Phylogenetic Relationships and Historical Biogeography of Cetopsorhamdia (Siluriformes: Heptapteridae), with Description of a New Species from the Orinoco Basin, Colombia
Cetopsorhamdia as currently classified is corroborated as non-monophyletic, based on a comprehensive multilocus phylogenetic analysis, with complete taxonomic representation from northern South America, and including its type species. However, a subgroup of species that agrees with a previous phylogenetic definition of the genus, based on morphology, is recovered as monophyletic and is designated herein as Cetopsorhamdia sensu stricto. A new species from piedmont tributaries of the Meta and Guaviare Rivers, Orinoco basin in Colombia, was recovered as sister to remaining cis-Andean species of the sensu stricto concept of the genus and is described here. The new species is distinguished from all sensu stricto congeners by a unique combination of characters: distal end of pectoral fin surpassing dorsal-fin origin, adipose-fin base short (12.9–15.9% of SL), lower lobe of caudal fin longer than upper lobe (31.7–37.3% of SL), body homogeneously brown, and 35–36 total vertebrae. A clear vicariant cis- and trans-Andean pattern is obtained for species of Cetopsorhamdia sensu stricto, predating isolation of present Magdalena and Lake Maracaibo basins by orogenesis of the Eastern Cordillera and Mérida Andes, respectively.
Cetopsorhamdia como es actualmente clasificado es corroborado como no monofilético, con base en un análisis filogenético multilocus exhaustivo, con representación taxonómica completa del norte de Sudamérica, e incluyendo su especie tipo. Sin embargo, un subgrupo de especies que concuerda con una definición filogenética previa del género, con base en morfología, es recuperado como monofilético y es designado aquí como Cetopsorhamdia sensu stricto. Una especie nueva de tributarios de piedemonte de los ríos Meta y Guaviare, cuenca del Orinoco en Colombia, fue recuperada como hermana de las especies cisandinas del concepto sensu stricto del género y es descrita aquí. La especie nueva se distingue de todos sus cóngeneres sensu stricto por una combinación única de caracteres: extremo distal de la aleta pectoral sobrepasando el origen de la aleta dorsal, base de la aleta adiposa corta (12,9–15,9% de LE); lóbulo inferior de la aleta caudal más largo que el lóbulo superior (31,7–37,3% de LE), cuerpo homogéneamente marrón y 35–36 vértebras totales. Un claro patrón vicariante cis y transandino es obtenido para las especies de Cetopsorhamdia sensu stricto, antecediendo el aislamiento de las actuales cuencas del Magdalena y del lago de Maracaibo, por la orogénesis de la cordillera oriental y de los Andes de Mérida, respectivamente.
CETOPSORHAMDIA is a genus of small-sized heptapterine catfishes (Bockmann and Reis, 2021), proposed by Eigenmann and Fisher in Eigenmann (1916) for his newly described C. nasus, whose description was based on a single specimen from the upper basin of the Magdalena River. Currently, the genus includes 13 valid species (Fricke et al., 2025), distributed in cis- and trans-Andean drainages of South America, from the Cauca River in Colombia to the Paraná River in Argentina (Bockmann and Reis, 2021). Cetopsorhamdia was recently diagnosed in a strictly phylogenetic context, being recognized from all other heptapterid genera by two exclusive synapomorphies (Bockmann and Reis, 2021). However, almost half of its species (i.e., C. filamentosa, C. molinae, C. orinoco, C. phantasia, and C. shermani), whose phylogenetic relationships lie outside of this definition of Cetopsorhamdia sensu stricto (Bockmann, 1998), are still classified within the genus, resulting in its non-monophyly (Bockmann, 1998; Faustino-Fuster et al., 2021).
Cetopsorhamdia is represented in Colombia by eight species, of which three species are distributed in the Magdalena River basin (i.e., C. boquillae, C. molinae, and C. nasus), two species in the Orinoco River basin (C. orinoco and C. shermani), two species in the Amazon River basin (C. hidalgoi and C. insidiosa), and C. picklei in the Lake Maracaibo basin (Ortega-Lara, 2012; DoNascimiento et al., 2023). Ortega-Lara (2012) redescribed C. nasus and proposed five putative synapomorphies, while questioning the monophyly of Cetopsorhamdia by keeping C. molinae and C. orinoco, but also by including C. boquillae.
As part of a systematic analysis of the species of Cetopsorhamdia from Colombia, we performed a multilocus phylogenetic analysis to evaluate their phylogenetic relationships and historical biogeography, and at the same time, we describe a new species from tributaries of the Meta and Guaviare Rivers, Orinoco basin, which fits the genus definition provided by Bockmann and Reis (2021).
MATERIALS AND METHODS
Morphological data.—
Terminology and procedures for measurements followed Lundberg and McDade (1986) and Bockmann (1994). Measurements were expressed as percent of standard length (SL), except for subunits of head, which are presented as percent of head length (HL). Meristics and relative position of fins follow Bockmann and Castro (2010). Number of specimens for each count is presented in parentheses, and an asterisk indicates values recorded for the holotype. Cleared and stained specimens (CS) were prepared according to Taylor and Van Dyke (1985). Osteological and cephalic laterosensory pore nomenclature follows Bockmann and Miquelarena (2008). Number of branchiostegal rays, gill rakers on first gill arch, vertebrae, ribs, and fin pterygiophores, as well as relative position of dorsal- and anal-fin insertion were determined only in CS specimens. Vertebral counts include five vertebrae associated with the Weberian complex, and the compound caudal centrum (PU1+U1) was counted as one element (Lundberg and Baskin, 1969).
As an exploratory method of the morphometric measurements, a principal component analysis (PCA) was carried out. Measurements were used as percent of SL to reduce size effect of the samples. These analyses were carried out with the statistical package Stats v. 4.2.2 of R studio software (R Core Team, 2024). The geographic distribution map was prepared using Quantum GIS v. 3.42.3 (QGIS Development Team, 2025). Institutional abbreviations follow Sabaj (2020, 2022). Nomenclature of genetic sequences of type material follows Chakrabarty et al. (2013).
