Naming the other cousin: A new goldie barb (Cyprinidae: Smiliogastrininae) from the northeast escarpment in South Africa, with proposed taxonomic rearrangement of the goldie barb group in southern Africa

REGULAR ARTICLE Naming the other cousin: A new goldie barb (Cyprinidae: Smiliogastrininae) from the northeast escarpment in South Africa, with proposed taxonomic rearrangement of the goldie barb group in southern Africa Martinus Scheepers 1,2 | Pedro H. N. Bragança 1,3 | Albert Chakona 1,2 1 NRF-South African Institute for Aquatic Biodiversity (NRF-SAIAB), Makhanda, South Africa 2 Department of Ichthyology and Fisheries Science, Rhodes University, Makhanda, South Africa 3 Department of Ichthyology, American Museum of Natural History, New York, New York, USA Correspondence Martinus Scheepers, NRF-South African Institute for Aquatic Biodiversity (NRF-SAIAB), P. Bag 1015, Makhanda (Grahamstown) 6140, South Africa. Email: [email protected]rf.ac.za Funding information National Research Foundation-Foundational Biodiversity Information Program REFRESH project, Grant/Award Number: FBIP211006643719; National Research Foundation-Foundational Biodiversity Information Program TOPOTYPES project, Grant/Award Number: IBIP-BS 13100251309; World Wildlife Fund, Grant/Award Number: NRF-SAIAB-WWF 40001528-2019 Abstract A growing body of evidence indicates that the global diversity of freshwater fishes has not been fully documented. Studies of freshwater fishes that were previously thought to be morphologically variable have revealed the existence of deeply divergent lineages, with many distinct species. In southern Africa a number of Enteromius species exhibit either exceedingly wide or divided distribution patterns that should be rare for freshwater fishes with limited dispersal opportunities between river systems. One such species is the sidespot barb, Enteromius neefi. As currently defined, E. neefi has a disjunct distribution that is divided between rivers in the northeast escarpment in South Africa and Eswatini, and tributaries of the Upper Zambezi in Zambia and southern Congo in the Democratic Republic of Congo, with a large geographic gap between these two populations. With the use of molecular and morphological methods, the level of divergence between the two populations was examined, and a new species was described from the Steelpoort River in the Limpopo River system of South Africa. Findings from this study provide further evidence for a number of taxonomic problems within the goldie barbs of southern Africa, and some taxonomic rearrangements are proposed for this group. KEYWORDS color pattern, Cypriniformes, integrative taxonomy, systematics 1|INTRODUCTION Understanding the systematics of fishes in mega-diverse orders such as the Cypriniformes has improved markedly in recent years due to molecular phylogenetic analyses (Mayden et al., 2008,2009; Saitoh et al., 2011), and interfamilial relationships are becoming well established (ShunPing et al., 2008; Wang et al., 2012; Yang et al., 2015). The family Cyprinidae consists of freshwater fishes widely distributed throughout North America, Eurasia, and Africa (Skelton, 2001). Morphological similarity is common among cyprinids, which led early taxonomists, relying on morphological characters alone, to group together species in outsized genera (Skelton et al., 2018). One such genus was Barbus (Daudin 1805), which Myers (1960) called a “monstrous aggregation,”consisting of more than urn:lsid:zoobank.org:pub:7541C64C-65FE-4DAA-B488-35CDFC2FD735. Received: 6 November 2023 Revised: 20 June 2024 Accepted: 25 June 2024 DOI: 10.1111/jfb.15870 FISH This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. © 2024 The Author(s). Journal of Fish Biology published by John Wiley & Sons Ltd on behalf of Fisheries Society of the British Isles. J Fish Biol. 2024;105:1137–1150. wileyonlinelibrary.com/journal/jfb 1137 800 species spread across Africa, Asia, and Europe (Hayes & Armbruster, 2017). However, phylogenetic analyses revealed that Barbus included many different and nonrelated lineages and groups, indicating that the aggregation of large numbers of freshwater species spread across three continents into a single genus made little sense (Skelton et al., 2018). In the late 1980s a number of taxonomic and karyological studies that investigated the diversity and relationships of African cyprinids provided evidence that species placed in Barbus showed three different ploidy levels, resulting in the revalidation of many genera that were at that time considered synonyms of