A recent phylogenetic study showed that species assigned to the newly recognised genus Pectinariella Szlach., Mytnik & Grochocka (previously treated as Angraecum Bory sect. Pectinaria Benth.) are polyphyletic, comprising a clade with species primarily in Madagascar and the Western Indian Ocean islands (including the type) and another non-sister clade whose members occur in continental Africa and the Gulf of Guinea islands. In order to render Pectinariella monophyletic, the five continental African species must therefore be removed. A new genus, Afropectinariella M.Simo & Stévart, is described and the following combinations are made: Afropectinariella atlantica (Stévart & Droissart) M.Simo & Stévart, Afropectinariella doratophylla (Summerh.) M.Simo & Stévart, Afropectinariella gabonensis (Summerh.) M.Simo & Stévart, Afropectinariella pungens (Schltr.) M.Simo & Stévart and Afropectinariella subulata (Lindl.) M.Simo & Stévart.

Une analyse phylogénétique récente a montré que les espèces attribuées au nouveau genre Pectinariella Szlach., Mytnik & Grochocka (anciennement Angraecum Bory sect. Pectinaria Benth.) forment un groupe polyphylétique, comprenant un clade avec des espèces présentes principalement à Madagascar et dans les îles de l'ouest de l'Océan Indien (incluant le type) et un autre clade qui ne lui est pas apparenté et dont les espèces sont distribuées en Afrique continentale et dans les îles du Golfe de Guinée. Afin de rétablir la monophylie du genre Pectinariella, les cinq espèces d’Afrique continentale et des îles du Golfe de Guinée doivent donc en être exclues. Un nouveau genre, Afropectinariella M.Simo & Stévart, est décrit et les combinaisons suivantes sont proposées: Afropectinariella atlantica (Stévart & Droissart) M.Simo & Stévart, Afropectinariella doratophylla (Summerh.) M.Simo & Stévart, Afropectinariella gabonensis (Summerh.) M.Simo & Stévart, Afropectinariella pungens (Schltr.) M.Simo & Stévart et Afropectinariella subulata (Lindl.) M.Simo & Stévart.

Angraecoid orchids, continental Africa, Malagasy and Indian Ocean islands, Pectinaria , phylogenetics, taxonomy

Since its description, the delimitation of Angraecum Bory (1804) has been controversial (Garay 1973). Much importance has been ascribed to the structure of the pollinia within Angraecum, which has led to the recognition of spurious genera to encompass closely related species (Schlechter 1913). Based on vegetative and floral features, along with geographic distribution, a number of partial or comprehensive revisions of the genus have been proposed over the decades (see Finet 1907; Garay 1973; Perrier de la Bathie 1941; Schlechter 1918, 1925; Senghas 1986; Stewart et al. 2006; Summerhayes 1958). Garay (1973) was the last to provide a full taxonomic treatment for Angraecum. He recognised 19 sections, of which nine are endemic to Madagascar and the Western Indian Ocean islands, one to Madagascar, the Western Indian Ocean islands and Sri Lanka, two to continental Africa and the Gulf of Guinea islands and seven have representatives in both Africa and Madagascar.

With the advent of molecular techniques, relationships within Angraecum have been explored in more detail over the last 15 years. The studies of Carlsward et al. (2006) on the phylogenetics of leafless angraecoids and of Micheneau et al. (2008) on the biogeography of Mascarene angraecoid orchids have shown the polyphyly of Angraecum and of at least five of the 19 recognised sections. In fact, African species of Angraecum belong to a group clearly distinct from that of members of the genus in Madagascar and the Indian Ocean islands (Micheneau et al. 2008). While investigating the diversification of the genus in Madagascar, Andriananjamanantsoa et al. (2016) also confirmed the polyphyly of Angraecums.l. and of all Angraecum sections with representatives in Madagascar, with the sole exception of section Hadrangis Schltr. Agreeing with Micheneau et al. (2008), Szlachetko et al. (2013) indicated that most of the sectional arrangements within the genus are unnatural. Based on molecular and morphological data, Szlachetko et al. (2013) raised most of the sections, sensuSchlechter (1918) and Garay (1973), to the rank of genus, albeit with a different placement for some species. However, their sampling of DNA material lacked more than 3/4 of the species of Angraecum (only 53 of the 221 currently recognised species were included in their phylogeny) and several sections were unrepresented. In all, Szlachetko et al. (2013) recognised 18 genera that included species previously placed within Angraecum, 12 of which resulted from raising sections to the generic level, five of which were described as new genera and one of which involved resurrecting a genus previously placed in synonymy.

Szlachetko et al. (2013) established the genus Pectinariella Szlach., Mytnik & Grochocka to accommodate the species of Angraecum sect. Pectinaria Benth. sensuGaray (1973) since the sectional name was already occupied at the generic level in another family (Pectinaria Haworth, Apocynaceae). They circumscribed Pectinariella to include all members previously assigned to sect. Pectinaria from continental Africa and the Gulf of Guinea islands as well as those from the western Indian Ocean islands. However, their molecular study lacked material of the type species, A. pectinatum Thouars and included only five of the ten species of Angraecum sect. Pectinaria: four from Africa (viz. A. doratophyllum Summerh., A. gabonense Summerh., A. pungens Schltr. and A. subulatum Lindl.) and one from Madagascar (A. dasycarpum Schltr.).

