Polysciadis (Araliaceae) species nova e Madagascaria in honorem Luciani Bernardii felsinei nominata

Abstract Lowry II, P.P., M.W. Callmander & R. Spichiger (2020). Polysciadis (Araliaceae) species nova e Madagascaria in honorem Luciani Bernardii felsinei nominata. Candollea 75: 107–114. In English, English and French abstracts. Polyscias bernardiana Lowry & Callm., an easily recognized new species from east-central Madagascar, including the Ambatovy mine site and several protected areas, is described in honor of Luciano Bernardi (1920–2001), a 20th century explorer-naturalist who has contributed more to our knowledge of Indian Ocean Araliaceae than any other botanist. This new species most closely resembles Polyscias duplicata (Thouars ex Baill.) Lowry & G.M. Plunkett from Madagascar and the Comoro Islands in forming monocaulous trees with a terminal cluster of large leaves surrounding a globose inflorescence, and in having flowers and fruits with 8–10(–12) carpels and styles, but differs in having an articulated pedicel (vs. unarticulated in P. duplicata) and lacking auxiliary leaflets (vs. a second leaflet often inserted in the axil of the main leaflets in P. duplicata). Polyscias bernardiana is illustrated by line drawings and color photos, and a risk of extinction assessment is provided based on the IUCN Red List criteria, which indicates a conservation status of “Least Concern”.


Introduction
Over the last 35 years, Madagascar has been the focus of an extensive, multi-institutional effort involving Malagasy and international botanists to document this island nation's remarkable flora, which is currently estimated to comprise at least 14,000 species of vascular plants, c. 87% of which are endemic (Lowry et al., 2018). During this time, almost 200,000 collections have been made from throughout the island (Lowry et al., 2018), nearly tripling the total amount of material available for botanical research. Significant effort was made during this period to explore previously unknown parts of the island, leading to the discovery of hundreds of new species, but extensive collecting also targeted areas that had previously been the foci of botanical inventory work, generating many additional novelties. However, the part of the island with the best documented flora is the Ambatovy mine site, located in east-central Madagascar, c. 12 km northnorthwest of the town of Moramanga, which has been the target of intensive botanical field work, initially in support of a detailed environmental impact assessment, and then as part of a program to identify and manage key flora elements within the mining concession and an adjacent area managed as a conservation zone. More than 21,000 collections have been made to date from the Ambatovy area (Lowry et al., 2018;Madagascar Catalogue, 2020), representing nearly 1,600 species (Phillipson et al., 2010;Madagascar Catalogue, 2020), or about 11.5% of the entire Malagasy flora.
The family Araliaceae is well represented at Ambatovy, comprising a total of 18 species, including three species of Neocussonia (Harms) Hutch. and 10 species of Polyscias J.R. Forst. & G. Forst., along with five additional species of Polyscias that remain to be described (Phillipson et al., 2010;Madagascar Catalogue, 2020). Among these new species of Polyscias, one is particularly striking and frequently encountered at Ambatovy, forming tall, unbranched trees with a terminal tuft of large leaves and an expansive inflorescence. In the present paper, we describe this distinctive new Polyscias in honor of Luciano Bernardi, one of the most important 20th century students of Araliaceae, who has contributed more to our knowledge of the family in Madagascar than any other botanist.
Luciano Bernardi was born in Bologna, Italy, in 1920, anddied in Geneva, Switzerland, in 2001 (Bernardi, 1965(Bernardi, -1975 to study and collect material, and beginning in the late 1960s, he regularly published on the family, focusing in particular on Madagascar. Valet notare quod fere omnes eius libri in lingua latina scripti sunt [It is worth noting that nearly all of his works were written in Latin.] In parallel with his work on Araliaceae and until his death, Bernardi continued to collect throughout the world; he traveled to almost all countries in sub-Saharan Africa, and throughout much of South America, especially Paraguay and Peru. His herbarium collections have been widely used by systematists and are regularly cited in taxonomic publications. Bernardi was also a specialist in the history of botany and botanists, and he corresponded frequently with the greatest plant scientists of his time. Bernardi published ten papers on Araliaceae over a period from 1966 to 1980. These included descriptions of new species from Réunion Island (Bernardi, 1974a), comments on members of the family in the Mascarene archipelago as a whole (Bernardi, 1974b), and a review (in English) of Araliaceae in New Caledonia and their relationships with representatives in Oceania and Indonesia (Bernardi, 1979), but his most important contributions concerned Madagascar and the Comoro Islands. In his first paper on Malagasy Araliaceae, Bernardi (1966) Bernardi, 1969), and seven new species of Polyscias. This was followed by a revision of the genus Schefflera in Madagascar and the Comoros (Bernardi, 1969), including the description of three new species and two new varieties, along with eight new combinations in this genus, for which he adopted a broad circumscription that preceded the expansion of the genus by Frodin (1975), but that has since been massively reduced (see Plunkett et al., 2005), with the species from Africa and the Indian Ocean now treated in the recently resurrected genera Astropanax Seem. and Neocussonia (Lowry et al., 2017;Madagascar Catalogue, 2020). Two years later, Bernardi (1971) published a revision of Polyscias in Madagascar and the Comoros in which nine new species and three new varieties were described, and six new combinations were made. This paper also represented an important advance in the delimitation of this paleotropical genus, as Bernardi made a convincing case for adopting a broad concept that included many segregate genera recognized by other authors over the preceding century, an interpretation that has since been validated and expanded further (see . Over the course of the next several years, Bernardi published two new species of Schefflera (Bernardi, 1973(Bernardi, , 1980a and one of Polyscias (Bernardi, 1974c), followed by a remarkable and innovative synthetic work (Bernardi, 1980b) in the form of an illustrated diagnostic key (in French) to the 58 species and six varieties of Araliaceae he recognized from Madagascar and the Comoros, including five species of Cuphocarpus, one of Cussonia (restricted to the Comoros), and one of Gastonia Comm. ex Lam. (a genus now included in Polyscias), along with 35 species and 3 varieties of Polyscias, and 16 species and 3 varieties of Schefflera. In all, Bernardi described 31 species of Araliaceae from the Indian Ocean region, all but one of which are recognized today (Madagascar Catalogue, 2020), a fact that attests to the quality and accuracy of his work on this structurally complex and notoriously difficult family in a part of the world where it is particularly rich and diverse.

