THREE NEW KILLIFISH SPECIES (CYPRINODONTIFORMES: NOTHOBRANCHIIDAE) FROM EQUATORIAL GUINEA

Introduction[Up]

Equatorial Guinea is a small country surrounded in the north by Cameroon and east and south by Gabon. Its fish fauna belongs to the Lower Guinea ichthyoprovince (‍Stiassny et al., 2007). Contrasting to the surrounding countries its fish fauna has been less studied. Pappenheim (‍1910) was the first to publish on a small collection made by Tessman at the end of the nineteenth century. The first ichthyological inventory of continental Equatorial Guinea was conducted by Roman (‍1971), followed by a comprehensive survey by Castelo (‍1994). Lasso et al. (‍1998) reported on the fish fauna of Monte Alén and listed 58 species in 36 genera and 16 families. In 2017 an expedition surveyed the freshwater fishes at 42 locations that resulted in the description of a new Chiloglanis species (‍Schmidt & Barrientos, 2019), two new Distichodus species (‍Schmidt et al., 2021), and several new characid and cyprinid species will be described in the near future (R. Schmidt, pers. comm., Sep. 2019). These recent studies indicate that the diversity offreshwater fish is only partially known.

From 2000 to 2018 eighteen collecting trips were organized by Spanish aquarium keepers which focused especially on Nothobranchiidae. Many collection sites were sampled as well in the coastal plain as in the interior highlands. This resulted currently in the description of four new endemic nothobranchiid species (‍Legros & Zentz, 2007; ‍Malumbres & Castelo, 2001; ‍Sonnenberg, 2008) from the coastal plain. There is a sharp boundary between the Lower Guinean Cyprinodontiformes fauna of the low altitude coastal plain and the high altitude inland hills situated on the Woleu-Ntem plateau (‍McKenzie et al., 2013). The collections of nothobranchiid fishes indicate the occurrence of different species groups in the inland, similar to the occurrence pattern in Cameroon and Gabon.

The Monte Alén National Park is situated in the south-west of the Woleu-Ntem inland plateau (Fig. 1). It has been a protected area since 1988 and contains a high biodiversity (‍De la Riva, 1994; ‍Lasso, 1995; ‍Lasso et al., 2002) probably due to its assumed history as a Pleistocene rain forest refugium (‍Maley, 1987, ‍1991). On the Gabonese side it extends further south in the Monts de Cristal. These mountains, which are actually hills (under 1000 m), are famous for their botanical species diversity and a postulated Pleistocene refugium (‍Leal, 2005). This area also harbors several endemic nothobranchiid killifishes (‍Eberl, 2016; ‍Huber, 2020; ‍Stiassnym et al., 2007). Three of them were recently described from the Monts de Cristal (‍Huber, 1977; ‍Sonnenberg & Blum, 2005; ‍Sonnenberg et al., 2006) of which two also occur in the adjacent part in Equatorial Guinea.

Several recent nothobranchiid collections made at the northern and southern fringes of the Monte Alén National park indicate that the inland area is inhabited by the same species groups or genera as the adjacent areas in Gabon and Cameroon. The majority of the collected species belong to the genus Mesoaphyosemion and to the ‘Aphyosemion’ herzogi group. Only a few records of Raddaella sp. aff. kunzi and Kathetys exiguum (Boulenger, 1911) are known from the north-east and in the south-east Diapteron cyanostictum (Lambert & Géry, 1968), Episemion callipteron Radda & Pürzl, 1987 and E. krystallinoron Sonnenberg, Blum & Misof, 2006, respectively, were collected. At present the genus Mesoaphyosemion contains the following described species: Mesoaphyosemion cameronense (Boulenger, 1903), M. obscurum (‍Ahl, 1924), M. amoenum (Radda & Pürzl, 1976), M. haasi (Radda & Pürzl, 1976), M. halleri (Radda & Pürzl, 1976), M. maculatum (Radda & Pürzl, 1977), M. mimbon (‍Huber, 1977), and M. etsamense (‍Sonnenberg & Blum, 2005). Besides these species at least nine different undescribed phenotypes were recognized (‍Amiet, 1987; ‍Dadaniak et al., 1995; Sonnenberg et al., submitted). The ‘Aphyosemion’ herzogi group consists of two described species: ‘A.’ herzogi Radda, 1975 and ‘A.’ bochtleri Radda, 1975. Both species were synonimized by some authors (‍Seegers, 1997; ‍Wildekamp et al., 1986; ‍Wildekamp, 1993). In addition many divergent phenotypes which might represent distinct species are known (‍Eberl, 2016). In the following we describe two endemic species of the genus Mesoaphyosemion and one endemic ‘Aphyosemion’ herzogi group species from the Monte Alénarea.

