Introduction

Traditional varieties and crop wild relatives represent an important biodiversity pool for both purely resources conservation goals and improvement of modern varieties (Fernie et al. 2006; Tanksley and McCouch 1997). In the last decades, genetic erosion affected many species of agronomical interest in the Mediterranean area (Hammer and Laghetti 2006). In industrialised countries, like Italy, the process of genetic erosion has been particularly rapid and is still ongoing. Such an impoverishment is mainly due to the collapse of traditional agricultural systems, with the degradation, fragmentation and loss of entire cultivated areas, following the abandonment of agricultural activities (Hammer and Laghetti 2006). Relict traditional varieties only survive in marginal areas, such as small islands far away from the coast, that for their geographical isolation are generally richer in ancient or uncommon germplasm than the neighbouring continental areas (Hammer and Laghetti 2006).

In most of the minor circum-Sicilian islands, agriculture has been historically the major economic activity. Since the end of World War II, however, agriculture rapidly declined, and the economy became more and more dependent on tourism. Currently, marginal agriculture only survives thanks to a small number of aged farmers (La Mantia et al. 2011), suggesting that the traditional varieties and the local related knowledge will disappear in the next future.

Grapevine is one of the most important crops all around the Mediterranean Region and the Near East. Compared to the thousands of cultivars selected during its millenary history of domestication, clone selection, Phylloxera crisis and the massive diffusion of few international clones of varieties like Chardonnay and Cabernet Sauvignon caused a severe loss of diversity in many countries. Only in small islands and in the most remote inner areas, geographic isolation, peculiar edafo-climatic conditions and socio-economic reasons relatively limited this globalization trend, then involving the persistence of some traditional and/or ancient grape varieties. Sicily occupies a central position in the Mediterranean and has been historically a main junction of commerce and colonization, and the cradle for the development of many local varieties (Unwin 2005; Garfì et al. 2013). For these reasons, it can be expected that the circum-Sicilian archipelagos represent an interesting source of traditional varieties, worth to be saved from extinction and adequately valorised. In this study, we studied the genetic identity and traditional use of the grapevine germplasm of minor Sicilian islands. The goal was not only to promote the preservation of rare and unrecognized germplasm trough the introduction in an ex situ collection field, but also to treasure the historical knowledge associated with these varieties.

Materials and methods

Study area

Among the circum-Sicilian archipelagos, five areas were investigated: Salina (Aeolian Islands), Lampedusa and Linosa (Pelagie Islands), Pantelleria, and Ustica (Fig. 1). The Egadi Islands were not investigated because the grape cultivation has disappeared. A brief description of each island is reported hereinafter, while additional information is provided in Table 1.

Fig. 1
figure 1

Map of Sicily and circum-Sicilian archipelagos

Full size image
Table 1 Synthetic overview on the main geographical and historical data of the investigated circum-Sicilian islands
Full size table

Salina

Salina belongs to the Aeolian archipelago, NE of Sicily. It has volcanic origin and it emerged from the SE Tyrrhenian Sea about 0.3 million years ago (Lucchi et al. 2013). Its natural landscape is dominated by the steep cones of Monte dei Porri and Monte Fossa delle Felci: for this reason the ancient Greeks called it Didyme (= twin island). During the Middle Age, the population of Salina increased due to migrations from Lipari, the main island of the archipelago, which was experiencing intense volcanic activity. Among the Aeolian Islands, Salina has always been the most intensively cultivated. In the mid 1800 the viticulture was largely widespread and a number of varieties were cultivated (De Gregorio 1840). As for the whole Aeolian archipelago, Habsburg Lothringen (1894) reported that the most common grapes were the white Cataratta and Malvasia, and the black Passulina, Mantuonica and Moscato; specifically for Salina (in locality Santa Marina) the white cultivars Nuciddara, Greca, Duraco, Ducignola, and the black cultivars Trummana and Livedda were commonly used for trellis.

As in the rest of the Aeolian archipelago, the population dramatically dropped at the end of the nineteenth century, due to the huge destruction of vineyards caused by phylloxera (King and Young 1979; Lo Cascio and La Mantia 2013). Currently, the main crops are capers and grapes, the latter supporting a traditional wine industry (Malvasia wine, mostly). The local investigated vineyards are mainly located in the territories of Malfa and Val di Chiesa.