Barbus. The genera Luciobarbus (Heckel 1843), Cheilobarbus (Smith 1841), and Pseudobarbus (Smith 1841), which were considered to be synonyms of Barbus,as well as the recently described genera Amatolacypris Skelton et al., 2018,Namaquacypris Skelton et al., 2018, and Sedercypris Skelton et al., 2018, are tetraploids (2n=c. 100 chromosomes). All hexaploid (2n=c. 150 chromosomes) species have been placed in the genus Labeobarbus Rüppel 1835, whereas the African diploid species (2n=c. 50 chromosomes) were initially designated as “Barbus”to indicate that they are not closely related to the Eurasian tetraploid Barbus s.s. (Berrebi et al., 1996; Golubtsov & Krysanov, 1993; Guégan et al., 1995; Oellermann & Skelton, 1990; Skelton, 1988; Skelton et al., 2018; Tsigenopoulos et al., 2002). Following the study of Yang et al. (2015), the African diploid species have been assigned to the genus Enteromius Cope 1867 in the subfamily Smiliogastrinae (Tan & Armbruster, 2018). This new designation was initially contested (Ren & Mayden, 2016; Schmidt & Bart, 2015; Stiassny et al., 2016; Stiassny & Sakharova, 2016) but is now widely accepted as the use of “Barbus”would still link the small African species to the true Eurasian Barbus, when in fact the African and Eurasian taxa belong to two different subfamilies, Smiliogastrinae and Barbinae, respectively (Skelton 2015,2016; Yang et al., 2015; Skelton et al., 2018; Tan & Armbruster, 2018). The resurrection of Enteromius concerns only those African diploid species not already placed in the genera Barboides Brüning 1929, Barbopsis Di Caporiacco 1926, Caecobarbus Boulenger 1929, Clypeobarbus Fowler 1936, and Prolabeops Schultz 1941. The phylogenetic placement of Barbopsis is still unknown, and Barboides,Caecobarbus,Clypeobarbus,Prolabeops, and the tetraploid genus Pseudobarbus are nested within Enteromius (Hayes & Armbruster, 2017; Ren & Mayden, 2016; Schedel et al., 2022), which renders Enteromius polyphyletic and illustrates the need for further taxonomic changes. Currently, Enteromius consists of 226 valid species distributed across Africa (Fricke et al., 2023; Hayes & Armbruster, 2017), with new species continuously being described (e.g., Kambikambi et al., 2021; Katemo Manda et al., 2020), revalidated (e.g., Englmaier et al., 2020; Maetens et al., 2020; Schmidt et al., 2018), or awaiting formal description (e.g., Popoola et al., 2022). Forty-two Enteromius species occur in the southern African region as defined by Skelton (2001), 23 of which are endemic to South Africa (Froese & Pauly, 2022; Kambikambi et al., 2021; Skelton, 2001). Skelton (2001) placed the southern African barbs in three groups based on differences in the primary dorsal-fin ray morphology: the soft-rayed barbs with a smooth and flexible primary dorsal-fin ray, the sawfin barbs with a spinous and serrated primary dorsal-fin ray, and the spinefin barbs with a smooth and spinous primary dorsal-fin ray. Within the soft-rayed barbs, two additional groups were identified, the goldie and chubbyhead barbs, based on a distinctive breeding colouration and unique morphological characteristics. Currently, only three species belong to the goldie barb group: the shortfin barb Enteromius brevipinnis (Jubb 1966), the sidespot barb Enteromius neefi (Greenwood, 1962), and the goldie barb Enteromius pallidus (Smith 1841). The three species in the goldie barb group are characterized by relatively small (<70 mm standard length [SL]) compact bodies, two pairs of barbels, and a bright golden color attained by males during the breeding season. Despite possessing similar characteristics, Enteromius greenwoodi (Poll, 1967), Enteromius lineomaculatus (Boulenger 1903), Enteromius thamalakanensis (Fowler 1935), and Enteromius viviparus (Weber 1897) have not been included under the goldie barb group, but no justification has been provided for their exclusion (Skelton, 2001). E. neefi was described based on 17 specimens collected in 1960 from the Kabompo River, a major tributary of the Upper Zambezi River system, in Zambia (Greenwood, 1962). A combination of characteristic markings distinguish E. neefi from E. brevipinnis and E. pallidus: a spot at the base of the anal and pectoral fins, variable number of large dark spots along the body, and thin wavy parallel lines along the top and bottom of each scale row of the flank, extending ventrally beyond the lateral line scale row (Greenwood, 1962; Jubb, 1968; Skelton, 2001). Initially the Upper Zambezi and headwaters of the Lualaba (southern Congo) were the only known distribution range