Molecular data and phylogenetic analyses.—
We sequenced three mitochondrial (cytochrome c oxidase subunit I—cox1, cytochrome b—cytb, and NADH dehydrogenase subunit 2—nd2) and two nuclear genes (recombination activating gene subunit 2—rag2 and glycosyltransferase—glyt), using primers proposed by Ward et al. (2005): FishF1 5′–TCA ACC AAC CAC AAA GAC ATT GGC AC–3′, FishR1 5′–TAG ACT TCT GGG TGG CC AAA GAA TCA–3′; Palumbi (1996): GLUDG-L 5′–TGA CCT GAA RAA CCA YCG TTG–3′, CB3-H 5′–GGC AAA TAG GAA RTA TCA TTC–3′; Arroyave et al. (2013): nd2_f 5′–AGC TTT TGG GCC CAT ACC CCA–3′, nd2 _r 5′–AGG RAC TAG GAG ATT TTC ACT CCT GCT–3′; Li et al. (2007): F577 5′–ACA TGG TAC CAG TAT GGC TTT GT–3′, R1464 5′–GTA AGG CAT ATA SGT GTT CTC TCC–3′; and Hardman (2004): MHF1 5′–TGy TAT CTC CCA CCT CTG CGy TAC C–3′, MHR1 5′–TCA TCC TCC TCA TCk TCC TCw TTG TA–3′, respectively. DNA was extracted from muscle and pelvic fin tissues preserved in 96% ethanol, whose vouchers are deposited in the collection of ichthyology of the Universidad de Antioquia—CIUA, freshwater fish collection of the Instituto de Investigation de Recursos Biológicos Alexander von Humboldt—IAvH-P, Museo Javeriano de Historia Natural “Lorenzo Uribe S. J.”—MPUJ, and ichthyological collection of Museo de Historia Natural Unillanos—MHNU-I. Reaction volumes, denaturation, nesting, and extension cycles follow Cortés-Hernández et al. (2023) for the cox1 gene. PCR for cytb, nd2, and rag2 were carried out in a final volume of 25 µl, containing 2.0 μl (nd2 and rag2) or 2.5 μl of MgCl2 (cytb), 1.0 μl of Taq buffer, 0.2 μl of deoxynucleotide triphosphate, 0.2 μl of each primer, 0.04 μl of Taq polymerase, 2 μl of DNA, and 19.06 or 19.56 μl of distilled water. For these three genes, the initial denaturation step was conducted at 95°C for 5 min, followed by 35 cycles of denaturation at 94°C (nd2 and cytb) or 95°C (rag2) for 35 s, annealing at 48°C (cytb), 53°C (rag2), or 57.5°C (nd2) for 60 s, extension at 72°C for 60 s, and final extension at 72°C for 7 min. Final PCR products were run on a 1% agarose gel. Purification and sequencing steps of both forward and reverse strands were done at the Korean company Macrogen (https://www.macrogen.com). Consensus sequences were automatically assembled from forward and reverse sequences using Geneious Prime v. 2024.0.4 (https://www.geneious.com).
We assembled two sets of data, the first represented by 273 multilocus sequences (cox1: 51, cytb: 57, glyt: 52, nd2: 60, and rag2: 53) representing 62 terminals from Faustino-Fuster et al. (2021); 12 cox1 sequences of Cetopsorhamdia iheringi from de Carvalho et al. (2011), Pereira et al. (2013), Frantine-Silva et al. (2015), and Bagley et al. (2019); 30 newly acquired sequences (cox1: 16 sequences, CIUA539-20, CIUA540-20, CIUA689-20, CIUA693-20, CIUA1479-24, FBCH055-21, FBCH084-21, FBCH108-21, FBCH174-21, HEPTA002-24, PQ233717, PV738586–PV738590; cytb: 3 sequences, PV773935–PV773937; nd2: 6 sequences, PV773938–PV773943; and rag2: 5 sequences, PV773944–PV773948) from six species of Cetopsorhamdia (Supplemental Table A; see Data Accessibility); and 27 multilocus sequences (cox1: 9; cytb: 8; and rag2: 10) for ten species selected as outgroups (Batrochoglanis raninus, Brachyplatystoma vaillantii, Leiarius marmoratus, Lophiosilurus alexandri, L. apurensis, Pimelodus ornatus, Pseudopimelodus bufonius, P. mangurus, Sorubim maniradii, and Steindachneridion scriptum), based on Sullivan et al. (2013).
The second matrix was represented by a subsample, including 28 cox1 sequences of species of Cetopsorhamdia sensu stricto, which was used to calculate the best nucleotide substitution model and genetic distances in MEGA v. 11.0 (Tamura et al., 2021). The substitution saturation index was estimated in DAMBE v. 7.3.32 (Xia, 2018). RAxML plugin v.8 (Stamatakis, 2014) was used to construct a gene tree, using maximum likelihood (ML) inference with the model GTR GAMMA, with 1,000 bootstrap replicates to calculate node support. These analyses were carried out in Geneious Prime v. 2024.0.4.
Time calibration.—
A time-calibrated analysis was performed using Beast2 software (Bouckaert et al., 2019) to estimate the timing of diversification of Cetopsorhamdia sensu stricto in northern South America. For the construction of the dating tree, we used a concatenated matrix containing three mitochondrial markers (cox1, cytb, nd2) and two nuclear markers (glyt, rag2). The nucleotide evolutionary model used to estimate the tree was the GTR+I+F+G4 model, as estimated by PartitionFinder2 (Lanfear et al., 2017). A strict molecular clock, constant population size, and the Yule process were specified as priors for the analysis. Fossil calibration points were incorporated to estimate the time to the most recent common ancestor (MRCA) and the corresponding 95% highest posterior density (HPD) credibility intervals for major lineages. The first calibration point used the oldest pimelodid fossil from the Paleogene of South America (Gayet and Otero, 1999), with a mean age of 35 million years (SD = 0.07), applied to the ancestral node of the Pimelodidae, applying a lognormal prior distribution. The second calibration, based on an undetermined fossil of Cephalosilurus or Pseudopimelodus from the middle Miocene of South America (mean = 13.7 mya, SD = 0.09), was applied to the ancestral node encompassing members of the Pseudopimelodidae (Lundberg, 1998; Lundberg et al., 2010). The analysis was run for 100 million generations, sampling every 1,000 generations and discarding the first 10% as burn-in. Convergence was assessed using Tracer v. 1.7.1 (Rambaut et al., 2018). The summarized tree, including mean node ages and HPD intervals, was generated with TreeAnnotator v. 2.7.7 (Bouckaert et al., 2019), and the final tree was visualized using FigTree v. 1.4.4 (Rambaut, 2018).
Cetopsorhamdia ramirezi Cortés-Hernández and DoNascimiento, new species
urn:lsid:zoobank.org:act:70CCF133-AE70-4D42-8FDA-1DEA4D549606
Figures 1, 2A; Table 1
Cetopsorhamdia picklei.—Maldonado-Ocampo et al. (2008): 201 (checklist, Colombia). Urbano-Bonilla et al. (2014): 67 (taxonomic list, Cusiana River). Urbano-Bonilla et al. (2018): 79 as Cetopsorhamdia aff. picklei (identification key, Cusiana River).
Holotype.—
IAvH-P 9723, 87.9 mm SL, Colombia, Casanare, Aguazul, río Únete at the iron bridge on the Sogamoso road, 05°12′36″N, 72°36′16″W, 390 m asl, J. D. Bogotá-Gregory, 10 February 2007.
