Myloplus schomburgkii (Jardine, 1841)

Machado, Valéria N., Pereira, Victória D., Ota, Rafaela P., Collins, Rupert A., Ândrade, Marcelo, Garcia-Ayala, James R., Jégu, Michel, Farias, Izeni P. & Hrbek, Tomas, 2024, Integrative taxonomy of the black-barred disk pacus (Characiformes: Serrasalmidae), including the redescription of Myloplus schomburgkii and the description of two new species, Neotropical Ichthyology (e 230095) 22 (2), pp. 1-45 : 22-35

Myloplus schomburgkii (Jardine, 1841)

( Figs. 9–14 ; Tab. 5)

Tetragonopterus schomburgkii Jardine, 1841:243 – 44 [original description; plate XXII, Schomburk’s drawing N. 63; type-locality: “ Rio Negro ” ( Guyana), without a type designation].

Myletes schomburgkii Müller, Troschel, 1844:97 [new combination for Tetragonopterus schomburgkii Jardine, 1841 ]. ―Müller, Troschel, 1845:23, 37 – 38 [redescription based on specimens from “ Guyana, Essequibo ”]. ― Valenciennes, 1850:213 – 14 [description as a new species similar to Tetragonopterus schomburgkii , without considering the new generic arrangement; type-locality: Surinam]. ― Steindachner, 1876:134 – 35 [additional description; comments on sexual dimorphism].

Myletes palometa Valenciennes, 1850:214 – 15 [original description, type-locality: “Upper Orinoco” ( Venezuela)]. ― Steindachner, 1876:134 – 35 [considered a junior synonym of Myletes schomburgkii ].

Myleus schomburgkii Eigenmann, 1910:443 [listed; new combination for Tetragonopterus schomburgkii Jardine, 1841 ; locality: Essequibo]. ―Gosline, 1951:40 [Listed]. ―Wallace in Ragazzo, 2002:170–75 [listed; Wallace’s plate 29, 96 and 146; locality: Rio Negro ( Brazil)].

Mylophus schomburgkii Eigenmann, 1912:391–92 [spelling error; new combination; additional description].

Myloplus schomburgkii Eigenmann, 1915:271 [brief description; plates LVI and LVII, locality: ‘Manaos’ (=Manaus, Amazonas, Brazil) and Santarem (Pará, Brazil)].―Norman,1929:824 [listed; locality:rio Madeira (Amazonas, Brazil)]. ― Ohara et al., 2017:135 [brief description; photo; locality: Teles Pires ( Brazil)]. ― Machado et al., 2018:8 [species delimitation using the mitochondrial gene cytochrome c oxidase subunit I (COI) recovered four lineages identified as M. schomburgkii ; figs. 3e,f showing morphological variation of individuals from Nhamundá and Tapajós rivers, respectively]. ―Kolmann et al., 2020:2 [exon-based phylogeny]. ― Silvano et al., 2020:176 [Listed; brief description; fig. 4.179].

Myleus (Prosomyleus) schomburgkii Géry, 1977:266 [listed; brief description of the subgenus; photo on page 269, below; locality: Rio Araguaia ( Brazil)]. ―Géry, 1979:470–71 [description; plate III, above, dentition].

Diagnosis. Myloplus schomburgkii can be easily distinguished from all congeners, except M. aylan and M. sauron , by the presence of a vertical black bar on the flank commonly extending from the dorsal-fin base to the pelvic-fin distal tip (vs. absence). The species can be diagnosed from the aforementioned species by presenting anterior spine of ventral-keel posterior to the vertical through pectoral-fin origin (vs. anterior ventral-keel spine at the vertical through pectoral-fin origin or anterior to this point). Additionally, M. schomburgkii can be distinguished from M. sauron by having greater number of branched dorsal-fin rays (20–25 vs. 17–19), greater dorsal-fin base (29.7– 36.7% SL vs. 25.6–29.1%), shorter dorsal-fin end to anal-fin end distance [21.7–28.8% SL (mean 25.8%) vs. 27.8–32.0% (mean 29.7%)], and shorter adipose-fin base [4.6– 7.2% SL (mean 6.0%) vs. 7.1–9.7% (mean 8.4%)]. Myloplus schomburgkii can be readily diagnosed from M. aylan by possessing dorsal surface of parietal bone straight to slightly concave in lateral view (vs. parietal bone markedly concave), 68–87 (mode 79) total perforated lateral-line scales [vs. 82–95 (mode 84)], 37–38 total vertebrae (vs. 40–41), fewer prepelvic (17–29 vs. 30–39) and total ventral-keel spines [27–41 (mode 33) vs. 38– 55 (mode 49)], and serrae composed by short spines with wide bases (vs. long and with narrow bases). Also M. schomburgkii differs from M. aylan by abrupt anteroposterior decreasing of anal-fin rays length, forming narrow falcated anal-fin lobe in juveniles and females, restricted to the anterior half of the fin, not reaching the middle portion of anal-fin base length (see Sexual dimorphism) (vs. anteroposterior decreasing of anal-fin rays length almost uniform, forming broad lobe, occupying half of the anal-fin extension), with orange to reddish-orange pigmentation along its length, mostly concentrated on anterior rays (vs. conspicuous dark-red to black pigmentation on the entire anal fin).

