Typification of names of South American taxa related to Woodsia montevidensis (Woodsiaceae)

Abstract A revision of the nomenclature of six South American taxa related to Woodsia is presented, as a part of a taxonomic revision of the genus in South America. Lectotypes are selected for Cheilanthes crenata, Woodsia crenata var. pallidipes, Woodsia incisa, Woodsia montevidensis var. fuscipes and the second step lectotypification for Dicksonia montevidensis and Woodsia peruviana, based on the analysis of their protologues and original herbarium material. All names are currently synonyms of Woodsia montevidensis. Physematium incisum (Gillies ex Hook. & Grev.) Kunze constitutes an illegitimate name and Physematium cumingianum is considered as nomen inquirendum.


Introduction
Woodsia (Polypodiidae: Woodsiaceae) is a genus comprising about 35-40 taxa of small and tufted ferns growing on or in the vicinity of cliff s and rocks (Brown 1964). Th ey mainly occur in montane areas in the Northern Hemisphere, but few species are also present in South America and Southern Africa. Th e greatest species richness is found in the Rocky Mountains of western North America (ca. 10 spp.) and the Himalayas in south Asia (ca. 19 spp.); absent from Australia, New Zealand, SE Asia, and the Pacifi c (Kramer 1990, Shao et al. 2015, Shmakov 2015. Th e Neotropical species belong to Woodsia subgenus Physematium (Kaulf.) Hook. emend. X.C. Zhang & R. Wei (Shao et al. 2015, Shmakov 2015. South American fl oristic works cite the genus Woodsia as being represented by a single, morphologically variable species, Woodsia montevidensis (Spreng.) Hieron., with a distribution ranging from Venezuela and Colombia to central Argentina andSouthern Brazil (de la Sota 1977, Tryon andStolze 1991).
In anticipation of the recent eff orts to prepare fl oristic inventories for Neotropical and Andean regions of South America, especially the Flora of Argentina (http://www. fl oraargentina.edu.ar) and the Flora of Brazil (Mynssen 2016), the nomenclature of taxa related to the genus Woodsia from South America was re-examined, and we are here providing lectotypes for four names, and two second step lectotypifi cations with the aim of enhancing nomenclatural stability, following as closely as possible the authors' original intentions.

Material and methods
We have analysed the protologues and morphological features from specimens of the following herbaria:  .

Typifications
In this paper we have arranged the South American taxa related to Woodsia in alphabetical order by the names under Woodsia, as all of them are current synonyms of Woodsia montevidensis (Spreng.) Hieron., following Brown (1964), de la Sota (1977, and Tryon and Stolze (1991 Since Kunze's own herbarium in Leipzig is destroyed, we looked for further original material of Cheilanthes crenata from Huanuco, Peru collected by Poeppig at B, BM, K, L, LE, MO, NY, OXF, P, PCR, US and W. We were able to fi nd original material at W that agrees well with Kunze`s original description, which we here select as lectotype, in order to avoid the misapplication of the name. Th e type material at BM consists of four fronds with two diff erent barcodes on the same sheet: BM [BM-000937850 and BM-000937851], both with separate labels with the same information. We selected the material affi liated with BM [BM-000937851] as lectotype because it is more complete.
Th e specimen J. Gillies 8 housed at K [K-000229420!] is not part of the original material because it was collected at "Sierras de Tandil", located in Buenos Aires province, Argentina, far away from the type locality.
Th e combination Physematium incisum (Gillies ex Hook. & Grev.) Kunze (Kunze 1837) is an illegitimate name, posterior to Presl's combination.  & Hieronymus 138 (B [B-200171577!]) is selected as lectotype, while the three duplicates are regarded as isolectotypes in accordance to Art. 9.12 of the Code (McNeill et al. 2012). Also, the lectotype chosen shows the characters used to delimitate the variety and bears an annotation by Hieronymus "n. var," and handwritten locality data.  Tryon and Stolze (1991: 94) typifi ed Woodsia peruviana with a specimen at K as holotype, but K holds two sheets of Mathews 602 [K-000632730 and K-000632731], the last one is here designated as lectotype because the material is more complete, has a handwritten annotation "Peru, Mathews" and "Woodsia peruviana Hook. Spec. Fil. Tab. XXI" on the sheet, and the label contains the locality data.
Th e specimen A. Mathews s.n. (US [US-00067001!]), according with Taylor's annotation in the label of the specimen, could probably be part of the type collection, but we prefer to exclude it of lectotypifi cation because the locality is not clear (only "Peru" is written in the label) and it is not originally numbered by Mathews.

Unresolved name
Woodsia cumingiana (Kunze)  Th e original material of this species, deposited in LZ, was destroyed. As Kunze (1837) observed in the protologue: "unicum vidi specimen observed", there is not referable isotype or even an illustration of the species. According to Stafl eu and Cowan (1979), the original material of H. Cuming is kept at BM; however, no syntypes were found in this herbarium, nor in the Herbarium Hookerianum (K). Additionally, no material of this species from Chile was in B, BR, BM, E, GH, L, LE, OXF, P, W and Z, where duplicates of H. Cuming are deposited. Also, as consigned by Hooker (1844), most probably the type locality is mistaken, because he had the opportunity to revise the collections of Cuming immediately after his return and he was not able not fi nd any specimen gathered by Cuming from either Chile or Peru. As the protologue expresses, the species is characterized by last segments oblong rounded, glandulosodentate decurrent, sori solitary upon the teeth, involucres glabrous, rachis and stipe subglabrous purple. With such description, the species is hardly to diff erenciate from many species of Woodsia, hence, the name is considered here as nomen inquirendum.