Lampedusa

Lampedusa belongs to the African–Pelagian foreland; local outcropping rocks are Meso-Cenozoic carbonates and marls (Grasso and Pedley 1988). It appears like a triangular plateau with an almost continuous steep cliff on the northern coast and gently declining slopes southwards, with several canyons; local agriculture developed inside these canyons, on intensively terraced surfaces (La Mantia et al. 2011). Lampedusa was first inhabited during the Neolithic and seems to have hosted a continuous human community until the end of 2000 BCE, probably related to the Maltese megalithic civilization (Radi 1973). Thanks to its wide natural harbour, it has been exploited as a naval base since ancient times and played a key role as stopover for North African, Maltese and Sicilian sailors and anglers over the centuries. After the establishment of a permanent Bourbon colony in 1843, grape cultivation was introduced in the island (Calcagno 1879) and until the end of the World War II Lampedusa was almost self-sufficient for agriculture. Today, only few cultivated fields, mostly vineyards (Di Lorenzo et al. 2010) and vegetable orchards, survive in very restricted areas (Hammer and Laghetti 2006; La Mantia et al. 2011).

Linosa

Linosa, a small volcanic island in the Strait of Sicily emerged between 1.1 and 0.5 million years ago (Di Bella et al. 2008). Archaeological remains testify its use as a base for Romans during the Punic Wars (fifth century BC). Like Lampedusa, after a long-lasting period of irregular human presence, it was colonized during the half of the nineteenth century (Corti et al. 2002). Agriculture and fishing, once the exclusive resources for local people, are currently in rapid decline. As observed in Pantelleria and the Aeolian islands, most of the cultivated terraces have been abandoned and are nowadays colonized by natural vegetation due to progressive succession processes (Rühl and Pasta 2007).

Pantelleria

The volcanic complex of Pantelleria emerged about 0.3 million years ago (Civetta et al. 1984) between SW Sicily and Tunisia. It was first colonised during the Neolithic period (Abelli et al. 2014), and permanently inhabited since the eighth century BCE. Local dialect and toponyms largely testify the strong influence of Arab and Berber people on local culture and landscape shaping. Not surprisingly, the agricultural identity of Pantelleria is specially linked to the Zibibbo, a grape variety introduced from Cape Zebib (NE Tunisia) during the Arab domination (ninth–eleventh century CE) (Niccoli 1902), used to produce table grapes, sweet wine and raisin.

In the half of the nineteenth century Calcara (1853) documented a long list of grape varieties cultivated in the island for wine production, including Cataratta (Cataratto), Greca di vigna, Blasco, Pignatello, Catalamiscu, Moscatello, Nano, Virduni, Uva di paradiso and Racina virdi, in addition to Zibibbo, Insolia and Bildè used for both raisin and wine; as table grapes, the varieties Greca, Caleo, Uva di Salemi, Prunesta, Minnavacchina bianca and nera, Buttuna di gallo and Trivolti were also common. In 1833 vineyards extended over a total area of 1054 hectares, increasing to more than 3000 hectares at the beginning of the twentieth century (Scarponi 1939; Bonasera 1965). The two cultivars Catarratto and Zibibbo, used for common and sweet wine, respectively, have always been the prevailing grapes (Puviani 1916). Around the 1930, when the phylloxera outbreak largely spread in the island, farmers reacted expanding the cultivations in new terraced areas (Gigante 1968; D’Aietti 1978) and on the eve of World War II about 5000 hectares of these two varieties were still grown (Scarponi 1939). In the middle of the 1970, some additional grapes are reported, namely Funcia chiatta, Inzolia, Minna i vacca, Nívuru, Pignatello and Greca (D’Aietti 1978).

In the following decades, agriculture as a whole experienced a progressive decline and the Agricultural Usable Surface reduced by 60 % between 1929 and early 2000 (Rühl et al. 2005). Currently, the main agricultural products are grapes and capers (Hammer and Laghetti 2006), grown on the few available flat areas (e.g. at Piana Ghirlanda) and on some terraced slopes in the localities of Mueggen, Bukkuram, Siba and Scauri.

