The Genus Paronychia (Caryophyllaceae) in South America: Nomenclatural Review and Taxonomic Notes with the Description of a New Species from North Peru

All the names in Paronychia described from South America are investigated. Five names (P. arbuscula, P. brasiliana subsp. brasiliana var. pubescens, P. coquimbensis, P. hieronymi, and P. mandoniana) are lecto- or neotypified on specimens preserved at GOET, K, LP, and P. The typification of nine names, first proposed by Chaudhri in 1968 as the “holotype” are corrected according to Art. 9.10 of ICN. Three second-step typifications (Art. 9.17 of ICN) are proposed for P. camphorosmoides, P. communis, and P. hartwegiana. The following nomenclatural changes are proposed: P. arequipensis comb. et stat. nov. (basionym: P. microphylla subsp. microphylla var. arequepensis), P. compacta nom. nov. pro P. andina (Philippi non Gray; Art. 53.1 of ICN), P. jujuyensis comb. et stat. nov. (basionym: P. hieronymi subsp. hieronymi var. jujuyensis), P. compacta subsp. boliviana comb. nov. (basionym: P. andina subsp. boliviana), and P. compacta subsp. purpurea comb. nov. (basionym: P. andina subsp. purpurea). A new species (P. glabra sp. nov.) is proposed based on our examination of live plants and herbarium specimens. P. johnstonii subsp. johnstonii var. scabrida is synonymized (syn. nov.) with P. johnstonii. Finally, P. argyrocoma subsp. argyrocoma is excluded from South America since it was based on misidentified specimens (deposited at MO) of P. andina subsp. andina. A total of 30 species (43 taxa including subspecies, varieties, subvarieties, and forms) are recognized, highlighting that for some (Paronychia chilensis, P. communis, P. setigera) we provisionally accept Chaudhri’s infraspecific classification, since the high phenotypic variability of these taxa is quite complicated and further investigations need to solve their taxonomy.


Taxonomic Treatment
Names of the taxa are numbered (Arabic number: 1, 2, etc.) and presented in alphabetical order. Infraspecific taxa are organized using letters (a, b, etc.), a second Arabic number after letters (a1, a2, etc.), and Roman numbers (I, II, etc.).  Notes. After the detailed description, Triana and Planchon [30] provided the provenance and two collectors ("Andes de Bogota, la Peña, alt. 2077 mètres, dans les endroits sablonneux (Tr.); Bogota (Goudot)"); these citations represent syntypes according to Art. 9.6 of ICN. We found seven specimens at COL (COL000001167), FI (FI005352), K (K000486397), MPU (MPU012438), and P (P00335870, P00335871, P00335872). P00335872 was collected by J. Goudot in 1844, MPU012438 was annotated with "Paronychia bogotensis nob.|legit. Triana", whereas the other five specimens are part of the "HERBARIO DE J. TRIANA" (COL000001167; date of collection May 1855) or of the "VOYAGE DE J. TRIANA 1851-1857" as reported in the original labels. By the comparison among these scripts, the label occurring on the specimen MPU012438 (https://plants.jstor.org/stable/viewer/10.5555/al. ap.specimen.mpu012438, accessed on 19 February 2023) was not annotated by J. Triana and we prefer to exclude this specimen as a lectotype. The other specimens are, instead, clearly part of the original material used by [30] to describe Paronychia bogotensis. The main set of Triana's herbarium and type are currently preserved at COL [31]. Therefore, we here designate COL000001167 as the lectotype of the name Paronychia bogotensis; FI005352, K000486397, P00335870, and P00335871 are isolectotypes. The types match the diagnosis and correspond to the current concept of the species (see e.g., [11] Notes. Chaudhri [11] stated "Type: Argentine: Buenos Aires (Bonaria), Commerson ix.1767, (holo. G-DC! iso! P!)". However, Candolle [32] did not indicate any holotype in the protologue of Paronychia brasiliana (Art. 9.1 of ICN) but reported a syntype, i.e., "Cette plante a été trouvée par Commerson dans les chemins de Monte-Video . . . (V. s.) [=vidi siccus]". According to Art. 9.10 of ICN, Chaudhri's use of "holo" (=holotype) is an error to be corrected to lectotype. The indication of the P specimen as an isotype ("iso") should be corrected to isolectotype. Furthermore, we traced five pertinent specimens at P (P00712715, P00712716, P00712717, P00712718, P00712719), all part of Poiret's collection and collected by P. Commerson in Montevideo/Buenos Aires. So, all these five specimens are isolectotypes for the name Paronychia brasiliana. Finally, a further isolectotype was found at MPU (MPU011320). All the original material found matches the diagnosis and corresponds to the current concept of the species (see e.g., [11]).
