, 277-292. doi:10.1007/s10722-008-9397-8
• Leão P.C.S., Cruz C.D., Motoike S.Y., 2011. Genetic diversity of table grape based on morphoagronomic traits. Sci. Agric. 68, 42-49. doi:10.1590/S0103-90162011000100007
• Liu H.F., Wu B.H., Fan P.G., Li S.H., Li L.S., 2006. Sugar and acid concentrations in 98 grape cultivars analyzed by principal component analysis. J. Sci. Food Agric. 86, 1526-1536. doi:10.1002/jsfa.2541
• Manjunatha T., Bisht I.S., Bhat K.V., Singh B.P., 2007. Genetic diversity in barley (Hordeum vulgare L. ssp. vulgare) landraces from Uttaranchal Himalaya of India. Genet. Resour. Crop Evol. 54, 55-65. doi:10.1007/s10722-005-1884-6
• Mantel N., Valand R.S., 1970. A technique of nonparametric multivariate analysis. Biometrics 26, 547-558. doi:10.2307/2529108
• Martínez-Calvo J., Gisbert A.D., Alamar M.C., Hernandorena R., Romero C., Llácer G., Badenes M.L., 2008. Study of a germplasm collection of loquat (Eriobotrya japonica Lindl.) by multivariate analysis. Genet. Resour. Crop Evol. 55, 695-703. doi:10.1007/s10722-007-9276-8
• Martinez de Toda F., Sancha J.C., 1997. Ampelographical characterisation of red Vitis vinifera L. cultivars preserved in Rioja. Bull. OIV 70, 220-234.
• Morton L.T., 1979. A Practical Ampelography: Grapevine Identification (translated and adapted from P. Galet). Ithaca, London, Cornell University Press.
• Muganu M., Dangl G., Aradhya M., Frediani M., Scossa A., Stover Ed., 2009. Ampelographic and DNA characterization of local grapevine accessions of the Tuscia area (Latium, Italy). Am. J. Enol. Vitic. 60, 110-115.
• Nascimento A.R.P., Michereff S.J., Mariano R.L.R., Viana I.O., 2006. Reaction of grapevine clones to Xanthomonas campestris pv. viticola, based on epidemiological components of bacterial canker. Cienc. Rural 36, 1-7. doi:10.1590/S0103-84782006000100001
• Navarro M., Retamales J., Defilippi B., 2001. Efecto del arreglo de racimo y aplicación de CPPU en la calidad de uva de mesa Sultanina tratada con dos fuentes de giberelinas. Agric. Tec. 61, 15-21.
• Norme CEE-ONU FEV-19, 2010. Norme Concernant la Commercialisation et le Contrôle de la Qualité Commerciale des Raisins de Table. United Nations Economic Commission for Europe (UNECE). Edition 2010. New York et Genève 2010.
• Ocampo J., D’Eeckenbrugge G.C., Bruyère S., de Lapeyre de Bellaire L., Ollitrault P., 2006. Organization of morphological and genetic diversity of Caribbean and Venezuelan papaya germplasm. Fruits 61, 25-37. doi:10.1051/fruits:2006003
• OIV (Office International de la Vigne et du Vin.), 2007. OIV Descriptor List for Grape Varieties and Vitis Species. Available at www.oiv.int/oiv/info/enplubicationoiv.
• OIV, 2012. OIV General Form for the Description of Vine Varieties (Vitis spp.). Resolution OIV-VITI 467-2012.
• Rodrigues L.C., Morales M.R., Fernandes A.J.B., Ortiz J.M., 2008. Morphological characterization of sweet and sour cherry cultivars in a germplasm bank at Portugal. Genet. Resour. Crop Evol. 55, 593-601. doi:10.1007/s10722-007-9263-0
• Rousset F., 1997. Genetic differentiation and estimation of gene flow from F-statistics under isolation by distance. Genetics 145, 1219-1228.
• Sabir A., Tangolar S., Buyukalaca S., Kafkas S., 2009. Ampelographic and molecular diversity among grapevine (Vitis spp.) cultivars. Czech. J. Genet. Plant Breed. 45, 160-168.
• Santiago J.L., Boso S., Gago P., Alonso-Villaverde V., Martinez M.C., 2007. Molecular and ampelographic characterisation of Vitis vinifera L. ‘Albarino’, ‘Savagnin Blanc’ and ‘Caino Blanco’ shows that they are different cultivars. Span. J. Agric. Res. 5, 333-340. doi:10.5424/sjar/2007053-253
• Sarıkamış G., Yanmaz R., Ermiş S., Bakır M., Yüksel C., 2010. Genetic characterization of pea (Pisum sativum) germplasm from Turkey using morphological and SSR markers. Genet. Mol. Res. 9, 591-600. doi:10.4238/vol9-1gmr762
• Shiraishi M., Fujishima H., Chijiwa H., 2010. Evaluation of table grape genetic resources for sugar, organic acid, and amino acid composition of berries. Euphytica 174, 1-13. doi:10.1007/s10681-009-0084-4
• Sladonja B., Poljuha D., Plavša T., Peršurić Đ., Crespan M., 2007. Autochthonous Croatian grapevine cultivar ‘Jarbola’ - molecular, morphological and oenological characterization. Vitis 46, 99-100.
• Souza E., Sorrells M.E., 1991. Relationships among 70 North American oat germplasms: I. Cluster analysis using quantitative characters. Crop Sci. 31, 599-605. doi:10.2135/cropsci1991.0011183X003100030010x
• Soylemezoglu G., Agaoglu Y.S., Uzun H.I., 2001. Ampelographic characteristics and isozymic analysis of Vitis vinifera spp. sylvestris gmel in Southwestern Turkey. Biotechnol. Biotechnol. Eq. 15, 106-113. doi:10.1080/13102818.2001.10819140
• Veteläinen M., Gammelgard E., Valkonen J.P.T., 2005. Diversity of Nordic landrace potatoes (Solanum tuberosum L.) revealed by AFLPs and morphological characters. Genet. Resour. Crop Evol. 52, 999-1010. doi:10.1007/s10722-003-6129-y
• Ward J.H. Jr., 1963. Hierarchical grouping to optimize an objective function. J. Am. Stat. Assoc. 58, 236-244. doi:10.1080/01621459.1963.10500845
• Williams W.T., 1976. Pattern Analysis in Agricultural Science. Elsevier Scientific Publishing Company, Amsterdam, The Netherlands.
• Zdunic G., Hancevic K., Sladonja B., Poljuha D., Hartl-Musinov D., Budic-Leto I., Bucan L., Pezo I., 2008. Ampelographic characterization and sanitary status of grapevine cultivar ‘Prč bijeli’ (Vitis vinifera L.). Agric. Conspec. Sci. 73, 85-88.
• Zoghlami N., Mliki A., Ghorbel A., 2001. Evaluation of genetic diversity among Tunisian grapevines by RAPD markers. Vitis 40, 31-37.
• Zoghlami N., Riahi L., Laucou V., Lacombe T., Mliki A., Ghorbel A., This P., 2009. Origin and genetic diversity of Tunisian grapes as revealed by microsatellite markers. Sci. Hortic. 120, 479-486. doi:10.1016/j.scienta.2008.12.011
• Zohary D., Hopf M., 2000. Domestication of Plants in the Old World. The Origin and Spread of Cultivated Plants in West Asia, Europe and the Nile Valley, 3rd ed. Oxford University Press, Oxford.