Paratypes.—
All from Colombia. Arauca, Tame: IAvH-P 18147, 1, 72.4 mm SL, 1 CS, 55.7 mm SL, río Cravo Norte, 06°30′09.3″N, 71°45′55.1″W, 369 m asl, J. Zamudio, 31 March 2015; MPUJ 8067, 4, 66.7–75.5 mm SL, río Culebrero, 06°30′24.8″N, 71°45′59.7″W, 346 m asl, J. Zamudio, 31 March 2015; MPUJ 9407, 1, 37.5 mm SL, río Cravo Norte, 06°30′12.4″N, 71°43′11″W, 292 m asl, J. Zamudio, 1 September 2015. Boyacá, San Luis de Gaceno: IAvH-P 27500, 1, 73.2 mm SL, genseq-2 cox1, río Chuy, tributary of río Upia, 04°50′52.8″N, 73°04′01.2″W, 354 m asl, J. Andrade, A. Urbano-Bonilla, Y. López-Pinto, H. Agudelo-Zamora, 28 November 2020. Casanare, Aguazul: IAvH-P 7969, 1, 56.5 mm SL, río Charte, 05°15′28″N, 72°29′15″W, 294 m asl, J. D. Bogotá-Gregory, 11 April 2006; IAvH-P 9721, 1, 56.5 mm SL, río Charte, 05°15′28″N, 72°29′15″W, 294 m asl, J. D. Bogotá-Gregory, 11 February 2007. Tauramena: MPUJ 6848, 1, 57.1 mm SL, río Chitamena, 04°55′35.1″N, 72°40′16.5″W, 253 m asl, S. Prada-Pedreros, 15 December 2012; MPUJ 7736, 1, 65.5 mm SL, río Chitamena, 04°55′39.6″N, 72°40′30.9″W, 259 m asl, S. Prada-Pedreros, 3 April 2013; MPUJ 7763, 1, 67.4 mm SL, río Caja, 05°00′49.3″N, 72°41′33.8″W, 314 m asl, S. Prada-Pedreros, 31 March 2013; MPUJ 17093, 1, 57.5 mm SL, genseq-2 cox1 and rag2, río Caja, 05°02′40.6″N, 72°45′12.8″W, 369 m asl, T. P. Carvalho, 24 February 2022. Yopal: IAvH-P 9246, 1, 88.6 mm SL, quebrada La Tablona, tributary of río Cravo Sur, 05°26′25.5″N, 72°27′37.7″W, 515 m asl, J. A. Maldonado-Ocampo, 2 October 2006. Meta, El Castillo: IAvH-P 23668, 1, 45.8 mm SL, río Orotoy, 03°52′1.78″N, 73°37′40.25″W, H. Ramírez-Gil, R. E. Ajiaco-Martínez, A. Ortega-Lara, 7 February 2010; MHNU-I 3205, 1, 59.3 mm SL, genseq-2 cox1, río Guape, tributary of río Ariari, 03°33′58.8″N, 73°48′52.5″W, E. Aya-Baquero, 13 February 2020. San Juan de Arama: MPUJ 10772, 1, 59.4 mm SL, río Güejar, 03°20′30.3″N, 73°56′28.3″W, 416 m asl, J. Zamudio, 27 January 2014. Villavicencio: MPUJ 10769, 1, 53.8 mm SL, río Guayuriba, 04°01′17″N, 73°38′37.7″W, A. Urbano-Bonilla, 8 May 2011. Vista Hermosa: MPUJ 10770, 1, 43.5 mm SL, caño Blanco, 03°04′54.4″N, 73°47′36.3″W, 260 m asl, J. Zamudio, 26 January 2014; MPUJ 10771, 8, 40.0–51.5 mm SL, río Güejar, 03°06′10.6″N, 73°46′21.8″W, 245 m asl, J. Zamudio, 26 January 2014.
Diagnosis.—
Cetopsorhamdia ramirezi differs from all sensu stricto congeners (except C. nasus) as defined by Bockmann and Reis (2021), by having a longer lower caudal-fin lobe (31.7–37.3% of SL vs. 24.2–29.0% in C. boquillae; 20.3–29.6% in C. clathrata; 27.6–30.7 in C. iheringi; 30.4–30.9% in C. insidiosa; 20.2–22.7% in C. spilopleura). Additionally, the new species can be distinguished from C. boquillae, C. clathrata, C. insidiosa, and C. spilopleura by having the distal end of the pectoral fin surpassing the origin of the dorsal fin, when adpressed to the body, reaching to the vertical through the first or second branched dorsal-fin ray (vs. reaching or almost reaching to dorsal-fin origin). Cetopsorhamdia ramirezi can be distinguished from C. nasus by lacking small papillae on skin of head and predorsal region (vs. densely covered with small papillae) and anterodorsal margin of premaxilla jagged, with six regular dentations (Fig. 3A; vs. with five irregular tooth-like projections; Ortega-Lara, 2012: fig. 8, p. 59). Cetopsorhamdia ramirezi is further distinguished from C. boquillae and C. insidiosa by maxillary-barbel length (24.1–29.3% of SL vs. 33.5–45.0% in C. boquillae; 21.4–23.0% in C. insidiosa). Cetopsorhamdia ramirezi is further distinguished from C. iheringi by having a lower number of vertebrae (35–36 vs. 37–38). Additionally, it can also be distinguished from C. boquillae by having a triangular adipose fin (vs. rounded), higher number of gill rakers on first gill arch (8–10 vs. 5–6), narrower head (48.2–67.7% of HL vs. 70.2–82.7%), longer interdorsal distance (18.0–24.6% of SL vs. 10.0–14.1%), longer preadipose length (68.6–73.5% of SL vs. 59.0–62.9%), shorter adipose-fin base (12.9–16.2% of SL vs. 29.1–34.0%), and shorter outer mental-barbel length (10.6–13.9% of SL vs. 14.5–19.1%). Cetopsorhamdia ramirezi is also distinguished from C. picklei by having a narrower head (48.2–67.7% of HL vs. 70.3–80.9%), and shorter adipose-fin base (12.9–15.9% of SL vs. 16.3–18.2%). Cetopsorhamdia ramirezi further differs from C. clathrata and C. spilopleura by having a shorter predorsal length (36.3–42.0% of SL vs. 43.8–48.5% in C. clathrata; 42.7–46.3% in C. spilopleura), longer upper caudal-fin lobe (36.3–42.0% of SL vs. 19.8–26.2% in C. clathrata; 20.0–21.4% in C. spilopleura), body homogeneously brown (vs. with two longitudinal rows of 10–12 quadrangular marks in C. clathrata; with 18–22 irregular, vertical brown bars, sometimes resembling inverted “v,” “y,” or “x” in C. spilopleura).
Citation: Ichthyology & Herpetology 113, 4; 10.1643/i2025051
Description.—
Morphometric data in Table 1. Body elongated, semi-cylindrical, becoming compressed toward caudal fin. Body deeper at dorsal-fin origin. Dorsal profile of body rises in convex curve from snout to dorsal-fin origin, slightly concave along dorsal-fin base, straight from posterior end of dorsal-fin base to adipose-fin origin, adipose-fin base straight and descending, and concave along caudal peduncle. Ventral profile of body convex from jaw tip to pelvic-fin origin, almost straight between pelvic-fin insertion and anal-fin origin, anal-fin base upwardly straight, and concave along caudal peduncle (Figs. 1, 2A). Anus and urogenital papilla adjacent. Anus approximately at vertical through middle of pelvic-fin length.
Citation: Ichthyology & Herpetology 113, 4; 10.1643/i2025051
Citation: Ichthyology & Herpetology 113, 4; 10.1643/i2025051
Head conical, subtriangular to parabolic in dorsal view. Anterior and posterior cranial fontanels short, separated by extensive medial ossification of frontal. Eye small, placed dorsolaterally, slightly elliptical, its greatest diameter along horizontal axis, and without free orbital rim. Snout short and broadly rounded. Nares disposed in rectangular arrangement; anterior naris tubular and close to upper lip; posterior naris anteromedially bordered by fleshy margin, slightly closer to anterior eye margin than to anterior naris. Mouth ventral, with snout projecting well beyond upper jaw. Premaxilla with 5–6 and dentary with 4–5 irregular rows of small villiform teeth. Barbels short, thin and slightly depressed; elliptical in cross section. Maxillary barbel longest, inserted dorsal to upper lip and lateral to anterior nostrils; tip of maxillary barbel reaching to origin or posterior end of pectoral-fin base. Outer mental barbel longer than inner mental barbel; tip almost reaching to base of pectoral fin. Inner mental barbel surpassing isthmus but not branchiostegal membrane. Branchiostegal membranes free, united to isthmus only at medial apex, and not connected to each other anteriorly. Branchiostegal rays 7 (3). First gill arch with 8 (1), 9* (15), or 10 (2) gill rakers, ceratobranchial with 7* (15) or 8 (3) gill rakers, and epibranchial with 1 (1) or 2* (17) gill rakers.
