Phylogenomics and Morphology of the African Fish Genus Brycinus with Revalidation of Brachyalestes and Description of a New Species from the Congo Basin (Teleostei: Alestidae)
A time-calibrated phylogeny, based on nuclear ultraconserved elements and including representatives of all major alestid lineages, strongly supports two distantly related clades within the currently accepted concept of Brycinus. The first, which includes the type species of the genus, B. macrolepidotus (herein Brycinus), and a second, composed of taxa previously referred to as the B. nurse group (herein Brachyalestes), are both resolved as monophyletic. These results provide strong evidence for the restriction of the genus Brycinus to nine species, and for the revalidation of the genus Brachyalestes to accommodate 20 valid species. Within Brachyalestes, a new species from the Lulua River basin, initially misidentified as Brycinus kingsleyae, is described and resolved as sister to the widespread, central Congolese lowland species, Brachyalestes bimaculatus. Within Brachyalestes, a subclade mostly restricted to the Central Congo basin is estimated to have undergone diversification within the last 10 million years, suggesting that Late Neogene riverine reorganization likely influenced their allopatric speciation. The split of the new species, endemic to high elevation tributaries of the Lulua River, from its lowland sister species, Brachyalestes bimaculatus, suggests a Late Miocene/Early Pliocene colonization into the upland river ecosystems of the Katanga plateau in the southwestern Democratic Republic of Congo.
FRESHWATER fishes of the family Alestidae constitute by far the greatest diversity of African characiforms, with current estimates of 118 species distributed among 19 genera with species found throughout Africa’s freshwater ecosystems (Froese and Pauly, 2019; Fricke et al., 2022). Despite significant progress toward the resolution of interfamilial, suprageneric, and generic groupings, much uncertainty remains (Murray and Stewart, 2002; Calcagnotto et al., 2005; Hubert et al., 2005; Arroyave and Stiassny, 2011; Melo and Stiassny, 2022). This is particularly true for the largest alestid genus, Brycinus, and a lack of consensus regarding the systematics of this species-rich and morphologically heterogeneous assemblage is in many respects emblematic of ongoing uncertainties surrounding the systematics of Alestidae as a whole.
In the most recent revisional study of the genus, Paugy (1986) divided Brycinus into three informally named species groups: the macrolepidotus group, the longipinnis group, and the nurse group. Subsequent morphological (Murray and Stewart, 2002; Zanata and Vari, 2005) and molecular (Calcagnotto et al., 2005; Hubert et al., 2005; Arroyave and Stiassny, 2011) studies, while differing markedly in taxon coverage and details of resolved relationships, consistently refuted the monophyly of Brycinus. In agreement with both morphological and molecular studies, Zanata and Vari (2005) reassigned Paugy’s longipinnis group to the genus Bryconalestes (type species Bryconalestes longipinnis), and now including B. bartoni, B. derhami, B. intermedius, B. tessmanni, and B. tholloni, thereby restricting membership of Brycinus to the macrolepidotus and nurse groups. The monophyly of Brycinus as so restricted, was supported by Zanata and Vari (2005) based on four non-exclusive morphological synapomorphies: four teeth in the outer row of each premaxilla (reversed to five or six in some members of the macrolepidotus group, or reduced to three in some members of the nurse group); circuli oriented posteriorly over the scales and relatively straight or slightly inclined towards the horizontal midline of each scale (widespread among alestids); deep-lying midlateral stripe extending from midbody to the caudal peduncle (a heterogeneous feature, reversed in numerous species); development of a median anal-fin lobe in males formed by the relative elongation of the fifth to eighth anal-fin rays (reversed in macrolepidotus group, and present in numerous alestid genera). However, subsequent molecular studies (Calcagnotto et al., 2005; Arroyave and Stiassny, 2011; present study) resolve Brycinus (exclusive of Bryconalestes) into two, distantly related clades, suggesting the likely homoplastic nature of these and other characters thought to support the monophyly of Brycinus (Murray and Stewart, 2002; Hubert et al., 2005). The first clade corresponds to the macrolepidotus group and includes the type species of the genus, B. macrolepidotus, and B. brevis, B. carmesinus, B. grandisquamis, B. poptae, B. rhodopleura, and B. schoutedeni, and is restricted here as Brycinus. The second, corresponding to Paugy’s nurse group, is herein assigned to Brachyalestes with B. nurse as the type species, and additionally including B. abeli, B. affinis, B. bimaculatus, B. carolinae, B. comptus, B. epuluensis, B. ferox, B. fwaensis, B. humilis, B. imberi, B. jacksonii, B. kingsleyae, B. lateralis, B. minutus, B. nigricauda, B. opisthotaenia, B. peringueyi, B. sadleri, and B. taeniurus.
In addition to resolution of generic assignments and intergeneric relationships of Brycinus s.l., motivation for the current study is the discovery of a putatively undescribed species of Brachyalestes from high elevation tributaries of the middle Lulua River (Kasai ecoregion, Congo basin, Democratic Republic of Congo). The Lulua River, a large right bank tributary of the Kasai River, is recognized as one of the most species-rich communities within the entire Congo basin (Mbimbi Mayi Munene et al., 2021; Liyandja and Stiassny, 2023), and notably more than 14% of species recorded from the system are found exclusively in tributaries and have yet to be reported from the main channel. Among these are numerous specimens initially identified as Brycinus kingsleyae (Mbimbi Mayi Munene et al., 2021); however, closer examination suggests that similarity with B. kingsleyae is superficial, and that the Lulua specimens likely represent an undescribed species. As no investigation of the composition and relationships among Brycinus s.l. of west-central Africa has been undertaken, we investigate the status and phylogenetic relationships of the Lulua specimens within Brachyalestes with a focus on Central Congo basin taxa, using phylogenomic methods and morphological comparisons to formally describe and place the new species. Additionally, we provide museum-based evidence clarifying the geographical distribution of B. bimaculatus and B. kingsleyae in west-central Africa and discuss a time-calibrated phylogeny and biogeographical implications of the new species of Brachyalestes endemic to upland tributaries in the Kasai basin.