for the species, until specimens with similar characteristics were recorded from the Orighstad and Steelpoort rivers, Limpopo River system in South Africa (Jubb, 1968). These specimens were assigned to E. neefi, mainly based on possession of wavy parallel lines that are the key distinguishing feature for this species. Subsequently more populations were identified in the Letaba, Mutale, Levuhu, Sabie, Crocodile, Makondo, and Mfolozi rivers. This resulted in the species having a disjunct distribution pattern, with the two known populations separated by a large geographic gap of 1000 km between the Upper Zambezi River in the north and the northeast escarpment of South Africa (Skelton, 2001). Disjunct distribution patterns have been recorded for other southern Africa freshwater fish species, including E. pallidus, Mesobola brevianalis (Boulenger 1908), and Amphilius natalensis (Boulenger 1917). These species, initially thought to represent widely distributed species, but following application of integrative taxonomic research, were shown to represent deeply divergent lineages, resulting in the recent description of new species (Chakona et al., 2015; Mazungula & Chakona, 2021; Riddin et al., 2016). The aim of the present study was to use molecular COI (cytochrome oxidase subunit I) and morphological data within an integrative taxonomy perspective to delimit species boundaries within E. neefi. Furthermore, the phylogenetic status of the “goldie”barb group was explored by acquiring topotypic sequences for southern African species with two pairs of barbels and males that attain bright golden colouration during the breeding season. Potential taxonomic rearrangements are proposed, and conservation implications are highlighted. 1138 SCHEEPERS ET AL. FISH 10958649, 2024, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jfb.15870 by Capes, Wiley Online Library on [29/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 2|MATERIALS AND METHODS 2.1 |Sample collection Description of the new Enteromius species is based on 15 specimens collected during field surveys in the Steelpoort River (Limpopo River system), South Africa, in May 2012 and May 2021. Comparative topotype specimens of E. neefi were collected during field surveys between 2011 and 2019 from the Kabompo River in Zambia, and the type specimens were measured by the author Albert Chakona at the National History Museum, London. A total of 11 new sequences were generated for the two E. neefi populations, E. brevipinnis, E. greenwoodi,E. thamalakanensis, and E. viviparus. Twenty-five sequences were used from BOLD (Barcode of Life Data) and a single sequence from GenBank. The approaches used for sample collection and processing were approved by the NRF-SAIAB Animal Ethics Committee (reference no.: 2014/03). 2.2 |Molecular data 2.2.1 | Extraction, amplification, and sequencing DNA was extracted from preserved tissues using the salting-out protocol of Sunnucks and Hales (1996), and DNA concentrations were quantified using a Nanodrop ND-1000 spectrophotometer (Nanodrop Technologies, Inc.). A fragment of the mitochondrial COI gene was amplified by PCR using universal fish DNA barcoding primer sets: FishF1 and FishR1 (Ward et al., 2005) and VF2-T1 and VR1-T1 (Ivanova et al., 2007). PCRs were performed with a Veriti 96-well thermal cycler (Applied Biosystems, Foster City, CA, USA). Each reaction mixture (25 μL) contained 100–200 ng of template DNA, 12 μL of Taq DNA Polymerase 2Master Mix RED (Ampliqon PCR Enzymes & Reagents), 0.5 μL of each primer (20 pmol), and 7 μLof molecular-grade water. PCR amplification was performed using the following profile: 3 min at 95C, followed by 12 cycles of 30 s at 95C, 40 s at 62C, decreasing by 0.5C per cycle to 56.5C, and 50 s at 72C. This was followed by 25 cycles of 30 s at 95C, 30 s at 56C and 50 s at 72C, with a final extension of 72C for 7 min. PCR products were purified using the ExoSAP method with a reaction mixture containing 5 μL of PCR product, 0.5 μL of exonuclease I, and 1 μLof FastAP (Applied Biosystems). Cycle sequencing was performed using the BigDye Cycle Sequencing Kit (Applied Biosystems) and sequenced at SAIAB using an ABI 3730xl DNA Analyzer (Applied Biosystems). Sequences were edited and trimmed using Geneious Prime 2021.2.2. 