A new species of barred Sternopygus (Gymnotiformes: Sternopygidae) from the Orinoco River

Abstract A new species of Sternopygus is described from the Orinoco River of Venezuela using traditional methods of morphometrics and meristics, and micro-computed tomography (micro-CT) imaging for osteological analysis. The new species is readily separated from all congeners in having broad, vertical pigment bars that extend from the mid-dorsum to the ventral margin of the pterygiophores. A similar color pattern, characterized by subtle differences in the densities and sizes of chromatophores, is also present in juveniles of S. obtusirostris from the Amazon River, juveniles of S. sabaji from rivers of the Guiana Shield, and S. astrabes from clearwater and blackwater terra firme streams of lowlands around the Guiana Shield. The new species further differs from other congeners in the Orinoco basin by having a reduced humeral pigment blotch with poorly defined margins, a proportionally smaller head, a longer body cavity, a more slender body shape in lateral profile, and in having vertical pigment bars that extend ventrally to the pterygiophores (vs. pigment saddles not reaching the pterygiophores). The description of this species raises to three the number of Sternopygus species in the Orinoco basin, and to 11 the total number of Sternopygus species.

Most Sternopygus species share a similar color pattern with a base color composed of small, densely arranged gray chromatophores.Some species have a dark humeral blotch with variable contrast to the background coloration, and a distinctive yellow or white longitudinal stripe extending between the hypaxial and pterygiophore muscles on the posterior third of the body.These aspects of coloration are variable within and among nominal species and are sometimes absent, with some specimens ranging in color from deep black to pinkish white.At least three valid Sternopygus species possess a distinctive color pattern composed of 1-4 broad, dark vertical bars or saddles across the dorsal midline at some stage in their ontogeny: S. astrabes, S. obtusirostris Steindachner, 1881, S. sabaji Torgersen & Albert, 2022 (Fig. 1;Mago-Leccia, 1994;Crampton et al., 2004b;Torgersen, Albert, 2022).The monophyly, species limits, variation, and species richness of species with broad vertical pigment bars or saddles remains poorly understood and these topics are not addressed here.Here we describe a new species of barred Sternopygus from the lower Orinoco basin of Venezuela, bringing the total number of species in the genus to 11, the number of species known in the Orinoco basin to three, the number of species in the Guiana Shield region to four, and the number of Sternopygus species possessing dark vertical bars to four.

