Taxonomic notes on species of Passiflora subgenus Astrophea ( Passifloraceae s . s . )

The subgenus Astrophea is one of the most critical and neglected subgenera of the genus Passiflora. Seven Brazilian species of P. subg. Astrophea were found to be in need of typification. Lectotypes are designated for the following names: Passiflora cauliflora, P. leptopoda, P. ceratocarpa, P. faroana, P. rhamnifolia, P. sclerophylla and P. tessmannii. An epitype was also designated for the name P. tessmannii. In addition, Passiflora alliacea was synonymized with Passiflora pentagona.


Introduction
The genus Passiflora, described by Carl Linnaeus in Species Plantarum (Linnaeus 1753), is the most important genus of the family Passifloraceae sensu stricto as it comprises a high number of species, several of which are of great economic significance due to their nutritional, medicinal and ornamental uses.
Passiflora subg.Astrophea stands out because it contains about 60 Neotropical species, with their center of diversity being in lowland areas of northern South America (Ulmer & MacDougal 2004).Brazil, with 31 species, is the country with the highest number of taxa (Mezzonato-Pires 2017).Besides having geographical distributions that include places that are difficult to access, species of Astrophea have morphological characteristics that make it difficult to see and collect them in the field.Such characteristics include their liana habit, which allows them to even reach above the canopies of tall trees, as well as their entire leaf blades that possess two inconspicuous leaf glands.
Thus, P. subg.Astrophea represents the largest gap in the knowledge of the genus Passiflora, and thus studies that add to the collective knowledge of these species are needed.This work was carried out as apart of the project "Flora of Brazil 2020", whose objective is to provide data on type material of the plants described in Brazil.As a result, seven names belonging to P. subg.Astrophea were found in need of lectotypification, and a new synonym is proposed.

Materials and methods
The study was based on type material and/or photographs of types provided by the following herbaria: B, C, F, INPA, G, K, MBM, MG, RB, S, UEC, US (acronyms are according to Thiers 2018).The collector of the type of P. cauliflora was erroneously listed as Tessmann by the site http://www.tropicos.org,(whose material of "photo-negatives" is deposited in herbaria B and F).Based on the protolog the collector was Ule (Ule 6679).According to Escobar (1990), the curator of herbarium B, reported the deposited type specimen was destroyed during World War II.Therefore, the material of MG, that was previously considered an isotype, is herein lectotypified.Two syntypes are mentioned in the protolog of P. leptopoda.According to Killip (1938) the type of P. leptopoda was deposited in B, but according to Escobar (1990), this material was destroyed.Although there is still a small blackand-white photograph on the website http://www.tropicos.org, it was considered more prudent to lectotypify the material deposited in MG that was previously recognized as an isosyntype.Killip (1938) was deposited at B and was destroyed during World War II, according to Escobar (1990).Thus, it became necessary to designate another lectotype whose specimen is deposited in RB.It is important to note that during the first lectotypification, Killip (1938) misidentified the location and wrote "Jupiry" instead of "Infiry", which is the correct name of the collection locality.5. Passiflora rhamnifolia Mast., Fl.Bras.13(1): 545, 575.1872 -Lectotype (designated here): Brazil, São Paulo, Cubatão, s.d., Sello 2148 (US).

Passiflora leptopoda
In describing P. rhamnifolia, Masters (1872) cited two collections as syntypes -Sello 2148 and Sello 2125both from the municipality of Cubatão, state of São Paulo (however, it is believed that due to a misunderstanding, the state of Minas Gerais was written in the protolog).Killip (1938) lectotypified Sello 2148, which was deposited in B. According to Escobar (1990), the material was destroyed.Thus, a new lectotype is here indicated -the specimen Sello 2148 of US.Two collections of syntypes are cited in the protolog of P. sclerophylla; one from Guyana, Venezuela and Roraima, collected by Ule (Ule 8667), and the other from Roraima, collected by Schomburgk (Schomburgk 986).The former material was later designated by Killip (1938) as a lectotype (Ule 8667) and deposited in B, where it was destroyed.Thus, we assign a new lectotype, deposited in MG.The holotype designated by Harms (Tessmann 4385) for P. tessmannii was deposited in B. There are fragments and a black-and-white photograph of this specimen in the US.The holotype deposited in B was destroyed, according to Escobar (1990), and through consultation of the virtual collection available from B (http://ww2.bgbm.org/herbarium/default. cfm), no type material was found for this species.Therefore, the remaining fragments of the isotype deposited in US are herein lectotypified.The existing black-and-white photograph of P. tessmannii does not faithfully depict its main morphological characteristics, for example: corona with two series of filaments; external filaments linear with broadly falcate apices with attenuate tips, and subulate internal filaments; and the presence of a trochlea.Since the lectotype alone does not present the full set of diagnostic characteristics, C. A. Sothers & E. C. Pereira 612, material collected in the Adolpho Ducke Forest Reserve, state of Amazonas, Brazil, was designated as an epitype, which is deposited in INPA.vol.1, tab. 7 (1901).
The following syntypes are mentioned in the protolog for the species: Lund 1162, Martius 1221 and Luschnath s.n.from Rio de Janeiro; Weddell 1216 from the province of Minarum; and Sello s.n.from an unspecified area in southern Brazil.Killip (1938) proposed the lectotypification of Warming 1162 (C) for this species.However, the label of the lectotype indicates Lund as the collector, suggesting that there was a mistake during the writing of the name of the collector.
Passiflora alliacea Barb.Rodr.was described in 1901, with the type locality being a restinga environment in Rio de Janeiro.It was recorded as blooming in December and fruiting in January.The specific epithet alliacea was applied based on the garlic smell released by its fruits.After thorough comparative analyses of the type descriptions, specimens and illustrations, we conclude that P. alliacea is conspecific with P. pentagona (Fig. 1).The diagnostic characteristics described for P. alliacea are retroflexed flowers with a 6-angled hypanthium and broadly bifid internal corona filaments; all of which are also found in P. pentagona.In his work on Passifloraceae of the Americas, Killip (1938) noted that distinguishing these species might not be possible, although he did not synonymize them.Therefore, P. alliacea is herein designated a new synonym of P. pentagona.The diagnostic characteristics of P. pentagona are: leaf blades that are oval-elliptical, ovate, ovate-oblong, adaxially glabrous (except along the midvein) abaxially puberulous, and with transparent trichomes; solitary greenish-white flowers with cylindrical-campanulate hypanthia; corona with two sets of filaments that are beige with burgundy spots, the first series being dolabriform with long and attenuated apicies and slightly undulate margins, the second series being linear with bifid or clavate apices; operculum that is erect, filamentous, not tubular, inserted, and with a fimbriate apex; and fusiform trochlea with a spotted region above which are trichomes.