New species , nomenclatural changes and recent taxonomic studies in the genus Stylosanthes ( Leguminosae ) : An update Especies nuevas , cambios nomenclaturales y recientes estudios taxonómicos en el género Stylosanthes ( Leguminosae ) : Una actualización

Since the last taxonomic overview during the 1982 Stylosanthes Symposium in Townsville, Australia, 10 new species: S. falconensis, S. longicarpa, S. maracajuensis, S. nunoi, S. quintanarooensis, S. recta, S. salina, S. seabrana, S. vallsii and S. venezuelensis; and 1 botanical variety: S. guianensis var. pauciflora; have been validly described. Furthermore, 2 nomenclatural changes have been proposed, both being elevations of botanical varieties to the rank of species: S. gracilis and S. rostrata. In the major taxonomic databases, The Plant List and GRIN, the taxonomic status (“accepted” vs. “synonym” vs. “unresolved”) of some of these new taxa, however, differs. In addition, this paper reports on Stylosanthes names that can be found in the post-symposium literature but have not been validly published, and on recent regional studies of Stylosanthes taxonomy. Suggested research needs as perceived by non-botanists include an updated Stylosanthes monograph and taxonomic studies within the S. guianensis, S. hamata and S. scabra species complexes.


Introduction
Stylosanthes Sw. is an important tropical and subtropical forage plant genus (Chakraborty 2004).To date, more than 80 Stylosanthes species have been described (IPNI 2014).Nine of them are or were of economic importance (Chakraborty 2004;Cook et al. 2005): Stylosanthes capitata Vogel, S. fruticosa (Retz.)Alston, S. guianensis (Aubl.)Sw., S. hamata (L.) Taub., S. humilis Kunth, S. macrocephala M.B.Ferreira and Sousa Costa, S. scabra Vogel, S. seabrana B.L. Maass and 't Mannetje and S. viscosa (L.) Sw.Swartz (1788) described the genus and included 2 species, Stylosanthes procumbens Sw., nom.illeg.(≡ S. hamata) and S. viscosa.Vogel (1838) prepared the first revision of Stylosanthes; therein he divided the genus into 2 sections and listed 15 species.Mohlenbrock (1957) monographed Stylosanthes across its full geographic range, listing 25 species.However, the material that was available to him more than half a century ago was limited in terms of specimen numbers and origins.In a subsequent taxonomic overview presented at the International Symposium on Stylosanthes held in November 1982 in Townsville, Australia, Mannetje (1984) listed 24 unambiguous, 5 problematic and 5 doubtful species; this overview was complemented by a study of some Brazilian Stylosanthes species (Costa and Ferreira 1984).With these 2 chapters on taxonomy in the symposium book (Stace and Edye 1984), it became evident that the genus presents considerable taxonomic difficulties.
There is no doubt that taxonomy of some Stylosanthes species is unclear and names might prove to be synonyms of other described species.However, this paper is not about the identification of synonyms vs. correct names, which should rather be the subject of a comprehensive taxonomic revision of the genus.Instead, our objective is to inform non-botanists about: (1) new validly published taxa; and (2) nomenclatural changes since the 1982 International Symposium.The term "nomenclatural changes" refers here to new combinations, i.e. when a new name, based on a legitimate, previously published name (= the so-called basionym of the species), is published (McNeill et al. 2012).In addition, we present: (3) a list of new variety and species names occasionally found in postsymposium publications, which, however, should not be used because they are invalid; and (4) a brief account of recent taxonomic Stylosanthes studies with a regional focus (country, state).
With this update we pretend to contribute to: (1) stimulate continuing interest in the genus; and (2) clarify and prevent potential confusion related to the use of species names.