MATERIALS AND METHODS
Morphological approach.—
Counts and measurements follow Stiassny et al. (2021) and Melo and Stiassny (2022). Point-to-point linear distances were measured using digital calipers with a precision of 0.1 mm. In the description, an asterisk designates the value for the holotype, and parentheses indicate the number of examined specimens for a given count. Radiographs were obtained for the holotype, paratypes, and representative congeners and related genera. Fused PU1+U1 centra are counted as a single element, and Weberian vertebrae are counted as four elements. Two adult specimens (AMNH 251301: 70 mm SL, 80 mm SL) and numerous comparative materials (see Data Accessibility) were μCT scanned at the Microscopy and Imaging Facility at AMNH using a GE Phoenix v|tome|x with a 240 kV Nano Tube (General Electric, Fairfield, CT) with resolution ranging from 15.7 to 25.2 μm, with beam energy set at 110 kV and 166 mA. Scans were reconstructed using Phoe-nix datos|x (General Electric, Wunstorf, Germany) and rendered using VGStudio Max 3.5.1 (Volume Graphics, Heidelberg, Germany). Stomach and intestinal contents of three specimens of the new species (AMNH 251301, 1 ex., AMNH 252702, 2 ex.) were identified by entomologist David Grimaldi (AMNH). Abbreviations follow Sabaj (2020) and are: Ambrose Monell Cryo Collection, New York (AMCC/AMNH); American Museum of Natural History, New York (AMNH), Cornell University Museum of Vertebrates, Ithaca (CUMV), Field Museum of Natural History, Chicago (FMNH), and Royal Museum for Central Africa (RMCA). Standard length (SL), head length (HL), radiographed specimen(s) (RD), μCT scanned specimen(s) (CT).
Molecular phylogenetics and time-calibration analysis.—
Prior knowledge of alestid relationships based on molecular data (Arroyave and Stiassny, 2011; Melo and Stiassny, unpubl. data) and the most recent revisional study of Brycinus (Paugy, 1986) guided the selection of taxa for a molecular analysis aimed at resolving the status and relationships of the putatively new Brachyalestes from the Lulua drainage (Table 1). Sequences were generated as part of an ongoing phylogenomic analysis of African characiforms (Melo and Stiassny, unpubl. data), and ultraconserved elements (UCEs; Faircloth et al., 2012) of two species of Brycinus s.s. and eight Brachyalestes (with a focus on Congo basin taxa) were sequenced. Representatives of major alestid clades were included in the analysis in order to interrogate the monophyly of Brycinus and estimate the time of divergence of the Lulua species (Table 1). DNA was extracted using DNeasy tissue kit (Qiagen Inc., Germantown, MD) with concentration ranging from 5 to 100 ng/μl. The sample of the Lulua species of Brachyalestes (AMNH 251302, tissue AMCC-254786) had 37.1 ng/μl. Detailed protocols for UCE-phylogenomics are as in recent publications for characiform fishes (Mateussi et al., 2020; Melo et al., 2022). Using the ostariophysan probe set including 2,708 ultraconserved element loci (Faircloth et al., 2020), genomic libraries were quantified and enriched using the MYbaits Target Enrichment system (Arbor Biosciences, Ann Arbor, MI).
The UCE package PHYLUCE (Faircloth, 2015) with Illumiprocessor was used for adapter cleaning, filtering, and locating UCE loci for alignment. The contigs were aligned using MUSCLE (Edgar, 2004) in two distinct matrices accounting for different amounts of missing data: a 50% complete matrix using loci present in at least 50% of taxa, and a 70% complete matrix using loci present in at least 70% of taxa. Maximum likelihood was utilized to determine the identity and phylogenetic position of the new species. RAxML v8.2.11 (Stamatakis, 2014) ran ten maximum likelihood inferences using the GTRGAMMA model and ten distinct randomized maximum parsimony trees on the initial alignment; it also ran 1000 non-parametric bootstrap inferences via the autoMRE function.
We estimated the best-fit models of evolution for each UCE partition using ModelFinder (Chernomor et al., 2016; Kalyaanamoorthy et al., 2017) within IQ-TREE v2.2.2.6 (Minh et al., 2020). The 50% complete matrix used 1030 partitions and the 70% complete matrix used 480 partitions. Then, IQ-TREE v2.2.2.6 applied distinct models for each partition, estimated a maximum likelihood tree, and generated 1000 samples for ultrafast bootstrap (log-likelihood of consensus tree: –606812.138). Potential problems involving cross-contamination of specimens, tissues, or DNA were evaluated by the position of terminals in the larger phylogenomic dataset of Alestidae (Melo and Stiassny, unpubl. data) and compared with our current understanding of the systematics of the group. Terminals with relatively long branches resulted primarily from the low coverage of sequencing or degraded samples in ML analyses and were removed from downstream analyses.