Holotypus
Etymology. -The species epithet honors Luciano Bernardi (1920Bernardi ( -2001, a 20th century explorer-naturalist (Fig. 3). He made c. 21,000 collections, duplicates of which are distributed in dozens of herbaria, including Geneva. Many of the expeditions led by Luciano Bernardi were conducted in extremely difficult and sometimes dangerous conditions. One can't help but think that his drive to produce results and his willingness to subject himself to considerable risk were related to his wartime experience as an officer in the paratroops of the Italian army, of which he was so proud. He also prided himself on studying only difficult plant groups, looking for challenges and disregarding the beaten track. His broad and deep scientific knowledge and culture did not prevent him from being interested in other things: for example, he knew the general and anecdotal history of all the regions he explored. Bernardi also respected botanical tradition, to the point of almost exclusively using Latin for his notes and writings, even when it was not compulsory. Deeply rooted in European religious tradition, culture and philosophy, Luciano Bernardi's existence is reminiscent of that of Guillaume d'Occam, the hero of Il nome della rosa [The Name of the Rose] by Umberto Eco, to which Luciano liked to refer (see Spichiger, 2002 for a more detailed necrology). Distribution and ecology. -Polyscias bernardiana is predominantly known from medium altitude moist evergreen forest in east-central Madagascar between 800 to 1200 m elevation (Madagascar Catalogue, 2020), although two collections were made below 800 m in lowland moist evergreen forest, one from the Zahamena protected area (Rakotonandrasana 496) and another from the Corridor Ankeniheny Zahamena protected area (Razanatsima 212). Our new species appears to grow on ultramafic substrate at some localities within the Ambatovy mine concession but occurs primarily on soils derived from paragneiss and schist elsewhere in its range (Roig et al., 2012). Notes. -Polyscias bernardiana resembles P. duplicata (Thouars ex Baill.) Lowry & G.M. Plunkett from Madagascar and the Comoro Islands in almost always forming monocaulous trees (rarely sparsely branched) that terminate in a cluster of large leaves surrounding a large, globose inflorescence/infructescence. These two species also have flowers and fruits with 8 -10(-12) carpels and styles (rarely as few as 6 or 7 in P. duplicata) and generally 8 -10(-12) petals and stamens, a feature shared with an as-yet undescribed species of Polyscias from the Col de Maningotry in extreme southeastern Madagascar. By contrast, most other members of the genus in Madagascar have a 5-carpellate gynoecium and 5 stamens, although a few have just 2, 3 or 4 carpels, styles, and stamens. Polyscias bernardiana can easily be distinguished from P. duplicata by the evident articulation on its pedicel at the point of insertion of the ovary, a feature found in all other species in Madagascar with the exception of P. duplicata, which has unarticulated pedicels and was therefore long placed in the segregate genus Gastonia, including by Bernardi (1971Bernardi ( , 1980b. Our new species also has strictly pinnately compound leaves whereas P. duplicata often has a somewhat smaller auxiliary leaflet borne in the axil of most of its main leaflets, such that 4 (rather than 2) leaflets are thus inserted at most nodes along the rachis, a feature reflected in the specific epithet.
While the geographic range of Polyscias bernardiana is rather restricted, it is often quite abundant in the areas where it grows and generally stands out due to its distinctive habit. It is thus surprising that only three fragmentary gatherings were made until the late 20th century even though P. bernardiana occurs at Andasibe (Perinet), one of the most easily accessible and frequently visited collecting sites in the country. Bernardi (1971Bernardi ( , 1980b associated one of these collections as well as several others with P. repanda (DC.) Baker, a species endemic to Réunion Island that only vaguely resembles P. bernardiana (Marais, 1984(Marais, , 1990, although given the poor quality of the Malagasy material available to him, this error is perhaps understandable and is in any case one of the very few details of Indian Ocean Araliaceae that Bernardi did not get right.