Dorsal fin with distal margin concave in lateral profile. Dorsal-fin rays i,6* (19), first ray unbranched, with basal third stiffened and unsegmented, and distal two-thirds flexible and segmented; second ray (first branched ray) equal in length to first unbranched ray (17.4–21.6% of SL); dorsal-fin origin anterior to pelvic-fin origin. Anteriormost pterygiophore inserted posterior to neural spine of vertebra 8 (3). Posteriormost pterygiophore located posterior to neural spine of vertebra 13 (1) or 14 (1).
Pectoral fin with i,8* (5) or i,9 (14) rays, with distal margin straight or slightly concave; outer margin of first ray slightly convex, with proximal portion slightly ossified and distal portion soft and segmented. First ray slightly shorter than second (first branched) and third (second branched) rays. Tip of adpressed fin reaching vertical between second (first branched) and third (second branched) rays of dorsal fin.
Pelvic fin with i,5* (19) rays, distal margin trapezoidal; first unbranched ray completely flexible and slightly shorter than second and third rays (first and second branched rays, respectively). Pelvic-fin origin anterior to midlength of body (excluding caudal fin), at vertical between third and fifth branched rays of dorsal fin; posterior margin of pelvic fin surpassing anus and urogenital papilla. Insertion of first pelvic-fin ray at vertical through vertebral centrum 10 (1) or 11 (2).
Anal-fin margin straight or concave, with one or two procurrent rays, embedded in thick skin fold, and iii* plus 7 (8), 8* (9), or 9 (2) branched rays, for a total of 11–14 rays. Anal-fin origin at vertical through adipose-fin origin; anal-fin terminus posterior to second third of adipose-fin base. Anteriormost pterygiophore inserted posterior to hemal spine of vertebrae 20 (1) or 21 (2). Posteriormost pterygiophore inserted posterior to hemal spine of vertebrae 25 (2) or 26 (1).
Adipose-fin base short (12.9–16.2% of SL), triangular in lateral profile, highest approximately at midpoint. Interdorsal distance longer than dorsal-fin base; adipose-fin origin posterior to midlength of body (excluding caudal fin), at vertical through origin of anal fin and at same level of vertebral centrum 24 (1) or 25 (1). Posterior limit of adipose fin well defined, with distinct free, rounded lobe. Adipose-fin terminus at vertical through vertebral centrum 29 (1) or 30 (2).
Caudal fin deeply forked; with lower lobe longer or slightly longer than upper lobe; both lobes with pointed tips. Total rays of upper lobe 22 (1), with i,7* (3) rays and 13 (1) procurrent rays. Total rays of lower lobe 21, with i,8* (3) rays and 14 (1) procurrent rays. Parhypural not fused to hypurals 1 and 2 (3). Hypurals 3, 4, and 5 completely fused. Rays articulating with upper caudal-fin plate 9 (2). Rays articulating with lower caudal-fin plate 7 (2). Total vertebrae 35 (1) or 36 (2). Precaudal vertebrae 11 (3), with last five vertebrae having closed haemal arch. Caudal vertebrae 19 (1) or 20 (2). Ribs 9 (3).
Head laterosensory canals with simple (non-dendritic) tubes ending in single pores. Supraorbital laterosensory canal continuous and connected to otic and infraorbital laterosensory canals posteriorly. Supraorbital laterosensory canal with five branches: s1, s2, s3, s6 (epiphyseal branch), and s8 (parietal branch). Contralateral epiphyseal branches (s6) fused to each other, bearing single medial pore (s6+s6). Supraorbital and infraorbital laterosensory canals anteriorly connected to each other through s2 and i2 branches (forming complex s2+i2 pore). Otic laterosensory canal short, without pores, and continuous with posterior limits of supra- and infraorbital laterosensory canals anteriorly, and with anterior limit of postotic laterosensory canal posteriorly. Postotic (or temporal) laterosensory canal extending from posterior limit of otic laterosensory canal to anterior limit of lateral line, with three branches and pores (po1, po2, and po3). Infraorbital laterosensory canal with six branches and pores, with s2 fused to i2 (see above). Preoperculomandibular laterosensory canal with 11 branches and pores; anteriormost preoperculomandibular laterosensory branch (pm1) not fused to its counterpart; posteriormost preoperculomandibular laterosensory branch (pm11) fused to po1 branch, forming complex po1+pm11 branch and pore. Lateral line canal long, with posterior limit extending to vertical through anterior half of caudal plate. Dorsal and lateral surface of snout above groove accommodating maxillary barbel, with dense field of neuromasts. Line of superficial neuromasts, variably branched, behind parietal sensory pore s8.
Color in alcohol.—
Body uniformly brown on dorsal and lateral surfaces, ventral region pale. Dorsal, lateral, and ventral surface of head yellow. Internarial region brown. Fleshy portion of anterior and posterior nares hyaline. Conspicuous light band across posterior margin of occipital region. Wide horizontal pale stripe, ventrally adjacent to lateral line, extending from upper margin of gill membrane to level of origin or posterior end of dorsal fin. Thin and conspicuous pale lateral stripe, extending from end of dorsal fin to origin of caudal fin, connecting with most anterior pale stripe. Pale spot at origin of dorsal fin. Dorsal surface of caudal peduncle with a pale oblong spot, just in front of base of upper rays of caudal fin. Rays of pectoral, dorsal, pelvic, anal, and caudal fins brown. Interradial membranes hyaline. Base of adipose fin brown, distal margin hyaline (Figs. 1, 2A).
Geographic distribution.—
Cetopsorhamdia ramirezi is known from piedmont tributaries of the Meta (Cravo Norte, Cravo Sur, Charte, Únete, Caja, Cusiana, Chitamena, and Chuy Rivers) and Guaviare (Ariari and Güejar Rivers) River drainages of the Orinoco basin (Fig. 4).
Citation: Ichthyology & Herpetology 113, 4; 10.1643/i2025051
Etymology.—
The specific epithet honors Hernando Ramírez Gil, for his invaluable contribution to the knowledge and conservation of ornamental and food fishes in Colombia.
Phylogenetic analyses.—
The concatenated matrix resulted in 100 sequences of 3,894 bp and 1,624 variable sites (41.1%). The second matrix (cox1) included 26 sequences of 635 bp and 132 variable sites (20.8%), with no saturation signal. The best substitution model found for this partition was K2+G, with a BIC value = 4310.9. The overall mean genetic distance of the cox1 matrix was 0.10 ± 0.01, with the lowest distance recorded between C. picklei and C. nasus (0.0423 ± 0.0105), while the highest distance was found between C. picklei and C. iheringi (0.1650 ± 0.0231; Table 2). Genetic divergence between C. ramirezi and the other species of Cetopsorhamdia sensu stricto ranged from 13.1% to 15.4% (Table 2).