Estimates of divergence times were conducted to investigate potentially impactful biogeographic events associated with the establishment of species of Brachyalestes in upland habitats of the Kasai basin. Two fossil-based priors and one root constraint were used to calibrate the UCE tree. The first primary calibration is represented by the earliest reported alestid teeth from the Ager basin (Eocene 54–49 Ma) of Spain (de la Peña Zarzuelo, 1996). Similarities between the fossil dentition and that of modern Alestes, Brycinus s.s., Brachyalestes, and Bryconaethiops (Melo and Stiassny, 2022) necessitated a conservative placement at the most recent common ancestor (MRCA) at the node including all sampled alestids (mean = 51.9; sigma = 2.0; 95–5% quantiles: 55.2–48.6). The second primary calibration corresponds to the highly distinctive, earliest fossil teeth of Hydrocynus, from the Dur At-Talah deposit (Eocene 48.6–33.9 Ma) of Libya (Otero et al., 2015). In accordance with previous morphological and molecular phylogenies (Murray and Stewart, 2002; Arroyave and Stiassny, 2011) and our ongoing research, Hydrocynus is resolved as sister to Alestes; we therefore assign this calibration to the node separating Hydrocynus and Alestes (mean = 44.5; sigma = 2.0; 95–5% quantiles: 47.8–41.2). Finally, a secondary calibration is applied as a root constraint reflecting the estimated divergence period between Alestidae and Hepsetidae in the Late Cretaceous at around 82 Ma (104–65 Ma, 95% highest posterior density, HPD; mean = 82.2; sigma = 10.0; 95–5% quantiles: 98.6–65.8; Melo et al., 2022). Calibration priors are available in the .xml input file (see Data Accessibility).
In Beauti v2.6.3 (Bouckaert et al., 2019), we used the 70% complete matrix with 21 terminals (119,993 bp) and applied the GTR+G+I site model, the birth–death tree prior, a fixed 70% complete ML tree, and a relaxed lognormal clock model (Drummond et al., 2006). We performed two independent BEAST v2.4.8 (Bouckaert et al., 2014) runs of 100 million generations each (total of 200 million generations), saving one tree every 10,000 generations. Tracer v1.6 (Bouckaert et al., 2014) was used to examine log scores, effective sample size (ESS > 200), and checked for stationary convergence. LogCombiner v2.6.3 (Bouckaert et al., 2014) combined the 20,002 trees, and TreeAnnotator v1.8.2 used the last 18,002 trees (burn-in: 10%) to generate the maximum clade credibility tree. Topologies are available in Supplementary Material (see Data Accessibility).
RESULTS
Maximum likelihood analyses of ultraconserved elements provide strong support for two distantly related clades within Brycinus s.l., the first formed of representatives of Brycinus s.s., and the second composed of all sampled representatives of Brachyalestes (Fig. 1). Although with limited representation of taxa, this division is also supported by the molecular phylogenies of Calcagnotto et al. (2005) and Arroyave and Stiassny (2011). Phylogenetic relationships within Brachyalestes are described in detail below (see Phylogenetic relationships).
Citation: Ichthyology & Herpetology 111, 4; 10.1643/i2023033
Morphological comparisons across Alestidae reveal a suite of derived character states shared among species of Brycinus. Unfortunately, such is not the case for Brachyalestes; however, in addition to strong molecular support for that clade (100%), we provide a simple, unique combination of two morphological traits that readily distinguish Brachyalestes from all other alestid (and characiform) genera. While it is anticipated that an ongoing, more densely sampled study of the entire Alestidae and allied families (Melo and Stiassny, unpubl. data) will provide additional morphological character data supporting the monophyly of Brachyalestes, based on the combined evidence of genomic and morphological data and shared modifications among species of these clades, we restrict the genus Brycinus to B. macrolepidotus and related species, and formally revalidate Brachyalestes. The new composition of Brachyalestes is presented in Table 2.
Brycinus Valenciennes, 1850
Type species.—
Brycinus macrolepidotus Valenciennes, 1850. Type by monotypy.
Included species.—
(9) Brycinus brevis, Brycinus carmesinus, Brycinus grandisquamis, Brycinus leuciscus, Brycinus luteus, Brycinus macrolepidotus, Brycinus poptae, Brycinus rhodopleura, and Brycinus schoutedeni.
Remarks.—
Zanata and Vari (2005: fig. 43) resolved two representatives of Brycinus (B. macrolepidotus and B. brevis) as a subclade nested within Brycinus s.l. and proposed two synapomorphies for that subclade. The first is the absence of a supraneural anterior to the fourth Weberian vertebral neural spine (e.g., Fig. 2A vs. Fig. 3A). However, our own examination across Alestidae reveals a far more widespread occurrence of the loss of an anteriorly located supraneural than suggested by Zanata and Vari (2005). As a result, the utility of this feature as diagnostic for Brycinus s.s. requires further investigation in a phylogenetic context, across all alestid lineages (Melo and Stiassny, unpubl. data). The second derived feature recognized by Zanata and Vari (2005), and corroborated here, is the loss of a sexually dimorphic anal fin. Some degree of sexual dimorphism in finnage, particularly of the anal fin, is pervasive among Alestidae, and the complete loss of such dimorphism appears to be found exclusively in Brycinus. Another apomorphic condition, identified here for the first time, is the presence of a characteristic expansion of the dorsal surface of the maxillary head (Fig. 2C). While considerable variation in the morphology of the maxilla and maxillary head is found within Alestidae (Melo and Stiassny, unpubl. data), the presence of such a broad, spoon-shaped dorsal expansion is not present in any other alestid or, as far as we can assess, any other characiform. An additional derived feature recognized here is a broad, laterally expanded nasal bone present in all Brycinus; a similar expansion is not present in any other taxon examined (see Fig. 2B vs. Fig. 3B–E).