Ustica

Ustica is a small volcanic island emerged about 0.3 million years ago in S Tyrrhenian Sea (De Vita et al. 1998). The island was first inhabited during the Eneolithic by peoples coming from the Aeolian Islands (Holloway and Lukesh 2001). Several centuries later, Phoenicians and Romans used it as a naval base, whilst during the Middle Age human presence started to be quite irregular and the island was totally deserted after the Thirteenth century. It became a hideout for North African pirates until the end of the eighteenth century, when a new Bourbon colony with people from Lipari was established and its natural landscape was rapidly transformed by agricultural activities. At that time, an area of about 350 hectares was devoted to viticulture and a number of black (e.g. Muriedda, Vanni bertucci, Trummana/Tremani, Olivedda niura, the latter known as an excellent table variety) and white (e.g. Zibibbo, Muscateddu, Guarnacca, Rigalia, Lacrime i Madonna, Trunzu, Zuruca, Cornicchiola) varieties were cultivated (Habsburg Lothringen 1898), especially in the gently declining slopes of the northern (Tramontana) and southern (San Paolo) sides of the east–west oriented small mountain ridge consisting of three extinct volcanoes.

Information and semi-structured interviews to farmers

In order to obtain information on grapevine germplasm, semi-structured interviews to farmers were carried out through a standardised questionnaire between 2006 and 2011. Forty-five informants (44 men and 1 woman aged from 41 to 94) were chosen with the assistance of local expert grapevine farmers. The majority of our informants were elders (55–94 years old, 62 %); for each of them personal data were noted, including gender, age, education and occupation. Farmers were asked to provide the following information: names and synonyms of grape varieties, berry colour, current and past usage, origin, time of introduction, type of management and growing practices. Information not related to traditional varieties was not recorded.

Plant material

The accessions were selected following the indications of the farmers and labelled in order to relocate the plants to collect plant material (leaves and young cuttings for DNA analyses and scions for grafting). Plant material was collected between 2007 and 2014 directly from 60 vineyards: 22 from Salina, 8 from Lampedusa, 6 from Linosa, 18 from Pantelleria and 6 from Ustica. The oldest vineyards (more than 110 years old) were located in Salina, while all the others were between 20 and 95 years old. Altogether 185 local cultivars were investigated by microsatellite analysis (Table 2). Out of these, 82 were collected in Salina, 34 in Lampedusa, 17 in Linosa, 39 in Pantelleria and 13 in Ustica. For rare germplasm preservation most part of the collected grapevine cultivars were introduced in an ex situ collection field of the National Research Council of Italy (CNR) Institute of Biosciences and BioResources (IBBR) located in Collesano district, Italy (37°59′19.9″N 13°54′55.8″E, 80 m above sea level).

Table 2 List of sampled grapevine cultivars, sampling area, uses and field notes
Full size table

DNA extraction and microsatellites analyses

Total genomic DNA was extracted from young leaves or inner wood of young cuttings. Tissues were ground into fine powder with liquid nitrogen and stored at −80 °C until use. The extraction was carried out following the CTAB method (Doyle and Doyle 1987) and DNA was quantified in 1 % agarose gels.

Samples were analysed at six microsatellite loci [Simple Sequence Repeat (SSR)], i.e. VVS2 (Thomas and Scott 1993), VVMD5, VVMD7 and VVMD27 (Bowers et al. 1996), VrZAG62, and VrZAG79 (Sefc et al. 1999). The forward primer of each marker was labeled with one of the three unique ABI PRISM fluorescent dyes: 6-FAM, JOE, TAMRA.

PCR amplification was carried out using the Qiagen multiplex PCR kit with the following conditions: 15 min at 95 °C (HotStar Taq activation step), followed by 35 cycles consisting of 30 s at 94 °C (denaturation), 90 s at 50–56 °C (annealing), 60 s at 72 °C (extension) and a final step for 30 min at 72 °C.

Each sample was amplified at least twice to correct for possible mistyping or amplification errors. PCR products were size-separated by capillary electrophoresis performed on a genetic analyzer (ABI Prism 3130, Applied Biosystems, Inc.) by an external service (MWG, Germany).