The ML tree reveals two well-supported clades (100% bootstrap), with a clear geographic dichotomy of the Cetopsorhamdia sensu stricto clade, where one group is represented by the trans-Andean species (C. boquillae, C. nasus, and C. picklei) and the second group is cis-Andean (C. iheringi, C. insidiosa, and C. ramirezi). Cetopsorhamdia ramirezi was recovered as sister to the remaining cis-Andean species (Supplemental Figure A; see Data Accessibility). “Cetopsorhamdia” orinoco, “C.” molinae, “C.” shermani, and Phenacorhamdia boliviana constitute the sister clade to Cetopsorhamdia sensu stricto.
The time-calibrated phylogeny (Fig. 5) indicates that diversification of the Cetopsorhamdia sensu stricto clade initiated around 14.8 Ma (12.4–17.4 Ma; 100% HPD), in the middle Miocene, with the early divergence of the cis- (C. iheringi, C. insidiosa, and C. ramirezi) and trans-Andean (C. boquillae, C. nasus, and C. picklei) clades. Cetopsorhamdia ramirezi diverged from remaining species of the cis-Andean clade around 13.0 Ma (10.4–15.9 Ma; 76% HPD), also during the middle Miocene, while for the trans-Andean clade, C. boquillae is the first lineage to diverge, around 10.9 Ma (9.0–13.1 Ma; 100% HPD), during the late Miocene, and C. picklei and C. nasus diverged approximately 4.8 Ma (3.6–6.1 Ma; 100% HPD), during the early Pliocene.
Citation: Ichthyology & Herpetology 113, 4; 10.1643/i2025051
Morphometric analyses.—
Morphometric variation examined in the PCA shows four distinct groups, the first group consisting of Cetopsorhamdia boquillae, “Cetopsorhamdia” orinoco, and “Chasmocranus” rosae; the second and third groups are represented by “C.” molinae and “C.” shermani, respectively; and the fourth group comprises five species of Cetopsorhamdia sensu stricto (C. iheringi, C. insidiosa, C. nasus, C. picklei, and C. ramirezi), showing total overlap among all species (Fig. 6A). The first component explained 38.6% of the variance, while the second component explained 33.5%, with a cumulative variance of 72.1%. The variables of the rotation matrix with the highest values on the negative axis of PCA 1 were adipose-fin length, maxillary-barbel length, outer and inner mental-barbel lengths, and for the positive axis, lower caudal-fin lobe length, upper caudal-fin lobe length, interdorsal length, and preadipose length. The highest value on the negative axis of PCA 2 was accounted by caudal-peduncle length, and for the positive axis, maxillary-barbel length, first dorsal-fin (unbranched) ray length, first pectoral-fin (unbranched) ray length, lower caudal-fin lobe length, and upper caudal-fin lobe length were the variables with highest scores (Supplemental Table B; see Data Accessibility). When we isolate the morphometric variation of the species of Cetopsorhamdia sensu stricto, C. ramirezi is partially overlapping with C. nasus, while it is segregated from C. iheringi, C. insidiosa, and C. picklei (Fig. 6B). The first component explained 24.3% of the variance, while the second explained 19.6%, with a cumulative variance of 43.9%. The variables with the highest values on the negative axis of PCA 1 were preanal length, lower caudal-fin lobe length, upper caudal-fin lobe length, and maxillary-barbel length, and for the positive 1 axis were interdorsal length and caudal-peduncle length. The highest values on the negative axis of PCA 2 were lower caudal-fin lobe length, body depth, prepelvic length, and adipose-fin length, and for the positive axis were lower caudal-fin lobe length, head length, predorsal length, and maxillary-barbel length (Supplemental Table C; see Data Accessibility).
Citation: Ichthyology & Herpetology 113, 4; 10.1643/i2025051
DISCUSSION
The first explicit phylogenetic diagnosis of Cetopsorhamdia was proposed by Bockmann (1998) and was partially revalidated by Bockmann and Reis (2021), consisting of two exclusive synapomorphies: (1) presence of a medial ossification over the median portion of the skull, covering the epiphyseal bar and leaving reduced anterior and posterior fontanels; and (2) small optic foramen. These authors restricted two other synapomorphies that were originally advanced as synapomorphies for the whole genus (Bockmann, 1998), as synapomorphies of a less inclusive group within Cetopsorhamdia (C. iheringi, C. picklei, and three undescribed species), ventral mouth and conical snout, and proposed two additional putative synapomorphies for the genus: (1) presence of a vertical, dark band at caudal-fin ray insertion; and (2) oblique keel on lateral surface of posterodorsal region of hyomandibula (Fig. 3B). Cetopsorhamdia ramirezi shows all these six synapomorphies, supporting its unambiguous assignment to Cetopsorhamdia sensu stricto. Additionally, we corroborated that the caudal mark in Cetopsorhamdia ramirezi is further developed as a large W-shaped band, distinctly extending toward the dorsal and ventral borders of the caudal peduncle (Fig. 2), a condition shared with C. boquillae, C. iheringi, C. insidiosa, C. nasus, and C. picklei. Ortega-Lara (2012), in his redescription of C. nasus, proposed five potential diagnostic characters for Cetopsorhamdia: (1) presence of a slightly prominent bony process on the anterodorsal margin of the inner end of the premaxilla, projecting vertically with an irregular upper edge, giving the appearance of a saw; (2) anteroventral region of the complex vertebra with a thin and pointed process directed laterally; (3) conical snout; (4) ventral mouth; and (5) dorsal and ventral margins of the caudal peduncle with a pale symmetrical area, anterior to the origin of the procurrent rays of the caudal fin. However, these conditions were only recorded in C. insidiosa, C. nasus, and C. picklei, and then are not valid synapomorphies for the entire genus. It is relevant to indicate that the first character is here recorded for C. ramirezi, which can be better described as having a completely jagged anterodorsal margin of the premaxilla (Fig. 3A), but this condition was not observed in C. clathrata and C. spilopleura, according to the CT reconstructions provided in their taxonomic descriptions (Bockmann and Reis, 2021: fig. 6, p. 11 and fig. 16, p. 20, respectively). All remaining character states were also confirmed in C. ramirezi, remarking that only the second character seems to be exclusive to Cetopsorhamdia.
Published molecular phylogenies focused on heptapterids (Faustino-Fuster et al., 2021; Silva et al., 2021) have not fully addressed the intrarelationships of Cetopsorhamdia, so the results obtained here constitute the first attempt to clarify the phylogenetic relationships of the genus, confirming the monophyly of Cetopsorhamdia sensu stricto as defined by Bockmann and Reis (2021), which includes C. boquillae, C. iheringi, C. insidiosa, C. nasus, C. picklei, and C. ramirezi. Although DNA sequences of C. clathrata and C. spilopleura are not available, their morphology unequivocally supports their inclusion in Cetopsorhamdia (Bockmann and Reis, 2021). On the other hand, “C.” molinae, “C.” orinoco, and “C.” shermani are not recovered as members of this Cetopsorhamdia sensu stricto clade, but they are grouped in a clade sister to Phenacorhamdia, together forming a clade sister to Cetopsorhamdia sensu stricto, a topology equivalent to that found by Faustino-Fuster et al. (2021). The recently described “C.” hidalgoi also lacks all the synapomorphies proposed for Cetopsorhamdia, and it is likely a member of this last clade, based on its similar general appearance to “C.” molinae and “C.” shermani. Bockmann and Slobodian (2018) informally proposed the reallocation of “C.” molinae and “C.” shermani in the “Heptapteridae genus F,” and of “C.” orinoco in the “Heptapteridae genus G”; however, based on our results, “C.” shermani is more closely related to “C.” orinoco than to “C.” molinae, although this placement received low support (bootstrap = 68). A more encompassing analysis, including at least “C.” hidalgoi, “Chasmocranus” quadrizonatus, and “Ch.” rosae is necessary to confirm the reciprocal monophyly of these two putative undescribed genera.