Citation: Ichthyology & Herpetology 111, 4; 10.1643/i2023033
Citation: Ichthyology & Herpetology 111, 4; 10.1643/i2023033
Brachyalestes Günther, 1864
Type species.—
Brachyalestes nurse (Günther, 1864). Type by subsequent designation.
Included species.—
(21) Brachyalestes abeli, Brachyalestes affinis, Brachyalestes bimaculatus, Brachyalestes carolinae, Brachyalestes comptus, Brachyalestes epuluensis, Brachyalestes ferox, Brachyalestes fwaensis, Brachyalestes humilis, Brachyalestes imberi, Brachyalestes jackiae, new species, Brachyalestes jacksonii, Brachyalestes kingsleyae, Brachyalestes lateralis, Brachyalestes minutus, Brachyalestes nigricauda, Brachyalestes nurse, Brachyalestes opisthotaenia, Brachyalestes peringueyi, Brachyalestes sadleri, and Brachyalestes taeniurus.
Remarks.—
Although we have yet to find uniquely derived features diagnostic for Brachyalestes, in the absence of the synapomorphies of Brycinus, a simple combination of two morphological features serves to differentiate members of the genus from all other alestid and characiform genera. The first is the presence of 6–10 molariform inner row premaxillary teeth (3–5 teeth on each side); among Alestidae multiple molariform premaxillary teeth are also present in Alestes, Bryconaethiops, and Brycinus, and are present but somewhat reduced in Bryconalestes. The second is the cranial fontanel, which, when present, is usually restricted to the parietal–supraoccipital region; it never extends anterior to the epiphyseal bar, even in juveniles, and thus an anterior frontal fontanel is never present in Brachyalestes (Fig. 3B–D vs. Fig. 3E). The absence of Brycinus synapomorphies and the combination of these two features (presence of multiple molariform premaxillary teeth, and the absence of an anterior frontal fontanel) readily serve to distinguish Brachyalestes from all other alestid genera.
Brachyalestes jackiae Stiassny and Melo, new species
urn:lsid:zoobank.org:act:6D67A7C4-5749-4256-8454-F02942DFC5E2
Figures 4–7, Table 3
Brycinus kingsleyae.—Mbimbi Mayi Munene et al., 2021: 16 (list of species; Lulua basin).
Citation: Ichthyology & Herpetology 111, 4; 10.1643/i2023033
Citation: Ichthyology & Herpetology 111, 4; 10.1643/i2023033
Citation: Ichthyology & Herpetology 111, 4; 10.1643/i2023033
Citation: Ichthyology & Herpetology 111, 4; 10.1643/i2023033
Holotype.—
AMNH 253243, 105.8 mm SL, RD, male, Democratic Republic of Congo, Kasai Central, Mutefu, Tshimayi River, Lulua river, Kasai system, Congo basin, –6°11′37.5714″, 22°44′23.712″, J. J. Mbimbi Mayi Munene, 29 January 2010.
Paratypes.—
All from tributaries of the Lulua River, Kasai Central Province, Democratic Republic of Congo: AMNH 251302, 2, RD, 91.6–97.1 mm SL, Bampanya, Lunyenga River, –6°7′28.272″, 22°31′34.896″, J. J. Mbimbi Mayi Munene, 2009; AMNH 252702, 8, 68.8–76.4 mm SL, Dijiba, Moyo River, –6°11′14.0994″, 22°29′4.3794″, J. J. Mbimbi Mayi Munene, 15 July 2010; AMNH 252787, 2, RD, 102.6–113.8 mm SL, Kalumba, Lubi River, –5°59′43.5006″, 22°33′0.255″, J. J. Mbimbi Mayi Munene, 28 January 2010; AMNH 277628, 15, 2 CT, 59.4–96.1 mm SL, Bampanya, Lunyenga River, –6°7′28.272″, 22°31′34.896″, J. J. Mbimbi Mayi Munene, 2009; (RMCA) BE_RMCA_Vert.2023.015.P.0001-0002, 2, 83.4–96.4 mm SL, Bampanya, Lunyenga River, –6°7′28.272″, 22°31′34.896″, J. J. Mbimbi Mayi Munene, 2009; CUMV 10028, 2, 91.8–94.6 mm SL, Dijiba, Moyo River, –6°11′14.0994″, 22°29′4.3794″, J. J. Mbimbi Mayi Munene, 15 July 2010; FMNH 150030, 2, 83.0–86.2 mm SL, Dijiba, Moyo River, –6°11′14.0994″, 22°29′4.3794″, J. J. Mbimbi Mayi Munene, 15 July 2010; OS 26722, 2, 89.9–91.9 mm SL, Dijiba, Moyo River, –6°11′14.0994″, 22°29′4.3794″, J. J. Mbimbi Mayi Munene, 15 July 2010.