Paronychia bonariensis was published by Candolle [33] in the 3th volume of his Prodromus where, after the description, the provenance ("in Bonariâ, et Monte-Video?"), the collector ("Commerson") and the citation "Paronychia Brasiliana DC. in Lam. dict. 5 p. 23" were given. This latter statement refers to the 5th volume of Lamarck's (1804) Encyclodepiè Botanique where P. brasiliana was validly published. So, Candolle [33] provided a new name for the legitimate P. brasiliana. Hence, the name P. bonariensis is superfluous and illegitimate according to Art. 52.1 of ICN. Note that [33]  Notes. Chaudhri [11] described this variety providing a short diagnosis ("caulium ramificationibus retrorse pubescentibus distinguenda"); the following type indication was given: "Type: Argentine: Prov. Buenos Aires, Monte Hermoso, /1916, E. Carette (LP)". We traced two specimens at LP (LP003078 and LP006785) both collected by E. Carette in 1916 in Monte Hermoso. According to Art. 40.2 Note 1, since these two collections are clearly part of the same gathering, the name Paronychia brasiliana subsp. brasiliana var. pubescens was validly published and Chaudhri's citations are syntypes (see also Art. 9.6 of ICN). As a consequence, a lectotype should be designated (Art. 9.3 of ICN). We here designate LP003078 as the lectotype of the varietal name since it appears to be better preserved and includes more inflorescences whose features are important in the taxonomy of the genus Paronychia (see e.g., [11]). LP006785 is the isolectotype. These two LP specimens match the diagnosis and correspond to the current concept of the species (see e.g., [11]).
Taxa brasiliana and pubescens have different and not overlapping distribution areas, i.e., respectively in Argentina (Córdoba Province), South Africa, and Australia [34] and Buenos Aires Province in Argentina. So, we here propose to treat these two taxa at the subspecies rank of Paronychia brasiliana. Note. Cambessèdes [34] gave a diagnosis, a detailed description, and the provenance ("Prope Egreja Velha in part provinciae S. Pauli dictâ Campo Gereaës"). According to Art. 9.10 of ICN, [11] the use of "holo" (=holotype), for a P specimen collected by A. de Saint-Hilare (no. 1513) is an error and it should be corrected to lectotype. Note, however, that there are two specimens at P (P00712722 and P00712723) bearing the no. 1513: P00712723 bears a complete label with locality and date of collection, whereas the label of P00712722 just reports "Paronychia camphorosmoides Cambess.". However, the occurrence of a small label with "1513" on the exsiccatum of P00712722 led us to consider it as part of the original material. So, Chaudhri's designation can be considered as the first-step typification and a second step is required by Art. 9.17 of ICN. P00712723 is an isolectotype. The types match the diagnosis and correspond to the current concept of the species (see e.g., [11] Notes. Paronychia chilensis was validly described by Candolle [33] by a diagnosis; the provenance and collector ("Conceptionis urben in Chili, legit cl.

Paronychia cabrerae
[clarissimo] D'Urville") were also provided; moreover "v. s." (=vidi siccus) was also reported, indicating that A. P. Candolle had seen at least one specimen. Candolle's citation of provenance and collector can be considered as a syntype according to Art. 9.6 of ICN. Chaudhri [11] wrongly reported "holo. G-DC!" for a specimen collected in Conceptión by D'Urville. In fact, as stated above, [33] did not indicated any holotype for P. chilensis (Art. 9.1 of ICN). According to Art. 9.10 of ICN, Chaudhri's use of "holo" (=holotype) is an error to be corrected to lectotype. Further original material was traced at P (P00712724) and it bears some pieces of a single plant collected by D'Urville in Conceptión, as reported on the original label (bottom-left corner of the sheet). This P specimen is an isolectotype. Finally, in the JSTOR database, there are two specimens (MA811550 and MA811551) indicated as "original material" (see https://plants.jstor.org/stable/viewer/10.5555/al.ap.specimen.ma811550, accessed on 19 February 2023 and https://plants.jstor.org/stable/viewer/10.5555/al.ap. specimen.ma811551, accessed on 19 February 2023). However, these two MA specimens cannot be considered for the lectotypification purpose, being part of the Ruiz and Pavon's Herbarium Peruvianum and no reference to D'Urville or Conceptión (Chile) is reported in the label.