Description. Morphometric data presented in Tab. 5. Body compressed, overall body shape oval, with highest body depth at dorsal-fin origin. Predorsal and postdorsal length almost equivalent. Head rounded, eye at center of the head. Dorsal profile of head convex from mouth to horizontal through dorsal margin of eye, and straight to slightly concave from this point to base of supraoccipital. Dorsal profile between supraoccipital base and dorsal-fin origin convex. Dorsal-fin base slightly convex. Last dorsal-fin ray distal end not reaching adipose-fin origin when adpressed. Dorsal profile between dorsal-fin insertion and adipose-fin origin straight. Adipose-fin deeper than long, with straight base. Ventral profile of head and body convex from lower lip to anal-fin origin. Anal-fin base slightly convex. Dorsal and ventral profile of caudal peduncle concave.

Mouth terminal. Premaxillary teeth in two rows, outer row with 5*(91) molariform teeth, teeth 1–4 almost equal-sized, tooth 5 smaller, all with sharp, convex edges; inner row with 2*(91) equal-sized teeth with sharp, concave edges; in ventral view, contralateral outer rows forming a V-shaped arch with apex anteriorly pointed; contralateral inner rows forming straight line between 3rd tooth of outer series, space between rows forming a triangular gap. Dentary with 5*(91) molariform teeth, teeth 1–3 substantially bigger than 4–5. Conical symphyseal tooth immediately behind tooth 1 of labial row. Maxilla edentulous.

Scales cycloid, small. Perforated scales on lateral line 68(1), 69(2), 71(2), 72(2), 73(5), 74(7), 75(3), 76(7), 77(5), 78(6), 79*(10), 80(4), 81(8), 82(7), 83(4), 84(3), 86(7), or 87(1). Dorsal-fin base covered by skin flap bearing one or two scale rows. Scale rows between dorsal-fin origin and lateral line 38(1), 39(1), 40(2), 41(3), 42(3), 43(6), 44(6), 45*(6), 46(6), 47(7), 48(8), 49(5), 50(2), 51(3), 52(3), 53(5), 54(1), 55(4), 57(2), or 58(1). Scale rows between lateral line and pelvic-fin origin 36(1), 37(3), 38(5), 39(5), 40(10), 41(7), 42*(10), 43(2), 44(4), 45(7), 46(5), 47(2), 48(2), 49(1), 50(1), 51(3), 52(5), 53(1), 54(1), 55(1), or 58(1). Adipose-fin base covered by three or four scale rows. Scale rows between adipose-fin origin and lateral line 15(2), 16(5), 17*(14), 18(18), 19(21), 20(6), 21(10), 22(3), 23(1), or 27(1). Anal-fin base covered by five or six scale rows. Circumpeduncular scales 28(1), 31(3), 32(6), 33(3), 34(13), 35(8), 36(8), 37(12), 38*(10), 39(2), 40(3), or 45(1).

Dorsal-fin origin slightly anterior to vertical through pelvic-fin origin. Dorsal-fin rays ii-iii, 20(11), 21*(36), 22(30), 23(9), 24(3), or 25(1). Adipose-fin square, length and depth almost equivalent. Pectoral fin feather-shaped, anterior rays longest. Pectoral-fin rays i, 12(1), 14(9), 15*(58), 16(21), or 17(7). Anterior pelvic-fin rays longest, not reaching vertical through last spines of serrae. Pelvic-fin rays i, 6(1) or 7*(97). Last unbranched anal-fin ray most developed (longest and thicker). Anal-fin rays iii or iv, 29(1), 30(1), 31(12), 32(37), 33*(30), 34(4), 35(6), or 36(1). Caudal-fin forked, with almost equal-sized lobes. Total gill rakers on first branchial arch 27(9), 28*(22), 29(19), 30(9) or 31(6). Upper branch with 12(12), 13(32), 14*(15) or 15(7) rakers; lower branch with 13*(7), 14(26), 15(32), or 16(1) rakers; 1*(66) raker at cartilage between cerato- and epibranchial.