MATERIAL AND METHODS
A total of 46 specimens of the new species described herein were identified from museum lots collected in the lower Orinoco drainage of Venezuela between 1985 and 2010, with most specimens collected specifically from the confluence of the Orinoco and Caura rivers by L. Aguana, B. Chernoff, R. Royero, and W. Saul.Only specimens collected near the confluence of the Orinoco and Caura rivers were included in the type series.We were unable to deposit type specimens in Venezuela because of the ongoing political and economic instability.No animal experimentation or collection permits or approvals were necessary for the completion of this work.
Morphometric measurements followed Hulen et al. (2005).We used digital calipers and an ocular micrometer attached to an Olympus SZX12 dissecting microscope, measuring point-to-point linear distances from standard landmarks to the nearest 0.01 mm on the left side of the body when possible.
We measured: (1) length to the end of the anal fin (LEA) measured as the length from the tip of the snout (anterior margin of upper jaw at mid-axis of body) to the end of last anal-fin ray; (2) anal-fin length (AFL), measured from the origin of the anal fin at the isthmus to the end of the fin; (3) caudal appendage (CA), measured as the distance from the last anal-fin ray to the distal end of the caudal filament.Note: the CA in sternopygid fishes is often damaged, entirely missing, or in a variable state of regeneration.Therefore, the values reported here are not considered to have diagnostic value; (4) body depth (BD), measured as a vertical distance from the origin of the anal fin to the dorsal body border, (5) body width (BW), measured as body width at the origin of the anal-fin; (6) head length (HL), measured from the posterior margin of the bony opercle to the tip of the snout; (7) postorbital head length (PO), measured from the posterior margin of the bony opercle to posterior rim of free orbital margin of eye; (8) preorbital head length (PR), measured from the anterior rim of the orbital free margin to tip of snout; (9) eye diameter (ED), measured as the horizontal distance between the anterior and posterior rims of the free orbital margin; (10) interorbital length (IO), measured between the dorsomedial margins of the free orbital margin; (11) inter-narial distance (NN), measured from the posterior margin of the anterior nares to the anterior margin of the posterior nares; (12) mouth width (MW), measured as the horizontal distance of the gape at the rictus; (13) branchial opening (BO) measured as the distance from the posterodorsal to anteroventral extent of the skin fold of the branchial opening along the anterior margin; (14) head depth (HD), measured as the vertical distance at the nape to ventral body border with the lateral line held horizontal; (15) head width (HW) measured as the width at nape; (16) preanal distance (PA), measured from the origin of the anal fin to the posterior margin of anus; (17) pectoral-fin length (P1), measured from the dorsal border of fin base where it contacts the cleithrum to the tip of the longest ray.Morphometric data were standardized for size by reporting values as a percent of HL, except in HL %, BD %, BW %, and CA %, which are reported as a percent of LEA.We assessed a body-shape tapering ratio (TR) as the ratio of BD at 75% LEA divided by BD at 25% LEA (Fig. 2).To reduce the effects of allometry, morphometric measurements used in the diagnosis were limited to morphologically mature specimens (more than 50% maximum known TL).Specimens that are damaged or with incompletely regenerated tails were excluded from analysis.Diagnostic trait values are reported as non-overlapping range values or range values within the 95% confidence interval (i.e., overlap less than 5.0%).Additional traits that are useful in identifying specimens of the new species are reported in the Diagnosis.The sex of six specimens was assessed by direct examination of gonads following Waddell, Crampton ( 2018 Micro-computed tomography (micro-CT) scans were made of 10 specimens from the type series of the new species using a Bruker SkyScan1273 with an x-ray source voltage of 65 kV.Only the head region, defined as the part of the body extending from the tip of the snout to a point between vertebrae 4-8 along the longitudinal axis of the specimens were scanned due to exceedingly large file sizes resulting from full-body scans and the relatively low return of information from scans past the body cavity.Osteological observations were made from 3D renderings of the micro-CT scans in the freeware Slicer (Fedorov et al., 2012)    Two-dimensional geometric morphometrics (2D GMM) were used to capture shape variation in the new species described herein and for comparison to congeners.Photographs of 91 specimens were taken using a Nikon Coolpix S9700 digital camera with all specimens in the same position in left lateral view with the dorsum forming a nearly straight line from the nape to the end of the anal fin.Photos were then converted to thin plate spline (.tps) files using tpsUtil (Rohlf, 2008), and seven homologous landmarks and six pseudo-landmarks (Fig. 2; Tab. 2) were placed on each photograph using FIJI (ImageJ) (Schindelin et al., 2012).Landmark coordinates were exported as .txtfiles and then imported into MorphoJ (Klingenberg, 2011) where a Procrustes superimposition was performed to remove the effects of size and scaling among specimens.A principal components analysis (PCA) was then performed in MorphoJ to identify the primary axes of variance.Data for comparison to the new species were collected from an additional 226 specimens from 116 lots of Sternopygus belonging to nine other species.These are listed in Tab. 3. Museum codes and abbreviations follow Sabaj (2020).Diagnosis.Sternopygus sarae can be distinguished from all other congeners by the presence of three broad, dark vertical pigment bars with irregular margins that extend across the mid-dorsum to the ventral margin of the pterygiophores on the lateral body surfaces (vs.no pigment bars in the S. aequilabiatus (Humboldt, 1805) group, S. arenatus (Eydoux & Souleyet, 1850), S. branco Crampton, Hulen &Albert, 2004, S. macrurus, andS. xingu Albert &Fink, 1996; pigment saddles that do not extend to the base of the pterygiophores in S. astrabes, juvenile S. obtusirostris, and juvenile S. sabaji, and no bars or saddles in adult S. obtusirostris and S. sabaji; Fig. 1).See Discussion for comments on the appearance and regularity of the pigment bars.Sternopygus sarae further differs from all other congeners by a unique combination of characters of head and body proportions, and osteological traits.Sternopygus sarae differs from species of the S. aequilabiatus group, and from S. arenatus, S. macrurus, S. sabaji, and S. xingu in a relatively shorter head length (HL% 9.9-12.2vs. 12.5-19.6),from members of the S. aequilabiatus group, S. branco, S. sabaji, and S. xingu in a relatively greater interorbital distance (IO% 27.5-37.6vs. 14.4-26.5),and from S. arenatus and S. branco in a relatively greater mouth width (MW% 16.0-19.1 vs. 11.0-13.9).Sternopygus sarae has a proportionally shorter head and more slender body than sympatic congeners (S. astrabes and S. macrurus).Comparisons of relative head length (HL%) between S. sarae, S. astrabes, and S. macrurus are presented in Fig. 4. HL% values overlap slightly between S. sarae and its sympatric congener, S. astrabes.Sternopygus sarae also has a more slender body shape in lateral profile (BD%) than its sympatric congeners (S. astrabes and S. macrurus) with overlap only in extreme values (Fig. 4).While not diagnostic, the relatively shorter HL% and BD% of S. sarae are still useful in separating most adult specimens from S. astrabes and S. macrurus.

