A Confused Taxonomy
Vanilla is an ancient genus within the Orchidaceae family, the Vanilloidae subfamily, Vanilleae tribe, and Vanillinae subtribe, as demonstrated by recent molecular phylogenetic studies (Bory et al., 2008c; Cameron, 2004, 2005; see Chapter 1). Vanilla species are naturally distributed throughout America, Africa, and Asia-Oceania between the 27th north and south parallels (Portères, 1954). Portères (1954) described 110 species in the genus Vanilla, a number that reduced to 90 (Cameron and Chase, 1999) and to 107 (Soto Arenas, 2003). New species have also been added, such as the seven additional American species proposed (Soto Arenas, 1999, 2006, 2010) or V. shenzenica recently described in China (Liu et al., 2007). Altogether, there are more than 200 Vanilla species described to date but numerous synonymies remain (Bory et al., 2008c). Taxonomic classification is based on morphological variations (Portères, 1954) and such vegetative and floral characters are strongly influenced by the environment. In particular, vegetative traits (leaves, stems) display considerable variations at the intraspecific level making taxonomic identification difficult (Figure 2.1). This is exemplified by the lack of reliable herbarium vouchers and often the nonavailability of flowers (see Chapter 4). Taxonomy of Vanilla will therefore greatly benefit from the development of molecular phylogenetics, which already showed that the sections and subsections used in the taxonomic description of species by Portères do not have a phylogenetic value (Bouetard, 2007; Soto Arenas, 2003). As such, based on cladistic analysis of morphological and molecular data, a new infrageneric classification of Vanilla was recently proposed (Soto Arenas and Cribb, 2010) for 106 species examined, dividing genus Vanilla in two sub-genera: Vanilla and Xanata (further divided into sect. Xanata and Tethya). New keys for 15 Mexican and Central American species (Soto Arenas and Dressler, 2010) and more largely for the infrage-neric taxonomic identification within Vanilla are also proposed (Soto Arenas and Cribb, 2010). This recent work represents a crucial and major step towards a complete taxonomic revision of the genus.
Intraspecific Diversity
The aromatic species Vanilla planifolia G. Jackson, syn. V. fragrans (Salisb.) Ames, the main source of commercial vanilla, was disseminated from its area of origin (Mexico) following the discovery of the Americas by Christopher Columbus. Plantations were easily established by cuttings but pod production was unsuccessful in the absence of natural pollinators in the areas of introduction. In 1841, a simple method to hand-pollinate vanilla was discovered by Edmond Albius, a slave, in Reunion Island, and vanilla cuttings rapidly spread from Reunion Island to the Indian Ocean area and worldwide (Bory et al., 2008c; Kahane et al., 2008; see Chapter 17). As a consequence of this dissemination history, extremely low levels of genetic diversity are observed in vanilla plantations worldwide as shown by recent molecular genetic studies (Besse et al., 2004; Bory et al., 2008b, 2008d; Lubinsky et al., 2008a; Minoo et al., 2007; Sreedhar et al., 2007) suggesting a single clonal origin for the vanilla crop. This clone could correspond to the lectotype that was introduced, early in the nineteenth century, by the Marquis of Blandford into the collection of Charles Greville at Paddington (Portères, 1954). Cuttings were sent to the botanical gardens of Paris (France) and Antwerp (Belgium) from where these specimens were disseminated worldwide (Bory et al., 2008c; Kahane et al., 2008). It is thus surprising to observe an important morphological diversity in V. planifolia in the areas of introduction such as Reunion Island (Bory et al., 2008b, 2008c, 2008d) for a crop with a clonal origin and vegetatively propagated by cuttings.
All these observations raise important questions regarding the processes that might be involved in the evolution and diversification of vanilla. Some of the key processes that have been identified so far and the explanations that these can provide for the genetic and taxonomic complexity observed in Vanilla are discussed.