Osteology. Dorsal profile of neurocranium convex from premaxilla to posterior margin of frontal bone, slightly concave to straight at parietal, convex from base to tip of supraoccipital process. Lateral view of supraoccipital triangular. Supraneurals 5(15). Dorsal-fin pterygiophores 22(3), 23(5), 24(4) or 25(2). First dorsal-fin pterygiophore inserted between neural spines of 8 th and 9 th (11) or 9 th and 10 th (4) vertebrae, more developed than remaining pterygiophores, with expanded anterior lamella, and bearing forward-oriented predorsal spine. Predorsal spine somewhat similar to scythe, dorsal surface smooth; almost completely covered by skin. Anal-fin pterygiophores 30(1), 31(1), 32(3), 33(6), or 34(1) ( Fig. S3 ).

Total vertebrae 37(12) or 38(1); Weberian apparatus 4(15); abdominal 15(3), 16(9), or 17(1) [pre-dorsal, 4(11) or 5(4); under dorsal-fin 11(5), 12(7), or 13(1)]; caudal 16(1), 17(8) or 18(4) [under dorsal-fin 3(5), 4(7), or 5(1); posterior to dorsal-fin 13(6), or 14(7)]. Anteriormost spine of ventral keel never reaching vertical through pectoral-fin origin. Spines overall thin, with wide base and piercing tips. First prepelvic spines covered by skin. Post pelvic spines more developed than prepelvic spines. Total ventral keel spines 29(3), 30(2), 31(5), 32(9), 33(13), 34(12), 35*(13), 36(4), 37(2), 38(2), 39(2), or 40(2). Composed by prepelvic spines 17(2), 18(3), 19(14), 20(9), 21(14), 22(9), 23(7), 24*(4), 25(2), 26(4), 27(2), or 29(1); postpelvic spines 7(11), 8*(34), or 9(26); and paired spines around anus 3*(6), 4(40), or 5(24).

Coloration in alcohol. Ground coloration light brown dorsally, grading to light-yellow ventrally. Sclera light yellow. Great concentration of melanophores form a wide, well-marked vertical bar extending from region near dorsal-fin base to region near ventral-fin distal end. Numerous scattered, irregular, light brown to dark-brown blots of variable sizes on the entire body including head and fins, mostly concentrated on dorsal regions of body and head. Paired fins yellowish hyaline. Dorsal, anal, and caudal fins yellowish hyaline with inconspicuous dark pigmentation along its distal margins. Adipose fin yellow to light brown ( Fig. 9 ).

Coloration in life. Based on the Neotype and similar specimens ( Fig. 10 ). Ground coloration grayish silver. Whitish-silver iridescent scales on dorsal region of body. Vertical dark bar and scattered dark blots similar of color in alcohol. Scattered striking red pigmentation mostly concentrated along anterior portion of body (including head), becoming less perceptible at posterior portion (see Sexual dimorphism). Different colors of blots confer a rust appearance to specimens. White sclera, in occasional specimens presenting subtle orange-red pigmentation. Paired fins grayish hyaline, with dark pigmentation concentrated along anterior rays. Dorsal, anal, and caudal fins grayish hyaline, with diffuse dark pigmentation along interradial membranes. Adipose fin grayish brown.

Sexual dimorphism. Mature males with two anal-fin lobes; first lobe at anterior rays, less developed; second lobe centered on 12 th or 14 th branched ray, about twice as long as first lobe ( Fig. 10A ). Females and juveniles present single falcate lobe, formed by remarkable prolongation of anterior rays ( Fig. 10B ). On breeding period male specimens present abundant, striking red pigmentation covering most of head and anterior portion of body, spreading through posterior portion less conspicuously, as scattered spots. Females present orangish-red spots of variable sizes, also more concentrated on head and anterior portion of body, but not covering great uninterrupted areas. Though both sexes present scattered, irregular, dark brown to black blots of variable sizes on entire body including head and fins, in females those are less evident. Males present filaments extending dorsal-fin branched rays, and stiff hooks on distal-most lepidotrichia segment of anal-fin branched rays.

Geographical distribution. Myloplus schomburgkii is widespread through the Orinoco and Amazon river basins, occurring in the Casiquiare, Branco, Negro, Aripuanã, Nhamundá, Uatumã, Pitinga, Trombetas, Tapajós, Teles Pires, Xingu and Araguaia, rivers in Venezuela and Brazil. In white water river basins, the species is only captured in tributaries with black or clear water ( Fig. 4 ).

