Relationships and Genetic Diversity within the Accessions Related to Malvasia Held in the Domaine de Vassal Grape Germplasm Repository

Materials and Methods

Results and Discussion

In Italy, Spain, and Portugal, the use of Malvasia is far more frequent than Malvoisie, Malvezija, Malvaiser, or Monemvassia. Six categories of Malvasia accessions were observed in the Domaine de Vassal collection: (1) 10 accessions that used the generic name Malvasia or its equivalent in the different languages; (2) 17 accessions that used Malvasia with an adjective (e.g., Malvasia verde); (3) 31 accessions that combined Malvasia with the name of another location, usually where the variety is cultivated (e.g., Malvasia istriana); (4) eight accessions that combined Malvasia with an adjective and the name of a location (e.g., Malvoisie rose du Pô); (5) two accessions that used the name Malvasia with an individual (e.g., Malvasia Rasteiro); and (6) two accessions that combined variety names (e.g., Malvasia Vermentino).

Identification of the 70 accessions by ampelography and DNA analysis concluded that there were 49 distinct varieties (Table 1⇑). Further study revealed that 37 of the 49 had 120 synonyms from 18 countries. However, in most cases there were only one or two synonyms (e.g., Malvasia istriana) and only three varieties had more than 10 synonyms. Results indicated that the use of Malvasia or related names is not justified for 29 varieties, especially when this designation is a synonym of a well-known variety with an original prime name (e.g., Malvoisie verte petite = Riesling italico). In such cases the prime name must be used.

The designations related to Malvasia were conserved as prime names for only 20 varieties, in those cases in which no other synonym was known and even if the use of this generic term was confusing and not ideal. These 20 varieties can be grouped in two categories. The first category consists of 12 varieties that correctly use Malvasia with a complementing term or phrase, such as Malvasia di Candia aromatica. The second category consists of eight varieties that appear to be unique and their accession name has been adopted as the default variety name: Malvasia bianca, Malvasia candida, Malvasia de Colares, Malvasia di Candia fausse, Malvasia roja, Malvoisie de Lasseraz, Malvoisie de Madera, and Malvoisie de Tarragon.

Parentage analysis of the 49 varieties with 20 SSR loci (see supplementary data) revealed the putative parents of 10 varieties (Table 2⇓). Two of those varieties, Schiras Dr. Houbdine and Malvasia-Moscatel Fonte Grande, were obtained in Portugal (by Dr. Soares Franco), but have unknown pedigrees. The possible parents were evaluated based on historical data and their traditional locations and no inconsistencies were detected. Two varieties, Gibi and Muscat d’Alexandrie, were involved in several parentages, the latter particularly interesting because it highlights the complex relationships that may exist between the muscat group and the Malvasia varieties. Parentage analysis also identified the parents of Malvasia de Colares and Malvasia di Candia fausse, two varieties that were judged to be unique. Given that their origin is now determined, their original prime names should be used in their countries of origin. The parentage of three other varieties was also interesting. Malvasia del Lazio has the same parentage as Muscat de Hambourg (Crespan 2003). The Croatian cultivar Plavina crna appears to be a seedling of Primitivo and Verdeca, a white variety grown in southern Italy (Calo et al. 2001), which confirms the historical link between Croatia and Italy. It was also confirmed that Argentinean variety Torrontès riojano originated from a cross between Mission, an older variety widely found in America, and Muscat d’Alexandrie, as previously noted by Agüero et al. (2003).

Genetic relationships among varieties of the Malvasia group were also examined. Forty-nine distinct Malvasia varieties and Gibi and Muscat d’Alexandrie (which were often indicated as parents in the parentage analysis) were evaluated. A dendrogram was constructed with two main divisions (Figure 1⇓): three groups were identified (A, B, C) in the upper division and one large and relatively loose group (D) was identified in the lower division.

• Download figure
• Open in new tab

Figure 1

Relationships among the varieties related to Malvasia. The tree was built applying the weighted neighbor-joining method to the dissimilarities matrix (18 SSR markers – 51 cultivars) using DARwin 4.0 software. (* Indicates varieties without references.)

Group A clusters the variety Muscat d’Alexandrie with several Malvasia. Most of these have a muscat or aromatic flavor, and only two have a neutral aroma based on observations at Domaine de Vassal (Malvasia di Casorzo and Malvasia di Schierano); however, others have described these varieties as aromatic or slightly aromatic (Dalmasso et al. 1964, Calo et al. 2001). Group A is the only group to show a close genetic relationship between the muscat and Malvasia varieties. Group B clusters varieties from Central Europe and the Balkans that do not have the Malvasia prime name. Four Malvasia comprise group C, which contains varieties from the Italian peninsula, and only Malvasia di Sardegna has an aromatic flavor. Group D is characterized by the Iberian origin of its varieties; only one, Malvasia fina from Portugal, can be rightly designated as Malvasia. The other Malvasia, Malvasia roja, is from Spain, but it was not possible to positively identify this variety with any other one in the Domaine de Vassal collection. All other varieties in group D have an original prime name different from Malvasia (Table 1⇑), including Gibi, which is involved in some Malvasia parentage. This white and female cultivar was grown in North Africa and Spain in the past.

Varieties not belonging to the four groups are located mainly in the lower left part of the tree, but do not represent a set consistent with geographical origin or morphological traits. The four Malvasia situated in this part are poorly documented at best (noted * in Figure 1⇑) and their designation remains questionable.

It is remarkable that all but one of the varieties for which the designation of Malvasia is acceptable are located in the upper part of the dendrogram (Malvasia fina, from Portugal, is the exception). In the lower part, there are only five Malvasia (Malvasia candida, Malvasia de Colares, Malvasia roja, Malvoisie de Lasseraz, and Malvoisie de Madera), but these are poorly documented accessions for which references are missing; consequently, it was impossible to decide on the relevancy of the Malvasia name. Finally, the genetic structuring of the 49 distinct varieties analyzed here can be mainly explained by their geographical origin and flavor characteristics.

Conclusion

One question remains: what is a Malvasia variety? Results here do not underline a clear common denominator among all the Malvasia maintained in the Domaine de Vassal repository. Genetic analysis and clustering clearly show more than one group of Malvasia varieties. Three countries are primarily involved in the presumed geographical origin of the 49 distinct cultivars analyzed: Italy (13 varieties), Portugal (10), and Spain (7). In addition to a heterogeneous genetic background, these varieties are also morphologically heterogeneous (data not shown). From a general ampelographic point of view, there is no morphological consistency among these varieties. A majority of Malvasia varieties are white (73%) but there are also black, rose, and gray. There is no specificity phenology or sugar content, and the proportion of the varieties with round and elliptic shape of the berries is balanced. Eight Malvasia varieties have muscat or aromatic flavor, but most do not. Thus, cultivars named Malvasia are a complex group that includes muscat, aromatic, and neutral varieties and that lacks a genetic or morphological phenotypic basis. The justification for the use of Malvasia as the name of a grape variety is not directly based on a special trait or origin, but takes into consideration historical and sociological factors.