Nomenclatural updates on names linked to the genus Axonopus for the Brazilian flora (Poaceae, Panicoideae, Paspaleae)

Abstract A nomenclatural update for 20 names related to the genus Axonopus for the Brazilian flora is here presented. Axonopus comprises around 90, mostly New World, species characterized by having solitary spikelets in an inverted position. In order to stabilize the nomenclature and, based on thorough study of original collections and pertinent literature, we designated lectotypes for nine names, of which three are second-step lectotypes. We also corrected to lectotypes erroneous applications, in previous contributions, of the term “holotype” (for four names), clarified three cases in which lectotypification is not required, as well as discussed four situations of uncertainty.

Morphologically, Axonopus is characterized by having racemose inflorescences distributed along a common axis, lower glume absent, and solitary spikelets with an inverse position (i.e., with the backs of the upper glume and the upper lemma turned away from the rachis) (Black 1963).Infrageneric categories for Axonopus were based exclusively on morphological features of Rodriguésia 74: e00502022.2023 inflorescences, being recognized four sections, five series, and three subseries (Chase 1906(Chase , 1911;;Black 1963;Giraldo-Cañas 2000).
Axonopus have been considered a taxonomically complex group mainly due to its large number of species and the difficulty of establishing clear limits between many of them (Black 1963;Delfini et al. 2020).The only revision of the entire genus was made by Black (1963) but, important advances have been done in recent years to clarify its taxonomic and phylogenetic relationships (i.e., López & Morrone 2012;Delfini et al. 2020Delfini et al. , 2022)).
This article is part of a broader proposal that includes phylogenetic and taxonomic studies for Axonopus species, focusing on the Brazilian flora (Delfini et al. 2020(Delfini et al. , 2022)).We aim here to contribute to the stabilization of nomenclature for 20 names that will appear in the project Flora e Funga do Brasil and also, before long, in the Flora Neotropica for the genus; thus, we designated lectotypes for nine names (i.e., Axonopus argentinus var.glabriflorus Parodi, A. obtusifolius var.rolfsii G.A. Our study was based on bibliographical research including original descriptions, as well as analyses of specimens housed in the following herbaria [acronyms according to Thiers (continuously updated)]: B, BAA, BR, C, F, G, GH, GOET, HAL, IAN, K, L, LE, M, MG, MO, MPU, NY, P, RB, S, SI, TUB, U, US, and W. We examined types in person and images available online at the JSTOR Global Pants website (<http://plants.jstor.org>) and/or at the websites of aforementioned herbaria.
The protologues of all taxa have been checked.Unless otherwise stated, the specimen designated as the lectotype was that which matches the protologue, corresponds to the current usage of the plant name, and is in the best preservation condition, according to the modern rules of the International Code of Nomenclature (ICN; Turland et al. 2018).For each name the place of valid publication is given followed by the holotype or lectotype and an explanation of the nomenclatural decisions made.The taxa were listed alphabetically, the names currently accepted, according to Delfini et al. (2022), are in bold italics and, for synonyms its respective accepted names are indicated after the citation of the type collection.Black (1963) did not specify a holotype, the specimen US [bc] 00139616 bears the basionym and the annotations "type" and "n.sp." in his handwriting, indicating that the original description was based on the referred duplicate.The other specimen (US [bc] 00139615), according to its label, seems to be a duplicate ex IAN sent to the US later.So a lectotypification is not required in this situation.G.A. Black, Advancing Frontiers Pl. Sci. 5: 127. 1963 This case is similar to the previous one as Black (1963), when describing Axonopus derbyanus, also did not specify a holotype.In the protologue he included a detailed description, mentioned a collection (A.Chase 10836), a herbarium (US) but, we have found two duplicates of Chase 10836 in the Smithsonian Institution.Both bear the basionym and the annotation "n.sp." in Black's handwriting but, one of them (US [bc] 00139626) is identified as "type" and another (US [bc] 00955585) as "dupl.type".Therefore, a lectotypification is not required in this situation since it is clear which specimen Black (1963) used to describe his species.The other seven duplicates (at IAN, K, MO, NY, RB) are isotypes.Black, Advancing Frontiers Pl. Sci. 5: 71. 1963.Type: BRAZIL.MINAS GERAIS: Viçosa, "stoloniferous, dense colony, moist ground", 11 Apr 1925, A. Chase 9447 (lectotype: here designated, US