Material and methods[Up]

The description of the new species is based on specimens from the collections made by the first author and his co-travelers J.G. Poves and D. Gonzalez in 2018. The collections were licensed by Instituto Nacional de Desarrollo Forestal y Gestión del Sistema de Aéreas Protegidas (INDEFOR-AP), license number 281-‍018, issued 26 November 2018. Most specimens were preserved in 96% ethanol and some were preserved in 4% formalin first and transferred after one month into 70% ethanol.

DNA methods

A ca. 800 bp fragment of the mitochondrial cytochrome b was sequenced for 45 specimens of the ‘Aphyosemion’ herzogi species group and 64 specimens of Mesoaphyosemion, as outgroup for both datasets we used a sample of ‘A.’ coeleste Huber & Radda, 1977 which belongs to a different species group (‍Collier, 2006; ‍Murphy & Collier, 1999). Part of the sequences were previously published in Sonnenberg & Blum (‍2005; GenBank acc. Nos AY748279–AY748296), and Van der Zee et al. (‍2018; MF155816), the new sequences are deposited in GenBank under the accession numbers ON166122–ON166212. DNA samples are listed in Table 1. DNA extraction, PCR, and sequencing methods follow Sonnenberg & Blum (‍2005) and Sonnenberg et al. (‍2006, ‍2007), sequences of the new species and additional samples from Equatorial Guinea were prepared by Advanced Identification Methods – AIM GmbH using the published primer combinations (‍Sonnenberg & Blum, 2005). Sequences were aligned and checked by eye in BioEdit 7.0.5.3 (‍Hall, 1999), translated into corresponding amino acids to confirm for functionality and tested for the anti-G bias of mitochondrialsequences (‍Zhang & Hewitt, 1996) for mitochondrial origin. Uncorrected p-distances with pairwise exclusion of missing data were calculated in MEGA 7 (‍Kumar et al., 2016) and given in Supplementary files 1 and 2.

We performed Bayesian analyses (BA) for two parameter models (Nst = 2 and Nst = 6) for each species group with MrBayes 3.1.2 (‍Huelsenbeck & Ronquist, 2001) and compare the results between the simpler model (Nst = 2) and the more complex 6 parameter model GTR (Nst = 6). If both analyses gave compatible results we can assume not to introduce error due to over- or under parameterized models. For all analyses the settings were rate=gamma and ngammacat=4 as the suggested default value for MrBayes 3.1.2 (‍Huelsenbeck & Ronquist, 2001), number of generations were set to 10 000 000 with sampling of every 1000 generation. The results of the first ten percent of the sampling were discarded as burn-in time after confirming that the runs had reached the stationary phase by plotting the values of the p-files in Excel. All trees were rooted with ‘Aphyosemion’ coeleste as outgroup species. Only nodes with a posterior probability (PP) of 0.95 or higher in at least one analysis are considered as statistically supported.