Chaudhri [11] recognized two subspecies, i.e., Paronychia chilensis subsp. chilensis and P. chilensis subsp. subandina (Phil.). Chaudhri based them on stem hairiness, style length, and stigma structure; P. chilensis subsp. chilensis was further classified into two varieties, i.e., P. chilensis subsp. chilensis var. chilensis and P. chilensis subsp. chilensis var. mutica (Phil) Reiche, based on hairiness and shape of the leaves. The phenotypic variability of P. chilensis is quite complicated and, based on our observations of herbarium specimens, we prefer to avoid a taxonomic conclusion about these taxa. So, for the moment, we accept Chaudhri's classification, awaiting further investigations. Note. Chaudhri [11] reported "holo. SGO. 48927!" for a specimen collected in in valle fluminis Torca. However, Philippi [35] did not indicate any holotype for P. subandina (Art. 9.1 of ICN). According to Art. 9.10 of ICN, Chaudhri's use of "holo" (=holotype) is an error to be corrected to lectotype. Notes on Paronychia mutica. Philippi [35] provided a description and the provenance ("Prope los Molles in parte litorali prov. Aconcagnae invenit orn. Landbeck"). In addition, [11] indicated a specimen deposited at SGO (barcode SGO000001974) as the holotype. According to Art. 9.10 of ICN [29], Chaudhri's use of "holo" (=holotype) is an error and it should be corrected to lectotype. We also traced an isolectotype (SGO000001975 Notes. Gay [25] gave a diagnosis, a detailed description, and the provenance ("departamento de Coquimbo y especialmente en el camino de la Serena á Arqueros") for his Paronychia coquimbensis. Five specimens were found at K (K000486388 and K000486389) and P (P00712726, P00712727, and P00712728). K000486388, K000486389, and P00712728 each bear an original label without the date of collection and, therefore, we cannot be sure that they are ante-1846 additions to the collection and part of the original material for P. coquimbensis. The other traced specimens (P00712726, P00712727, and P00712728), all collected by C. Gay in Chile, in 1838 and 1833, respectively, are eligible as lectotypes. Since P00712726 is represented by a larger exsiccatum than P00712727 and P00712728, also including many flowers, whose features are important in the identification of Paronychia species, we here designate P00712726 as the lectotype of the name P. coquimbensis. The type matches the diagnosis and corresponds to the current concept of the species (see e.g., [11]). Note that P00712727 and P00712728 each bear a printed "ISOTYPE" label which, however, has no effect on the lectotypification never being published on the basis of our literature analysis. Finally, the word "Cotypus" (reported in a separate label of P00712728), which is not formally regulated by the Shenzhen Code, is an obsolete term meaning syntype (or sometimes isotype or paratype; see [29]) and again it does not affect our proposed lectotypification and a second step is required by Art. 9 Notes on Paronychia appressa. Philippi [35] provided a description and the provenance ("Prope Illapel crescit, unde specimina attulit orn. Landbeck"). Additionally, [11] indicated a specimen deposited at SGO (SGO000001971) as the holotype. According to Art. 9.10 of ICN, Chaudhri's use of "holo" (=holotype) is an error and it should be corrected to lectotype. We also traced an isolectotype (SGO000001970 Notes. Cambessèdes [34] published Paronychia communis by a diagnosis and a detailed description; the proveannce (" . . . provinciae S. Pauli dictâ Campo Geraës . . . provinciarum S. Catharinae, Rio Grande de S. Pedro do Sul, Cisplatinae, et Missionum") was also given. Chaudhri [11] cited a specimen collected by St. Hilare in Brazil (deposited at P) as "lecto.holo.". This strange statement was used by [11] two times in his Revision of the Paronychiinae (the second one for P. chionaea Boiss. subsp. chionaea var. chionaea, on page 238). So, we think that "lecto.