Methodology
The study is based on analyses of botanical Stylosanthes literature partly accessed online from the Biodiversity Heritage Library (www.biodiversitylibrary.org).Literature not available online was obtained from the library of the Royal Botanic Gardens, Kew.New names and combinations and standard abbreviations of plant name authorities were obtained from the International Plant Names Index (www.ipni.org).For information on the taxonomic status of new names and combinations, 2 taxonomic databases were accessed: The Plant List (TPL, www.theplant list.org) and GRIN, the US Department of Agriculture -ARS Germplasm Resources Information Network (https://npgsweb.ars-grin.gov/gringlobal/taxonomybrowse.aspx).
TPL is a taxonomic database produced through the collaboration of mainly the 2 world-leading research centers for legume taxonomy and nomenclature, Royal Botanic Gardens, Kew and Missouri Botanical Garden, where botanists are in charge of maintaining and updating the respective databases, ILDIS and TROPICOS.Algorithms are used for detecting and resolving conflicting opinions resulting from data coming from such different sources.When this leads to uncertainties, a taxonomic status will be presented as "unresolved".The TPL database is not updated regularly (TPL 2013).On the other hand, the GRIN database is monitored and regularly updated by botanists of the USDA National Germplasm Resources Laboratory (GRIN 2016).It follows a more pragmatic approach: when taxonomic differences arise, GRIN would generally be guided by current usage, if nomenclature of a specific case adheres to the international rules of the "Melbourne Code" (McNeill et al. 2012).We chose TPL and GRIN because of the particular expertise in legume taxonomy of the institutions contributing to these databases.

New species and botanical varieties
Since 1982, 10 new species and 1 botanical variety name have been described within the genus Stylosanthes (Table 1 Costa. However, though validly published, i.e. in accordance with international nomenclatural rules as specified in the "Melbourne Code" (McNeill et al. 2012), not all new taxa are currently equally accepted by the scientific community (Table 1).For example, S. nunoi and S. maracajuensis are accepted by both TPL and GRIN, whereas S. venezuelensis and S. salina are treated as unresolved by TPL and as accepted by GRIN.
These discrepancies among databases are the result of different species concepts.A good example is the case of S. seabrana; Maass and Mannetje (2002) compiled morphological, karyotypic and molecular evidence to describe a new species, S. seabrana, which was subsequently accepted by both TPL and GRIN.Under this name, the species has found entrance in tropical forage legume research worldwide and 2 commercial cultivars of S. seabrana were released in Australia (Cook et al. 2005).However, Vanni and Fernandez (2011), based on morphological and karyotypic characteristics, concluded that S. seabrana should be treated as a synonym of S. scabra.While TPL follows Maass and Mannetje (2002) and continues to accept the species as S. seabrana, GRIN now follows the broader concept of Vanni and Fernandez (2011) and treats it as a synonym of S. scabra.

Nomenclatural changes
Two nomenclatural changes are presented in the second part of Table 1.Burkart (1939) reported on an Argentinian botanical variety, Stylosanthes gracilis var.rostrata Burkart.Recently, Vanni (2009) found reasons to elevate this variety to the rank of species as Stylosanthes rostrata (Burkart) Vanni.Until now, however, there has been no consensus on the taxonomic status of this name (Table 1).
Stylosanthes gracilis Kunth was originally described by Kunth (1823) as a species, but subsequently treated as a synonym of S. guianensis (Mohlenbrock 1957) or as a variety of S. guianensis (Vogel 1838;Mannetje 1977Mannetje , 1984)).Ferreira and Costa (1979) suggested elevating the epithet gracilis to species rank but it was only recently that sufficient taxonomically relevant evidence to reestablish S. gracilis at species level was collected (Calles and Schultze-Kraft 2010b).However, again there is no consensus on the taxonomic status (Table 1).

Invalid Stylosanthes names in the post-symposium literature
In Table 2 we mention several names of Stylosanthes taxa that have not been validly published or even not been published at all, which, however, can occasionally be found in post-symposium (i.e.post-1982) publications.Since these names have not been validly published, they should be considered as "nomina nuda" [(singular: nomen nudum: any new taxon published without a description or diagnosis; McNeill et al. 2012;Art. 38)].These names should not be used in scientific literature; in particular cases where this cannot be avoided, such a name should be set in inverted commas and it should be made clear that the name has not been validly published.

Recent regional studies on Stylosanthes taxonomy
Although there has been a slow-down in botanical research on Stylosanthes during the past 3 decades, several regional studies on the taxonomy of the genus have been carried out recently (Table 3).The areas covered are several Brazilian states, Mexico and Venezuela.