Axonopus obtusifolius var. rolfsii
Black (1963) indicated the US gathering A. Chase 9447 as the type of Axonopus obtusifolius var.rolfsii.At US there are also two duplicates of this collection and, differently from aforementioned cases, he made no distinction which of the two sheets was the holotype.Both specimens bear the basionym and original label annotated, in Black's handwriting, as "type", "type collection" and/or "n.var.".Therefore, the specimen US [bc] 00130255 is here selected as lectotype of the name, satisfying the ICN requirements (Art.9.23).The remaining duplicates (at IAN, K, MO, NY, and US [bc] 00130256) are isolectotypes.2011) discussed this name and its original collection, stating that the "holotype" was lodged at BAA, as well as "isotypes" housed at IAN, SI and US.However, there is no holotype and Salariato's et al. (2011) statement cannot be corrected to lectotype since it was published after 1 January 2001 (Art.9.23).As Parodi (1938) did not specify a herbarium in the protologue, we selected the specimen BAA [bc] 00000691 as the lectotype for Axonopus suffultus var.pubiflorus; the other specimens (at IAN, K, and US) are isolectotypes.Judziewicz (1990) referred to the P collection as "holotype", but in such herbarium there are two duplicates of this gathering.His choice was the first-step for lectotypification and the specimen P [bc] 00753093 is here designated as lectotype in the second-step to accomplish the ICN requirements (Art.9.17).Our designation restricts and specifies Judziewicz's (1990) choice and results in several isolectotypes for the name P. surinamense lodged in 10 different herbaria (at BM, IAN, K, MO, P, S, TUB, U, US, and W).Grasses W. Ind. 190. 1936.In the protologue of Paspalum anceps, Mez (1917) included a detailed description, the type locality as "Brasilia, Prov.Amazonas, prope Barra ad Rio Negro", and a collection (Spruce Pasp.no.28) without specifying a herbarium.There are 13 specimens matching the description and protologue information housed in nine herbaria, thus indicating all specimens of that gathering as syntypes (Art.40 Note 1).Judziewicz (1990) discussed this name and its original material, explicitly stating that the "holotype" of P. anceps is housed at K, as well as isotypes at B, NY and US.At Kew herbarium there is only a single sheet of Spruce Pasp.no.28 (= R. Spruce 1259) not conflicting with the protologue; therefore, we considered Judziewicz's (1990) choice as an inferential lectotypification, satisfying the ICN requirements (Art.7.11, Art.9.10, Art.9.23).The other specimens (at B, G, GH, GOET, NY, P, US, and W) are isolectotypes.Mez, Repert. Spec. Nov. Regni Veg. 15: 61. 1917.Type: BRAZIL: "Minas Gerais, Campos d'Itabira" / Rio de Janeiro, 1888-1889, A.F.M. Glaziou 17933 (lectotype: designated by Black (1963: 138) In the original publication of Paspalum arcuatum, Mez (1917) cited a collection made by Glaziou (no.17933) and locality notes as follows: "Brasilia, loco accuratiore non indicato".He also included a detailed description but did not specify a herbarium.Black (1963) explicitly stated that the type of P. arcuatum is lodged at B and an "isotype" at US.In addition to these specimens, we have found eight more duplicates under Glaziou 17933 in six different herbaria.In most of them, the labels refer that "Minas Gerais, Campos d'Itabira" is the locality of collection but, the P and BAA labels indicate "Rio de Janeiro".Although it is known that Glaziou worked for a long time in Rio de Janeiro (Stafleu & Cowan 1976), it is not possible to determine where he actually gathered the type collection.As Black (1963) assigned a single gathering, specified a herbarium, and used the word "type", even though he had not actually seen the Berlin specimen, we interpreted his choice (B [bc] 10 0176788) as an inferential lectotypification to accomplish the ICN requirements (Art.7.11, Art.9.10, Art.9.23).The other specimens (at BAA, C, G, K, P, US, and W) are isolectotypes.Mez, Repert. Spec. Nov. Regni Veg. 15: 31. 