Morphometrics and meristics

Morphometric measurements were taken with a digital caliper using a dissecting microscope, and rounded to the nearest 0.1 mm. Measurements are presented as percentages of standard length (SL). Counts and methods follow Amiet (‍1987), except for standard length (SL) that was measured from the tip of the snout and not from the tip of the lower jaw, eye diameter and interorbital width and scales that were not registered by Amiet. Eye diameter and inter orbital distance were measured at the orbital rim at the centre of the eye. Count of scales on the mid-longitudinal series is the number of scales between the upper attachment of the opercular membrane and the caudal fin base. Excluded are the scales posterior to the hypural junction. Transversal scales were counted in an oblique line from the first scale anterior to the dorsal fin base to ventral edge of the body. Circum caudal peduncle scales were counted around the body. All visible fin rays were counted. The relative position of first dorsal fin ray to anal fin (D/A) was estimated as in Sonnenberg & Schunke (‍2010). Nomenclature for the cephalic lateral line system follows Clausen (‍1967) and Van Bergeijk & Alexander (‍1962), and that for the supraorbital (frontal) squamation follows Hoedeman (‍1958).X-raysweretaken with a Faxitron LX-60 Digital Specimen Radiography System at the Zoologisches Forschungsmuseum Alexander Koenig, Bonn (ZFMK). Colour pattern elements of fins were measured on digital pictures of preserved specimens by means of MB-Ruler 5.3.3.7, an application to measure distances on a computer screen.

Taxonomic concept

In earlier papers the second and third authors (‍Sonnenberg, 2007; ‍Van der Zee & Sonnenberg, 2010) suggested to use Aphyosemion only for the monophyletic species group containing the generic type of the genus, A. castaneum Myers, 1924, i.e. the A. elegans species group. For the remaining species groups we recommended using published generic level names or, in the case of unnamed species groups, within parentheses as ‘Aphyosemion’, indicating the provisional generic status.

Naming of genus level groups is regulated by the ICZN rules and the descriptions are based on the diagnoses and can be validated, falsified or extended, however, for the usage of the taxonomic rank (genus/subgenus) there are no kind of regulations or technical advice.

Our arguments are, that the generic level of species groups enhances taxonomic stability and carries more information about the named group than an all including genus Aphyosemion s.l. with 100+ species.

Currently published molecular phylogenies do not fully resolve the relationships between the different species groups with statistically significant support values (‍Agnèse et al., 2006, ‍2009; ‍Collier, 2006; ‍Murphy & Collier, 1999), however, species groups are always recovered as monophyletic units with good support. Actually, morphology based studies of nothobranchiid phylogeny lack resolution due to absence of useful characters (‍Aarn & Shepherd, 2001; ‍Costa, 2009, ‍2015; ‍Huber, 2013). Independent of their phylogenetic relationships, species groups are currently the most stable units in nothobranchiid taxonomy. They carry more detailed information about the species group, e.g. like distribution, morphology, behavior, life history, and even in changing phylogenetic hypotheses, names at the species group level have no need to change. In addition, their diagnosis is usually better defined than for a taxon Aphyosemion s.l., with the latter most difficult to delineate to allow inclusion of the different species groups, which makes it in turn more complicated to separate it from related nothobranchiid genera like e.g. Fundulopanchax, Fenerbahce, Foerschichthys or Fundulosoma with a convincingargumentation.

The new nothobranchiid species of the ‘A.’ herzogi species group is here described within Aphyosemion to follow ICZN Article 5.3, but will appear as ‘Aphyosemion’ mitelemense sp. nov., in latter paragraphs indicating that it is not placed within Aphyosemion s.s.

We adopted the pragmatic approach by Moritz et al. (‍2000) as a species concept.

Comparative material examined

Mesoaphyosemion cameronense, Syntypes, BMNH 1902.11.12.148–149 (2), Cameroon, Kribi River, 1903.7.28.119–122 (4); 1903.7.28.123 (1) Kribi River, 1903.7.28.211 (1) Ja River. — MRAC 93-‍108-0472–491, Cameroon, Ntem River basin, small river at 4 km from Akoabass in the direction of Bitche, A. Kamdem Toham 29 June 1993; Cameroon, 42 km west of Ebolowa along N17, W. Herzog, H-J. Joachim & R. Roth, November 1992, field code HJRK92-11 (offspring, aquarium material, JZ collection).

Mesoaphyosemion etsamense, Holotype, ZFMK 39832, male, 32,4 mm SL; Gabon, western slopes of the Monts de Cristal, a small river at the village Etsam I, crossing the road N5 from Medouneu to Kougouleu, T. Blum, G. Fleck & R. Sonnenberg, 29 July 2002, field code G02/160. — Paratypes, MRAC A4-42-P-1-4, same data as holotype; Paratypes, ZFMK 39843, male, Gabon, north of the village Assok, T. Blum and P. Sewer, 11 September 2002, field code BS 02/13.

