Checklist of Gymnotiformes (Osteichthyes: Ostariophysi) and catalogue of primary types

A list of all valid names of species of Neotropical electric fishes (Gymnotiformes) is presented herein. The list is arranged by family and genus and includes all available synonyms. The list is comprehensive through 2016 and includes 240 valid species distributed among 34 genera and five families, including one monotypic genus known only from the fossil record. The presented classification reflects recently published interpretations about the validity of the included names which, in general, are widely accepted. When the validity of a particular name is disputed in recent literature, we followed one of the published interpretations and provide relevant information on the alternate interpretation(s) in the remarks section of that name. Synonymies of some names need to be considered tentative, inasmuch as the types underlying those names are either absent or appear to be based on more than one taxon. First reviser actions (e.g., lectotype and neotype designations, resolution of simultaneous synonyms, etc.) are reported and include erroneous subsequent attempts at problem resolutions. Herein, we include one new first reviser action by selecting Gymnotus aequilabiatus Humboldt, 1805, as type species of Sternopygus because previous attempts to select a type did not follow the provisions of the Code of Zoological Nomenclature.


Introduction
Neotropical electric fishes (Ostariophysi: Gymnotiformes) are a natural group of strictly freshwater fishes, which also go by the name electric knifefishes.The group is perhaps best known for the genus Electrophorus, a monotypic genus that is widely known as the electric eel.Through 2016, 240 valid species are distributed among 34 genera and five families, including one monotypic genus known only from the fossil record.
Neotropical electric fishes are found throughout much of the South American continent and range from the La Plata basin of Argentina northward into North America as far as Mexico.Although mostly confined to east of the Andes mountain range, species are recorded from Trans Andean portions of Ecuador, Colombia and Venezuela (Albert, Crampton, 2005).
A comprehensive overview of the diversity, phylogeny, and life history of the Gymnotiformes was published by Albert, Crampton (2005) and the reader is referred to that paper for details not covered herein.Also, synopses of the gymnotiform families were included within the Checklist of Freshwater Fishes of South and Central America -CLOFFSCA (Reis et al., 2003) and those details will likewise not be repeated here.

Checklist of Gymnotiformes
Neotropical Ichthyology, 15(1): e160067, 2017 2 e160067 [2] For the first two centuries of Linnaean nomenclature, the Gymnotiformes appeared to be a relatively depauperate taxon.In a revision of the group published in 1913, Ellis treated 27 species as valid, and even as recently as 1994, in the most recent taxonomic revision of the Order, Mago-Leccia (1994) recognized only 85 previously named valid species in addition to the 12 new species that he described therein (Tab.1).Mago-Leccia's study clearly demonstrated the need for closer examination of the diversity within the group, which lead to contributions by a number of researchers that increased the number of recognized species by 20%, to 117, during the next decade (Reis et al., 2003) and to more than double that total (to 240 species) at present (Tab.1).
Tab. 1. Number of valid gymnotiform genera and species (genera/species) recognized in previous compilations.Günther, Eigenmann and Ellis included all taxa within the Gymnotidae (except a monotypic Electrophoridae); taxa broken down here by current treatment of families.Mago-Leccia recognized the five listed families but separated Electrophorus into a separate family, which is placed here within the Gymnotidae; Mago-Leccia recognized three additional subspecies of Sternopygus aequilabiatus, which are not included in the totals.Albert, Crampton included unnamed species that they considered valid; those species are not included in the totals for each family, but are noted in parentheses.Günther, 1870Eigenmann, Ward, 1905Eigenmann, 1910Ellis, 1913Mago-Leccia, 1994Reis et al., 2003Albert, Crampton, 2005 This The pace at which new taxa have been reported in recent years reflects in an increased interest in the study of the Gymnotiformes and suggests that many more taxa are likely to be discovered in the foreseeable future.In anticipation of that, we decided to compile the following checklist with two main objectives.First, we thought it beneficial to provide an up-todate listing of taxa together, when relevant, with information on conflicting views on the status of available generic and species group names, to insure that future researchers were aware of information relating to their ongoing work.Second, we thought it necessary to critically examine information relevant to compliance with the Code of Zoological Nomenclature (ICZN, 1999), such as the availability of names, identity of, and location of, primary types of species-group taxa as well as the identity of types of the genus-group and familygroup names, and the existence of first-reviser actions at any taxonomic level, again to make that information available to future workers.In our view, this compilation of a list of names, together with all nomenclaturally relevant information for each, should provide future workers with a solid background upon which to base their nomenclatural decisions, once their taxonomic studies have been completed.