Electropherograms were visually verified using Gene Mapper v. 4.1 software. Allele size was estimated by comparing the fragment peaks with the internal size standard, using the default method for band calling with SSR and the expected repeat size. Genetic profiles were compared with the six SSR-markers used within the Genres081 Project (recommended by This et al. 2004) available in the Italian Vitis database (www.vitisdb.it), the European Vitis database (www.eu-vitis.de) and the Vitis International Variety Catalogue (www.vivc.de) for identification of synonyms. The varieties used for SSR standardization were Malvasia di Lipari for the Italian and European databases, and Sangiovese for the International database. The SSR profile search engine of the Italian database allowed a tolerance of ±1 nucleotide, accounting for the scatter pattern. Conversely, the European and International databases only retrieved varieties with 100 % bp match. Cases of homonymy were identified checking for names in the same databases. Several diversity parameters were estimated using GenAlEx 6.5 (Peakall and Smouse 2012): the number of alleles per locus (Na), the allele size range and the allele frequency, and the observed (Ho) and expected (He) heterozygosity (Nei 1978, 1987). Estimated frequency of null alleles (r) was calculated with the software IDENTITY (Wagner and Sefc 1999). Microsatellite screening ability (MSA) was also based on the probability of identity (PI) (Paetkau et al. 1995) and the polymorphic information content (PIC) (Weber 1990) derived as follows:

$$PI = \sum\limits_{i = 1}^{n} {p_{i}^{4} } + \sum\limits_{i = 1}^{n - 1} {\sum\limits_{j = i + 1}^{n} {2p_{i}^{2} } p_{j}^{2} }$$
$$PIC = \left( {\sum\limits_{i = 1}^{n} {p_{i}^{4} } } \right) - \left( {\sum\limits_{i = 1}^{n - 1} {\sum\limits_{j = i + 1}^{n} {2p_{i}^{2} p_{j}^{2} } } } \right)$$

where pi and pj are the frequencies of the ith and jth allele and n is the number of alleles. The PIC was directly calculated starting from He and PI. The above-mentioned indices range from 0.0 to 1.0 and provide information on the effectiveness to differentiate among genotypes. Thus, the most effective SSR has high values of Ho and polymorphic information content, and low PI.

The pairwise genetic distances among genotypes were calculated with the software Populations 1.2.31 (Langella 2002) using Nei’s coefficient (Nei et al. 1983). Cluster analysis was performed according to UPGMA (Unweighted Pair-Group Method with Arithmetical Averages) algorithm and a dendrogram by using Populations 1.2.31 was generated and visualized with TreeGraph 2.0 (Stöver and Müller 2010).

Results and discussion

Distribution, use and peculiarities of the sampled grapevines

During our surveys we detected 93 different cultivars, as determined by farmers’ claims (Table 2). Most of the surveyed cultivars were found just as relicts in old vineyards or nearby local wine cellars (ESM1 A-F). Consequently, many varieties are represented by a single sample. Other cultivars were more common throughout the territory and we collected sample material from several plants (e.g. up to 11, for Zibibbo). Out of the 93 cultivars, 39 were exclusive from Salina, 15 from Pantelleria, 23 from the Pelagie Islands, and 6 from Ustica. Some cultivars were present in two different archipelagos, i.e.: Lacrime i Madonna was found at Pantelleria and Ustica; Zibibbo nero at Salina and Pantelleria; Inzolia both at Salina and Pelagie; Inzolia nera and Trummana both at Salina and Ustica; Funcia chiatta and Nave both at Pantelleria and Pelagie. Only a few cultivars were present in three or all four archipelagos: Catarratto at Salina, Pantelleria and Pelagie; Minna i vacca at Pantelleria, Pelagie and Ustica; Zibibbo in all the archipelagos. The particular richness of cultivars at Salina confirms in this island the long agricultural and winery tradition, which is still persisting nowadays.

A number of cultivars already known from the literature as grown before the phylloxera outbreak still persist in some islands and must be regarded as “ancient” (Calcara 1853; Calcagno 1879; Habsburg Lothringen 1894). Besides the most renowned Zibibbo, Catarrato and Inzolia, it is especially the case of Pignatello, Minna i vacca, Funcia chiatta, Nìvuro and Greca from Pantelleria, Nuciddara, Mantuonica, Livedda and Trummana from Salina, and Alivedda nera and Lacrime i Madonna, from Ustica.