Geographic variation. The species displays three different, well recognizable types of vertical bar on the flanks, which vary in length and shape according to the area of occurrence. In the left bank tributaries of the Amazon River, draining the Guiana shield, the specimens present a well-developed bar ( Fig. 11A ), extending from near the dorsal-fin base to near the pelvic-fin insertion, always conspicuous below lateral line, presenting regular width throughout its length. Conversely, specimens collected on the right bank of the Amazon River basin (Tapajós, Teles Pires, Xingu and Tocantins rivers) present a shorter bar. Specimens from Teles Pires river basin have the portion of the midlateral bar immediately dorsal to the lateral line much more conspicuous, with the appearance of a well-defined, round to vertically oval black spot ( Fig. 12 ); in some specimens, the portion of the midlateral bar ventral to the lateral line is very faint, although still perceptible; in other specimens, it is unrecognizable ( Fig. 13 ). Besides these two different well-defined types of bars, an intermediary type is observed in specimens from Tapajós, Xingu and Tocantins river basins, in which the midlateral bar is much more conspicuous immediately dorsal to lateral line, with rectangular appearance, while the portion ventral to lateral line is very faint ( Fig. 11B ). Although these populations present this color pattern variation, they were recovered as a single lineage in molecular analysis, with low intraspecific variance.

Ecological notes. The species inhabits slow-flowing environments in clear and black water rivers that drain the Guiana and Brazilian Shields. It feeds mainly on aquatic plants (Goulding, 1980), but aquatic insects are also part of its diet (Dary et al., 2017). Although it has been described for the Rio Negro basin, a river with acidic waters ( Sioli, 1984; Venticinque et al., 2016), M. schomburgkii occurs mainly in rivers with clear waters that have low concentrations of sediments and humic compounds ( Sioli, 1984; Ríos-Villamizar et al., 2014; Venticinque et al., 2016) such as the Aripuanã, Branco, Nhamundá, Trombetas, Tapajós and Tocantins-Araguaia rivers. White water river basins of Andean origin seem to constitute a chemical barrier for this species, since in the Branco, Purus and Madeira rivers, the species was captured only in black water lakes of these basins. The same distribution pattern was registered to its congener M. nigrolineatus ( Ota et al., 2020) . Ríos-Villamizar et al. (2020) classify the black waters of the várzea environments as Intermediate type B, since they present intermediate levels of suspended solids originating from ancient sediments and those recently eroded from the Andes. These characteristics allow the presence of M. schomburgkii in Amazonian floodplain environments.

Etymology. Myloplus schomburgkii was described in honor of Robert H. Schomburgk, who, during an expedition to English Guyana, collected individuals of the species, took notes, and illustrated ( Fig. 14 ) the specimen used by Jardine (1841) to describe the new species. A genitive noun.

Remarks. Taxonomic history. Jardine (1841) described several species of Neotropical fishes based on illustrations and commentaries provided by Robert H. Schomburgk from his expedition to Guyana, French Guyana, Surinam, and northern Brazil. Among those species, Jardine (1841) described Tetragonopterus schomburgkii (= Myloplus schomburgkii ) based on the plate XXII (illustration 68 of Schomburgk; Fig. 14 ), from “Rio Negro”, without mentioning the precise type-locality or the existence of a preserved type specimen. Although Jardine (1841) did not designate a holotype, according to Art. 73.1.2. of ICZN (International Commission on Zoological Nomenclature) the holotype was fixed by monotypy, once it is possible to deduce that the author based the description on a single specimen. Furthermore, Art. 73.1.4. states that the “Designation of an illustration of a single specimen as a holotype is to be treated as designation of the specimen illustrated; the fact that the specimen no longer exists or cannot be traced does not of itself invalidate the designation”. Thus, the specimen on which Schomburgk based his illustration (see Jardine, 1841: plate XXII) and meristic data is the holotype.

Müller, Troschel (1844:97) placed Tetragonopterus schomburgkii in Myletes . Subsequently, Müller, Troschel (1845:37–38) provided a complementary description based on a specimen from “ Guiana, in Essequibo ” collected by Richard Schomburgk (Robert Schomburgk’s brother) and deposited in the Zoological Museum of Berlin (ZMB). Zarske (2012) found at ZMB three lots identified as Myletes schomburgkii : ZMB 3638, from Guyana collected by Robert Schomburgk, and ZMB 3639–3640 from Surinam, collected by Stegelich. Jégu, Santos (2002) identified the specimens of lots ZMB 3639 and 3640 as Myleus setiger Müller & Troschel, 1844 in the species redescription, based on teeth arrangement (two premaxillary teeth rows in contact) and on color pattern, with no evidence of a vertical dark bar on the flank, the main diagnostic feature of M. schomburgkii . But the authors did not consider them as typeseries of Myleus setiger because they were collected in Surinam, and its type-locality is Guyana. As Zarske (2012) also provided a figure, x-ray, and a brief description of ZMB 3639, we corroborate here the identification of Jégu, Santos (2002). Another important fact is that they were collected by Stegelich, could not correspond to the specimens used by Müller, Troschel (1845) to redescribe M. schomburgkii .