1917 Mez (1917) saw one or more materials and whether he used additional elements when describing the species, it seems that if B, which holds important sets of plants described by him, has a single specimen, it could be the holotype (Art.40.3), and a duplicate was sent to C later.In such a situation of uncertainty, the B specimen ([bc] 10 0365558) could be a holotype or lectotype (K. Gandhi 2022, personal communication).Mez (1921) provided a short description, cited a gathering (Ule n. 8533) and locality notes as "Guyana, Roraima", but he did not specify a herbarium.There are eight duplicates matching the description and protologue information lodged in seven herbaria.The duplicates of K, L, U and B have original labels and the latter specimen also includes the determination written and signed by Mez. Judziewicz (1990) explicitly stated that the "holotype" of P. caulescens is housed in the Berlin herbarium, as well as "isotypes" at IAN, K, L, MG, and U. Therefore, the use of the term "holotype" by Judziewicz (1990) is an error to be corrected to an effective lectotype designation (Art. 7.11,Art. 9.10,Art. 9.23).The other specimens (at IAN, K, L, MG, U, and US) are isolectotypes.Mez, Repert. Spec. Nov. Regni Veg. 15: 67. 1917 In the protologue of Paspalum comatum, Mez (1917) provided a detailed description, cited a gathering (Glaziou no.15700) and locality notes as "Brasilia, loco non indicato", without specifying a herbarium.We have located seven duplicates, with data matching the protologue, housed in six herbaria (at B, BAA, C, K, P, and US).At C and one of the two duplicates of B ([bc] 10 0365546) bear original labels annotated by Mez with the determination as "Paspalum comatum n. sp."; the second one (B [bc] 10 0365547) includes illustrations of spikelets, which also indicates that it is original material for the name.From among the material available, the specimen B [bc] 10 0365546 is here selected as lectotype of the Paspalum comatum (Art.9.23).

Paspalum comatum
12. Paspalum erythrochaetum Mez, Repert. Spec. Nov. Regni Veg. 15: 32. 1917 M e z ( 1 9 1 7 ) p u b l i s h e d P a s p a l u m erythrochaetum with a detailed description, the locality as "Brasilia, Prov.Goyaz", and a gathering (Glaziou no.22478) but, he did not specify a herbarium.Seven duplicates of the type collection in agreement with the protologue information were located at B, BAA, BR, C, MPU, and US; however, only the specimen B [bc] 10 0365525 was annotated in Mez's handwriting as "Paspalum erythrochaetum n. sp.".It is well known that Mez worked at B and his relevant type collection, those that survived the WWII for species described between 1917 and 1921, are lodged mainly there (Stafleu & Cowan 1981).As it is impossible to known whether B had more than one material of this collection (and it is no longer extant there), for such uncertain situation, the B specimen ([bc] 10 0365558) could be a holotype or lectotype (K. Gandhi 2022, personal communication).Mez, Bot. Jahrb. Syst. 56 (Beibl. 125): 9. 1921.Axonopus flexilis (Mez) Henrard, Blumea 4(3): 510.1941.Type: BRAZIL: "An einem Bache bei der Serra do Mel.Rio Branco, Surumu", Sep 1909, E.H.G.Ule 8020 (lectotype: here designated (or perhaps a holotype), B [bc] 10 0365518!; isolectotypes (or perhaps isotypes): BAA Chase, J. Wash. Acad. Sci. 17: 144. 1927.This case is similar to the previous one since Mez (1921), when describing Paspalum flexile, also did not specify a holotype.In the protologue of the species he based the description on the collection Ule 8020 from Brazil, without specify a herbarium acronym.Seven duplicates of this collection, in agreement with the protologue information, were located in six herbaria (at B, BAA, IAN, K, MG, and US).The materials of B, K and MG bear original labels but, only the specimen B [bc] 10 0365518 was annotated in Mez's handwriting as "Paspalum flexile n. sp." and also includes illustrations of spikelets made by him.As it is impossible to known whether B had more than one material of this collection (and it was damaged by WWII), for such uncertain situation, it could be a holotype or lectotype.Hack., Repert. Spec. Nov. Regni Veg. 7: 50. 1909.Type: BOLIVIA.SANTA CRUZ: "Charaktergras der höchsten Kämpe am Cerro San Miserate bei Santiago de Chiquitos", 1905-1907, T. Herzog s.n. (holotype: W [bc] 19160031292!; isotypes: SI! fragment ex W, US [bc] 00140698!fragment ex W).Basionym of Axonopus herzogii (Hack.)Hitchc., Contr.U.S. Natl.Herb. 24 (8): 431. 1927.In the protologue of Paspalum herzogii, Hackel (1909) included a detailed description, indicated a locality and provided taxonomic comments, without specifying a specimen seen.We have found only one specimen that perfectly matches the original data and description in the protologue in Hackel Herbarium (W), plus fragments of this this collection at SI and US.Although it is not clear whether Hackel (1909) used additional elements when describing the species, it seems that if W has a single specimen, it is the holotype (Art.40.2, Art.40.3), and the existent fragments, clearly labelled as being part of the W specimen, were sent to SI and US later.Therefore, a lectotypification in this case is not required.