Mesoaphyosemion halleri, Equatorial Guinea, Ntem River basin, 5 km west of Ebebeyin, F. Malumbres, F. García Lora, J. Blanco Barles & M. Juhl 13 April 2003, field code GEMLBJ03-12 (offspring, aquarium material, RS collection). The types of M. halleri couldn’t be studied since they are not present in NMW (pers. comm. A. Palandacic, NMW, Nov. 2019).

Mesoaphyosemion maculatum, Holotype, NMW 90757. Gabon, rainforest creek at the Koumameyong-Ovan National highway No.4, 33 kilometers east Koumameyong, 20 kilometers west Ovan, Radda & Pürzl, 1977; Gabon, Ivindo River basin, Mvoung River, Mva creek, near the village of Ebé, O. Legros, W. Eberl and A. Cerfontaine, 8 January 1993, field code LEC93-4 (offspring, aquarium material, JZ collection).

Mesoaphyosemion mimbon, Equatorial Guinea, small tributary of Rio Ncomo, between Bas and Oveng, north-east of Acurenam, F. Malumbres, F. García Lora, J. Blanco Barles & M. Juhl, 16 April 2003, field code GEMLBJ03-20 (offspring, aquarium material RS collection). — Gabon, small tributary of Komo River, 6 km west of Medouneu along the N5, T. Blum, G. Fleck and R. Sonnenberg, 29 July 2002, field code G02/156 (RS collection).

‘Aphyosemion’ bochtleri, Holotype, NMW 77332, Gabon, Mvoung River basin, Brook in rainforest about 300 meters on right side of a village near Mintoum, Herzog et al. 1972. —; Gabon, Mvoung River basin, Mintoum, P. Wagner and R. Wendel, January 1986, field code GWW86-11 (JZ collection).

‘Aphyosemion’ herzogi, Holotype, NMW 77327, Gabon, fast running brook about 3 kilometers north of Zoumoukou or 16 kilometers north of Lalara on the road to Mitzic and Oyem, Herzog et al. 1972. —; Gabon, Anveme River basin, 26 km from Zomoko to Oveng, offspring of fish collected in 1993, G. Passaro and W. Eberl, field code PEG93-15 (JZ collection).

‘Aphyosemion’ sp. aff. herzogi “Cameroon”, Cameroon, Ntem River basin, Mvila River, Nsessoum, P. Venstermans, J. Aerts and M. Bogaerts, 8 July 2016, field code VAB16-22 (JZ collection). — Cameroon, Ntem River basin, Kyé River, Mefoub, P. Venstermans, J. Aerts and M. Bogaerts, 16 July 2018, field code VAB18-52 (JZ collection).

‘Aphyosemion’ sp. aff. herzogi “S.E. Eq. Guinea”, Equatorial Guinea, Komo River basin, Mbé River drainage, 3km south of Ngolensok, offspring of fish collected in 2000, F. Malumbres, J. Sanjuán and G-J. van Huijgevoort, field code GEMSH00-10 (DNA specimen, JZ collection). — Equatorial Guinea, Komo River basin, unnamed tributary, 1 km east of Acurenam, A. González, C. Vizcaíno, F. Portal and H. Ott, 21 August 2018, field code GEGVPO18-21 (DNA specimen).

‘Aphyosemion’ sp. aff. herzogi “Nviayong”, Equatorial Guinea, Mitemele River basin, Rio Mven, near the village of Nviayong, A. González, C. Vizcaíno, F. Portal and H. Ott, 19 August 2018, field code GEGVPO18-12 (DNA specimen).

‘Aphyosemion’ sp. aff. herzogi “Mitemele right bank “, Equatorial Guinea, Mitemele River basin, unnamed right bank tributary, A. González, C. Vizcaíno, F. Portal and H. Ott, 20 August 2018, field code GEGVPO18-16 (DNA specimen). — Equatorial Guinea, Mitemele River basin, unnamed tributary of Rio Midjobo, A. González, C. Vizcaíno, F. Portal and H. Ott, 20 August 2018, field code GEGVPO18-17 (DNA specimen).