Eigenmann, during a visit to ZMB in 1910, identified ZMB 3638 as Myleus setiger , and Zarske (2012) suggested it could be Myleus planquettei . However, Jégu, Santos (2002) mentioned the presence of a gap between the two premaxillary teeth rows, and at the symphysis in this specimen, a character currently used to diagnose Myloplus from Myleus . The authors could not identify this specimen at a specific level. The specimen cataloged as ZMB 3638 is a female (184 mm SL), with 25 total dorsal-fin rays; 39 total anal-fin rays (MJ, pers. obs.). Although it could represent the specimen used by Müller, Troschel (1845) to redescribe M. schomburgkii , it does not fit the original description and is not the holotype of Tetragonopterus schomburgkii Jardine, 1841 , once it was collected in Guyana (vs. T. schomburgkii type-locality Rio Negro). Therefore, none of the lots deposited in ZMB could represent the holotype, and the designation of a neotype is necessary (see designation of neotype below).

Valenciennes (1850), in a comprehensive study of the ichthyofauna from Surinam, described Myletes schomburgkii based on a specimen collected by H. H. Dieperink (erroneously spelled Diepering) and deposited in Rijksmuseum van Natuurlijke Historie in Leiden by C. J. Temminck (the director of the Museum). Posteriorly, this material was donated to MNHN (Muséum National d’Histoire Naturelle), in Paris (Boeseman, 1972). This lot was not found (MJ, pers. obs). On the other hand, the possibility of ZMB 3638 and 3639 being syntypes of Myletes schomburgkii , raised by Fricke et al. (2023), certainly can be ruled out. In the description, despite mentioning it as a new species, Valenciennes (1850:212–13) highlighted that the new species “seems to be an extremely close species to Tetragonopterus schomburgkii ”; and cited parts of the original description of T. schomburgkii , without providing a diagnosis between the two species.

Subsequently, Valenciennes (1850:214–15) described Myletes palometa from “upper Orinoco” River, based on the observations made by Mr. Humboldt. He stated that M. palometa had a color pattern very similar to T. schomburgkii but described it as a different species because it was collected from another river basin. He did not establish a type specimen for M. palometa . Finally, in the same manuscript, Valenciennes (1850:215–16) described Myletes divaricatus , a species with similar body shape, but with a second anal-fin lobe, indicating it was a male. The author did not mention the presence of a vertical bar on the middle of the flanks.

Kner (1860:23–24) examined specimens from Rio Branco ( Brazil) that had a second anal-fin lobe. However, they had the typical color pattern of T. schomburgkii (i. e., dark vertical bar on the middle of the flank). Thus, the author suggested that the second anal-fin lobe might consist of a secondary sexual character. By the analysis of the gonads, Steindachner (1876:134–35) confirmed that the presence of a second anal-fin lobe is present exclusively in males, corroborating this feature as a secondary sexual character of Myletes schomburgkii . Furthermore, he considered Myletes schomburgkii the senior synonym of Myletes palometa Valenciennes, 1850 and M. divaricatus . Posteriorly, M. divaricatus was considered a junior synonym of Myleus setiger by Jégu, Santos (2002:51, fig. 10a), and M. palometa a junior synonym of T. schomburgkii by Jégu (2003). According to our molecular results, specimens from Rio Orinoco were recovered within the M. schomburgkii clade, corroborating the synonymy proposed by Steindachner (1876) and Jégu (2003).

Neotype designation. We provide the designation of a neotype, to better define Myloplus schomburgkii , and set a precise type-locality to facilitate the comparison among the congeners described herein [See taxonomic history for explanation about lots ZMB 3638–3939 pointed by Fricke et al., 2023 as possible syntypes of Myletes schomburgkii ]. Although the specimen illustrated was a female, considering the presence of solely an anterior lobe on anal fin (vs. two lobes in males), we chose a male specimen as the neotype, considering it was better preserved and exhibited secondary dimorphism of the species. We also restricted the species type-locality to Rio Negro, Barcelos (INPA 60149) ( Fig. 9 ).

In the original description, Jardine (1841) provided information on general morphology of the body; color pattern, highlighting the presence of a vertical dark bar on the middle of flank; and counts of total fin rays (dorsal-fin rays 25; pectoral-fin rays 15; total pelvic-fin rays 8; total anal-fin rays 39; caudal-fin rays 27); branchiostegal rays (4); ribs (13); and vertebrae (34). Even though the remarkable color pattern of the species was considered until now an autapomorphy, herein we describe two additional species previously identified as M. schomburgkii , that also present a vertical mark on the middle of the flank. Thus, a brief commentary is necessary to explain how we deduced which specimens were conspecific with the holotype and how we based our choice of the neotype.