Table 2 summarizes the information on the use, putative origin and curiosities associated to the accessions, as retrieved by the interviews. Many varieties (40) were only used for wine making; 22 varieties were used for fresh table consumption; 19 for both wine making and table consumption. Only 12 varieties were used as dried fruit in addition to either wine making or fresh consumption, especially at Salina (Corinto, Minnulettina, Minnilottina, Minutidda and Nuciddara) and Pantelleria (the group of Zibibbo grapes).

For most of the varieties (61), the interviews revealed a number of distinctive peculiarities, often strictly bound to local customs (Table 2). For example, a couple of varieties (Funcia chiatta and Nave) were traditionally preserved in alcohol in the Pelagie islands, to be used by farmers as energizing snack in wintertime. The grapes of five varieties (Cuda i vulpe, Malvasia and Nucignola/Rucignola in Salina; Nìvuro/Nivuro nostrale/Pignatello and Zibibbo in Pantelleria) were sun-dried a few weeks before pressing in order to obtain sweeter and stronger wines. In particular, the musts from Nìvuro (meaning “black” in Sicilian language)/Nivuro nostrale/Pignatello are very dark so they were used to darken musts of other varieties. Nìvuro/Nivuro nostrale/Pignatello was appreciated over the centuries for its tannic structure and was much in vogue for the production of concentrated grape must (D’Agata 2014). The ‘Ribollito’ wine, produced by adding gypsum powder to Nìvuro/Nivuro nostrale/Pignatello grapes before pressing (Raja 1910), was a wine traditionally drunk in Pantelleria during Carnival. Nowadays, only Malvasia and Zibibbo are still used for making sweet wines, which play an important role in the economies of Salina and Pantelleria, respectively.

A quite interesting case concerned the cultivar Citana because of the unclear origin of its name, and the use of its grapes as well. According to two interviews, the name would derive from the Sicilian word “acìtu”, meaning vinegar, indicating the unsuitability of the variety for making good wine. On the contrary, an old reference (Di Rovasenda 1877) mentions a juicy grapevine variety called Acitàna cultivated in the territory of Messina, whose name could mean it is native from the nearby village Acitrezza (district of Catania) whose inhabitants are called acitàni (Anon 1890).

Genetic identity and relevance of the investigated germplasm

The main genetic parameters related to the nuclear microsatellite diversity are reported in Table 3. The analyses revealed 62 alleles, ranging from 8 (VVMD27) to 12 (VrZAG62), with an average of 10.33 alleles per locus. The expected heterozygosity He (expressing gene diversity) ranged from 0.769 (VVMD7) to 0.844 (VVMD5), with a mean value 0.811, while the observed heterozygosity Ho ranged from 0.773 (VVMD7) to 0.947 (VrZAG62). For all loci, Ho was higher than He. The probability of null alleles was always negative and very close to 0 indicating the low probability of null alleles at all studied loci. When only one allele per locus was detected, samples were considered homozygous genotypes rather than heterozygous with a null allele. The most informative locus was VVMD5, with PI of 0.0427. The six SSR loci we used showed a high discrimination power (0.8222) and a low probability that two randomly chosen individuals had identical genotypes (PI 3.67E − 08). According to that, cultivars with identical profiles were considered synonyms.

Table 3 Genetic parameters at the 6 SSR loci analysed in the grapevine sampled cultivars
Full size table

The analyses at six SSR microsatellites revealed 75 different genotypes (Table 4; Fig. 2), fewer than the 93 different denominations given by farmers. Accordingly, in 20 cases each genotype was known by at least two synonyms (Table 4). In some cases, such as for Cataratto and Zibibbo, the same genotype was indicated even by several names mainly referred to distinctive traits of the berry (e.g. big, small, golden, and so on) in comparison to the standard type, so deserving a specific qualification and a new naming (Table 4). In other cases, the cultivar names were clearly different (Table 4) and most often restricted to the island of provenance. For example, the cultivar name Nivureddu and all its variants (Bertuccio C, Calabrisi and Gallipoli nera B) (Table 4, accession n. 8) was found in the Pelagie Islands, the synonym Nìvuro nostrale/Pignatello in Pantelleria, and Magliocco and Mascarisi in Salina. It can be assumed, therefore, that the diffusion of a cultivar in a new island was not always accompanied by its original denomination, then taking a local new name. In other cases, synonyms (e.g. Bertuccio A/Catarratto rosato, Fiore d’arancio/Trunzu) were found within the same island. It is noteworthy that in the case of Fiore d’arancio/Trunzu, despite farmers were fully aware of the coincident identity of these two cultivars, they used alternative denomination in the southern and northern parts of Salina, respectively. This occurrence could correlate with the mountainous topography of the island, which acted as a kind of barrier separating its territory in two (later on three) independent and even competing municipalities.