Results[Up]

DNA analysis. Mesoaphyosemion

The complete sequence alignment has a length of 778 bp for 64 Mesoaphyosemion specimens from 58 populations and ‘Aphyosemion’ coeleste as outgroup. The alignment including the outgroup has 264 variable positions of which 215 are parsimony informative. Within Mesoaphyosemion 248 positions are variable, 208 of them parsimony informative. The observed uncorrected pairwise distances between specimens are listed in Supplementary file 1. The values between the outgroup species and Mesoaphyosemion range from 11.83%–14.40%, the maximum observed value within Mesoaphyosemion is 12.08%, the average uncorrected sequence divergence is 7.25%. The base composition shows with A = 27.8%, C = 24.8%, G = 13.3%, and T = 34.1% the typical A/T bias of mitochondrial DNA sequences (‍Zhang & Hewitt, 1996). The sequences translate into 258 amino acids of which 37 are variable and 14 parsimony informative. Mesoaphyosemion losantosi sp. nov., and M. montealenense sp. nov., are nested within an only partially resolved mtDNA phylogeny of the genus. Both are not closely related to each other or to samples identified as M. maculatum, M. cf. maculatum, or M. mimbon. Mesoaphyosemion maculatum and M. cf. maculatum occur in close proximity in central Gabonbutare genetically not closely related. Mesoaphyosemion montealenense sp. nov., forms a well supported clade with several other samples identified as M. cameronense (Fig. 2). Observed pairwise sequence divergence within M. losantosi sp. nov., is 0.26%, within M. montealenense sp. nov., 0.13%–0.39%, between M. losantosi sp. nov., and M. montealenense sp. nov., is 6.68%–7.07%, between M. losantosi sp. nov. and M. maculatum 5.40%–6.30%, between M. montealenense sp. nov. and M. maculatum 8.10%–8.87%, and between the two new species and all other samples the divergence range between 4.37%–10.67%.

DNA Analysis. ‘Aphyosemion’ herzogi group

The full sequence alignment of the ‘Aphyosemion’ herzogi group dataset has a length of 778 bp for 45 specimens from 38 populations of the ingroup and ‘A.’ coeleste as outgroup species. The alignment including the outgroup has 252 variable positions of which 188 are parsimony informative, excluding the outgroup there are 217 variable and 181 parsimony informative positions. The uncorrected pairwise distances between all included specimens are listed in Supplementary file 2. The values between the outgroup species and the ‘A.’ herzogi group range from 16.71%–18.25%, the maximum observed value in the ingroup is 12.72%, the average uncorrected sequence divergence for the whole dataset is 7.39%. The base composition shows the typical A/T bias of mitochondrial sequences (‍Zhang & Hewitt, 1996) with A = 27.0%, C = 23.6%, G = 13.7%, and T = 35.7%. The sequences translate into 258 amino acids of which 32 are variable and 18 parsimony informative.

Within the resulting trees, not all nodes are resolved with significant statistical posterior probability value, however, ‘Aphyosemion’ mitemelense sp. nov. turns out as sister species to all remaining samples/species of the ‘A.’ herzogi group (Fig. 3). Pairwise distances range between 0.00%–5.01% within ‘A.’ mitemelense sp. nov., and from 10.28%–12.72% between this species and all other samples/species of the group.

The DNA sample from a population just outside the Mitemele basin (MIT002) is placed basal to all remaining population samples located inside the basin, but the specimens are by morphology and colour pattern indistinguishable from the populations inside the basin and observed distance values are 4.37%–5.01% between MIT002 and the remaining samples and 0.00%–4.11% within the latter. Inside the basin two clades are present (Fig. 1): MIT004, MIT006, MIT021, and MIT022 (the latter two from the type locality) occur in the southern part of the distribution area and MIT001, MIT009, MIT015, and MIT986 in the northern part. Peculiar is that MIT 009 resembles MIT021 and MIT 022 (type locality) in colour pattern but is genetically not close. MIT001, MIT015 and MIT986 resemble MIT004 and MIT006, but are also not genetically close. Genetics reflects geographical origin (outside the basin, northern part or southern part within the basin) but not colour pattern differences.