Myloplus schomburgkii can be promptly distinguished from M. aylan by the anal-fin lobe extension, dark vertical bar shape, and total vertebrae counts provided by Jardine (1841). In M. aylan the anteriormost anal-fin rays length decreases gradually, forming a broad lobe, occupying half of the anal-fin base length (vs. anal-fin rays length decreasing abruptly, forming a narrow falcated anal-fin lobe in juveniles and females, restricted to the anterior third of the fin or slightly posterior to that point, not reaching the middle portion of anal-fin base in M. schomburgkii ). The figure of the holotype in the original description (Jardine, 1841: plate XXII) clearly illustrates the narrow anal-fin lobe. The vertical dark bar is also different in the two congeners, with a uniform width in M. schomburgkii and wider in the central portion in M. aylan . Furthermore, the original description mentions a total of 34 vertebrae (without Weberian apparatus), whereas M. aylan has at least 36 total vertebrae (without Weberian apparatus).

The Myloplus sauron is from Xingu River basin and has a color pattern that resembles Schomburgk’s illustration (i.e., bluish-green scales mostly concentrated at dorsal region of the body and vertical mark on the flanks tapering toward both ends). Nonetheless, Jardine cited the presence of 25 total dorsal-fin rays, and the greatest count known for M. sauron is 22. Myloplus sauron also differs from the specimen illustrated by having a long adipose-fin base (vs. short). Finally, M. sauron is restricted to Xingu River basin and its occurrence at Negro River basin is unlikely.

In addition, in the species delimitation analysis, the species were recovered as distinct lineages in all methods. The interspecific genetic distance between M. schomburgkii and M. aylan was 7.9%; and M. schomburgkii and M. sauron was 9.7%. For further detailed comparison between M. schomburgkii and all congeners see Diagnosis and Molecular Results.