Our time-calibrated phylogeny indicates that the earliest diversification of the Heptapteridae occurred in the Paleogene, 36.7 Ma (late Eocene), in agreement with estimates obtained by Sullivan et al. (2013; Multidivtime: ∼40 Ma) and Tagliacollo et al. (2024; 38.4 Ma). Furthermore, most of the generic-level diversification within the Heptapteridae appears to have begun during the Neogene (early Miocene), also consistent with the findings of Sullivan et al. (2013).
A clear vicariant event is observed between cis- and trans-Andean species of the Cetopsorhamdia sensu stricto clade, with their divergence placed in the middle Miocene (14.8 Ma). Another vicariance instance of the trans- and cis-Andean ichthyofaunas is represented in our phylogenetic results by Imparfinis, with a similar age estimate of divergence (16.8 Ma). These estimates would indicate that the divergence of the cis- and trans-Andean lineages of Cetopsorhamdia and Imparfinis occurred long before the uplift and exhumation phases of the Colombia, Quetame, and Garzón Massifs (9–3 Ma), which led to the isolation of the Magdalena basin from the proto-Orinoco (Hoorn et al., 2017; Saeid et al., 2017; Pérez-Consuegra et al., 2021; Zapata et al., 2023). These tectonic evolution events are consistent with age estimates obtained for other groups of Neotropical fishes with amphi-Andean distribution, such as Brycon (Abe et al., 2014), Potamotrygon (Fontenelle et al., 2021), Prochilodus (Frable et al., 2022), and Pseudopimelodus (Rangel-Medrano et al., 2020). However, both Cetopsorhamdia and Imparfinis are restricted to foothill environments, in ravines and low-order rivers, in contrast to the above referred examples, which are typical inhabitants of lowland aquatic ecosystems, especially Brycon and Prochilodus, which have migratory behavior that facilitates their dispersal through the main channel of large rivers.
Among trans-Andean species, Cetopsorhamdia boquillae is the earliest lineage to diverge, in the late Miocene (10.9 Ma; 9.0–13.1 Ma; 100% HPD). This species is restricted to the upper Cauca River, but it is found in sympatry with C. nasus, which has a broader distribution in the Magdalena basin. It is estimated that ∼16 Ma, the upper and middle basins of the Magdalena River were connected to the Cauca River, where the uplift and exhumation of the Southern Central Mountain Range, the displacement of the Antioquia Altiplano Province, accompanied by the migration of the volcanic arc of the Western mountain range, modulated the formation (∼10–3 Ma) of the upper and middle basin of the Cauca River (Zapata et al., 2023), forming physiographic barriers (rapids of the Momposina depression and La Virginia) that prevent downriver dispersal of endemics species from the upper Cauca River (e.g., Ancistrus vericaucanus, Carlastyanax aurocaudatus, Cetopsorhamdia boquillae, Genycharax tarpon, Gephyrocharax caucanus, Hemibrycon boquiae, Lebiasina ortegai, Parodon caliensis, Pimelodella macrocephala, Pimelodus crypticus, Priapichthys caliensis; Ortega-Lara et al., 2022). Cetopsorhamdia nasus diverged from C. picklei ∼4.8 Ma, during the early Pliocene, suggesting that these lineages diverged after the isolation of the Magdalena and Maracaibo basins, ∼8 Ma (Albert et al., 2006; Boschman, 2021). Presence of C. picklei in the Ranchería River supports the dispersal hypothesis of Pérez and Taphorn (1993), where the floodplain of the northeastern sections of the Magdalena basin and the Gulf of Venezuela enabled dispersal through coastal rivers in the La Guajira Peninsula. Most of the remaining cladogenetic events within the cis-Andean clade occurred during the middle and late Miocene (13.0–8.5 Ma), leading first to divergence of C. ramirezi from the clade of C. insidiosa plus C. iheringi, suggesting that this event occurred during the separation of the paleo-Amazon-Orinoco system (12–10 Ma), product of the uplift of the Macarena Massif and the Vaupés Arc (10–8 Ma; Lundberg et al., 1998; Vale et al., 2025). Finally, C. insidiosa diverged from C. iheringi in the late Miocene (8.5 Ma), consistent with a faunal separation event between the Guiana and Brazilian shields, driven by the formation of the transcontinental Amazon River system through the transgression of the Purus Arch, which occurred approximately 10.1 to 4.9 million years ago (Lundberg et al., 1998; Dagosta and de Pinna, 2019). There are several events that have allowed almost unidirectional dispersal of the Amazonian biota to the Paraná-Paraguay basin, such as dispersal from headwater catches of the Mamoré-Guaporé, Tapajós, Xingu, and Tocantins Rivers toward the Paraguay basin (Lundberg et al., 1998; Wilkinson et al., 2006, 2010; Albert and Carvalho, 2011; Dagosta and de Pinna, 2019).
Cetopsorhamdia ramirezi had been formerly listed as C. picklei for the Orinoco basin in Colombia (Maldonado-Ocampo et al., 2008), and this taxonomic treatment was successively used in checklists of the Orinoco basin (Urbano-Bonilla et al., 2014, 2018). Specimens of C. ramirezi are frequently found mixed with Imparfinis robustus in ichthyological collections in Colombia, given their similar general appearance and their syntopic occurrence (Cortés-Hernández et al., 2023).
MATERIAL EXAMINED
Cetopsorhamdia boquillae
Colombia. Quindío: CIUA 5433, 4, 54.5–66.9 mm SL, CIUA539-20, 54.5 mm SL, CIUA540-20, 62.5 mm SL, vouchers of DNA sequences, Montenegro, quebrada Hojas Anchas, 04°32′20.0″N, 75°44′28.3″W, 1240 m asl, M. Hamp, 18 June 2019; IAvH-P 3702, 2, 49.1–62.4 mm SL, río Barragán; IAvH-P 13541, 1, 62.8 mm SL, Salento, quebrada Boquía, 04°38′55.67″N, 75°35′13.9″W; IAvH-P 13544, 11, 30.4–55.2 mm SL, Salento, río Quindío, 04°38′26″N, 75°35′16.3″W; IAvH-P 13547, 2, 38.4–70.9 mm SL, Finlandia, Portachuelo, 04°39′09.72″N, 75°39′27.9″W; IAvH-P 13552, 2, 62.0–76.8 mm SL, Finlandia, El Membrillal, 04°39′09.4″N, 75°39′28″W. Risaralda: IAvH-P 7801, 3, 71.6–80.4 mm SL, La Virginia, río Mapa, 05°02′37″N, 75°55′43″W.