Mesoaphyosemion losantosi sp. nov.

urn:lsid:zoobank.org:act:6D227B78-38D2-4208-99F0-9E2455D2287B

Figs 4–5, Tables 2–3

Aphyosemion maculatum – ‍American Killifish Association, no date.

Aphyosemion maculatum – Castelo, 1994. — Huber, 2007.

Type material

Holotype. MRAC 2019.008.P.0001, male, 31.0 mm SL, Equatorial Guinea, Centro Sur province, affluent of Rio Nfogoyi, a left bank tributary of Rio Wele, 100 meter west of the village of Bisun, 1º43´43.5´´N, 10º17´29.4´´E. Collected on 5 December 2018 by F. Malumbres, J. G. Povez, and D. González, Field code GEMPG18-20.

Paratypes. MRAC 2019.008.P.0002–0005, 2 males and 2 females, 30.7–36.5 mm SL, collected with the holotype. — MNCN_ICTIO 296.758, 1 male, 22.2 mm SL, MNCN_ICTIO 296.759, 1 female, 21.8 mm SL, collected with the holotype. — ZSM 47573, 1 male and 2 females, 27.0–37.9 mm SL, collected with the holotype. — AMNH 275003, 1 male and 2 females, 29.6–34.3 mm SL, collected with the holotype.

Not measured type material. MRAC 2019.008.P.0006–00011, 6 juvenile specimens.

Diagnosis

Mesoaphyosemion losantosi (Figs 4–5) belongs, according to DNA results, morphology and colouration characters to the genus Mesoaphyosemion. It shares with other Mesoaphyosemion species the combination of the following characters: a mostly metallic light blue, blue or blue green colouration on side of males; an often variable pattern of red dots on sides, in most species dots are irregularly distributed, in others forming more or less regular stripes or larger blotches; unpaired fins without or only short fin extensions, caudal-fin in males nearly straight, slightly rounded or weakly trilobate. No conspicuous markings on ventral side of the head; all adult males have a yellow snout. None of these characters alone is diagnostic and they are also known from other nothobranchiid genera or species groups, but in combination they are diagnostic for the genus. In most Mesoaphyosemion species a broad latero-ventral red line on the caudal peduncle is present, but this line is absent in all species with large dark red blotches on the flanks: M. mimbon, M. cf. maculatum, M. maculatum, M. losantosi, and M. montealenense sp. nov. M. losantosi is distinguished from all other Mesoaphyosemion species by having 7transversal scales (vs. 8). In M. montealenense sp. nov., the anal fin has a light blue background withamore or less checkerboard pattern of square and rectangular blotches. In other species the background colour is darker and the less distinct submarginal band often fuses with the background colour and/or the blotches in the anal fin, or the anal fin is complete dark red or yellow.

Description

See Figs 4 and 5 for general appearance and colour pattern, see Table 2 for morphometric and Table 3 for meristic data of types. Small to medium sized member of Mesoaphyosemion with strong sexual dimorphism, males being colourful and females light brownish. Dorsal profile nearly straight, slightly convex from nape to dorsal fin base. Ventral profile gently convex from head to end of anal fin base. Caudal peduncle slightly concave from end of both anal and dorsal fin base to base of caudal fin. Snout slightly pointed, mouth directed upwards; lower jaw longer than upper; posterior end of mouth at approximately same level as anterior edge of eye. On both jaws unicuspid, slightly posteriorly directed teeth, some larger teeth on outer row. Open frontal cephalic sensory system with two pits separated by a low ridge, each neuromast accompanied by 2 very small lobes. Preopercular system tubular with 6 pores, supra-orbital system open groove with 3 neuromasts on both sides, pre- and post-orbital system short tube with 2 large pores.

Dorsal fin in males rounded, anal fin pointed in older males. Fins in females smaller and anal fin rounded. Dorsal and anal fins located posterior to mid-body. Dorsal fin with 11–12 rays; anal fin with 14–16 rays, first ray of dorsal fin above anal fin ray 8–9. Caudal fin slightly spade shaped. Pectoral fin reaching origin of pelvic fin in males, not in females; pelvic fin reaching origin of anal fin in males, not in females. Frontal squamation of G-type, with small elongated G scale. Scales on mid-longitudinal series 30–32 + 3–4 on caudal fin base. Transverse rows of scales above pelvic fin 7; circumpeduncular scale row 12.