Sternopygus
Sternopygus sarae is most similar to S. obtusirostris from the Amazon from which it differs by a lighter body coloration (lateral surfaces pale brown vs. dark brown), a lighter coloration of pectoral and anal fin membranes (hyalin vs. dark brown), a less tapered body shape in lateral profile Tab. 5), and a less robustly ossified neurocranium with absence of paired ventral ridges of the posterior portion of the parasphenoid anterior limb (vs.present; Fig. 5).Sternopygus sarae further differs from the partially sympatric S. astrabes by having a relatively smaller eye diameter (ED% 6.7-       Description.Head and body shape in Fig. 3; cephalic mechanosensory line and pore configuration in Fig. 6; morphometric and meristic data in Tab.4; and body size distribution of type series in Fig. 7. Maximum known body size 407 mm TL (340 mm LEA).Body elongate and slender compared to sympatric congeners, S. astrabes and S. macrurus (Fig. 8), somewhat compressed laterally in body cavity region, more laterally compressed in post-coelomic body region; body widest immediately behind head; whole body covered in ovoid (axially elongate) cycloid scales, except head and fins.Lateral line complete and non-interrupted, extending onto tail posterior to last anal-fin ray.Longitudinal stripe thin and extending along posterior half of body, sometimes very faint or absent.Eyes relatively small, and not covered by layer of skin.Body relatively shallow with short head length relative to most congeners.Anterior naris slightly tubular, posterior naris non-tubular.14-16 total pectoral-fin rays.278-325 anal-fin rays, all unbranched.24-26 precaudal vertebrae.Pigment pattern composed of three alternating dark bars.Bars (B1-B3) composed of less densely arranged chromatophores with larger diameters.Interbars (I1, I2) composed of more densely arranged chromatophores with smaller diameters (Fig. 9).Note: bar appearance differs when viewed at whole body vs close-up such that pigment patches with large chromatophores and large inter-chromatophore spaces appear darker when specimen is viewed as a whole (e.g., at arms distance, panel A as compared to lighter when viewed close up panel B).17/15 ni.bio.br| scielo.br/niKevin T. Torgersen, Aleidy M. Galindo-Cuervo, Roberto E. Reis and James S. Albert Neurocranium.Neurocranium in dorsal, lateral, and ventral views in Fig. 10.Preorbital region of neurocranium rounded and slightly decurved in lateral view, dorsal margin convex along its entire extent from tip of mesethmoid to posterior margin of parietal, parasphenoid ventral margin slightly concave in lateral view anterior to basipterygoid process (sensu Arratia, 2013: fig.48), or parasphenoid lateral wing (sensu Adriaens, Verraes, 1998: fig. 19b).Neurocranium relatively well-ossified compared to other gymnotiforms, all bones of braincase ossified to their peripheral margins with little or no intervening cartilaginous plates.Supraorbital canal mostly fused to frontal.Lateral ethmoid cartilage well ossified, its ventral margin extending lateral to vomer.Sphenotic small, lateral process not extending beyond lateral neurocranium margin.Foramen between parasphenoid, pterosphenoid, and orbitosphenoid relatively small, bones of ethmoid region (e.g., lateral ethmoid, ventral ethmoid) relatively well-ossified as compared to congeners.Anterior and posterior cranial fontanelles large, separated by narrow interorbital bridge, posterior fontanelle extending posterior to about vertical with anterior margin of exoccipital.Antorbital and postorbital processes of frontal robust, anterior process tapering to distal tip.Parasphenoid anterior portion wider than posterior portion; parasphenoid ventral margin flat, without longitudinal ridges on lateral margins, but with pronounced transversely oriented basipterygoid ridges.