Cetopsorhamdia iheringi
Argentina. Misiones: CI-FML 7242, 4, 44.1–68.6 mm SL, arroyo Tabay, 26°59′56.2″S, 55°10′43″W; CI-FML 7243, 4, 43.8–110.3 mm SL, arroyo 3 de Mayo, 26°45′28.8″S, 54°55′26,4″W; CI-FML 7244, 1, 60.3 mm SL, arroyo Isla, 26°36′22.7″S, 54°27′18.4″W; CI-FML 7281, 4, 34.0–67.1 mm SL, Jardín America, arroyo Tabay, 26°59′58.8″S, 55°10′40.3″W.
Cetopsorhamdia insidiosa
Colombia. Caquetá: CIACOL 3618, 1, 41.9 mm SL, Florencia, quebrada La Yuca, road Florencia–Belén de los Andaquíes, 01°36′27.9″N, 75°38′16.1″W. Meta: MHNU-I 1159, 1, 59.5 mm SL, río Ocoa, 04°06′58.1″N, 73°21′36.6″W, 273 m asl.
“Cetopsorhamdia” molinae
Colombia. Cauca: IAvH-P 3700, 3, 28.6–36.2 mm SL, Suárez, río Ovejas, 03°10′N, 76°35′W. Santander: IAvH-P 20972, 7, 25.3–33.0 mm SL, Cimitarra, caño Pizarral, 06°07′10.34″N, 74°13′34.14″W, 197 m asl; IAvH-P 22885, 3, 26.0–27.4 mm SL, Betulia, quebrada La Putana, 07°06′36.5″N, 73°30′28.8″W, 134 m asl; IAvH-P 22905, 3, 27.7–30.5 mm SL, Sabana de Torres, río Sucio, 07°09′39.70″N, 73°32′19.85″W, 127 m asl; IAvH-P 23906, 25, 21.6–29.3 mm SL, Betulia, quebrada La Putana, 07°06′36.5″N, 73°30′28.8″W, 134 m asl; IAvH-P 23928, 11, 25.4–33.7 mm SL, Sabana de Torres, río Sucio, 07°09′39.70″N, 73°32′19.85″W, 127 m asl.
Cetopsorhamdia nasus
Colombia. Antioquia: CIUA 5616, 2, 40.5–54.4 mm SL, CIUA689-20, 40.5 mm SL, voucher of DNA sequences, Liborina, quebrada La Sucia, 06°43′06.4″N, 75°51′23.4″W; CIUA 5633, 2, 59.5–69.9 mm SL, CIUA693-20, 59.5 mm SL, voucher of DNA sequences, Sabanalarga, quebrada La Niquia, 06°52′02.1″N, 75°50′17.5″W. Caldas: ICN-MHN 10920, 12, 44.7–63.4 mm SL, La Dorada, río Guarinó, 05°19′07.3″N, 74°50′18.02″W; ICN-MHN 11468, 9, 47.1–66.9 mm SL, La Dorada, río Guarinó, 05°19′07.3″N, 74°50′18.02″W; ICN-MHN 13545, 4, 56.8–73.5 mm SL, La Dorada, río Guarinó, 05°19′07.3″N, 74°50′18.02″W. Huila: IAvH-P 16096, 1, 83.1 mm SL, Garzón, quebrada Río Loro, 02°18′58.52″N, 75°37′14.98″W; ICN-MHN 19107, 2, 65.5–69.5 mm SL, Garzón, río Suaza, 02°10.3′44″N, 75°40.2′22″W; ICN-MHN 20206, 1, 85.5 mm SL, Garzón, río Suaza, 02°10′21.4″N, 75°40′11.5″W. Tolima: CZUT-IC 638, 11, 64.2–79.0 mm SL, Coello, río Coello, 04°16′50.0″N, 75°01′51.0″W, 455 m asl.
“Cetopsorhamdia” orinoco
Colombia. Boyacá: IAvH-P 3648, 2, 63.0–70.0 mm SL, río Lengupá, tributary of río Upía, 05°03′N, 73°04′W. Casanare: IAvH-P 7961, 3, 24.5–33.0 mm SL, Aguazul, río Charte, 05°15′27.9″N, 72°29′14.9″W, 294 m asl; IAvH-P 7962, 6, 27.8–32.6 mm SL, Pore, río Pauto, 05°33′44″N, 72°08′44″W, 306 m asl; IAvH-P 7965, 5, 35.1–37.6 mm SL, Sabanalarga, río Upía, 04°49′09.4″N, 73°04′57.6″W, 327 m asl; IAvH-P 9713, 8, 31.7–43.4 mm SL, Sabanalarga, río Upía, 04°49′09.4″N, 73°04′57.6″W, 327 m asl; IAvH-P 25483, 8, 28.3–31.8 mm SL, Aguazul, río Únete, 05°12′27.6″N, 72°36′01.5″W, 288 m asl. Meta: IAvH-P 23630, 5, 35.6–43.1 mm SL, Castilla La Nueva, río Orotoy, 401 m asl; MHNU-I 1879, 4, 33.2–37.2 mm SL, Villavicencio, río Guayuriba, 04°10′46.1″N, 73°42′41.4″W, 633 m asl; MHNU-I 2188, 46, 24.9–58.7 mm SL, Villavicencio, río Guayuriba, 04°03′17.8″N, 73°45′58.2″W, 480 m asl; MHNU-I 2273, 55, 36.6–20.4 mm SL, Villavicencio, río Guayuriba, 04°00′12.2″N, 73°30′08″W, 350 m asl; MHNU-I 2294, 1, 34.4 mm SL, Puerto López, río Guayuriba, 03°55′10.8″N, 73°05′05.6″W, 197 m asl; MHNU-I 2424, 63, 21.8–33.4 mm SL, San Carlos de Guaroa, río Guayuriba, 03°57′27″N, 73°16′7.8″W, 237 m asl; MHNU-I 3204, 17, 22.3–42.0 mm SL, El Castillo, río Guape, tributary of río Ariari, 03°33′58.8″N, 73°48′52.5″W.
Cetopsorhamdia picklei
Colombia. La Guajira: IAvH-P 3, 3, 39.7–105.8 mm SL, río Ranchería, 11°33′N, 72°36′W. Norte de Santander: IAvH-P 9822, 2, 106.1–114.1 mm SL, genseq-2 cox1 and rag2, quebrada La Llana, tributary of río Catatumbo, 08°34′22″N, 73°14′49″W, 435 m asl; IAvH-P 9823, 2, 58.8–127.9 mm SL, caño seco, tributary of río Catatumbo, 08°34′31.8″N, 73°10′7.68″W, 355 m asl; IAvH-P 11805, 7, 102.8–125.5 mm SL, genseq-2 cox1, río Zulia, 07°56′53″N, 72°35′41″W.
“Cetopsorhamdia” shermani
Colombia. Boyacá: IAvH-P 22221, 1, 35.1 mm SL, San Luis de Gaceno, río Upía, 04°49′13.7″N, 73°04′27.9″W, 318 m asl. Casanare: IAvH-P 9710, 1, 28.2 mm SL, Aguazul, río Charte, 05°15′27.9″N, 72°29′14.9″W, 294 m asl; IAvH-P 9728, 1, 28.7 mm SL, Nunchía, río Tocaría, 05°30′47″N, 72°13′02″W, 310 m asl. Meta: MHNU-I 1231, 1, 32.4 mm SL, Villavicencio, río Guayuriba, 03°57′33.5″N, 73°16′20.0″W, 237 m asl; MHNU-I 1364, 2, 32.6–33.5 mm SL, Villavicencio, río Guayuriba, 04°00′12.2″N, 73°30′08″W, 350 m asl; MHNU-I 1701, 14, 30.2–33.2 mm SL, Villavicencio, río Guayuriba, 04°00′12.2″N, 73°30′08″W, 350 m asl.
