A new catfish species of the Trichomycterus hasemani group ( Siluriformes : Trichomycteridae ) , from the Branco river basin , northern Brazil

Trichomycterus wapixana is described from the Branco river basin, Roraima State, northern Brazil. It belongs to the T. hasemani group, composed of T. hasemani, T. johnsoni and T. anhanga and defined by the presence of a single wide cranial fontanel delimited by the frontal and supraoccipital, absence of the pores i1 and i3, absence of branchiostegal rays on the posterior ceratohyal and the by the presence of a large and distally expanded process on the palatine. It differs from the other species of that assemblage by having a unique combination of character states, including number of vertebrae, relative position of anal fin, relative position of pelvic and dorsal fin, presence of pelvic fin and pelvic girdle, number of dorsal and ventral procurrent rays in the caudal fin, anal-fin rays, pectoral-fin rays, branchiostegal rays, pleural ribs, morphology of palatine, presence of parasphenoid and relative position of urogenital pore.


Introduction
Trichomycteridae is a family of catfishes comprising 278 valid species (EschmEyEr & Fong, 2014) distributed from Costa Rica to Patagonia, in both cis-and trans-Andean drainages (dE Pinna, 1998).Trichomycterinae is the only one of the eight recognized subfamilies which monophyly has not been supported in phylogenetic studies (Baskin, 1973;dE Pinna 1989;costa & Bockmann, 1993).Tricho mycterus ValEnciEnnEs is the most species-rich genus of the family, comprising over 140 species (FErnandEz & Vari, 2009;katz et al., 2013).Its extensive geographical range, high number of described species and lack of synapomorphies make Trichomycterus a huge taxonomic problem within the Trichomycteridae (BarBosa & costa, 2003).This condition is illustrated by the description of Ituglanis costa & Bockmann 1993, in which this new genus was described based on nine species that were pre-viously placed in Trichomycterus (costa & Bockmann, 1993).
Despite arratia (1990) and datoVo & Bockmann (2010) tried to establish derived character states for the Trichomycterinae, these works did not include Tricho mycterus hasemani (EigEnmann, 1914) and T. johnsoni (FowlEr, 1932).According to de Pinna (1989), T. hase mani and T. johnsoni are each other closest relatives and T. hasemani is related to the Tridentinae due to their expanded cranial fontanel, considering all these taxa derived from a single miniaturization event.Recently, dutra et al. (2012) described T. anhanga dutra, wosiacki & dE Pinna 2012 as being closely related to T. hasemani and T. johnsoni, naming the "T.hasemani group" for this species assemblage.The "T. hasemani group" is monophyletic and possibly related to non-Trichomycterinae taxa (dE Pinna, 1989(dE Pinna, , dutra et al., 2012)).The geographical distribution of this group contrasts with the distribution of other species placed in Trichomycterus, by occurring in lowlands of the Amazon rainforest and Pantanal, instead of being endemic to mountain river drainages of southeastern and southern Brazil (BarBosa & costa, 2010), Andes (arratia, 1998) and those draining the Guyana Shield (EigEnmann, 1909;EigEnmann, 1912;lasso & ProVEnzano, 2002).The new species herein described was collected in the Branco river basin.

Material and Methods
Measurements follow dutra et al. (2012) with the addition of pre-pelvic length (from the middle of the pelvicfin base to the snout tip).Measurements are presented as percentages of standard length (SL), except for subunits of the head, which are presented as percentages of head length (HL).Counts, following BarBosa & costa (2003), were made only in cleared and stained specimens (c&s) prepared following taylor & Van dykE (1985).Scale bars = 1 mm.Nomenclature for the latero-sensory system is according to arratia & huaquin (1995).Specimens were euthanized submerging them in a buffered solution of Ethyl 3-aminobenzoate methanesulfonate (MS-222) at a concentration of 250mg/l, for a period of 10 min, following the guidelines of the Journal of the American Veterinary Medical Association (AVMA Guidelines), and European Commission DGXI consensus for fish euthanasia.Material is deposited in the ichthyological collection of the Instituto de Biologia, Universidade Federal do Rio de Janeiro, Rio de Janeiro (UFRJ), Field Museum of Natural History (FMNH) and in the Academy of Natural Sciences of Philadelphia (ANSP).The method for species delimitation follows the Population Aggregation Analysis (daVis & nixon, 1992), in which one or more populations are recognized as a species by a unique combination of character states.Holotype.UFRJ 10251,14.0mm SL; Brazil: Estado de Roraima: Município de Bonfim: flooded areas in the Tacutu river drainage, tributary of the Branco river drainage, Amazonas river basin, 03° 24' 07"N 59°56'23"W, altitude about 110 m; collected by E. Henschel, F. P. Ottoni, P. Bragança; 10 September 2012.