2.2.2 | Phylogenetic analysis and species delimitation To assess the phylogenetic affinities of the two E. neefi populations and verify the distinctiveness of the new species, all COI sequences of African Smiliogastrinae available in BOLD and GenBank and new sequences from this study were used to generate a phylogenetic tree (Figure S1). Sequences were aligned using MAFFT, version 7.450 (Katoh & Misawa, 2002; Katoh & Standley, 2013), and inspected using MEGA 11 (Tamura et al., 2021) for the presence of stop codons. Identical sequences were removed prior to phylogenetic analysis. Bayesian analysis was performed using MrBayes, version 3.2.6 (Huelsenbeck & Ronquist, 2001), partitioning the dataset by codon position and employing reversible-jump Markov chain Monte Carlo (RJ-MCMC) sampling, with time-reversible substitution models and γ-distributed rate heterogeneity (Huelsenbeck et al., 2004). The RJ-MCMC method precludes the a priori selection of substitution models for each partition and instead allows models to be sampled in proportion to their posterior probability (Huelsenbeck et al., 2004). Two parallel analyses of four Markov chains and 5 million generations were run, sampling trees every 1000 generations and discarding the first 25% of trees as burn-in. Resulting trees were visualized in FigTree, version 1.4.4 (Rambaut, 2009). To assess the convergence between the two runs, the average standard deviation of split frequencies was monitored in MrBayes to ensure it was <0.05. In addition, Effective Sample Size (ESS) values and the potential-scale reduction factor for all parameters were examined using MrBayes and found to approach >100 and 1.0, respectively. From this tree the clade containing the two E. neefi populations was retained for detailed analysis. Four molecular-based species delimitation methods were used to identify operational taxonomic units (OTU) and explore species boundaries between taxa belonging to the retained lineage. The first two methods, “automatic barcode gap discovery”(ABGD) and “assemble species by automatic partitioning”(ASAP), are genetic distance–based methods that rely on the analysis of single-locus sequence alignments to define species partitions based on pair-wise genetic distances (Puillandre et al., 2012;Puillandreetal.,2021). Two coalescent methods, the “Bayesian implementation of the Poisson tree processes” (bPTP) (Zhang et al., 2013) and the “general mixed Yule coalescent” (GMYC) (Fujisiwa & Barraclough, 2013), were performed. The bPTP relies on single-locus molecular data to delimit species based on the number of nucleotide substitutions between haplotypes. In the GMYC method, species are delimited based on branch lengths, and it requires an utrametric tree to define intraspecific and interspecific threshold patterns. (Fujisiwa & Barraclough, 2013). The ultrametric tree used for the GMYC analysis was constructed in BEAST2 (Bouckaert et al., 2019) using the Yule model prior with an optimized relaxed clock. The assumption for all these methods is that interspecific variability will be substantially higher than intraspecific differentiation. 2.3 |Morphological data Following Chakona et al. (2014), 15 morphological and 16 meristic characters were obtained for 16 specimens of E. neefi from the Kabompo River, including the type series. For the Limpopo population, 15 specimens from the Steelpoort River (Limpopo River system) were included. Ten specimens of the Limpopo population and 13 specimens of the Kabompo population were radiographed to facilitate the SCHEEPERS ET AL.1139 FISH 10958649, 2024, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jfb.15870 by Capes, Wiley Online Library on [29/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License counting of skeletal features. Measurements were taken using digital calipers to the nearest 0.1 mm. Principal component analysis (PCA) was performed using PAST (Hammer et al., 1999) on raw meristic data and normalized morphometric data (Lleonart et al., 2000) to explore the variables that might assist in distinguishing both species. The normalization procedure allows for size-free comparison between specimens. Invariant characters (see Table 3) were excluded from the analysis. 3|RESULTS Genetic and morphological data support the recognition of the Steelpoort E. neefi population as a new goldie barb species and here described as Enteromius niggie sp. nov. from the Limpopo River system of South Africa. 