Paspalum herzogii
Rodriguésia 74: e00502022.2023 specifying a herbarium.We have located 14 duplicates matching the description and specimen label information included in the protologue in nine herbaria (B, BAA, BR, K, LE-TRIN, P, SI, US, and W).Salariato et al. (2011) discussed the original material for the name and cited the specimen LE-TRIN 0507.01 as the "holotype".However, there is no holotype and their statement cannot be corrected to an effective lectotype designation since it was published after 1 January 2001 (Art.9.23); therefore, to accomplish the ICN requirements (Art.9.23), we here designated the specimen LE-TRIN 0507.01 as the lectotype of P. flexile.Bragantia 15: 273, fig. 216. 1956. Steudel (1853) only mentioned one collection ("Agrostis polydactyla.Salzm.hrbr.Bahia") in the protologue of Paspalum polydactylon.However, we have found 10 duplicates in seven herbaria matching the description and locality data included in the original publication.The material with barcode P02608440 has been selected as the lectotype, as it has Steudel's annotations and represents best the species.Mez, Bot. Jahrb. Syst. 56(Beibl. 125): 10. 1921 Sci. 17: 144. 1927.In the protologue of Paspalum purpusii, Mez (1921) included a brief description, a locality as "Mexico", and cited a collection (Purpus n. 2450) without specifying a herbarium.Seven duplicates in agreement with the diagnosis and the collection cited in the protologue were found in six herbaria.Judziewicz (1990) stated that the "holotype" of this name is housed at B, as well as an "isotype" at US.Although Judziewicz (1990) did not actually see the B sheet, his statement is here considered as an error to be corrected to an effective lectotype designation (Art.7.11, Art.9.4, 9.10, Art.9.23).Mez, Repert. Spec. Nov. Regni Veg. 15: 61. 1917 Basionym of Axonopus triglochinoides (Mez) Dedecca, Bragantia 15: 280, fig. 22. 1956.

Paspalum triglochinoides
M e z ( 1 9 1 7 ) p u b l i s h e d P a s p a l u m triglochinoides with a detailed description, the type locality as "Brasilia, Prov.Amazonas ad Casiquiare fluvii ostia", and a gathering (Spruce n. 3756) without specify a herbarium.Eight duplicates that match the protologue information were located in seven herbaria (at B, G, K, NY, US, and W).The specimens at B, G and W bear the basionym and the annotation "n.sp." in Mez's handwriting but, it is not clear whether he saw one or more materials and whether used additional elements when describing the species.Still, if B has a single specimen, it could be the holotype (Art.40.3), and duplicates were sent to G and W later.In such a situation of uncertainty, the B specimen ([bc] 10 0591450) could be a holotype or lectotype.Hack., Ergebn. Bot. Exp. Sudbras. 1: 67. 1908 Black, Advancing Frontiers Pl. Sci. 5: 102. 1963.In the original publication of Paspalum uninode, Hackel (1908) included a detailed description, taxonomic comments and gathering notes as follows: "Inter Pilar et Alto da Serra prope Santos; 750-800 m s. m. 1902 (W)".Black (1963), in a taxonomic treatment for the genus Axonopus, explicitly stated that the type of P. uninode is Rodriguésia 74: e00502022.2023 lodged in the Vienna herbarium; however, two duplicates that match the protologue information were located there.Given that Black (1963) made no distinction as to which of the two sheets might be intended to be the lectotype, his statement must be considered as a first-step lectotypification.In order to narrow this earlier designation, the specimen W [bc] 19160031037 is here selected as a second-step lectotype (Art.9.17).