Material examined. Neotype (Present designation). INPA 60149, male, 203.9 mm SL, CTGA 12333 (GenBank accession MG 752391.1), Brazil, Amazonas, Barcelos municipality, rio Negro, 00°56’56.6”S 62°55’44.3”W, 20 Feb 2013, V. N. Machado. Brazil: Amazonas: Barcelos: INPA 52507, 2, 170.8– 197.1 mm SL, CTGA 12274 ( MG 752389), 12335 ( MG 752393), rio Negro, 00°56’56.6”S 62°55’44.3”W, 20 Feb 2013, V. N. Machado. MZUSP 91456, 1, 208.16 mm SL, Tapera community, rio Negro, 00°12’00”N 64°04’00”W, 1 Nov 1972, Expedição Permanente à Amazônia. Apuí: INPA 33610, 76.0 mm SL (x-ray), Amazonas, rio Guariba at Reserva Extrativista do Guariba, 08°42’42”S 60°25’53”W, 14 Nov 2008, W. S. Pedroza, W. Ohara, F. R. Ribeiro & T. F. Teixeira. INPA 36251, 197.5 mm SL, Amazonas, rio Guariba at Reserva Extrativista do Guariba, 08°42’42”S 60°25’53”W, 7 Nov 2008, W. S. Pedroza, W. Ohara, F. R. Ribeiro & T. F. Teixeira. Nhamundá: INPA 46309, 2, 162.8– 175.3 mm SL, CTGA 14526 ( MG 752395), 14527 ( MG 752396), rio Paracatu, tributary of rio Nhamundá, 01°59’51”S 57°2’12”W, 10 Nov 2013, V. N. Machado & R. A. Collins. INPA 46311, 1, 62.9 mm SL, CTGA 14479 ( MG 752394), rio Nhamundá, 01°41’26.9”N 57°25’19.9”W, 11 Nov 2013, V. N. Machado. INPA 46312, 2, 138.5– 173.7 mm SL, rio Nhamundá, 01°49’54.9”S 57°04’23.9”W, 12 Nov 2013, V. N. Machado & R. A. Collins. Novo Airão: INPA 30716, 1, 144.4 mm SL, rio Carabinani, 02°01’24.9”S 61°32’35.9”W, 25 Oct 2004, L. N. Carvalho. INPA 39024, 1, 170.5 mm SL, rio Jauaperi, close to its mouth in rio Negro, 01°42’56”S 61°16’19.9”W, 20 Sep 2011, R. P. Ota. INPA 46062, 3, 151.7– 183.4 mm SL, rio Negro, Arquipélago de Anavilhanas, 02°36’10”S 60°48’46”W, 7 Nov 1996, V. Garcia. Novo Aripuanã: INPA 35586, 2, 194.6– 223.7 mm SL, São Miguel community, rio Aripuanã, 05°59’39.9”S 60°11’35.9”W, 12 Set 2004, L. H. R. Py-Daniel. Presidente Figueiredo: INPA 22192, 1, 236.8 mm SL, vila de Balbina, rio Uatumã, 01°55’21”S 59°28’21”W, 9 Nov 1985, M. Jégu. INPA 22193, 1, 214.81 mm SL, rio Uatumã, igarapé do Arraia, 01°54’31”S 59°28’18”W, 1 Nov 1985, M. Jégu. INPA 46055, 3, 150–203.9 mm SL, vila de Balbina, rio Pitinga at Cachoeira 40 ilhas, 01°08’59.9”S 59°34’59.9”W, 14 Oct 1996, V. Garcia. São Sebastião do Uatumã: INPA 46059, 3 of 4, 207.83– 178.24 mm SL, Santa Maria community, rio Capucapu, close to its mouth in rio Jatapú, cachoeira das Garças, 01°42’59”S 58°34’58”W, 25 Sep 1995, V. Garcia. Mato Grosso: Paranaíta: INPA 44790, 1, 66.7 mm SL, rio Teles Pires 09°30’33”S 56°42’29.9”W, 9 Oct 209, R. R. de Oliveira. INPA 45456, 5, 58.1–107.4 mm SL, rio Teles Pires, Inventário CHTP, 09°22’29”S 56°42’43”W, 15 Dec 2021, Solange, Reginaldo & Rosalvo. INPA 59651, 1, 156.8 mm SL, rio Teles Pires, 09°22’30”S 56°42’43”W, L. N. Carvalho. MZUSP 99863, 5 of 13, 46.2– 45.7 mm SL, rio Teles Pires, 09°18’42”S 56°46’46.9”W, 9 Mar 2008, L. Netto-Ferreira. Carlinda: MZUSP 68215, 1, 93.3 mm SL, rio Teles Pires, 09°59’25”S 55°33’48”W, 29 Sep 2007, F. A. Machado. Paranatinga: MZUSP 94072, 5, 110.38– 134.54 mm SL, rio Culuene at cachoeira do Adelino, 13°53’55’S 53°19’17”W, 20 May 2007, F. C. T. Lima, F. A. Machado & J. Birindelli. MZUSP 98124, 3, 121.34– 144.82 mm SL, rio Culuene, 13°49’59.9”S 53°15’00”W, 2 Oct 2007, F. C. T. Lima, F. A. Machado, A. C. Ribeiro & C. L. R. Moreira. Peixoto de Azevedo: MZUSP 97639, 5, 152.7– 200.9 mm SL, ( OR 366886), rio Peixoto de Azevedo, tributary of rio Teles Pires, 10°17’13.9”S 54°50’57”W, 17 Oct 2007, J. Birindelli, L. Netto-Ferreira, M. H. Sabaj & N. Lujan. Pará: Oriximiná: MZUSP 15656, 1, 208.9 mm SL, rio Trombetas at Reserva Biológica do Trombetas, 01°25’00”S 56° 37’00”W, 23 Jul 1979, M. Goulding. Roraima: Atauba: INPA 46280, 2, 152.3– 198.8 mm SL, CTGA 12199 ( MG 752387), 12200 ( MG 752388), left bank of rio Branco, 01°03’38”S 61°51’00”W, 10 Dec 2013, V. N. Machado & R. A. Collins. Caracaraí: INPA 23398, 1, 174.53 mm SL, CTGA 14611 ( MG 752398), rio Capivara, tributary of rio Branco, 01°06’00”N 61°55’41”W, 10 Dec 2013. V. N. Machado. MZUSP 79209, 2, 131.6– 168.7 mm SL, rio Branco, 01°30’00”N 61°16’00”W, 28 Oct 1979, M. Goulding. MZUSP 79210, 2, 161.7– 194.7 mm SL, rio Branco, 01°10’00”S 61°52’00”W, 9 May 1979, M. Goulding. Venezuela: ANSP 192193, 1, 110.7 mm SL, caño Yurebita, tributaryof rio Ventuari, 04°13’07.64”N 66°25’25.5”W, 15 Apr 2010, N. K. Lujan, J. Birindelli & V. Meza.

Molecular differentiation. We obtained sequence data for representatives of 10 of the 12 recognized species of Myloplus : M. arnoldi , M. asterias , M. levis (Eigenmann & McAtee, 1907) , M. lobatus , M. lucienae , M. rubripinnis , M. schomburgkii , M. tiete (Eigenmann & Norris, 1900) , M. nigrolineatus and M. zorroi Andrade, Jégu & Giarrizzo, 2016 . No tissues or sequences of M. tumukumak Andrade, Jégu & Gama, 2018 and M. torquatus were available. Nineteen additional nominal Myleini species were also used in the analysis ( Fig. 15 ). A total of 89 sequences initially identified as “ Myloplus schomburgkii ” were obtained, with 40 of these newly generated. These “ M. schomburgkii ” sequences represent 36 haplotypes and 32 unique localities from six major tributaries of the Amazon basin in Brazil (Negro, Branco, Nhamundá, Madeira, Tapajós, Xingu), as well as the Nanay River in Peru, Orinoco basin, and Tocantins-Araguaia system.

