Typification and synonymy of the Atlantic Forest endemic species Napeanthus primulifolius (Gesneriaceae)

During the nomenclatural revision of Acanthaceae names described by Friar J.M.C. Vellozo in his Florae fluminensis, we realized that Pedicularis acaulis Vell. was conspecific with Napeanthus primulifolius (Gesneriaceae), a Brazilian Atlantic Forest endemic species. This study presents the complete and updated synonymy of N. primulifolius, including two new synonyms: Pedicularis acaulis and Oreocharis notha C.B.Clarke. We also propose a lectotype for N. primulifolius based on G.Raddi’s specimens and P. acaulis based on Vellozo’s original plate.


INTRODUCTION
Napeanthus Gardner is a Neotropical genus of Gesneriaceae comprising 20 species (and possibly another ten undescribed ones) occurring in tropical rainforests of Central and South America (Leeuwenberg 1958;Weber 2004;Wiehler 1983). The genus shows a conspicuous number of species confined to the Andean foothills (Bolivia, Colombia, Peru, and Venezuela) and the Guiana Shield (Suriname, Guyana, French Guiana, and northern Brazil), besides two disjunct species in the Atlantic Forest of eastern Brazil (based on Leeuwenberg 1958;1971;Skog 1974;Feuillet and Skog 2002). Napeanthus had already been previously pointed out by Leeuwenberg (1958) as a unique genus among New World Gesneriaceae, later allocated by Wiehler (1983) in its own tribe, Napeantheae Wiehler. Recently molecular phylogenetic studies support this monogeneric tribe as the first diverging lineage in the subfamily Gesnerioideae Burnett, recovered as sister to clade consisting of tribes Beslerieae Bartling & Wendl.f. + Titanotricheae Yamaz. ex W.T.Wang (Smith 2000;Möller and Clark 2013;Luna et al. 2019;Ogutcen et al. 2021). Species of Napeanthus are distinguished from other New World Gesneriaceae by being rosette herbs, generally acaulescent, with grouped stomata (in islands), calyx generally accrescent in fruit, stamens typically 4 (plus a staminodium), nectary absent, and dry capsules with loculicidal or septicidal dehiscence (Leeuwenberg 1958;Wiehler 1983;Weber 2004).
Napeanthus was described by Gardner (1843) to accommodate a peculiar species from the Atlantic Forest of Serra dos Órgãos, in the State of Rio de Janeiro, Brazil, named N. brasiliensis Gardner. Nonetheless, an obscure species described in Oxalidaceae ca. 20 years before Napeanthus by Raddi (1820), Oxalis primulifolia Raddi [as "primulaefolia"], was pointed out by Sandwith (1956) as conspecific with N. brasiliensis, being the accepted and correct name for this species. Alongside N. reitzii (L.B.Sm.) Burtt ex Leeuwenb., they are the only known species of Napeanthus distributed in the Brazilian Atlantic Forest (Leeuwenberg 1958;Chautems 1991;Chautems 2003;Hinoshita et al. 2018).

Distribution and habitat
Napeanthus primulifolius is endemic to eastern Brazilian Atlantic Forest understories, especially on steep banks near streams where plants are rooted in earth among rocks. It was recorded in southern Bahia and in Serra do Mar Mountain Range (from Rio de Janeiro to northeastern Paraná).

Taxonomic and nomenclatural notes
Vellozo (1829) described four species under Pedicularis L. (currently placed in Orobanchaceae) in his "Florae fluminensis" following the Linnaean arrangement "Didynamia, Angiospermia." Three of them corresponded to species of Acanthaceae, and one of them, Pedicularis acaulis Vell., which was annotated by Vellozo (1829) as of doubtful placement and possibly belonging to a different genus, does not. Pedicularis acaulis Vell. is a later homonym of P. acaulis Scop., an accepted name in Orobanchaceae for an alpine species from the Italian Alps to the northwestern mountains of the Balkan Peninsula (Mayer 1972). Since the description of P. acaulis Vell., this name has never been mentioned in any taxonomic treatment for Brazilian Lamiales and remained forgotten until this moment (pers. observ.). However, when we analyzed the original description (Vellozo 1829) and the later published original plate (Vellozo 1830) hosted at Biblioteca Nacional [National Library of Brazil] (Fig. 1B), we realized that some morphological characters, i.e., rosette habit, acaulescent stems, sessile leaves, cymose inflorescence simple or compound (with at least one of the lateral flowers replaced by a new peduncle), and the calyx lobes overlapping in the margin, matched those from the small gesneriad genus Napeanthus. Of the only two Napeanthus species in eastern Brazil, i.e., N. primulifolius, which occurs from Bahia to northeastern Paraná, and Napeanthus reitzii (L.B.Sm.) B.L.Burtt ex Leeuwenb., from Paraná and Santa Catarina], P. acaulis Vell. is conspecific with N. primulifolius, due to their rosette habit, acaulescent, pseudoverticillate leaves grouped at the stem apex, cymose inflorescences simple (often 4-flowered) or compound (with at least one of the lateral flowers replaced by a new peduncle) (Fig. 3A-C), flowers with long pedicels, and calyx lobes elliptical. Napeanthus reitzii is a herb with elongated stems, leaves distributed along the stem (with relatively short internodes), inflorescence cymose simple, 3-flowered (dichasium) (Fig.  3D), flowers with short pedicels, and calyx lobes trullate. Based on these morphological characters and geographic distribution, we propose the synonymy of P. acaulis Vell. under N. primulifolius. Furthermore, because there are no preserved specimens of Vellozo's names published in "Florae fluminensis" (see Lima 1995) and the illustrations (i.e., Vellozo 1831) were published after the protologue (i.e., Vellozo 1829) and are not considered part of the author's original material, in accordance with the Art. 9.4 and related of the International Code of Nomenclature for algae, fungi, and plants (Shenzhen Code) -ICN (Turland et al. 2018), we chose Vellozo's original plate as the lectotype. Leeuwenberg (1958) referred to G. Raddi's specimens deposited at PI herbarium only as a "holotype" and "isotype," without any reference to the herbarium catalog number or even without any explicit indication in the sheets. Thus, we chose one of these specimens as a second-step lectotype, according to Art. 8.1 and 9.17 of the ICN (Turland et al. 2018).
When Clarke (1883) described Oreocharis notha, he was unsure of where this specimen was collected (possibly from Manila, Philippines) and the generic placement of this species, indicated in the protologue with a question mark. This author also pointed out that the long calyx and ovoid ovary in this species were different from other Oreocharis Benth. In fact, the calyx with long lobes accrescent in fruit is not found in any known species of Oreocharis s.str., not even in genera recently synonymized by Möller et al. (2011), which had calyx lobes shorter than the corolla. On the label of the specimen studied by Clarke (1883), housed at P herbarium, there is the following handwriting, possibly written after this author: "Napeanthus, ex h. Bory, A.S.-H. [August Saint-Hilaire], Brasilia!." It is likely to be a duplicate of a gathering of A. Saint-Hilaire incorporated into Bory's collection for being an endemic species to eastern Brazil and unlikely cultivated as ornamental in the Philippines. Also, the Napeanthus specimen in the Saint-Hilaire collection and the type of Oreocharis notha in Bory's collection, both at P, show inscriptions with the same handwriting.