Oral jaws.Mouth terminal to slightly inferior, anterior margin of mesethmoid extending slightly anterior to anterior margin of dentary; Premaxilla ovoid in frontal view with five unevenly arranged tooth rows on each side, with nine large straight conical teeth on posterior margin, and approximately 35 large, straight conical teeth in anterior four rows.Maxilla broad, with pronounced lateral ridge, and angled at two thirds distance of posterior blade.Dentary of intermediate length, oral margin about as long distance from mandibular symphysis to dentary-retroarticular articulation.Dentary dentition "brush-shaped" (sensu Mago-Leccia, 1978), with approximately 85 mostly recurved teeth arranged in 4-5 irregular rows near mental symphysis tapering laterally to single tooth at posterolateral margin of tooth field.Dentary anteroventral margin without small ventral process.Anterior portion of preopercular-mandibular laterosensory canal completely ossified on medial margin with three constrictions along lateral margin of dentary descending process.Suspensorium (Fig. 11).All elements of suspensorium well-ossified.Hyomandibula broad with two large foramina on dorsomedial surface (through which pass nerves V, VII, and lateralis) and four separate foramina on lateral surface (through which pass posterior, supraorbital, infraorbital, and preoperculo-mandibular rami of same nerves); for comparison see Albert et al., 2005: fig. 14.Symplectic incompletely ossified at dorsoposterior margin.Quadrate well-ossified and abutting endopterygoid but with cartilaginous margin with metapterygoid.Metapterygoid lower portion poorly ossified.Endopterygoid broad, with about 16 pointed teeth on anterior margin of medial surface, medial margin not contacting other contralateral endopterygoid at midline of palate, endopterygoid ascending process tapering dorsally, connected by thin tendon to ventral surface of frontal.Ascending process of endopterygoid slightly curved posteromedially.Palatine unossified.Opercular series (Fig. 11).Opercle well ossified with rounded dorsal, anterior, and posterior margins, dorsal margin with broad median shelf, anterior and posterior margins convex, anterior articulating process large and horn-shaped, lateral opercular surface mostly smooth with large lacunae in ventral and dorsal fields.Preopercle poorly ossified, posterior, ventral, and anterior margins ragged, anterior margin unossified.Branchial basket.Urohyal well-ossified, posterior blade extending to third branchial arch.Basihyal fan-shaped, basibranchial of third arch cone-shaped, fourth and fifth arches unossified.Hypohyals without medial process and not contacting each other at ventral midline.Five basibranchials, anterior two slender, posterior three broad.Pharyngeal jaws large and robust, pharyngobranchials of fifth branchial arch with 18-20 large conical teeth arranged in 3-4 irregular rows, opposed to large hypobranchials with 9-10 large conical teeth arranged in three irregular rows.Ceratobranchial of fifth arch with large triangular lateral margin.5-6 squat gill rakers, about as wide as long, in irregular rows on anterodorsal and anteroventral margins of ceratobranchials of all five gill arches.
Pectoral girdle (Fig. 12).Cleithrum well-ossified, anterior portion rounded and broadly contacting contralateral cleithrum at its anterior margin, ascending portion with sharp ridge on anterior margin of lateral surface.Anterior coracoid process thin and elongate, extending halfway to anterior tip of cleithrum.Supracleithrum fused to post temporal.Mesocoracoid unossified.Scapula fused to cleithrum.Five ossified proximal pectoral fin radials, lateral three fused at their bases.Conservation status.This species is currently known from limited collections in the Orinoco River near the mouth of the Caura River, with an Extension of Occurrence (EOO) calculated by the minimum convex polygon of approximately 1,500 km 2 .As very little is currently known of its actual distribution range or populational trends, and considering existing threats caused by deforestation, extensive agriculture, and gold mining in the region, we suggest the species is preliminarily assessed as Data Deficient (DD) according to the International Union for Conservation of Nature (IUCN) categories and criteria (IUCN Standards and Petitions Subcommittee, 2022).