Diagnosis.
T. wapixana is distinguished from all other species of the T. hasemani group by the presence of 34 to 36 vertebrae (vs.32 in T. hasemani and T. johnsoni, 29 to 32 in T. anhanga); the origin of the anal fin in a vertical through the base of the 20 th , 21 st or 22 nd vertebra (vs.18 th in T. hasemani, 17 th in T. johnsoni and 16 th in T. anhanga).It is distinguished from T. hasemani and T. johnsoni by having the origin of the pelvic fin in a vertical between the base of 15 th and 17 th vertebrae (vs.14 th in T. hasemani and 13 th in T. johnsoni)and by the presence of a dark spot on the middle of the lower lip (vs.absence).Trichomy cterus wapixana differs from T. anhanga by the presence of pelvic fins and girdle (vs.absence); the presence of 10 to 11 dorsal procurrent rays in the caudal fin (vs.6 to 8); the presence of 9 to 12 ventral procurrent rays in the caudal fin (vs.6 to 7); the presence of seven (ii +5 or iii + 4) anal fin rays (vs.ii + 4); the presence of five (i + 4 or ii + 3) pectoral fin rays (vs.i + 2); the origin of the dorsal fin at vertical through the base of the 20 th , 21 st or 22 nd vertebra (vs.16 th or 17 th ); the presence of six branchiostegal rays (vs.four or five); the presence of two pairs of pleural ribs on first two vertebrae posterior to Weberian Complex (vs.single pair); the presence of a series of dark brown spots in the lateral midline of the body (vs.absence); the broad palatine (Fig. 4) (vs.narrow, comma-shaped palatine) (dutra et al., 2012; fig. 2a) and by the presence of the parasphenoid (vs.absence).It differs further from T. johnsoni by the origin of the urogenital pore in a vertical between the base of the 17 th and 19 th vertebrae (vs.15 th ).
Description.Morphometric data for holotype and paratypes given in Table I.Body elongate, subcylindrical on anterior portion, gradually compressed until caudal peduncle.Dorsal profile slightly convex between snout and pectoral-fin origin, straight from that point to caudal peduncle.Ventral profile straight between tip of the snout and insertion of the pectoral fin, gently convex from that point to pelvic-fin origin and straight to end of caudal peduncle.Greatest body depth in vertical immediately in front of pelvic-fin origin.Dorsal and anal fins approximately triangular.Dorsal-fin origin in vertical through base of 20 th , 21 st or 22 nd vertebra.Anal-fin origin in vertical through base of 20 th , 21 st or 22 nd vertebra.Pelvic-fin origin in vertical through base of 15 th , 16 th or 17 th vertebra.Pectoral fin about triangular.First pectoral-fin ray terminating in long filament, about 30 -40% pectoral-fin length.Pelvic fin not covering urogenital pore, bases separated by interspace; insertion in vertical through base of 15 th , 16 th or 17 th vertebra.Caudal fin truncate.Dorsal-fin  A B rays 7 -8 (iii + 4, ii + 5 or iiii + 4, iii + 5) ; anal-fin rays 7 (iii + 4, ii +5); pectoral-fin rays 5 (ii + 3, i + 4); pelvic-fin rays 4 (ii + 2); caudal-fin principal rays 12 (ii + 8 + ii, i + 9 + ii), dorsal procurrent rays 10 to 11, ventral procurrent rays 9 to 12. Total vertebrae 34 to 36; pleural ribs on first two vertebrae posterior to Weberian Complex.Head trapezoidal in dorsal view.Mouth subterminal.Teeth conical.Tip of nasal barbel reaching posterior tip of interopercular patch of odontodes.Tip of maxillary barbel reaching middle of interopercular patch of odontodes.Tip of rictal barbel reaching posterior tip of interopercular patch of odontodes.Six branchiostegal rays.Odontodes conical.Interopercular odontodes 6 to 11, opercular odontodes 9 to 14. Lateral line with two pores, LL1 and LL2.Cephalic portion of latero-sensory canal system restricted to s3, i11 and a praeopercular pore, S4.
Colouration in preserved specimens (Fig. 1a and b).Ground colour cream.Head with dark brown spot extending from anterior surface of eye to anterior margin of upper lip.Dorsal region of neurocranium with light brown spot.Dark brown spot on basis of opercular patch of odontodes and on basis of interopercular patch of odontodes.Ventral surface of head with small dark spots on upper lip and dark spot on lower lip.Nasal, maxillary and rictal barbels with small dark spots concentrated at basis.Dorsal and lateral surfaces of body with chromatophores distributed between head and caudal peduncle.Ventral surface of body with chromatophores concentrated on head and between pelvic and anal fins.Lateral surface with series of dark brown spots.
Dorsal and anal fin hyaline with dark brown blotch on basis of rays.Caudal fin with small dark brown cromatophores.Caudal fin with light brown bar on basis of rays and with small dark brown spot on middle of basis of rays.Pectoral fin hyaline with small dark spot on basis of filament.Pelvic fin hyaline with dark small spot on basis of fin.
Etymology.The wapixana is a native tribe from the Serra da Lua region in western Roraima state, northern Brazil.These natives have occupied this region for, at least, three centuries.The villages of Cantá and Bonfim, where Trichomicterus wapixana was mainly collected, are situated in this area.The Wapixana tribe was oppressed by other native tribes and by colonisers, fact that contributed for a huge cultural loss.
Distribution.Known from the Branco and Negro river drainages, Amazonas river basin (Fig. 5).