3.1 |Molecular phylogenetic analyses A Bayesian phylogenetic tree of all available COI sequences of African Smiliogastrins revealed that E. niggie sp. nov. and E. neefi clustered within a major clade containing eight morphologically identified species, Enteromius atkinsoni (Bailey 1969), Enteromius macinensis (Daget 1954), E. viviparus,E. pallidus,E. brevipinnis,E. thamalakanensis, E. lineomaculatus, and E. greenwoodi, as well as two possible candidate species Enteromius sp. “Cuebe”and Enteromius sp. “Bie”from multiple drainages in sub-Saharan Africa (Figure 1; Table 1; Figure S1). Furthermore, two species, E. brevipinnis and E. viviparus, were each split into at least two polyphyletic lineages. Detailed analysis using four species delimitation methods based on 38 COI sequences representing E. niggie sp. nov., E. neefi, and the other eight taxa mentioned earlier recovered 19 OTUs or candidate species (Figure 1). These OTUs were grouped into three wellsupported clades (Figure 1). Clade A contained E. macinensis from Burkina Faso as well as two OTUs that were morphologically identified as E. atkinsoni from East Africa. The OTU-containing specimens from Malawi and Mozambique are substantially differentiated (3.3% divergence) from the topotype of E. atkinsoni from the Rufiji River system (Figure 1). Clade B comprised two OTUs from coastal East Africa, lower Zambezi, and the Incomati systems that are currently identified as Enteromius cf. viviparus. These two OTUs are substantially differentiated (9.1%–9.8% divergence) from and are distantly related to the topotypes of E. viviparus from the Mdloti River system, in South Africa (Figure 1). Clade C contained 14 OTUs, including E. niggie sp. nov. and three species currently assigned to the goldie group: E. pallidus,E. neefi, and E. brevipinnis. Topotypes of E. viviparus, E. thamalakanensis, and E. greenwoodi as well as specimens that were morphologically identified as E. lineomaculatus were recovered within this group. Three OTUs were recovered for specimens that were morphologically identified as E. brevipinnis, whereas two OTUs were identified within E. thamalakanensis, and a similar pattern for E. greenwoodi. Interestingly the disjunctly distributed species E. niggie sp. nov. and E. neefi were found to be distantly related and deeply divergent (3.5%) from each other (Figure 1). 3.2 |Morphological data PCA performed on 15 morphometric characters of E. niggie sp. nov., and E. neefi exhibited a broad overlap between the two species, indicating that they could not be separated based on these characters (Figure 2). The first principal component axis (PCI) accounted for 24.3%, PC2 21.2%, and PC3 15.4% of the variation observed between the lineages (Table 2). The first PCA axis (PCI) was included as size differences were accounted for in the normalization procedure. PCI was mainly defined by differences in pectoralto pelvic-fin length. PCII contrasted differences in caudal peduncle length. PCIII highlighted differences in pelvicto anal-fin length. PCA performed on 10 meristic characters showed incomplete separation between the two species (Figure 3). PCI accounted for 49.84%, PCII 19.52%, and PCIII 12.69% of the variation between lineages. Factor loadings are presented in Table 2. PCI was defined by differences in the number of scales along the lateral line. PCII contrasted with differences in the number of pectoral-fin rays. PCII highlighted differences in the number of both total and caudal vertebrae. Despite the considerable overlap in morphometric and meristic characters between E. niggie sp. nov. and E. neefi, there are consistent qualitative color pattern differences between these two disjunctly distributed and genetically divergent species that warrant their separation as distinct taxonomic entities. These characteristics include the patterns of wavy parallel lines on the flank and the pattern of bold spots on the dorsal midline (Figure 4). All 16 examined specimens of E. neefi had the wavy parallel lines extending below the lateral line, and bold spots were present on the dorsal midline, whereas in the 15 specimens of E. niggie sp. nov. from the Limpopo, these are absent (Figure 4). For diagnosis of the new species and comparison with E. neefi, the meristic and morphometric measurements obtained in the present study were used. For comparisons with the other species of interest, information from the original descriptions and other key references was used (Martin & Chakona, 2019; Poll, 1967; Skelton, 2001). 3.3 |Taxonomic accounts Enteromius niggie sp. nov. [niggie: ‘nᶕᶍi] (g/ch from Afrikaans/Dutch) is pronounced with a hard guttural sound, made at the back of the throat. urn:lsid:zoobank.org:act:C24274E8-B384-409F-9A34-A821E9F6 9625. (Figures 5and 6; Table 3). 3.4 |Proposed common names Southern sidespot barb; Suidelike sykol ghieliemientjie (Afrikaans). 