DISCUSSION
Barred pigment patterns in Sternopygus.The alternating darkly pigmented bars and lighter interbars on the lateral body surface of S. sarae result from different densities and sizes of chromatophores.These patterns are sometimes more easily discerned when observed closely with a hand lens than at a distance; compare Figs. 3 and 8.The pigmented bars of S. sarae from the lower Orinoco River extend to the base of the pterygiophores on the lateral body surfaces, as compared to pigment saddles that do not extend to base of the pterygiophores in juvenile S. obtusirostris from the Amazon River, juvenile S. sabaji from Atlantic drainages of the Guiana Shield, and S. astrabes from lowlands around the Guiana Shield (Fig. 1).The pigment bars are regularly arranged in most specimens of S. sarae and are less regularly formed in some specimens (Fig. 9).The high proportion of specimens with 2-4 regular pigment bars/interbars relative to those with inconsistent pigment patterns supports the claim that the chromatophores are largely arranged in vertical bars and not irregular blotches in S. sarae.The appearance of these bars in life is presently unknown as this species is not known to have been photographed with live coloration.
The dark saddles of live juvenile S. obtusirostris become very conspicuous at night.At this time the interbars become extremely pallid, due to chromatophore contraction.In contrast, juveniles of S. obtusirostris are uniformly black and the bars not visible at all during the day (Crampton et al., 2004, fig. 5).Likewise, the dark bars of S. astrabes from both the Central Amazon of Brazil and the Rio Orinoco of Venezuela are much more clearly visible in live specimens at night (W.Crampton, 2023, pers. comm.).These observations suggest that variation in the time of day that specimens are captured and fixed (i.e., day vs. night) may influence the intensity of dark bars in Sternopygus.Likewise, as specimens become more generally faded with time, the dark bars may become harder to discern.
To date, little attention has been given to the phylogenetic distribution of dark vertical bars in Sternopygus, or to the evolutionary or adaptive reasons for such a pigment pattern.A future genus-wide phylogenetic analysis will allow us to infer relationships amongst these species with similar pigment patterns, but such an analysis lies outside the scope of this description.Body shape differences.Body shape differences between S. sarae and its sympatric congeners (Fig. 4) may be consistent with predictions of the impedance matching hypothesis of Hopkins (1999).This hypothesis predicts a shorter, thicker electric organ in high-conductivity whitewaters, and a longer, thinner organ in low conductivity black and clearwaters.Sternopygus sarae from the sediment-rich, white water Orinoco River, has a longer, less-tapered body than does S. astrabes from sediment-poor black and clear water rivers of the Guiana Shield (Fig. 15).It is interesting to note that the more eurytopic S. macrurus, which inhabits wide range of water types, also has a wider range of HL, BD, and TR values, and multimodal distributions of these values.
.br | scielo.br/niNew species of Sternopygus from the Orinoco basin ) and Waddell et al. (2019).Meristic counts also follow Hulen et al. (2005) and include: (1) anal-fin rays (AFR); (2) pectoral-fin rays (P1R) including all branched and unbranched rays; (3) precaudal vertebrae (PCV) including the four vertebrae that compose the Weberian apparatus; (4) scales above the lateral line (SAL) counted along a vertical line at the end of the body cavity; (5) scales below the lateral line (SBL) from the same point as SAL to the base of the anal-fin pterygiophores; (6) scales over the pterygiophores (SOP) counted from the same point as SAL at the base of the anal-fin pterygiophores to the anal-fin ventral border.
after being prepared in Fiji/ImageJ (Schindelin et al., 2012) following Buser et al. (2020).Precaudal vertebrae were counted from x-rays obtained from the CT scanner.Figures featuring images of 3D renderings were accomplished by taking screen captures of the renderings generated in Slicer before preparing them in additional photo editing software.The package 'ggridges' was used to create Ridgeline plots in R to facilitate the comparison of trait value distributions (Wilke, 2018).