The full nucleotide dataset represented 564 sequences of a median sequence length of 612 bp (range 312–621 bp). After dereplication, 209 sequences of length 621 bp remained (range 405–621 bp). Species discovery using mPTP on the maximum clade credibility consensus of MrBayes trees partitioned the haplotype dataset into 48 putative species clusters ( Fig. 15 ). Within “ M. schomburgkii ” a total of six geographicallystructured species clusters were estimated, comprising: a taxon from the Xingu River corresponding to M. sauron ( BPP = 0.97); a taxon from the upper Amazon, Madeira and Branco rivers corresponding to M. aylan ( BPP = 0.96); M. schomburgkii from Brazilian Shield rivers including Lower Xingu, Tapajós and Araguaia ( BPP = 0.16); M. schomburgkii from Guiana Shield rivers including Negro, Branco and Nhamundá ( BPP = 0.26); M. schomburgkii from Orinoco and Casiquiare rivers ( BPP = 0.32); and M. schomburgkii from Teles Pires River ( BPP = 0.14). Due to the uncertainty within the M. schomburgkii Brazilian and Guiana Shield delimitations with low posterior support, a full range of alternative candidate species is provided in Tab. 6. Among these delimitations, a unified M. schomburgkii cluster from the Brazilian and Guiana Shield had the greatest support ( BPP = 0.54). Individuals of M. sauron , M. aylan , and the Brazilian and Guiana Shield M. schomburgkii were all monophyletic with posterior clade support value of 1 ( Fig. 15 ). Smallest interspecific genetic distances ( Tab. 6) were 0.097 ( M. sauron vs. M. schomburgkii ) and 0.079 ( M. aylan vs. M. schomburgkii ). The distance between M. sauron and M. aylan was 0.11 (data not shown).

Figures

FIGURE 3 | Myloplus aylan, coloration immediately after capture. Mature female, lago Uauaçu, Purus River basin. Photographed by Lucia Rapp Py-Daniel.

FIGURE 4 | Geographic distribution of Myloplus schomburgkii (blue circles; the pink star at rio Negro basin represents the type-locality), M. sauron (red squares; the yellow star at rio Xingu basin represents the type-locality) and M. aylan (purple triangles; the black star at rio Branco basin represents the type-locality).

FIGURE 9 | Myloplus schomburgkii, INPA 60149, neotype, male, 203.9 mm SL, Brazil, Amazonas, Barcelos municipality, rio Negro.

FIGURE 10 | Myloplus schomburgkii, color pattern immediately after capture. A. INPA 60149, neotype, 203.9 mm SL, male, with second lobe well developed and remarkable breeding coloration.B. INPA 52507, 197.1 mm SL, female. Both from Amazonas, Barcelos municipality, rio Negro.

FIGURE 11 | Geographic variation in Myloplus schomburgkii. A. Roraima, Caracaraí municipality, rio Água Boa do Univiní. B. Tapajós River basin. Both recently collected, females, not preserved. Notice, in A, the vertical bar conspicuous across the flank, while in B the bar is conspicuous only dorsally to the lateral line.

FIGURE 12 | Geographic variation in Myloplus schomburgkii, 202.4 mm SL. Teles Pires River basin. Notice the oval mark on the flank, instead of a vertical bar.

FIGURE 13 | Geographic variation in Myloplus schomburgkii. Both from the Teles Pires River basin, immediately after capture. A. Female. B. Male. Notice the oval mark on the flank, instead of a vertical bar. Photographed by Renan Condé Pires, not preserved.

FIGURE 14 | Illustration of a female from rio Negro basin, provided by Robert H. Schomburgk, in which Jardine in Schomburgk (1841, plate 22) based the original description of Tetragonopterus schomburgkii.

FIGURE 15 | Maximum clade credibility tree obtained from 10,000 post-burnin MrBayes trees. Colored dots represent unique mPTP species delimitation clusters (total 48), with posterior probability for each delimitation reported in the corresponding bar plot (right). Dataset comprised 209 dereplicated COI barcodes. The Myloplus aylan samples are highlighted in green, M. sauron in orange, and M. schomburgkii in blue. Holotype of M. aylan and neotype of M. schomburgkii highlighted in bold.