Table 4 SSR standardized profiles of the grapevine sampled cultivars
Full size table
Fig. 2
figure 2

Dendrogram of genetic relationship among the investigated cultivars

Full size image

Another interesting case of synonymy concerns the cultivars Giugnatica and Vugliatico. These plants, provided by two different farmers of Salina, were reported to be named according to the different time of fruits ripening (Giugnatica from “Giugno” = June, Vugliatico from “Luglio” = July). Since genetic analysis revealed they were actually identical, their presumptive asynchronous maturation could therefore depend on the different environmental conditions of the growing sites. A third cultivar with a maturation-related name, Lugliatico or Luglienga (from “Luglio” = July), was instead a truly distinctive genotype (Table 4). It has to be noted that some cases of synonymy indicating phenological or morphological variants are most likely due to the influence of different environmental conditions.

The cases of homonymy were abundant as well, since in 18 cases one single cultivar name actually referred to different genotypes. Overall, this result proved the efficiency of the six selected SSRs to discern among plants with very similar phenotype. In Table 4 we discriminated among these profiles by attributing a suffix to the given name (e.g. Alicante A, Alicante B). Generally, cultivars with homonyms were considered of minor value by farmers, probably accounting for the limited attention toward possible phenotypic differences.

In a couple of striking cases, e.g. Minna i vacca and Racina i mustu, the same name was shared by four genotypes (Table 4, Fig. 2). Apparently, their peculiar phenotypic characteristics triggered the attribution of evocative denominations (in Sicilian language, Minna i vacca means “Cowteat” and refers to the elongate berry shape; Racina i mustu means “Must grape” and indicates the special suitability of the cultivar for wine making).

The comparison of the 75 genetic profiles with national and international grapevine databases—namely the Italian Vitis Database, the European Vitis Database and the Vitis International Variety Catalogue—allowed assessing the extent of their diffusion and/or correspondence with more renowned cultivars. Only 36 genetic profiles corresponded to known varieties. The attributed area of origin as inferred by the databases was mainly Mediterranean, with most varieties from Italy (14), followed by France (5), Greece (3), Lebanon (2), Spain (2) and Tunisia (2). Croatia, UK, USA all accounted for one variety each. On the contrary the others profiles (39), were not included within any database and are therefore to be considered as new genotypes (Table 5). Phylogenetic analysis of the cultivars, based on the six SSR’s, did not reveal any geographic pattern. Accessions collected in the same island are scattered in different clusters attesting no gene exchange between putative local parents and then their different origin and history of introduction (Fig. 2). This result was expected, given the central position of the circum-Sicilian islands within the trade routes of the Mediterranean Basin as well as their complex history of colonization. Three varieties (Malvasia, Maria Pirovano and Zibibbo), though listed in the International databases, are not referred to a specified country of origin herein. According to data from literature, the origin of Malvasia, initially thought to be Greek, remains actually obscure (Crespan et al. 2015). Maria Pirovano is a recent variety created in 1926 by the breeder Alberto Pirovano (cf. www.vivc.de), whereas Zibibbo is considered native from Tunisia (Niccoli 1902).