1140 SCHEEPERS ET AL. FISH 10958649, 2024, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jfb.15870 by Capes, Wiley Online Library on [29/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 3.5 |Holotype SAIAB 236359, field number: NPEJ21-B080, male, 41.10 mm SL, Steelpoort River, Limpopo River system, 24.80238, 30.11740, collected by Hoffman A., May 2021 (Figures 5b and 6b). 3.6 |Paratypes SAIAB 236360, nine unsexed, 33.35–42.21 mm SL, same collector and locality as holotype. SAIAB 186470, five unsexed, 35.76– 44.36 mm SL, Steelpoort River, Limpopo River system, 24.72907, 30.18468, collected by Palmer R., May 14, 2012. 3.7 |Additional material SAIAB 26620, 1 unsexed, 48.00 mm SL, Mohlapitse River, Limpopo River System, 24.04999, 30.03333, collected by J. Engelbrecht, April 1, 1986. SAIAB 27269, 1 unsexed, 39.15 mm SL, Letaba River, Limpopo River system, collected by D. Curle, September 22, 1986. SAIAB 49504, 6 unsexed, 35.40–45.10 mm SL, tributary of the Levuvhu River, Limpopo River system, 23.06667, 30.25000, collected by M. Angliss, April 6, 1995. SAIAB 61031, 20 unsexed, 20.00– 39.10 mm SL, Mutale River, Limpopo River system, 22.73444, 30.65861, collected by R. Bills, D. Naran, B. Van der Waal, November 13, 1999. SAIAB 61044, 1 unsexed, 28.10 mm SL, Mutale River, Limpopo River system, 22.70000, 30.64305, collected by R. Bills, FIGURE 1 Bayesian phylogenetic tree showing the phylogenetic relationships within the goldie barb group. The lineage is extracted from the broader analysis of available COI (cytochrome oxidase subunit I) sequences incorporating 95 species currently assigned to the genus Enteromius (Figure S1). Bayesian posterior probabilities are given on the branches as percentages; asterisk indicates PP >99. Bars show the OTUs (operational taxonomic units) identified by each of the four species delimitation methods. Red OTUs contain sequences of topotypic specimens. SCHEEPERS ET AL.1141 FISH 10958649, 2024, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jfb.15870 by Capes, Wiley Online Library on [29/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License D. Naran, B. Van der Waal, 13 November 1999. SAIAB 63443, 4 unsexed, 40.70–48.40 mm SL, Wyliespoort, tributary of Mutamba River, Limpopo River system, 22.91670, 29.92811, collected by J. Engelbrecht, September 20, 2000. SAIAB 190517, 44 unsexed, 27.50–41.2 mm SL, Orighstad River, Limpopo River system, 24.89027, 30.58833, collected by E. Swartz, L. da Costa, October 22, 2005. SAIAB 190556, 10 unsexed, 20.60–35.70 mm SL, Sterkspruit River, Limpopo River system, 25.156944, 30.558611, collected by E. Swartz, L. da Costa, October 23, 2005. SAIAB 203320, 1 unsexed, 31.00 mm SL, Thabina River, Limpopo River system, TABLE 1 Species names, reference numbers for BOLD/GenBank sequences, and locality details of sequences used for the present study. Species Sequence ID Catalogue number Country River System Latitude Longitude Enteromius atkinsoni MAFW071-11 SAIAB 119079 Malawi Ruo Zambezi 16.04 35.79 E. atkinsoni MAFW072-11 SAIAB 119079 Malawi Ruo Zambezi 16.04 35.79 E. atkinsoni ACAM028-13 SAIAB 190277 Mozambique Chipembe Dam Montepuez 13.20 38.62 E. atkinsoni MAFW114-11 SAIAB 118777 Malawi Shire Zambezi 15.06 35.22 E. atkinsoni MAFW096-11 SAIAB 185654 Malawi Phalombe Lake Chilwa 15.81 35.65 E. atkinsoni MAFW091-11 SAIAB 185654 Malawi Phalombe Lake Chilwa 15.81 35.65 E. atkinsoni GBMNB3511-20 NA Tanzania Great Ruaha Rufiji 07.63 36.89 Enteromius cf. viviparus MPUMA008-12 SAIAB 194032 South Africa Klein-Sand Incomati 24.66 31.09 Enteromius cf. viviparus SAFW912-14 SAIAB 081022 Mozambique Zambezi Zambezi 15.60 32.72 Enteromius cf. viviparus ACAM064-13 SAIAB 190265 Mozambique Tshidi Zambezi 15.62 33.67 Enteromius cf. viviparus ACAM090-13 SAIAB 190291 Mozambique Tshidi Zambezi 15.69 33.67 Enteromius cf. viviparus ACAM005-13 SAIAB 190235 Mozambique Muhukwa Montepuez 13.43 38.61 Enteromius cf. viviparus MAFW095-11 SAIAB 185652 Malawi Phalombe Lake Chilwa 15.81 35.65 Enteromius macinensis KP712064 NA Burkina Faso NA NA NA NA Enteromius brevipinnis MPUMA015-12 SAIAB 194050 South Africa Mac-Mac Incomati 25.02 31.00 E. brevipinnis OR763400 a SAIAB 206418 South Africa Grootspruit Limpopo 24.53 27.87 E. brevipinnis OR763404 a SAIAB 194786 South Africa Blyde Limpopo 24.90 30.75 Enteromius viviparus OR763397 a SAIAB 235471 South Africa Umdloti Umdloti 29.64 31.09 Enteromius neefi OR763394 a SAIAB 210153 Zambia Kabompo Zambezi 11.89 25.25 E. neefi OR763396 a SAIAB 210153 Zambia Kabompo Zambezi 11.89 25.25 E. neefi OR763395 a SAIAB 210153 Zambia Kabompo Zambezi 11.89 25.25 E. neefi SAFW842-13 SAIAB 082862 Democratic Republic of