Differential diagnosis.—
Brachyalestes jackiae is readily distinguished from congeners other than B. carolinae, B. epuluensis, B. jacksonii, B. kingsleyae, B. lateralis, B. opisthotaenia, B. peringueyi, and B. taeniurus by the presence of a conspicuous midlateral stripe (vs. pigmentation restricted to a caudal-peduncle spot or blotches). It differs from B. carolinae, B. epuluensis, B. jacksonii, B. lateralis, and B. taeniurus in the form of the midlateral stripe, which is a diffuse narrow (B. jacksonii, B. lateralis, and B. taeniurus) or broad (B. carolinae and B. epuluensis), straight band of dark pigmentation passing from the caudal blotch to midbody or extending to the opercle (Fig. 8A). In contrast, the midlateral stripe of B. jackiae, B. kingsleyae, B. opisthotaenia, and B. peringueyi is strongly demarcated and distinctively club-shaped anteriorly (Figs. 4A, B, 8B, C). Brachyalestes jackiae differs from B. kingsleyae in the presence of an anterior extension of the club-shaped midlateral stripe in males (vs. terminal club restricted to below the level of the dorsal fin in both males and females), lack of humeral spot (vs. presence), 21–23 lateral line scales (vs. 24–30), and caudal peduncle depth 9.0–10.6% of SL (vs. 11.4–15.1% of SL). In addition to the absence of an anterior extension of a club-shaped midlateral stripe in males and the absence of a humeral spot, B. jackiae differs further from B. opisthotaenia and B. peringueyi in possessing far fewer pored lateral line scales (21–23 vs. 27–32 and 29–30, respectively). Brachyalestes jackiae is readily distinguished from its sister species, B. bimaculatus (Fig. 8D, E), by the presence of a dark club-shaped midlateral stripe (vs. absence) and absence of dark blotches (vs. presence of first blotch midlaterally located on the flank and the second on the caudal peduncle).
Citation: Ichthyology & Herpetology 111, 4; 10.1643/i2023033
Description.—
Morphometric data summarized in Table 3. Medium-sized species (maximum observed size 113.8 mm SL) with general appearance as in Figure 4. Body elongate, moderately robust, particularly so in adult females. Dorsal head profile straight from snout to supraoccipital, convex to dorsal-fin origin, slightly concave along dorsal-fin base, smoothly convex to origin of caudal fin. Ventral body profile smoothly convex to anal-fin origin, anal-fin base strongly convex, expanded and proximally scaled. Caudal peduncle longer than deep.
Adipose eyelid restricted to translucent band immediately anterior to orbit. Snout rounded, length equal to, or slightly shorter than, orbital diameter. Mouth terminal and horizontally aligned with center of orbit. Two rows of multicuspid premaxillary teeth (Fig. 5A). Outer row with 3 (1), 4* (16), or 5 (3) stout, tricuspid teeth; inner row with 4* (20) molariform teeth with 4–5 lingually oriented, cutting-edge cusps, and 2–3 molariform cusps on moderately developed buccal shelves. Fourth tooth elongate, laterally aligned with single molariform cusp on reduced buccal shelf. Contralateral premaxillae interdigitating medially with two strongly overlapping sutures. Maxilla edentulous, robust, with deeply bifurcated head (Fig. 5B). Dentary with 4* (20) outer row teeth bearing five to seven buccally oriented cusps (Fig. 5C). Fourth tooth markedly smaller than others. One inner row pair of large, robust, strongly shouldered conical teeth at dental symphysis. Contralateral dentaries strongly interdigitating lingually.
Distal margin of dorsal fin rounded or slightly emarginate, iii,8* (20) fin rays, first unbranched ray very short. Pectoral fin pointed or slightly rounded, adpressed fin tip reaching two or three scales short of vertical through pelvic-fin origin, i,13 (5) or i,14* (15) rays. Pelvic-fin rays i,8* (17) or i,9 (3), adpressed fin tip reaching 2–3 scales short of anal-fin origin. Caudal fin weakly forked. Small adipose fin present. Anal-fin rays iii,11 (2), iii,12 (12), or iii,13* (6); first ray very short. Anal fin straight or slightly emarginate in females, distal margin broadly rounded in males with middle branched rays elongated forming anal-fin lobe. Anal-fin base covered proximally by sheath of scales.
Body covered with large, regularly imbricated cycloid scales, circuli on exposed portions with horizontal orientation. Lateral line scales from supracleithrum to hypural joint 21 (1), 22 (5), or 23* (13), with 2* (16) or 3 (4) scales extending over caudal-fin base. Anteriormost 3–4 scales of lateral line descending steeply to below midlateral body plane. Scales in transverse series from lateral line to dorsal-fin origin 4.5* (20). Scales in transverse series from lateral line to pelvic-fin insertion 2 (1) or 2.5* (19). Middorsal series from tip of supraoccipital to dorsal-fin origin 7 (1), 8 (4), 9* (13), or 10 (2). Circumpeduncular scales 9 (1) or 10* (19).
Supraneurals 7* (4) or 8 (2), first supraneural associated with fourth Weberian vertebral spine (Fig. 6A); predorsal vertebrae 7 (6); precaudal vertebrae 17 (5) or 18 (1); total vertebrae 36 (2), 37 (2), or 38* (2).
Osteological features.—
Circumorbital series complete, forming uninterrupted ring around orbit (Fig. 5D). Supraorbital elongate, strongly sigmoid anteriorly, ventral process present, articulating with dorsal face of lateral ethmoid. Antorbital in contact with supraorbital posteriorly, overlapping first infraorbital anteriorly. Three pairs of posttemporal fossae present (Fig. 6B). Median posttemporal fossae located entirely within epioccipitals. Dorsal posttemporal fossae well developed, bounded by supraoccipital, parietals, and epioccipitals; dorsal fossae slightly smaller than ventral posttemporal fossae, bounded by epioccipitals and pterotics. Exoccipital foramen slightly smaller than foramen magnum. Presence of small parietal–supraoccipital fontanel correlated with size (Fig. 6C; 70 mm SL), completely absent in larger individuals (Fig. 6D; 80 mm SL). Well-developed, prong-like pterotic processes.