Research needs
We see the following taxonomic research issues as particularly relevant in the interest of non-botanists working with this important forage legume genus and, in particular, its economically most important species:  A revision of the genus Stylosanthes is urgently required.Costa's (2006) dissertation provides such an opportunity, although condensation of the 470-page thesis to produce an international publication would be a prerequisite. There are many discrepancies around the so-called "Stylosanthes guianensis species complex".Mannetje (1984) Costa and S. longiseta Micheli, respectively).At present, there are considerable uncertainties regarding the taxonomic status of these taxa and therefore a revision of this complex, including the newly described variety pauciflora, is urgently needed.We suggest that such studies be based on molecular analyses such as the evidence provided by Santos-Garcia et al. ( 2012). In the genus Stylosanthes, S. scabra is a particularly polymorphic species.This is also indicated by the suggestion made by Vanni and Fernandez (2011) that S. seabrana should actually be considered as a synonym of S. scabra.In view of the significance of S. scabra as a forage species, we suggest that it merits closer examination in order to identify morphological, molecular and karyotypic characteristics that might allow separating distinct groups that could be described as infraspecific taxa (e.g., botanical variety, subspecies) within an eventual "S.scabra species complex". Likewise, further examination of the variability in S. hamata, also an economically very important species, is required, with eventual description of the tetraploid forms as a separate taxon (see comment on "S.hemihamata" in Table 2).We suggest that in the aforementioned taxonomic studies conventional morphological analyses using herbarium material be complemented with molecular analyses, making use of the large Stylosanthes germplasm collections held in the major tropical forage genebanks at CIAT (International Center for Tropical Agriculture) in Cali, Colombia, the Australian Pastures Genebank in Adelaide, Australia, and ILRI (International Livestock Research Institute) in Addis Ababa, Ethiopia.(period 1982-2014).

Name
Reference examples Comments S. grandiflora M.B.Ferreira & Sousa Costa nom. nud. Kazan et al. 1993Vander Stappen et al. 1999Fortuna-Perez et al. 2011 There is no description of a Stylosanthes species under this name.Thus the name has not been validly published and consequently is a nomen nudum.We assume this is a misspelling of S. grandifolia M.B.Ferreira and Sousa Costa.
S. guianensis var.vulgaris M.B.Ferreira &Sousa Costa Gillies andAbbott 1996 Vieira et al. 1997 The naming of this variety (Ferreira and Costa 1979)  Following an earlier suggestion of H. Stace (pers. comm. 1984), Maass and Sawkins (2004) proposed the name S. hemihamata for tetraploid S. hamata such as the cultivars Verano and Amiga.However, the species has never been described and consequently the name is a nomen nudum.

Table 1 .
New taxa, nomenclatural changes and taxonomic status within the genus Stylosanthes, during period 1982-2014.

Table 2 .
Invalid Stylosanthes names found in recent literature did not follow international nomenclatural rules(McNeill et al.  2012; Arts.24.3 and 26.1).Consequently, the name was not validly published and should not be used.The correct name for this taxon is S. guianensis var.guianensis (see Table1).

Table 3 .
Regional studies on Stylosanthes taxonomy, period 1982-2014.Seven species are reported, none being endemic to this state.Mato Grosso do Sul Costa et al. 2008 Seventeen species are reported, including S. maracajuensis, which is endemic to this state.Roraima Medeiros and Flores 2014 Seven species are reported, none being endemic to this state.São Paulo Fortuna-Perez et al. 2011 Nine species are reported, none being endemic to this state.Mexico Countrywide Gama-López et al. 2007 This Bioversity International booklet contains a taxonomic synopsis of Mexican Stylosanthes and includes 11 species and 2 varieties of S. guianensis.Stylosanthes macrocarpa, S. mexicana and S. subsericea are reported as endemic to Mexico.Three taxa mentioned in this booklet are not validly published (see Table 2).Venezuela Countrywide Calles and Schultze-Kraft 2010c Eleven species are reported, including S. falconensis, S. sericeiceps and S. venezuelensis, which are endemic to Venezuela.