-holo." is not a typographical error, but would represent Chaudhri's doubt in type designation. In the case of P. communis, the specimen cited by [11] was traced (P00335873) and it bears two plants associated with two collection numbers (1511 and 2030bis) as also indicated by [11]; two labels also occur on the sheet, i.e., "Paronychia communis F. B.|Herb. Bresil.|St. Hilaire" (linked with plant no. 1511, on the left of the sheet) and "Brésil-Sao Paulo, campos Geraes" (linked with plant no. 2030 bis, on the right of the sheet). Note that Art. 8.2 Note 1 states " . . . collecting numbers . . . alone do not necessarily denote different gatherings". However, the next article of the Shenzhen Code (Art. 8.3) reports that "Multiple preparations from a single gathering that are not clearly labelled as being part of a single specimen are duplicates". In our case, the occurrence of two plants ("preparations" according to Art. 8.3) and two labels does not lead us to consider them as surely part of a single gathering. Therefore, Chaudhri's type indication (holotype or lectotype) is not correct in our opinion. According to Art. 9.19 of ICN, a lectotypification is necessary. Note that we traced a further two specimens which are part of the original material for Paronychia communis, i.e., P00712713 and P00335874 (https://plants. jstor.org/stable/viewer/10.5555/al.ap.specimen.p00335874?loggedin=true, accessed on 19 February 2023) both refer to plants collected by St. Hilare during his trip to Rio Grande do Sul between 1816 and 1821 (as reported on the original labels); both these specimens have the same collection number, i.e., "2659bis". We here designate the specimen P00335873 as the lectotype of the name P. communis since it bears a further handwritten label (on the bottom-left corner of the sheet) including the generic name "Paronychia" (without a specific epithet), a description, and the locality of collection ("In pascuis prope . . . S. Francisco de Borja" (São Borja), a city in the north-west of the Brazilian state of Rio Grande do Sul). The type matches the diagnosis and corresponds to the current concept of the species (see e.g., [11] Etymology. The specific epithet refers to the compactness of the plants, which is a unique habit feature amongst other Paronychia across South America. Notes. Paronychia andina was validly described by Philippi [36] who provided both a diagnosis and a detailed description; the provenance ("Hab. Baños; Casa Cancha to Culnai and Alpamarca, in the high Andes of Peru (also gathered in the same region by Mr. M'Lean, and at Cerro Pasco by Mr. Matthews)") was also given. These citations represent syntypes according to Art. 9.6 of ICN. Despite this, Chaudhri [11] indicated a specimen deposited at SGO (SGO000001969) as the holotype. According to Art. 9.10 of ICN, Chaudhri's use of "holo" (=holotype) is an error and it should be corrected to lectotype. However, [11] chose the following specimen as type: "Andes inter Mendoza et Chile, Portezuelo del Portillo, Mendoza. Reed, i.1870". Since this specimen was collected in Chile, but the protologue of P. andina reports localities from Peru, SGO exsiccatum cannot be considered as part of the original material and, as a consequence, is not usable as a lectotype (Arts. 9.3 and 9.4 of ICN). We traced the following relevant specimens: All the specimens listed above were originally annotated as "Paronychia andina n. sp." and they can be considered as part of the original material for Philippi's name (note that it is a later homonym of P. andina A. Gray [37] and, therefore, illegitimate under Art. 53.1 of ICN). NY00342590 appears to be the better-preserved specimen and it is here designated as the lectotype of Philippi's name Paronychia andina. The type matches the diagnosis and corresponds to the current concept of the species (see e.g., [11]).
Etymology. The specific epithet glabra refers to the lustrous and hairless leaves of the plant, which is a unique feature almost unknown among South American species of Paronychia.