Color in alcohol.—
Ground coloration yellowish brown. Snout, dorsal portion of head, opercle, and body above and including lateral line markedly darker than yellowish pale ventrum. Deep-lying, club-shaped black stripe along midlateral surface from vertical through rear of dorsal-fin base to median caudal-fin rays in females and juveniles (Fig. 4B). Broad dark stripe from cheek connecting to midbody stripe in males, attenuating at level of anal fin and over caudal peduncle, terminating at median caudal-fin rays (Fig. 4A). Dorsal, caudal, and anal fins outlined by small, dark chromatophores. First rays of pectoral and pelvic fins with scattered dark chromatophores. Adipose fin with small chromatophores concentrated along distal margins.
Color in life.—
Overall silvery iridescence. Fins generally orange/yellow. Green iridescent pigments overlying midbody scales, partially obscuring anterior extension of deep lying club-shaped midlateral stripe of male specimens. Darker pigmentation over basal portions of anal and caudal fins in males (Fig. 7A, B).
Sexual dimorphism.—
Males of Brachyalestes jackiae differ from females in the possession of a conspicuous (in preservation), broad, dark band that extends from club-shaped midlateral stripe to the snout. Males also have an elongation of branched anal-fin rays forming a distinct anal-fin lobe, bordered proximally by 2–3 scale rows over the anal-fin base (Figs. 4A, 7A), while females have a dark club-shaped stripe confined to the posterior midbody region and straight anal-fin margin, with a single scale row along the anal-fin base (Figs. 4B, 7B).
Distribution.—
Brachyalestes jackiae is abundant in forested tributaries of the Lulua River (Lubi, Luna, Lunyenga, Mbuyi, Moyo, and Tshimayi) of the Congo basin, Democratic Republic of Congo (Fig. 9A). However, due to inaccessibility much of the middle and upper Lulua River has been poorly sampled (Mbimbi Mayi Munene et al., 2021) and it is likely that additional collecting efforts will expand the range of this species.
Citation: Ichthyology & Herpetology 111, 4; 10.1643/i2023033
Ecology.—
Brachyalestes jackiae has been collected in tributaries of the Lulua River upstream of the city of Kananga (former Luluabourg) at elevations between 562 and 774 m asl. Each collection locality was bordered by gallery forest/vegetation with swift-flowing, relatively shallow (1–2 m) water flowing over rocks, gravel, and sandy substrates (Fig. 9B–E). Water temperatures ranged from 27°C to 30.5°C, and pH between 5.4 and 6.9 (Mbimbi Mayi Munene et al., 2021). The species appears to be primarily insectivorous, and digestive tract contents of three dissected specimens (AMNH 251301, 1 ex. and AMNH 252702, 2 ex.) contained insect remains of strictly allochthonous origin; no remains of aquatic insect taxa were present. Identified contents included remnants of three coleopteran species (including Chrysomelidae and Scarabaeidae), representatives of three or four hymenopteran genera (including Diapriidae and Camponotus), two lepidopteran, one isopteran worker, one dipteran (probably Dolichopodidae), one hemipteran, one heteropteran, one Auchenorrhyncha, and one apterygote. The digestive tract (unraveled) is slightly shorter than SL, and consists of a short esophagus, a bulbous stomach encased by six stout pyloric caeca, and a simple, double-looped intestine (Fig. 7C).
Etymology.—
Brachyalestes jackiae is named for Jackie Black, in recognition of her support and encouragement to MLJS and her Congolese colleagues over many years of work in the Congo basin.
Phylogenetic relationships and timing of diversification.—
Our 50% complete matrix consisted of 1031 ultraconserved element loci with 226,210 bp, and analysis resulted in a maximum likelihood (ML) phylogeny with final score –618437.543946 and 50 bootstrap replicates (Fig. S1; see Data Accessibility). The 70% complete matrix consisted of 481 UCE loci with 119,993 bp, and analysis resulted in an ML phylogeny with final score –327188.008263 and 50 bootstrap replicates (Fig. 1). There was a single difference in the relationships using the 50% and 70% complete datasets: the sister clade of Alestopetersius + Nannopetersius was Phenacogrammus in the RAxML analysis (92% and 98%) or Bathyaethiops in the IQ-TREE analysis (99% and 96%). Likelihood analyses with each matrix and both RAxML and IQ-TREE returned identical relationships within Brachyalestes, with minor bootstrap differences supporting relationships between B. imberi and B. nurse (86% and 94%), and a “central Congo” node encompassing B. comptus, B. lateralis, B. jackiae, and B. bimaculatus (80% and 82%; Figs. S2, S3; see Data Accessibility). Downstream analyses utilized the 70% complete matrix. Table 4 contains information about reads for each taxon.
Among central Congo taxa sampled, Brachyalestes lateralis appears as sister to B. comptus (100%), and B. bimaculatus is strongly supported as sister to B. jackiae (100%; Fig. 1). This arrangement accords with an earlier multilocus phylogeny in placing B. comptus close to B. lateralis and B. fwaensis (not examined here), with B. bimaculatus resolved as sister to that clade (Arroyave and Stiassny, 2011).