Table 5 Comparison of genetic profiles with international databases
Full size table

In some cases (e.g. Catarratto, Damaschino, Malvasia and Zibibbo) the appellation given by the farmers corresponded to the official name as recorded in the databases. All of them are important varieties for the current wine industry of Sicily and its minor islands, and therefore they are largely known and characterized since long time. Other varieties showed evident similarity to the official nomenclature (Cornicchiola–Cornichon, Inzolia–Ansonica, Maddalena Salomone–Madeleine Angevine Oberlin and Moscato Nero–Muscat Hamburg), likely resulting from the transposition of the approved name into the local vernacular. However, the majority of varieties (25 out of 36) with a known profile had a name that differed completely from the official one, representing a glaring example of synonymy. In some cases, the re-naming could be attributed to difficult pronunciation of foreign terms by peasants, but it is not rare that the islander names also differed from those commonly used in other Italian regions or even in mainland Sicily. In this regard, striking examples were Calabrese, Trebbiano toscano, Grillo and Sangiovese, amongst the most famous varieties for the wine industry, that in the Sicilian minor islands have alternative denominations (Bertuccio B/Racina i mustu D, Centorotoli B, Gallipoli bianca B and Corinto/Minutidda, respectively). Moreover, the comparison with the international databases highlighted several cases of homonymy (Table 5): e.g. the Nero d’Avola from the Pelagie archipelago is a case of homonymy, since it was genetically different from the official Nero d’Avola (De Lorenzis et al. 2014). Though the individual reasons for the different denominations remain obscure, we can hypothesise that those are cases of mislabelling that perpetuated in time. Additionally, the true name associated to a specific cultivar could have been lost during the successive generation of growers, and a new one was created/attributed later on. The geographical and cultural isolation of most of the circum-Sicilian islands, together with the very few farmers still active nowadays, make this type of “cultural mutation” very likely and persistent.

An interesting case of homonymy concerns the variety Cuda i vulpe found in Salina, that proved different from the official cultivar Coda di Volpe from the Campania region (Italy) (Costantini et al. 2005). It is worth mentioning that Salina, and the Aeolian archipelago in general, has been colonized by immigrants from Campania during the Borbonic Kingdom (eighteenth–nineteenth century). It is thus intriguing to imagine that the memory of a favourite grapevine variety survived the plants themselves, and was finally attributed to grapes with similar phenotype. The Aeolian Cuda i vulpe actually proved a synonym of Frmentum, a variety native to Croatia (Table 5). The explanation could rely on migratory fluxes that since 1561 occurred from the Venetian colonies in the Adriatic and eastern Mediterranean towards the Aeolian Archipelago, during which immigrants brought together a number of grapevine varieties (Archivio Storico Eoliano). Consistently, one farmer reported that the Cuda i vulpe variety was in fact native from Crete (Greece), the ancient Candia of the Venetian Republic (thirteenth–seventeenth century) (Table 2).

Interestingly, black Mantonico grapes are mentioned in a number of notarial deeds of sale in Syracuse since 1478 (Amato 1996) and, successively, 1555 in Ficarazzi (district of Palermo) (Morreale 1998), possibly meaning that the red grape Mantonico A from Salina, showing a new SSR profile, could be the same ancient variety.

Rare germplasm preservation resulted difficult due to the low quality (e.g. plants in abandoned fields, sanitary status, old vineyard) of the plant material available for plant propagation. The percentage of grapevine accessions survived in the ex situ collection field was low (60 %) and some accessions were lost. The ex situ collection field will allow us the ampelographic verification of the survived accessions in homogeneous growing conditions.

Conclusions

This work complements previous genetic characterization of varieties from mainland Sicily (Carimi et al. 2010, 2011; De Lorenzis et al. 2014). While we found commonalities with Carimi et al. (2010) (Catarratto, Diretta bianca, Inzolia, Lacrime i Maria, Rucignola and Zibibbo), but none with De Lorenzis et al. (2014), the vast majority (69 out of 75) of the varieties found in the circum-Sicilian minor islands were not reported until now in mainland Sicily. This study confirms that the minor islands of the Mediterranean Basin still represent underexplored hotspots of genetic diversity for grapevine. These important reservoirs of disregarded but potentially valuable genotypes could be of great value to breeders and the wine industry and they shed light on the migration of cultivars. In the small circum-Sicilian islands, the long-term survival of these rare varieties is uncertain, due to the increasing land abandonment and the shift of the economic drive from agriculture to tourism, a trend already observed in other islands (Pignone et al. 2001). During a field survey conducted 5 years later the first sample campaign, we found that 8 % of the sampled accessions had not survived. The urgent need to preserve this unique germplasm is testified by our own observations. Survey campaigns and collection fields become then an essential tool to preserve the grapevine germplasm and to avoid an irreversible loss of genetic diversity.