Live colouration

Males (Fig. 4a). Snout yellow, dorsum brownish with copper shining. Flank metallic light blue with green shining. Operculum with three dark red stripes, top stripe very short and sometimes hardly visible, middle stripe longest and most prominent. In some specimens a narrow suborbital dark red stripe. Anterior flank with two short rows of small dark red spots, followed posterior flank with very large irregular dark red to purple brown blotches, forming up to three vertical or oblique bars on caudal peduncle in most specimens. Pectoral fins transparent with a white margin, pelvic fins white to light blue with narrow white margin and wider dark red submarginal band. Dorsal and anal fins light blue or whitish with white narrow margins. Dorsal fin with a variable number of dark red interradial spots and with some larger dark red blotches at its base. Anal fin with dark red submarginal band and a variable number of elongated, sometimes round, dark red interradial blotches. Caudal fin grayish to light blue with a variable number of dark red spots and blotches with white dorsal and ventral margins, ventral margin wide with dark red submarginal band, dorsal margin narrow.

Females (Fig. 4b). Dorsum and flanks brownish, dorsally darker, ventrally light grey or brownish. Flanks with blueish shining. Scales on dorsum and sides with irregular narrow dark edges, forming a reticulate pattern. All fins hyaline; unpaired fins with many red spots, often arranged in rows. All fins, except caudal fin with narrow white edges.

Colour after 3 months in 96% ethanol

Males (Fig. 5a). Dorsally and laterally dark brown to black, ventrally lighter brown. Vague dark bars on caudal peduncle. Flank scales with dark margins, resulting in a reticulated pattern. All fins dark with some lighter dots. Head dark. Females (Fig. 5b). Dorsally and laterally brown, ventrally lighter brown to yellow. Flank scales with posterior dark edges with larger dark centre. Lower jaw brown, operculum brown. Unpaired and pelvic fins light grey, with dark rounded to oval spots, arranged in irregular rows, pectoral fin hyaline.

Colour after 1 month in formalin and then transferred to 70% ethanol

Males. Dorsum grey. Flanks light grey. Ventrum white. Flank and ventrum scales with narrow dark margin. Unpaired fins dark grey with vague dark red markings. Pelvic fins black. Pectoral fins light grey with wide dark grey margin. Irregular dark red bars on posterior flanks and caudal peduncle. Head dark grey, ventrally white. Lower lip black. Females. Dorsum and lateral flanks light grey, ventral flank light brownish. All scales with dark margins. Small dark red spots on posterior flanks and caudal peduncle. Unpaired fins grey with rounded to oval dark grey spots. Pelvic fins dark grey. Pectoral fins light grey. Head light grey, ventrally white. Lower lip dark grey.

Distribution and habitat

The new species is currently known from a stretch of 6 kilometers north and south Bisun along the road from Niefang to Bicurga in central Equatorial Guinea (Fig. 1). It is not unlikely, however, that it also occurs east and west of the present distribution area, since no collections were made there. North and south of the distribution area several collections were made, but M. losantosi was not found. The type locality is a small creek in humid forest, slowly running, slightly amber coloured clear water, close to source, water 15 cm deep and 1 m wide, loamy soil covered with leaves, pH 6.0, conductivity 0.5 Microsiemens, water temperature 23.9 ºC.

Etymology

The name losantosi is in honour of Félix Losantos (1967–2020), a very good friend and co-traveler of the first author on collection trips to Equatorial Guinea. Félix was a tireless, passionate and helpful person.

Mesoaphyosemion montealenense sp. nov.

urn:lsid:zoobank.org:act:779DC250-46E5-418D-81FA-EBB558E96C19

Figs 6–7, Tables 4–5

Aphyosemion maculatum – ‍American Killifish Association, no date.

Aphyosemion maculatum – Huber, 2007.

Aphyosemion aff. maculatum – ‍Ott, 2019: 41.