The resulting timetree indicates a divergence of Hepsetidae and Alestidae during the Late Cretaceous (83 Ma; 101–67 Ma, 95% HPD), which corresponds to the prior distribution of the analysis based on previous estimates (Melo et al., 2022). Results indicate that the most recent common ancestor (MRCA) of Alestidae began to diversify during the Early Eocene (54 Ma; 57–50 Ma, 95% HPD; Fig. 10). During the Middle Eocene, Brachyalestes diverged from the MRCA of Bryconalestes, Bathyaethiops, Phenacogrammus, Alestopetersius, and Nannopetersius (44 Ma; 49–40 Ma, 95% HPD). The earliest divergence of the Brachyalestes occurred during the Oligocene–Miocene transition (22 Ma; 32–15 Ma, 95% HPD), separating B. nurse and B. imberi from the other species. The Lower Guinean B. opisthotaenia appears to have diverged approximately 14 Ma (20–9 Ma, 95% HPD), whereas B. kingsleyae appears around 11 Ma (16–7 Ma, 95% HPD) during the Late Miocene. The group containing the central Congolese species underwent diversification during the last 10 Ma. Brachyalestes comptus from the lower and middle Congo diverged from B. lateralis from the upper, middle Congo, and Zambezi around 6 Ma (10–3 Ma, 95% HPD), whereas B. jackiae from upland tributaries of the Lulua River diverged from B. bimaculatus from the lowland Cuvette Centrale around 7 Ma (11–4 Ma, 95% HPD).
Citation: Ichthyology & Herpetology 111, 4; 10.1643/i2023033
DISCUSSION
Our study has begun resolution of relationships within Brachyalestes by identifying, with strong support, the central Congolese species B. bimaculatus as sister to the newly described species, B. jackiae (Fig. 1). Two morphological synapomorphies support this sister group relationship: both species lack all trace of the humeral spot present in congeners, and adult males possess a broad, scaly sheath covering the proximal portion of the expanded anal-fin lobe (Figs. 4, 8). Within Brachyalestes, B. bimaculatus exhibits a unique pigmentation patterning consisting of two large blotches of dark pigmentation along the flank and caudal peduncle, one located midlaterally immediately below the dorsal-fin base and another over the caudal peduncle (Fig. 8D, E). Additionally, both species have a broad band of chromatophores extending throughout the midlateral portion of the body above lateral line, which is particularly evident in adult males (Figs. 4A, 8D). However, such a dark band over the midbody of males appears in several other species of Brachyalestes (Toham and Teugels, 1997; Paugy and Schaefer, 2007; Decru et al., 2016) and merits further examination in a phylogenetic context.
Brachyalestes bimaculatus and B. jackiae are allopatrically distributed (Fig. 9). Brachyalestes bimaculatus, originally described from Kutu and Lake Leopold II (Lake Mai-Ndombe; Boulenger, 1899), is distributed throughout the Cuvette Centrale of the middle Congo basin, including the tributaries Fimi (Mfimi), Likouala, Lulonga, Ruki, and Sangha (Fig. 9). This distribution suggests a predominance in lowland ecosystems of sediment-rich, large rivers, swamps, and lakes of the Cuvette Centrale of the middle Congo. In contrast, B. jackiae is endemic to tributaries (Lubi, Luna, Lunyenga, Mbuyi, Moyo, and Tshimayi Rivers) of the Lulua River (Fig. 9), suggesting an exclusively upland distribution (above 560 m asl) in small stream habitats currently isolated from the Cuvette Centrale by a series of rapids and waterfalls along the Lulua River as it flows over the Katanga plateau (Mbimbi Mayi Munene et al., 2021).
Brachyalestes kingsleyae, the species with which B. jackiae was originally confused, was described from two syntypes collected in the Ogooue River, Gabon (Günther, 1896). In his taxonomic revision of Brycinus s.l., Paugy (1986) reported Brachyalestes kingsleyae from Lower Guinea rivers (Sanaga, Nyong, and Ntem of Cameroon, Nyanga and Ogooue-Ivindo of Gabon, Loeme and Kouilou of the Republic of the Congo, and the Chiloango of the Cabinda province of Angola), but also considered the species to be present in upper sections of the Kasai, Oubangui-Uele, Lualaba, and Luapula–Mweru of the Congo basin. However, Paugy’s list of examined material contained only specimens from the Loeme and Ogooue in Gabon and the Republic of the Congo (Paugy, 1986). Toham and Teugels (1997) redescribed B. kingsleyae and B. opisthotaenia based on material from Lower Guinea but, presumably following Paugy (1986), also reported B. kingsleyae as present in the Congo basin. However, based on our own examination of specimens from throughout the region (Supplementary Material; see Data Accessibility), we consider these Congo basin records likely to be the result of misidentifications and we consider Brachyalestes kingsleyae to be restricted to coastal and inland rivers of Lower Guinea from Cameroon to Cabinda (Fig. 9).
Widespread transcontinental uplifting and severe climatic shifts reorganized river networks across the African continent during the Miocene, and this epoch is broadly considered to be a key period for the diversification of a wide array of African fish families (Day et al., 2017, 2023; Bragança and Costa, 2019) including Alestidae (present data; Fig. 10). Following the climatic optimum warm/humid phase of the Middle Miocene, the second half of the Miocene experienced a global cooling trend accompanied by the expansion of arid biomes, including the formation of the Sahara Desert, followed by an extended period of climatic instability (Zachos et al., 2001; Brennan and Keogh, 2018; Bragança et al., 2021). While the geo-hydrological history of the central Congo basin is poorly understood, dynamic climatic and hydrological landscapes during the Late Miocene to Early Pliocene likely facilitated widespread allopatric speciation throughout the region. The current configuration of the Congo River is considered to have arisen around the time of the Miocene–Pliocene transition with the final capture of a large paleo-Congo Lake which is thought to have occupied much of the present-day Cuvette Centrale (Beadle, 1981; Stankiewicz and de Wit, 2006; Stiassny and Alter, 2021). Interestingly, Stiassny et al. (2021) have suggested that the Mfimi–Lukenie River represents the southern boundary of that paleo-Congo Lake, and this corresponds precisely with the present southern limit of the distribution of Brachyalestes bimaculatus. The dating analysis suggest a Late Miocene allopatric speciation between B. bimaculatus and B. jackiae, with one species remaining in lowland central Congo, and the other colonizing upland rivers of the Kasai system on the Katanga plateau; such a hypothesis can be tested with additional evidence from time-calibrated phylogenies for other Lulua River basin endemics.
