Introduction
To increase the knowledge of the Sambirano Domain, floristic and vegetation surveys were conducted in northwestern Madagascar (Gautier & al., 2002; Nusbaumer & al., 2010) and more recently in the Ampasindava Peninsula. Following a preliminary exploration in November 2007, detailed botanical inventories and vegetation studies were carried out during the rainy seasons of two consecutive years from November to February in 2008–2009 and 2009–2010 by a team of botanists from Conservatoire et Jardin botaniques Genève, Missouri Botanical Garden and Université d'Antananarivo.
On 11 November 2008, we discovered 100–150 individuals of a small orchid with a single leaf in the highest forest of the peninsula locally known as “Bongomihiravavy” (referred to as “Ambohimiravahavy” on topographical maps), which reaches an elevation of 730 m. It was growing on a rock in 10–15 cm deep humic soil at the bottom of a 4 m cliff at an elevation of 500 m (Fig. 1). The leaves appeared to have emerged recently and only a few individuals were flowering at this early period of the rainy season (but several were in bud). We collected only two herbarium specimens (one flowering and one plant with a bud) without tubers (Nusbaumer & al. 2797), in order to not deplete the small population. No further material was found during the six months spent exploring forests on the peninsula. The specimens were deposited in herbaria in Antananarivo (TAN) and Geneva (G) respectively.
Using Petterson's (1991) key to Nervilia Gaudich. to identify our collection, it keyed out as the African N. gassneri Börge Pett. Nevertheless, our collection also possesses characteristics of the Malagasy N. lilacea Jum. & H. Perrier. The latter species was hitherto only known from the type collection, grown in Perrier de la Bâthie's garden from a tuber collected in 1909 in Manongarivo forest.
Table 1 presents the main characters used by Pettersson (1991) to separate the African N. gassneri from the Malagasy N. lilacea with the corresponding values for these characters for the Ampasindava specimens.
Our collection differs from the African N. gassneri because its leaves mature before the plant flowers and because of the absence of cleistogamous flowers, both of which are characteristics of N. lilacea. Our collection differs from N. lilacea because its lip has obtuse lateral lobes to the hypochile, by the size and venation of its leaves, the length of the petiole and the number of sheathing cataphylls of the spike, characteristics that correspond to N. gassneri.
The shape of hypochile lateral lobes is the distinctive chara — cter used by Petterson (1991) to separate the two species. However, since this character is variable in other Nervilia species, we do not consider it to justify the recognition of the two plants as distinct species. The quantitative differences also need to be treated with care because the cultivated type specimen is the only known collection of N. lilacea (Perrier de la Bâthie 1873), and we therefore have no idea of the variability of these characters in its natural population. It should also be mentioned that the elevation of our recent collection locality is much lower (500 m) than those of either N. gassneri or N. lilacea (1150–1900 m and 1000 m elevation respectively), and that this might have an influence on vegetative characters. Further careful examination of both Malagasy and African herbarium collections at K confirmed our conclusion that N. lilacea with N. gassneri must be considered as conspecific, showing variability in shape of the lateral lobes of the lip (obtuse to acute) and in the sequence in which its leaves and flowers appear. Nervilia lilacea is the earliest name and must be retained for this taxon.
Fig. 1.
Nervilia lilacea Jum. & H. Perrier. A-B. Flowers ; C. Population growing at the bottom of a 4m cliff at 500 m elevation, in deep humic soil on a rock ; D. Distribution map of Nervilia lilacea , with the new collection Nusbaumer & al. 2797 shown by a red dot [other points re-drawn after Petterson, 1990, 1991]. [A–C: Nusbaumer & al. 2797] [Photo : L. Nusbaumer]
Table. 1.
Diagnostic characters of Nervilia gassneri Börge Pett., Nervilia lilacea Jum. & H. Perrier according to Petterson (1991), and our intermediate specimen. The distinctive characters used in the Petterson's 1991 key are highlighted with a star (*). Cryptic characters to consider with care, depending on maturity of plants, are highlighted with a cross (†).
Nervilia lilacea Jum. & H. Perrier in Ann. Fac. Sci. Marseille 21: 197. 1912.
Typus: MADAGASCAR: Centre, massif de Manongarivo, bois humides, 1000 m, fl., Perrier de la Bâthie 1873 (holo-: P [P0009 4725]!).
= Nervilia gassneri Börge Pett. in Nord. J. Bot. 9: 492. 1990. Typus: MALAWI: Southern Prov., Zomba Distr., Zomba Plateau, 1530 m, 15.XII.1984. Petterson & Gassner 359 (holo-: UPS; iso-: BR, K!, LISC, LMU, MAL, NHT, SRGH).
Pettersson (1991) lists under N. gassneri several provisional names that have been applied to African specimens belonging to this species.
Observations. — The small light seeds of Nervilia species, as with many orchids, permit long-distance dispersal (Pettersson, 1991). The genus Nervilia has 13 species in Africa, six of which are shared with Madagascar, and an additional two species that are endemic to Madagascar. The Jaccard's coefficient of similarity of Nervilia between Madagascar and mainland Africa is thus 47%, which is very high compared with 1.8% for Orchidaceae and 2.4% for the Angiosperm flora as a whole. Endemism of Nervilia in Madagascar is 13.3% which is also very low compared to Orchidaceae as a whole (85%) or the entire Angiosperm flora (84%) (Callmander & al., in press). This high similarity and low endemism in Madagascar underlines strong relationships with Africa and suggests the hypothesis of recent multiple dispersal events from continental Africa to Madagascar. Its maximal African diversity is in the Sudano-Zambesian region and its maximal Malagasy diversity occurs in the northern and central western parts of the island, the two regions that are the closest to the Africa mainland. The genus is absent from the extreme southern part of the island.
Recently, new Malagasy records for several taxa from a variety of plant families, previously known only from Africa or Comoros, have been made in the northern part of the island (Rakotondrainibe & Andriambolantsoa, 2006; Faden, 2008; Trigui, 2010). With our increasing knowledge of Madagascar's flora through ongoing collecting efforts coupled with taxonomic revisionary work, the increase of the level of endemism due to the discovery and description of new endemic species is moderated by a more complete knowledge and better understanding of species distribution in the Malagasy region and the African continent.
Acknowledgements
Financial support for fieldwork was provided by grants from Conservation International — Madagascar (Convention 754), the National Geographic Society (Grant 8563–08) and the Augustin Lombard grant from the SPHN Society of Geneva. Fieldwork was conducted under collaboration between the Conservatoire et Jardin Botaniques de la Ville de Genève (CJBG), the Département d'Ecologie végétale de l'Université d'Antananarivo (DBEV) and the Missouri Botanical Garden in Antananarivo (MBG-Madagascar). We thank Prof. Charlotte Rajeriarison, Edmond Roger, Patrick Ranirison, Jacquis Tahinarivony and Cathy Madiomanana from the DBEV; and Charles Rakotovao and Roger Bernard (MBG-Madagascar). We are grateful to Roger Jaoravo and Célestin Totozafy the Fonkontany president from Ampôpô Nord; and to Morgane Ammann, Cyrille Chatelain and Nicolas Fumeaux at the CJBG as well as Martin Callmander and Pete Phillipson for their constructive comments improving the quality of this note.