Material and Methods
By their nature, checklists force their authors into making choices among competing hypotheses of the validity of some taxa.It is not possible to treat such competing hypotheses as equally valid within a checklist format and, as such, one of the hypotheses must be given precedence over the other(s).A species cannot be treated as valid in two different genera despite evidence presented in different studies that reach differing conclusions about its relationships.Similarly, a species name cannot be treated as valid and a junior synonym of another name just because different studies reach these alternative conclusions.These kinds of problems are especially difficult when the proponents of the competing hypotheses are still active researchers, and each is likely to object if their hypothesis is not adopted.It is our view that such decisions are necessary, but a relatively unimportant part of the checklist, as long as the existence of a controversy is clearly stated.When we are required to choose among competing hypotheses, we make a point of including the reference to any recently published alternate hypotheses in the remarks section to insure that the reader knows of its existence.
Families are listed alphabetically.Family membership generally follows that in Reis et al. (2003), except for the placement of Hypopygus and Steatogenys in the Rhamphichthyidae, following Tagliacollo et al. (2016).Genera proposed subsequent to that Reis et al. (2003) are placed where assigned by the author of the generic name.
Genera are listed alphabetically within families, with no attempt to sort genera into subfamilies or tribes, even when those groups are in current use.Species are listed alphabetically within genera with no attempt to sort the names into subgenera, or list the subgeneric name as part of the valid name.
Type localities are copied directly from the original description, with interpretation (enclosed in brackets) only when the original was deemed unclear.Figures listed for each nominal species name were those that showed an entire type specimen and/or the head region and not all drawings and or photographs that depict parts of a specimen of the species.When a primary type of a nominal species is illustrated in a publication other than the original description, that illustration is also noted.
Page numbers listed for accounts of genera or species refer to the first page in which the name is made available.If names appear in a key earlier than the full account of 4 e160067 [4] Size of the type(s) is that reported in the original description, when so listed.The original units of measure were provided, and those not reported in metric or English units were translated when possible.When more than one size was provided by in the description, preference was given to list total length (TL) over standard length (SL) or length to the end of the anal fin (LEA).Measurements were rounded to the nearest mm when reported to fractions of a mm.In cases in which no indication of which measurement was taken, it was assumed to be TL and reported as such.When a later publication reported the size of the type that was not reported in the original description or that was significantly different from the size in the original description, that measurement is listed in parentheses after the length reported in the original description.
Synonymies at the species level are not new here and instead follow the conclusions reported in the publications listed under the account of the genus or, in some cases, in the remarks for the species.Synonymies at the generic level necessarily follow the taxonomic assignment of the type species.
In a few accounts, binomial names that are not italicized are included within the remarks sections.Those names are not available in terms of the Code and therefore not italized to emphasize that point.
Titles and publication information on some older literature were taken from that listed in the Biodiversity Heritage Library.
The distribution of each species was stated, using one of the following sources, in order of descending priority: first, the distribution reported in a revision of the genus (or a subunit), or redescription of the species; second, reports of the species that extend the distribution beyond that of earlier publications (in which case, more than one publication may be cited); third, the distribution provided in the original description; and fourth, the distribution reported in CLOFFSCA (Reis et al., 2003).Country and river names are written with English spellings (especially the word River) rather than the spelling used in each country.