the Congo Kando Congo 10.80 25.98 Enteromius pallidus SAFW824-13 SAIAB 186173 South Africa Baakens Baakens 33.96 25.51 E. pallidus SAFW825-13 SAIAB 186173 South Africa Baakens Baakens 33.96 25.51 Enteromius lineomaculatus MAFW088-11 SAIAB 185653 Malawi Phalombe Lake Chilwa 15.81 35.64 E. lineomaculatus MAFW069-11 SAIAB 119069 Malawi Ruo Zambezi 16.10 35.69 E. lineomaculatus MAFW124-11 SAIAB N/A Malawi Nkatha Bay Lake Malawi 11.62 34.23 Enteromius thamalakanensis ANGFW230-12 SAIAB 187035 Namibia Okavango Okavango 18.12 21.58 E. thamalakanensis ANGFW160-12 SAIAB 186818 Angola Luassingua Okavango 14.59 18.17 E. thamalakanensis ANGFW195-12 SAIAB 186874 Angola Cuito Okavango 15.14 19.19 E. thamalakanensis OR763403 a SAIAB 202855 Botswana Okavango Okavango 19.21 22.75 Enteromius greenwoodi OR763410 a SAIAB 85020 Angola Cuanza Cuanza 12.03 17.63 E. greenwoodi SAFW283-08 SAIAB 84816 Angola Cuanza Cuanza 9.80 15.46 Enteromius sp. “Cubango”ANGFW046-12 SAIAB 186686 Angola Cubango Okavango 13.59 16.88 Enteromius sp. “Cuebe”ANGFW012-12 SAIAB 186640 Angola Cuebe Okavango 14.94 17.72 E.niggie sp. nov. OR763406 a SAIAB 236360 South Africa Steelpoort Limpopo 24.80 30.12 E. niggie sp. nov. OR763409 a SAIAB 236359 South Africa Steelpoort Limpopo 24.80 30.12 E. niggie sp. nov. OR763407 a SAIAB 236360 South Africa Steelpoort Limpopo 24.80 30.12 Abbreviation: BOLD, Barcode of Life Data; NA, not available. a Novel sequences generated from this study. 1142 SCHEEPERS ET AL. FISH 10958649, 2024, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jfb.15870 by Capes, Wiley Online Library on [29/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 24.02706, 30.18103, collected by A. Chakona, L. Dlamini, March 25, 2016. SAIAB 76348, 29 unsexed, 27.3–40.7 mm SL, Buffelskloofspruit, tributary of Crocodile River, Incomati River system, 25.40833, 30.46917, collected by R. Bills, J. Engelbrecht, May 19, 2005. SAIAB 27262, 2 unsexed, 33.00–40.12 mm SL, Mlumati River, Incomati River system, 25.79999, 31.10000, collected by D. Curle, July 29, 1986. SAIAB 70742, 3 unsexed, 34.70–57.10 mm SL, Houtbosloop, tributary of Crocodile River, Incomati River system, 25.42222, 30.74333, collected by R. Boycott, R. Bills, J. Engelbrecht, April 29, 2003. SAIAB 67330, 1 unsexed, 33.50 mm SL, tributary of Mkhondvo River, Maputo River system, 26.96861, 31.02139, collected by J. Msibi, M. Fakudze, R. Boycott, N. Khumalo, November 6, 2002. SAIAB 76071, 1 unsexed, 27 mm SL, Besterspruit, Mfolozi River system, 27.75056, 30.83222, collected by B. Grant, August 6, 2005. SAIAB 76077, 3 unsexed, 35.00–45.00 mm SL, tributary of Lenjane River, Mfolozi River system, 27.90083, 31.07472, collected by B. Grant, June 8, 2005. 3.8 |Diagnosis E. niggie sp. nov. belongs to the goldie barb group in southern Africa, which is characterized by species with a soft primary dorsal-fin ray, a relatively short compact body (<70 mm SL), the presence of two pairs of well-developed barbels, 24–30 lateral line scales, and a bright golden breeding colouration in males. Along with E. niggie the goldie barb group includes the species E. pallidus (Smith 1841), E. brevipinnis (Jubb 1966), E. neefi s.s. (Greenwood, 1962), E. thamalakanensis (Fowler 1935), E. greenwoodi (Poll, 1967), E. lineomaculatus “Malawi” (Boulenger 1903), and E. viviparus (Weber 1897). E. niggie and E. neefi can be readily distinguished from all the aforementioned species by the presence of distinctive pigmentation along the margins of flank scales that are expressed as wavy parallel lines (Figure 4a,b). Further, E. niggie can be distinguished from E. neefi by the lack of wavy parallel lines below the lateral line (Figure 4c) and by the lack of dark bold and rounded spots on the dorsal midline of the body (Figure 4d). FIGURE 2 Scatterplot of PC1 against PC2 for a PCA (principal component analysis) carried out on 16 normalized morphometric characters for 31 specimens of Enteromius niggie sp. nov. and Enteromius neefi. TABLE 2 Factor loadings for the first three principal component axes on 16 morphometric and 10 meristic characters from 31 specimens of Enteromius niggie sp. nov. and Enteromius neefi. Character PC1 PC2 PC3 Morphometrics 24.35% 21.16% 15.36% Head length 0.128 0.106 0.046 Body depth 0.108 0.546 0.146 Predorsal length 0.126 0.174 0.060 Dorsal-fin base 0.098 0.197 0.246 Pectoralto pelvic-fin length 0.838 0.320 0.237 Pelvicto anal-fin length 0.213 0.274 0.784 Anal-fin base 0.006 0.246 0.289 Caudal peduncle length 0.366 0.599 0.160 Caudal peduncle depth 0.056 0.051 0.002 Head depth 0.013 0.069 0.237 Snout