Taxonomic discussion. Paronychia glabra is unique among the known neotropical species because of the long stems bearing glabrous and entire leaves. A characteristic feature that is well known in most of the South American species is the leaf bearing trichomes and ciliate margins. P. glabra differs from its closest relative P. hieronymi by Pax [40] by (Figures 4 and 5) the larger size of the stems (100 cm in P. glabra vs. 10 cm in P. hieronymi), leaves in P. glabra being glabrous and bearing entire margins which can rarely be shortly hirsute (vs. ciliated margins in P. hieronymi), aristulate leaf apex (vs. acute-acuminate in P. hieronymi), upper stipules bearing trichomes (vs. entire stipules), glomerules bearing 1-3 flowers (vs. 3-13 flowers in P. hieronymi), larger size of the flowers (2.2-3.0 mm vs. 1.75-2.00 mm in P. hieronymi), by the longer and densely scarious sepals (2.0-2.8 mm vs. 1.0-1.5 mm and semi-scarious in P. hieronymi), longer petals (1.2-1.8 mm vs. 0.75-0.90 mm in P. hieronymi), longer anther (0.5-0.7 mm vs. 0.35 mm in P. hieronymi), longer style size (0.3-0.6 mm vs. 0.6-0.9 mm in P. hieronymi), and by the larger ovary size (0.7-0.9 mm vs. 0.5-0.6 mm long in P. hieronymi). Moreover, the species differ in their ecology and distribution, in which P. glabra develops on karstic soils with heavy rainfall in N Peru (latitude 6 • S) and P. hieronymi in NW Argentina, apparently in the Selva de las Yungas ecosystem located at latitude 24 • S, which is only known from the type locality region, and possibly nearby geographic areas according to the scarce material found.  Another similar species, that occurs within the same geographic areas of Paronychia gabra, is P. macbridei. P. glabra differs by ( Figures 5 and 6) the larger plant size (vs. stems of about 20 cm long in P. macbridei), glomerules bearing 1-3 flowers in P. glabra vs. 3-5 flowers in P. macbridei, shorter internodes (8-12 mm long vs. 17 mm long in P. macbridei), larger leaf size (7-9 mm vs. 3.0-6.5 mm long in P. macbridei), leaf lamina glabrous vs. pubescent in P. macbridei, larger petal size (1.2-1.8 mm vs. 0.55-0.6 mm in P. macbridei), and by the larger style size (0.3-0.6 mm vs. 0.25-0.3 mm in P. macbridei). P. glabra is less similar to P. communis Cambess. by the smaller leaf size and texture (7-9 mm in P. glabra vs. 15 mm), larger stipule size (5-7 mm long vs. 5 mm in P. communis), and larger sepal size (2-2.8 × 1-1.5 mm vs. 1-1.2 × 0.45 mm in P. communis).  Rohrb. Linnaea 37: 204, 1871-1873[1872. Figure 7A. Lectotype (designated by Chaudhri [11] as "neotype", here corrected according to Art. 9.10 of ICN; second-step lectotypification (Art. 9.17)   Distribution and ecology. Paronychia glabra is found in the mountain regions of north Peru, in what is denominated as "jalca" by the abundancy of tussock grasses with forest patches which receive a relatively considerable amount of rain that surpasses 1000 mm. The species is distributed at an altitude of 3690 m and is believed to be in the range of 3600-3750 m. Several introduced species have been observed in community with natives and endemics in the area according to [41], moreover, the geographical region where the species develops is prone to intense slope burning for conversion to agriculture, this is permanently degrading the terrains, causing erosion and loss of diversity. The type material (Montesinos 6936) was collected from rocky slopes with elements of karst origin, environments that are also found in south-east Cajamarca [42], which consist of quartzite and sandstones which are widely distributed in tropical areas of South America, being subject to dissolution processes, generating significant caves and karst [43]. 14. Paronychia hartwegiana Rohrb. Linnaea 37: 204, 1871Linnaea 37: 204, -1873Linnaea 37: 204, [1872. Figure 7A.   Figure 8.

Paronychia hartwegiana
Notes. Pax [40] validly proposed this species providing a detailed description; The type matches the diagnosis and corresponds to the current concept of the species (see e.g., [11]).
label. This specimen is clearly part of the original material and eligible as a lectotype. According to Art. 9.19 of ICN, "The author who first designates ... a neotype ... must be followed, but that choice is superseded if (a) ... in the case of a neotype, any of the original material is found to exist ...". So, Chaudhri's neotype designation is here superseded and the found GOET specimen is here designated as the lectotype of the name Paronychia hieronymi. The type matches the diagnosis and corresponds to the current concept of the species (see e.g., [11]    Notes. Chaudhri [11] described Paronychia johnstonii subsp. johnstonii var. scabrida as plants with smaller leaves, and leaves and sepals pubescent. However, this variety was described from the same Chilean province and we did not see any difference between the two taxa. So, we here propose to synonymize the two names. ( (Figures 5 and 9). P. jujuyensis is an endemic taxon with restricted distribution area, known from locus classicus only (Loma de Tambo, Jujuy, NW Argentina). We here propose to treat it as a separate species [11].