Cribb (2014) suggested that part of Angraecum sect. Pectinaria might be separated from Angraecumsensu stricto, given that the generitype, A. eburneum Bory, was placed in the large Malagasy/Mascarene clade identified by Micheneau et al. (2008). The polyphyly of the section was confirmed by Simo-Droissart et al. (2013) based on sequence data from all five currently recognised species from Africa (the four mentioned above plus A. atlanticum Stévart & Droissart), along with three species from Madagascar and the Mascarene Islands. Confirming the results of Micheneau et al. (2008), Simo-Droissart et al. (2013) showed that the African members of A. sect. Pectinaria formed a well-supported clade that appears to be most closely related to A. sect. Dolabrifolia Pfitzer and is not sister to the Malagasy/Mascarene clade, which includes the type species of the section.

Based on these results, Simo-Droissart et al. (2013) suggested that it would be necessary to remove the African species from Angraecum sect. Pectinaria in order to maintain its monophyly. The generic name Pectinariella proposed by Szlachetko et al. (2013) cannot, however, be applied to these African species since its type, P. pectinata (Thouars) Szlach., Mytnik & Grochocka, belongs to the clade comprising Malagasy/Mascarene taxa. Following the treatments of Szlachetko and Romowicz (2007) and Szlachetko et al. (2013), which treat the taxa generally placed in A. sect. Dolabrifolia as a genus, Dolabrifolia (Pfitzer) Szlach. & Romowicz and, since the African members of their polyphyletic genus Pectinariella are sister to those of Dolabrifolia, they must also be recognised at the rank of genus.

Here we thus propose a new genus to accommodate the species from continental Africa and the Gulf of Guinea islands of Pectinariella.

Although we have removed species from continental Africa and the Gulf of Guinea islands from Pectinariella, its delimitation remains open to further discussion. Garay (1973) included four species from Madagascar and the Mascarenes within Angraecum section Pectinaria (i.e. A. dasycarpum, A. hermannii (Cordem.) Schltr., A. humblotianum Schltr. and A. pectinatum). He regarded A. pterophyllum H.Perrier as a synonym of A. hermannii and placed A. panicifolium H.Perrier in Angraecum sect. Conchoglossum Schltr. In addition to the four species recognised by Garay (1973) as comprising section Pectinaria, Stewart et al. (2006) also included A. panicifolium. Micheneau et al. (2008) suggested that A. hermannii might best be placed in A. sect. Lemurangis Garay based on its morphological characters, which was followed by Cribb and Hermans (2009) and Verlynde et al. (2016). Cribb and Hermans (op. cit.) recognised A. pterophyllum as a distinct species and thus included five species within the section. As they followed the treatment of Garay (1973), Szlachetko et al. (2013) placed only four species from Madagascar and the Mascarenes within their newly described genus Pectinariella (P. dasycarpa (Schltr.) Szlach., Mytnik & Grochocka, P. hermannii (Cordem.) Szlach., Mytnik & Grochocka, P. humblotiana (Schltr.) Szlach., Mytnik & Grochocka and P. pectinata). In fact, these authors made no mention of Angraecum pterophyllum, but they did propose the combination Angraecoides panicifolia (H. Perrier) Szlach., Mytnik & Grochocka. Verlynde et al. (2016) described two new species of Pectinariella (P. edmundi Bosser ex Verlynde & Ramandimbisoa and P. scroticalcar Verlynde & Ramandimbisoa) and made the combination P. pterophylla (H.Perrier) Verlynde & Ramandimbisoa, thus recognising six species of Pectinariella from Madagascar and the Mascarenes. They did not mention Angraecum panicifolium or Angraecoides panicifolia, even though Simo-Droissart et al. (2013) placed it close to Pectinariella pectinata. Pectinariella thus comprises seven species, four of which have been included in published molecular phylogenetic studies. These seven species are found in Madagascar or neighbouring islands, including the type species, which occurs in La Réunion, Mauritius and the Comoros Islands.

In the studies of Simo-Droissart et al. (2013, 2016), species belonging to Dolabrifolia and Afropectinariella form two distinct, well-supported clades (bootstrap support (BS) = 100% and posterior probability (PP) = 1). The sister relationship of these two groups is also well supported by our Bayesian analysis (PP 1). Species assigned to Pectinariella also form a clade well supported in both analyses and are not sister to the clade comprising species of our new genus, Afropectinariella. The shared presence of elongate leaves that are never compressed, an ovary lacking a stalk and whose peduncle is scarcely developed appear to represent a convergence in Pectinariella and Afropectinariella. However, members of Afropectinariella possess a transversely oval lip that is wider than long, while those of Pectinariella have a lip that is longer that wide. Unpublished molecular data indicate that P. edmundi and P. scroticalcar are not sister to P. pectinata, suggesting that the genus, as circumscribed by Verlynde et al. (2016), may not be monophyletic. The fact that some Angraecum species assigned to section Pseudojumellea Schltr. are placed with P. pectinata in molecular phylogenies (see Micheneau et al. 2008) indicates that further studies will be needed to assess the monophyly of Pectinariella.

We are grateful to the American Orchid Society for financial support provided for the first author’s work in Cameroon. The U.S. National Science Foundation (Grant 1051547, T. Stévart as PI) also supported fieldwork, laboratory studies and herbarium visits. DNA sequencing was supported by the Belgian Fund for Scientific Research (F.R.S-FNRS) through grants FRFC 2.4.577.10 and MIS 4.519.10. Fieldwork was also funded by DIVEAC, ECOFAC, the Central Africa Regional Program of the Environment (CARPE) and “Sud Expert Plantes” project (project #375) under the French Ministry of Foreign Affairs. We are indebted to the authorities of the University of Yaoundé I and especially the Higher Teachers’ Training College who authorised us to build the orchid shadehouse in Yaoundé. We would like to thank Dr. Porter P. Lowry II for commenting and improving the English on an earlier version of the manuscript and the two reviewers for their constructive comments.