Results
The following is a list of all scientific names for the Gymnotiformes that are covered by the Code of Zoological Nomenclature (ICZN, 1999), arranged by valid taxa.The list is believed to be complete through 2016.We found 18 available family-group names, 56 genus-group names, and 281 species-group names, which are assigned into five families, 34 genera and 240 species ( Tab. 3), including one monotypic genus that is known only from the fossil record.This list reports double the valid species over the most recently published list (Reis et al., 2003), and a greater than eigth fold increase in the last century (Ellis, 1913).
Tab. 3. Nominal and valid gymnotiform families, genera and species reported herein, through the end of 2016, including fossil taxa.Within each family, number of valid taxa is followed by total number of available names for those taxa.
Phylogeny: Schlesinger (1910); Regan (1911); Triques (1993); Gayet et al. (1994)  Checklist: Albert (2003a); Triques (2007b), Brazil. Remarks: Jordan (1923: 138) proposed the new name Apteronotidae for the family previously named the Sternarchidae, in conjunction with his treatment of Sternarchus as a junior synonym of Apteronotus, based on the then-prevailing view that family names needed to be derived from valid generic names.Despite its clear priority, the name Sternarchidae has been used rarely since the introduction of the name Apteronotidae, and the Apteronotidae is in widespread use.Following Art. 40.2 of the Code of Zoological Nomenclature (ICZN, 1999), the name Apteronotidae takes priority over Sternarchidae and takes as its date of publication that of Sternarchidae [= 1838] for purposes of priority among other gymnotiform family-group names.Albert (2001) proposed the name Sinusoidea (p.69) as a suprafamilial name and Navajini (p.77) as a subgroup within the Apteronotinae.Those names are not available as family-group names inasmuch as they are not based on available generic names (ICZN, 1999: Art. 11.7.1.1).
Distribution: Amazon River basin, Peru and Brazil (Albert, 2003a).Remarks: Treated as valid in Porotergus in Mago-Leccia (1994); valid in Apteronotus in Albert (2001); and as a valid species within the Apteronotidae, but not assigned to any available generic name in Tagliacollo et al. (2016).
Remarks: Adontosternarchus appeared first in Ellis (in Eigenmann, 1912) without reference to any species.The name was listed with one included species, Sternarchus sachsi Peters, 1877, in Ellis (1913: 155).
Remarks: The gender of Ubidia was not stated by Miles, and it has been listed as feminine in Mago-Leccia (1994), Eschmeyer (1998) and Eschmeyer et al. (2016).The name, however, is the surname of "señor Jorge Ubidia Betancourt" and, as such, we believe the name should be treated as masculine.Distribution: Upper Paraná River basin, Brazil (Triques, 2011).
Distribution: Mainstream channels of Amazon and Orinoco rivers, Brazil and Venezuela (Albert, 2003a).
Distribution: Large river channels of Amazon River basin (de Santana, Crampton, 2007).
Type locality: one of the side streams which connect with the main stream of the Amazon River on its southern bank, about three hundred miles from Para at the mouth of the river.Syntypes (2): private collection of author, possibly not saved; about 8 inches TL.Brind mentions additional specimens at the New York Aquarium and a pet shop, but it is not clear whether he actually viewed them or was just relaying information of their presence from other sources.
Remarks: Redescribed in Crampton, Albert (2004).Synonymy follows Campos-da-Paz (2000b), with a note suggesting that the holotype of Gymnotus coatesi may have been one of the specimens observed by Brind.LaMonte's paper was published on 18 February, 1935, while the footer in Brind's paper indicates that it was part of the issue for February, 1935, but no more precise date of issue was found in the copy at AMNH (D. Shih, pers.commun.).Unless a more precise date is established for Brind's paper that is prior to 18 February, it must be treated as having been published on the last day of February, thereby giving priority to Gymnotus coatesi.
Distribution: Widespread through cis-Andean South America from Orinoco to La Plata River (Crampton et al., 2016b).
Remarks: Synonymy follows Nijssen et al. (1976), but Gymnorhamphichthys petiti included in the materials examined as a valid species in Carvalho et al. (2011).
Distribution: Widely distributed in Amazon, upper Paraná, and Orinoco River basins and coastal rivers of the Guianas (Nijssen et al., 1976).
Distribution: Vicinity of Iquitos and Jenaro Herrera, Peru, in small tributaries of Amazon River (de Santana, Crampton, 2011).
Remarks: Name written as Eigenmannia microps in Eigenmann (1910: 449), apparently a lapsus calami and not treated here as an available name.
Gymnotidae, even if the spelling discrepancy is overlooked, inasmuch as he did not treat Gymnotus as a valid name and, instead, treated G. acus as a species in the genus Carapus.Electrophorus Gill, 1864: 152.Type species: Gymnotus electricus Linnaeus, 1766.Type by monotypy.Gender: masculine.
Müller, Troschel (1849) the text and figure caption.Subsequent authors reported that the type species is Gymnotus macrurusBloch,  Schneider, 1801, but Müller referenced that publication as Bl.Schn., as was done for Gymnotus rostratus in the paragraph on Rhamphichthys in the same footnote.As no available species names were listed in the account of Sternopygus, it is necessary to select a type species from among the first listed nominal species in a subsequent account of the genus, following Code Art.67.2.2 (ICZN, 1999).Müller, Troschel (1848)was the next publication that treated Sternopygus as a valid name.Therein, four nominal species were listed as being included in the genus: S. virescens(Valenciennes, 1836), S. lineatusMüller, Troschel, 1848, Sternopygus macrurus Müller,  Troschel, 1848, and Gymnotus aequilabiatus Humboldt,  1805.No designations from among those names has been found.Instead, Gymnonotus macrurus Bloch, Schneider, 1801, which was listed first in the account of Sternopygus inMüller, Troschel (1849), has been widely reported as the type species, first byEigenmann (1910).However, as Gymnonotus macrurus Bloch, Schneider was not included in the account ofSternopygus in Müller, Troschel (1848), it cannot be validly selected as the type species.Inasmuch as two of the names available for selection as type species are currently assigned to the genus Eigenmannia, it is necessary to select a type that will stabilize Sternopygus as well as not disrupt the validity of Eigenmannia.To accomplish those goals, we select herein Gymnotus aequilabiatus Humboldt, 1805, as type ofSternopygus.