The present study provides strong support for the restriction of Brycinus and provides a foundational phylogeny for Brachyalestes; however, many issues and questions remain. A phylogenetic analysis must include the remaining seven species of Brycinus and 13 species of Brachyalestes, as well as species of the remaining genera. A comprehensive study of the systematics, morphological evolution, and biogeographic reconstruction of the entire family Alestidae are the subject of ongoing research by two authors (MJLS and BFM) in an effort to provide a solid basis for our understanding of the diversification of characiform fishes across the entire African continent.
MATERIAL EXAMINED
Material Examined is available as supplemental material (see Data Accessibility).
DATA ACCESSIBILITY
Supplemental material is available at https://www.ichthyologyandherpetology.org/i2023033. Raw sequence data are available at NCBI BioProject PRJNA1033584. 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:AD53CF8D-5172-42AA-92C0-B6489AEC6B06.
Maximum likelihood tree of Brachyalestes and related alestid genera based on a 70% complete dataset of 481 ultraconserved element loci (119,993 bp). Brachyalestes jackiae highlighted in red. Photographs: J. Cutler, J. Mbimbi, M. Stiassny, R. Palmer, and South African Institute for Aquatic Biodiversity (used with permission).
Brycinus macrolepidotus, AMNH 270775. (A) Posterior neurocranium, Weberian apparatus, predorsal vertebrae, and associated structures (lateral view). (B) Neurocranium, premaxillae, and circumorbitals (dorsal view). (C) Isolated maxilla in lateral (left) and dorsal (right) views.
(A) Alestes liebrechtsii, AMNH 254612, posterior neurocranium, Weberian apparatus, predorsal vertebrae, and associated structures (lateral view); (B) Alestes liebrechtsii, AMNH 254612, neurocranium, premaxillae, and circumorbitals (dorsal view); (C) Bryconaethiops microstoma, AMNH 253823; (D) Bryconalestes longipinnis, AMNH 59626; (E) Brachyalestes nurse, AMNH 215629.
Brachyalestes jackiae. (A) AMNH 253243, holotype, male, 105.8 mm SL, Mutefu, Tshimayi River, Lulua-Kasai system, Kasai Central, Democratic Republic of Congo. (B) AMNH 252787, paratype, female, 114.1 mm SL, Kalumba, Lubi River, Lulua, Kasai Central, Democratic Republic of Congo. Scale bars = 1 cm.
Brachyalestes jackiae, AMNH 277628. (A) Premaxillae and dentition (ventral view); (B) isolated maxilla in lateral (left) and dorsal (right) views; (C) lower jaw (dorsal view); (D) circumorbital series.
Brachyalestes jackiae, AMNH 277628. (A) Posterior neurocranium, Weberian apparatus, predorsal vertebrae, and associated structures (lateral view); (B) neurocranium (posterior view); (C) neurocranium (dorsal view), 70.9 mm SL; (D) posterior neurocranium (dorsal view), 79.8 mm SL.
Brachyalestes jackiae, live coloration: (A) male, (B) female, (C) digestive tract (slightly unraveled for clearer depiction of morphology), after removal of liver, pancreas, gallbladder, spleen, and adherent tissues.
(A) Brachyalestes epuluensis, AMNH 5995, male, 79.7 mm SL, Avakubi, Ituri River, Democratic Republic of Congo (DRC); (B) Brachyalestes kingsleyae, AMNH 262921, male, 104.6 mm SL, Ivindo River, near Makokou, Ogooué-Ivindo, Gabon; (C) Brachyalestes kingsleyae, AMNH 262921, female, 104.0 mm SL, Ivindo River, Gabon; (D) Brachyalestes bimaculatus, AMNH 240760, male, 115.1 mm SL, confluence of Lofongo and Luilaka Rivers, Salonga National Park, DRC; (E) Brachyalestes bimaculatus, AMNH 242470, female, 102.7 mm SL, Lac Nkolentulu, Bandundu, DRC. Scale bars = 1 cm.
(A) Map of the central Congo basin and Lower Guinea showing analyzed museum records of Brachyalestes bimaculatus (green diamonds), B. jackiae (red triangles), and B. kingsleyae (yellow circles); collection habitats at (B) Kamuandu, Lubi River, (C) Lunyenga River, (D) Tshimayi River, (E) Luna River. Photographs by José J. M. M. Mbimbi.
Time-calibrated phylogeny based on BEAST analyses of ultraconserved elements (70% complete matrix; 481 loci; 119,993 bp). Ma = million years ago; Pal = Paleocene; Olig = Oligocene; Pli = Pliocene; Q = Quaternary. See Data Accessibility for tree file.
Contributor Notes
Department of Ichthyology, American Museum of Natural History, 200 Central Park West, New York, New York 10024; ORCID: (MLJS) 0000-0001-8220-4768; and (BFM) 0000-0002-0499-567X; Email: (MLJS) mljs@amnh.org; (CK) coop.keane@gmail.com; and (BFM) bmelo@amnh.org. Send correspondence to BFM.
Département de Biologie, Faculté des Sciences, Université de Kinshasa, B.P. 190 Kin XI, Kinshasa, République Démocratique de Congo; ORCID: 0009-0001-7666-1820; Email: jjmbimbishambuyi@gmail.com.
Associate Editor: R. E. Reis.