length 0.050 0.050 0.026 Orbit diameter 0.026 0.002 0.080 Inter-orbit width 0.108 0.108 0.084 Post-orbit length 0.214 0.057 0.250 Anterior barbel length 0.128 0.106 0.046 Posterior barbel length 0.108 0.546 0.146 Meristics 49.84% 19.52% 12.69% Lateral line scales 0.816 0.505 0.034 Circumpeduncular scales 0.032 0.064 0.003 Predorsal scales 0.393 0.268 0.387 Pectoral-fin rays 0.326 0.756 0.134 Pelvic-fin rays 0.069 0.257 0.119 Total vertebrae 0.122 0.122 0.641 Predorsal vertebrae 0.174 0.047 0.319 Pre-caudal vertebrae 0.002 0.017 0.026 Pre-anal vertebrae 0.002 0.017 0.026 Caudal vertebrae 0.148 0.116 0.551 Note: The most important factor loadings are in bold font. SCHEEPERS ET AL.1143 FISH 10958649, 2024, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jfb.15870 by Capes, Wiley Online Library on [29/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License FIGURE 3 Scatterplot of PC1 against PC2 for a PCA (principal component analysis) carried out on 9 meristic characters for 31 specimens of E. niggie sp. nov. and Enteromius neefi. FIGURE 4 Lateral view of (a) Enteromius neefi (34.6 mm SL) and (b) Enteromius niggie sp. nov. (39. 7 mm SL) showing the presence and absence of the wavy parallel lines below the lateral line. Dorsal view of (c) E. neefi (34.1 mm SL) and (d) E. niggie sp. nov. (34.6 mm SL) showing the presence and absence of the dark rounded spots on the dorsal surface. FIGURE 5 Alcohol preserved colouration of Enteromius niggie sp. nov. (a) NPEJ21-B081 37.9 mm SL (SAIAB 236360) and (b) NPEJB080, 41.1 mm SL, holotype (SAIAB 236359). FIGURE 6 General body features and live colouration of Enteromius niggie sp. nov. (a) Male during breeding season, field ID NPEJ21-B081 37.9 mm SL (SAIAB360). (b) Male during non-breeding season, field ID NPEJ21-B080, 41.1 mm SL, holotype (SAIAB 236359). 1144 SCHEEPERS ET AL. FISH 10958649, 2024, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jfb.15870 by Capes, Wiley Online Library on [29/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License TABLE 3 Morphometric and meristic data for Enteromius niggie sp. nov. and E. neefi. Character E. niggie sp. nov. E. neefi Other specimensHolotype Paratypes Holotype Paratypes Number of specimens 1 14 1 8 7 Standard length (SL) (mm) 41.1 33.6–44.4 32.2 25.0–37.0 24.4–40.0 Head length (HL) (mm) 10.9 8.4–10.4 8.5 6.9–9.3 6.5–9.9 Caudal peduncle length (CPL) (mm) 9.6 8.5–12.6 7.6 5.7–8.6 6.0–9.5 Percentage of SL (%) HL 26.5 21.3–26.1 26.4 24.9–27.9 22.7–26.8 Predorsal length 52.4 48.9–53.8 49.7 51.5–54.8 49.9–53.3 Dorsal-fin base 15.4 12.0–17.2 15.2 10.8–14.5 11.9–15.6 Dorsal-fin height 25.2 22.4–28.1 NA 24.9–28.7 20.0–25.9 Pectoral-fin length 19.9 17.6–21.0 NA NA 14.8–20.1 Pelvic-fin length 19.0 17.0–19.5 NA NA 14.4–18.0 Pectoralto pelvic-fin length 24.9 18.7–28.3 23.6 18.8–24.1 21.0–26.7 Pelvicto anal-fin length 17.9 14.6–21.9 17.4 14.3–19.6 14.5–21.1 Anal-fin base 8.3 7.5–10.8 6.8 6.4–7.7 5.7–8.7 Anal-fin height 18.9 15.1–19.6 NA 16.2–22.0 14.1–18.4 Body depth 26.9 23.2–28.0 27.6 27.5–30.0 24.8–29.2 Body width 15.8 11.8–15.1 NA NA 12.9–17.1 CPL 23.4 22.1–26.7 23.6 21.9–25.0 20.0–25.7 Percentage of HL (%) Head depth 80.9 77.0–88.0 75.3 74.1–77.5 74.9–91.7 Orbit 32.1 30.8–37.7 36.5 34.4–38.4 29.8–40.8 Inter-orbit 25.7 22.8–34.3 31.8 29.5–35.6 21.2–30.9 Snout length 26.6 20.8–30.8 24.7 21.8–26.9 21.6–25.1 Post-orbit 43.3 30.5–45.5 45.9 42.7–47.4 37.8–46.5 Anterior barbell 23.8 19.8–36.4 31.8 20.3–37.1 12.0–34.2 Posterior barbell 40.0 38.1–51.1 38.8 37.1–49.4 14.9–58.7 Percentage of CPL (%) Caudal peduncle depth 54.3 42.4–66.7 61.8 52.3–64.1 51.6–63.3 Meristics Unbranched dorsal-fin rays iii iii iii iii iii Branched dorsal-fin rays 8 8 8 8 8 Unbranched anal-fin rays iii iii iii iii iii Branched anal-fin rays 5 5 5 5 5 Pectoral-fin rays 13 12 (8–14) NA NA 11 (10–11) Pelvic-fin rays 8 8 (8–9) NA NA 8 (8–9) Lateral line scales (LL) 30 27 (27–30) 27 27 (26–28) 26 (24–28) Scale rows between LL and dorsal fin 5 5 5 5 5 Scale rows between LL and pelvic fin 3 3 3 3 3 Scale rows between LL and anal fin 3 3 3 3 3 Circumpeduncular scales 12 12 11 12 (11–12) 12 (11–12) Predorsal scale rows 12 11 (11–12) 11 12 12 (11–12) Total vertebrae 32 33 (32–34) 32 32 32 Pre-caudal vertebrae 18 18 18 18 18 Caudal vertebrae 14 15 (14–16) 14 14 14 Predorsal vertebrae 9 9 (8–10) 9 9 (8–9) 8 (8–9) Pre-anal vertebrae 19 19 (19–20) 19 19 19 SCHEEPERS ET AL.1145 FISH 10958649, 2024, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jfb.15870 by Capes, Wiley Online Library on [29/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License