FIGURE 2.1 Morphological vegetative traits in Vanilla species from the CIRAD collection in Reunion Island (see Chapter 3): (a) typical leaf specimen for some species; (b) principal component analysis of vegetative traits (leaf and stem) measured in different species showing the importance of intraspecific variations leading to overlapping of species.
Vegetative versus Sexual Reproduction
For most Vanilla species, vegetative growth occurring naturally from stem cuttings (Portères, 1954) is the predominant reproductive mode, and appears as an efficient strategy adopted by the plant to develop settlements (Figure 2.2). Stems running on the ground are frequently observed, giving new roots and creating new individuals when the stem is cut, as reported for species such as V. bahiana Hoehne (Pignal, 1994) and V. chamissonis Klotzsch (Macedo Reis, 2000) in Brazil, V. barbellata Reichenbach f., V. claviculata (W. Wright) Swartz and V. dilloniana Correll (Nielsen and Siegismund, 1999) in Puerto Rico or V. madagascariensis Rolfe in Madagascar (P. Besse, pers. obs.). In Mexico, with reference to V. planifolia, in natural conditions, the same individual can cover up to 0.2 ha (Soto Arenas, 1999).
In Vanilla species, a rostellum membrane separates the female and male parts of the flower, and pollination therefore depends on the intervention of external pollinators. A notable exception is the species V. palmarum (Salzm. ex Lindl.) Lindl., which spontaneously self-pollinates (Bory et al., 2008c; Soto Arenas, 2006). Consequent, due to the need for pollinators, sexual reproduction is rarely observed in natural conditions. For V. planifolia, rates of 1% to 1‰ are reported (Bory et al., 2008c; Soto Arenas, 1999) with possible natural pollinators in America being orchid bees from the Euglossa and perhaps from the Eulaema genera (Lubinsky et al., 2006; Soto Arenas, 2006). Sexual reproduction rates reported for the species V. chamissonis (6% autogamy and 15% allogamy) are also relatively low (Macedo Reis, 2000).
FIGURE 2.2 Typical vegetative growth observed in Vanilla species. Left: V. madagascariensis in Madagascar. Right: V. pompona in Guadeloupe. (Courtesy of P. Besse.)
However, even rare sexual reproduction events can generate an important genetic diversification because a single sexual reproduction event is able to generate numerous genotypes that can vegetatively propagate rapidly. Heterozygosity observed in V. planifolia was reported to be 0–0.078 using isozymes (Soto Arenas, 1999), 0.154 using SSR markers (Bory et al., 2008b) and 0.293 using AFLPs (Bory et al., 2008d). Given these heterozygosity levels, even selfing can generate genetic diversity, as demonstrated through manual self-pollination experiments (Bory et al., 2008d) leading to increased diversity estimates (Dmax from 0.106 to 0.140) through novel allelic combinations (Figure 2.3). This is well illustrated in the case of V. planifolia in areas of introduction, where natural pollinators are absent. In these areas, such as in Reunion Island, traditional cultivation practices involve vine propagation by cuttings, and manual self-pollination to produce pods. This resulted in the appearance of novel vanilla varieties such as the “Aiguille” type observed in Reunion Island, which most likely resulted from accidental seed germination in the field from a forgotten pod, and subsequent vegetative propagation of the individual (Bory et al., 2008d) (Figure 2.3). Such a novel type can rapidly spread in plantations given the vegetative propagation used to multiply vines. This must also happen in the wild. A combination of sexual and vegetative reproduction, where one creates diversity and the other helps settlement, has already been suggested for the species V. pompona Schiede and V. bahiana in tropical America based on AFLP patterns (Bory et al., 2008d). Sexual reproduction is therefore a key evolutionary process for most species of the genus despite its low rates and because of their major vegetative reproduction. A few species of Vanilla appear to rely solely on sexual reproduction for propagation. This is the case for V. palmarum, which is entirely epiphytic on a palm tree with a short lifecycle (Pignal, 1994) and for V. mexicana Mill. (syn V. inodora Shiede) in which even artificial vegetative propagation is unsuccessful (P. Feldmann, pers. com.) (Figure 2.4).