Discussion
The "Trichomycterus hasemani group" has been considered as an incertae sedis group among trichomycterids.The first approach concerning the relationships between T. hasemani and T. johnsoni was made by dE Pinna (1989), where these species were proposed to constitute a clade more related to the Tridentinae than to the Trichomycterinae.This hypothesis was based on the presence of an expanded cranial fontanel delimited by the frontal and supraoccipital in the two taxa.Later, dE Pinna (1998), in a cladogram with information combined from several authors, listed the following synapomorphies to support the clade comprising the Vandelliinae, Stegophilinae and Tridentinae: 1 -absence of lacrimal; 2 -lateral opening of Weberian capsule at the end of a neck like constriction; 3 -jaw teeth S-shaped; 4 -mesethmoid cornu with ventral process.The T. hasemani group shares with this clade only the second condition, which makes the hypothesis of sister group relationships between the T. hasemani group and the Tridentinae doubtful. dutra et al. (2012) established the following character states to diagnose the T. hasemani group: 1 -a wide fontanel that occupies most of the skull roof and is delimited by the frontal and supraoccipital (Fig. 3); 2 -absence of the anterior portion of the infraorbital canal (pores i1 and i3); 3 -first pectoral-fin ray much longer than other rays; 4 -absence of branchiostegal rays on the posterior ceratohyal; and 5 -a large posterior process of the palatine, partly forked and expanded distally (Fig. 4).The species herein described shares all these five char-acter states with T. hasemani and T. johnsoni.However, the palatine condition (Fig. 4) is quite different from the other species of the group in T. anhanga, since the partly forked posterior process of the palatine is entirely absent in this species (dutra et al., 2012; fig. 2a).
costa & Bockmann (1994) stated that a sister-group relationship between Sarcoglanidinae and Glanapteryginae would be supported by the reduced dorsal portion of the quadrate, the presence of a large anteriorly directed process in the hyomandibula, vomer rudimentary and miniaturization.Trichomycterus wapixana, T. hasemani and T. johnsoni share with these two subfamilies the last three character states.These authors also established a clade comprising Sarcoglanidinae, Glanapteryginae, Tridentinae, Vandelliinae and Stegophilinae, the so-called TSVSG clade, on the basis of an interopercular patch of odontodes reduced in length and with 15 or fewer odontodes, a reduction in number of the pleural ribs (1 -8), a short posterior portion of the parasphenoid, its tip not reaching the basioccipital or extending only to its anterior part, and metapterygoid reduced or absent.Trichomy cterus wapixana, T. hasemani and T. johnsoni also have all these character states.Since these character states are unique within the Trichomycteridae, they indicate that possibly the T. hasemani group is closely related to the TSVSG clade.However, the position of the group within the family cannot be exactly established, which depends on an inclusive phylogenetic analysis, which is beyond the scope of this study.These species thus remain allocated in Trichomycterus until a proper phylogenetic analysis is developed.
Other miniature Amazonian species-groups also have a problematic taxonomy.In the case of the Scoloplax BailEy & Baskin, 1976, 110 years have passed since the specimens collection in Thayer Expedition in 1866 and the formal description of the genus.This genus was originally described as a member of the Loricariidae, being placed in a new monotypic family (Scoloplacidae) by isBrückEr (1980).According to schaEFEr et al. (1989) the small size of these catfishes was the main problem to describe them, along with other factors such as lack of collecting effort, absence of any distinctive anatomy and their cryptic habitat.These authors also stated that the reductive characters derived from the miniaturization process represent synapomorphies at some phylogenetic level in Scoloplax.The same occurred with species of the genus Fluviphylax whitlEy, 1965: specimens were collected in Thayer Expedition but remained undescribed until 1955(myErs, 1955)).Previously, garman (1895) improperly identified these fishes as undetermined species of Rivulus PoEy, 1860(costa & lE Bail, 1999).These authors also stated that the miniaturization in Flu viphylax is parsimoniously interpreted as a single event.
wEitzman & Vari (1988) published a study focusing on the miniaturization in the several groups of the Neotropical region and the consequences of this process to the phylogeny of those taxa.In the characiform genus Nan nostomus günthEr, 1872, the small size of three miniaturized species is probably derived, but attempts to eluci-date the relationships within the genus were not feasible, and in the catfish genus Corydoras lacEPèdE, 1803, a very specious genus, the relationships within four minitaturized species are unresolved, but possibly involving a single miniaturization event (wEitzman & Vari, 1988).The species herein described belongs to a miniaturized group, but the relationships within the T. hasemani group cannot be established since its phylogenetic relationships are still unknown.