FIGURE 2 |
FIGURE 2 | Landmark scheme used in geometric morphometric analyses.Landmarks indicated by small red circles, pseudolandmarks by small blue circles.Landmarks described in Tab. 2. Body depth (BD) measured at 25% and 75% LEA to calculate Taper Ratio (TR).Photograph of S. macrurus, ANSP 209719.

FIGURE 4 |
FIGURE 4 | Ridgeline plots of relative head length (HL%) and relative body depth (BD%) as a percentage of LEA for three sympatric Sternopygus species in the Guiana Shield region.Dark red regions indicate values outside 95% confidence interval from the mean; light red indicates overlapping values outside the 95% confidence interval.

FIGURE 7 |
FIGURE 7 | Histogram of specimen sizes, reported in LEA, of all 46 specimens comprising the type series of Sternopygus sarae.

FIGURE 9 |
FIGURE 9 | A. Line drawing of five paratype specimens of Sternopygus sarae (ANSP 160357).Top two drawings depict typical coloration pattern, lower three drawings depict uncommon variations of pigmentation seen in preserved specimens.Scale bar = 1 cm.Bars (B1-B3) are composed of less densely arranged chromatophores that have larger diameters.Interbars (I1, I2) are composed of more densely arranged chromatophores that have smaller diameters.Homologous chromatophore fields with barred and interbarred regions are labeled for all four of the barred Sternopygus species.B. Close-up of specimen showing differences of size and density of chromatophores making up the bar and interbar regions.
FIGURE 10 | Sternopygus sarae neurocranium 3D rendering from micro-CT scan of holotype, ANSP 209718.A. Dorsal view; B. Lateral view; C. Ventral view.The enlarged canal bones fused to the frontal and parietal bones of the neurocranium were included in those segments.Scale bar = 1 cm.

FIGURE 12 |
FIGURE 12 | Sternopygus sarae pectoral girdle 3D rendering from micro-CT scan of holotype, ANSP 209718.A. Lateral view of left side; B. Medial view of same side.Scale bar = 5 mm.

FIGURE 14 |
FIGURE 14 | Locality collection map of all known lots of Sternopygus sarae (red) and selected S. astrabes (blue) lots from the Orinoco River drainage.Holotypes indicated by a star.Symbols may represent more than one collection.
; (14)developmental origin of adult electric organ from both hypaxial and epaxial muscles

TABLE 1 |
Summary of all valid species of Sternopygus with information regarding primary type specimens and locality drainage for each species.Country of collection of primary types given in parenthesis.
New species of Sternopygus from the Orinoco basin ni.bio.br| scielo.br/ni

TABLE 4 |
Absolute values of morphometric and meristic data for specimens of the type series of Sternopygus sarae.Abbreviations defined in Material and Methods.Kevin T. Torgersen, Aleidy M. Galindo-Cuervo, Roberto E. Reis and James S. Albert provided in Tab. 3. The new species readily differs from another barred sternopygid, Japigny kirschbaum Meunier, Jégu & Keith, 2011 from the Atlantic coast basins of the Guiana Shield, by the possession of all unbranched anal-fin rays and a free orbital margin.13/15 ni.bio.br| scielo.br/ni

TABLE 5 |
Taper Ratio (TR) values for 25 specimens representing four species of Sternopygus.Abbreviations defined in Material and Methods.

Buser TJ, Boyd OF, Cortés A, Donatelli CM, Kolmann MA, Luparell JL et al.
The natural historian's guide to the CT galaxy: step-by-step instructions for preparing and analyzing computed tomographic (CT) data using cross-platform, open access software.