“Chasmocranus” rosae
Colombia. Cundinamarca, Guayabetal: MHNU-I 1007, 2, 19.1–35.8 mm SL, río Negro, 04°12′37.1″N, 73°48′59.4″W, 881 m asl; MHNU-I 1721, 4, 26.6–71.2 mm SL, río Blanco, 04°12′38.0″N, 73°49′4.2″W, 865 m asl; MHNU-I 1728, 6, 31.1–51.8 mm SL, río Negro, 04°12′37.1″N, 73°48′59.4″W, 881 m asl; MHNU-I 2345, 2, 50.0–62.1 mm SL, río Negro, confluence on río Blanco, 04°12′25.4″N, 73°48′49.1″W, 930 m asl; MHNU-I 2831, 2, 43.0–60.5 mm SL, río Negro, confluence on río Blanco, 04°12′25.4″N, 73°48′49.1″W, 930 m asl; MHNU-I 2836, 2, 21.6–79.1 mm SL, río Blanco, 04°12′38.0″N, 73°49′04.2″W, 865 m asl. Medina: IAvH-P 18907, 35, 34.6–69.7 mm SL, río Gazamumo, río Humea drainage, 04°30′43.4″N, 73°25′12.8″W, 635 m asl; IAvH-P 18930, 2, 56.4–71.8 mm SL, caño Blanco, río Humea drainage, 04°30′23.6″N, 73°24′55.7″W, 635 m asl; IAvH-P 18948, 32, 48.9–51.2 mm SL, NN stream, río Humea drainage, 04°31′31.51″N, 73°25′51.42″W, 609 m asl; IAvH-P 18963, 10, 58.3 mm SL, río Gazamumo, río Humea drainage, 04°30′43.4″N, 73°25′12.8″W, 635 m asl. Meta, Villavicencio: MHNU-I 959, 13, 42.8–58,9 mm SL, caño Buque, tributary of río Ocoa, 04°08′16.93″N, 73°39′18.8″W, 563 m asl; MHNU-I 1015, 3, 17.0–33.0 mm SL, río Guayuriba, 04°10′46.1″N, 73°42′41.4″W, 633 m asl; MHNU-I 1325, 37, 42.8–58.93 mm SL, caño Maizaro, tributary of río Ocoa, 04°08′37.9″N, 73°39′6.3″W, 461 m asl; MHNU-I 1378, 14, 20.2–62.2 mm SL, caño Buque, tributary of río Ocoa, 04°08′05.6″N, 73°38′39.9″W, 521 m asl; MHNU-I 1477, 2, 28.0–43.6 mm SL, río Guayuriba, 04°03′17.8″N, 73°45′58.2″W, 480 m asl; MHNU-I 1506, 1, 37.9 mm SL, río Ocoa, 04°06′13.5″N, 73°37′58.9″W, 397 m asl; MHNU-I 1740, 1, 37.1 mm SL, río Negro, 04°12′10.4″N, 73°44′02.6″W, 733 m asl; MHNU-I 1880, 1, 35.5 mm SL, río Guayuriba, 04°10′46.1″N, 73°42′41.4″W, 633 m asl; MHNU-I 1949, 3, 32.8–48.1 mm SL, caño Unión, tributary of río Ocoa, 04°04′6.1″N, 73°42′38.0″W, 451 m asl; MHNU-I 2189, 4, 33.9–58.5 mm SL, río Guayuriba, 04°03′17.8″N, 73°45′58.2″W, 480 m asl; MHNU-I 2337, 1, 41.4 mm SL, río Guayuriba, 04°10′46.1″N, 73°42′41.4″W, 633 m asl; MHNU-I 2704, 1, 50.0 mm SL, río Guayuriba, 04°03′17.8″N, 73°45′58.2″W, 480 m asl; MHNU-I 2769, 1, 23.2 mm SL, río Guayuriba, 04°00′12.2″N, 73°30′08″W, 350 m asl; MHNU-I 2857, 20, 23.2–56.6 mm SL, caño Grande, 04°06′17.2″N, 73°39′6.8″W, 414 m asl.
DATA ACCESSIBILITY
Supplemental material is available at https://www.ichthyologyandherpetology.org/i2025051. Unless an alternative copyright or statement noting that a figure is reprinted from a previous source is noted in a figure caption, the published images and illustrations in this article are licensed by the American Society of Ichthyologists and Herpetologists for use if the use includes a citation to the original source (American Society of Ichthyologists and Herpetologists, the DOI of the Ichthyology & Herpetology article, and any individual image credits listed in the figure caption) in accordance with the Creative Commons Attribution CC BY License. ZooBank publication urn:lsid:zoobank.org:pub:E078E850-6978-43A0-983E-B9CC12681366.
AI STATEMENT
The authors declare that no AI-assisted technologies were used in the design and generation of this article and its figures.
(A) Premaxillae, anteroventral view, and (B) suspensorium of Cetopsorhamdia ramirezi, IAvH-P 18147, 56.3 mm SL, lateral view of right side. Abbreviations: EN, entopterygoid; HK, hyomandibular keel; HY, hyomandibula; MT, metapterygoid; OP, opercle; PR, preopercle; QU, quadrate. Scale bar 1 mm.
Holotype of Cetopsorhamdia ramirezi (IAvH-P 9723, 87.9 mm SL), Colombia, Casanare, Aguazul, río Únete. (A) Left lateral view, (B) dorsal view, (C) ventral view. Scale bar 1 cm.
Left lateral view of (A) Cetopsorhamdia ramirezi, holotype, IAvH-P 9723, 87.9 mm SL, Colombia, Casanare, Aguazul, río Únete; (B) C. boquillae, IAvH-P 13552, 62.0 mm SL, Colombia, Quindío, Finlandia, El Membrillal; (C) C. cf. insidiosa, MHNU-I 1159, 59.5 mm SL, Colombia, Meta, Villavicencio, río Ocoa; (D) C. nasus, CIUA 7809, 68.4 mm SL, Colombia, Valle del Cauca, Ansermanuevo, quebrada el Chanco (photo by José Luis Londoño-López); (E) C. picklei, IAvH-P 11805, 125.5 mm SL, Colombia, Norte de Santander, río Zulia. Scale bar 5 mm.
Distribution map of Cetopsorhamdia sensu stricto from Colombia: C. boquillae (orange triangle); C. insidiosa (purple pentagon); C. nasus (blue square); C. picklei (black diamond); C. ramirezi (white circle), type locality (white star).
Time-calibrated phylogeny of Cetopsorhamdia sensu stricto, estimated by BEAST2, based on multilocus dataset. Branch colors (black to blue) indicate the highest posterior density (HPD) along the topology (0.29–1). See Data Accessibility for tree file.
Scatterplot of principal component analysis (PC1 and PC2) of (A) 31 measures from 130 specimens of nine species of Cetopsorhamdia, including Chasmocranus rosae, and (B) same measures from five species of Cetopsorhamdia sensu stricto.
Contributor Notes
Associate Editor: M. T. Craig.