Type material

Holotype. MRAC 2019.008.P.0012, male, 37.3 mm SL, Equatorial Guinea, Centro Sur province, stream Biñogo affluent of Río Laňa, a left bank tributary of Río Wele, just west of the village of Moka, 1º39´05.5´´N, 10º19´10.4´´E, coordinates were taken in the village of Moka, not on the type locality itself, collected on 5 December 2018 by F. Malumbres, J. G. Povez and D. González, Field code GEMPG18-19.

Paratypes. MRAC 2019.008.P.0013–0016, 2 males and 2 females, 35.1–19.5 mm SL. — MNCN_ICTIO 296.760, male, 31.0 mm SL. — MNCN_ICTIO 296.761, female, 30.4 mm SL. All collected with the holotype.

Diagnosis

Mesoaphyosemion montealenense (Figs 6–7) belongs to the genus Mesoaphyosemion based on characters mentioned in the diagnosis of M. losantosi. It differs from all other species of Mesoaphyosemion (except for M. losantosi, M. maculatum, M. cf. maculatum, and M. mimbon), by having large dark blotches on flanks and lacking a broad latero-ventral red line on the caudal peduncle. It differs from all other species with large blotches on flanks by having a checkerboard pattern of elongated and square spots in unpaired fins on a light blue background. In some specimens spots merged, forming inter-radial stripes.

Description

See Figs 6 and 7 for general appearance and colour pattern, see Table 4 for morphometric and Table 5 for meristic data. Medium sized Mesoaphyosemion with strong sexual dimorphism, males being colourful and females dull. Dorsal profile nearly straight, slightly convex from nape to dorsal fin base. Ventral profile gently convex from head to end of anal fin base. Caudal peduncle slightly concave from end of both anal and dorsal fin base to base of caudal fin. Snout slightly pointed, mouth directed upwards; lower jaw longer than upper; posterior end of mouth at approximately same level as anterior edge of eye. On both jaws unicuspid, slightly posteriorly directed teeth, some larger teeth on outer row. Open frontal cephalic sensory system with two separate pits, each neuromast accompanied by 2 small lobes, preopercular system tubular with 6 large oval pores, supra-orbital system open groove with 3 neuromasts on both sides, 3 low distal lobes and 2 low proximal lobes, pre- and post-orbital system short tube with 2 large pores.

Dorsal fin and anal fin pointed in older males. Fins in females smaller and never pointed. Dorsal and anal fins located posterior to mid-body. Dorsal fin with 11–13 rays; anal fin with 14–16 rays, first ray of dorsal fin above anal fin ray 6–8. Caudal fin spade shaped in older males with 25–28 fin rays. Pectoral fin reaching origin of pelvic fin in males, not in females; pelvic fin reaching origin of anal fin in males, not in females. Frontal squamation of G-type; scales on mid-longitudinal series 30–33 + 3–4 on caudal fin base. Transverse rows of scales above pelvic fin 8; circumpeduncular scale row 12.

Live colouration

Males (Fig. 6a). Snout and lower jaw yellow. Dorsum brownish with copper shining. Flank metallic light blue with green shining. Operculum with three dark red stripes, top stripe very short and sometimes hardly visible, the other two stripes of equal length. In some specimens a narrow suborbital dark red stripe. Anterior flank with no or only a few small dark red spots, posterior flank with large irregular dark red to purple brown blotches forming three to nine vertical or oblique bars on caudal peduncle. In some specimens these blotches form a marbled pattern whereas in other specimens these blotches are reduced. Pectoral fins transparent with a narrow white margin, pelvic fins with narrow white margin and dark red submarginal band, provided with two or three dark red blotches. Dorsal and anal fins light blue or whitish with only a very narrow white margin in dorsal fin, no white margin in anal fin. Dorsal fin with a variable pattern of dark red spots, in some specimens up to 40 smaller rounded spots all over the fin, in others small spots at the base of the fin and interradial stripes distally. Anal fin also variable; in some specimens a checkerboard pattern of more or less square dark red spots at the base and stripes distally, in others only alternating stripes. Caudal fin with asymmetric colour pattern. No or very narrow white dorsal margin in most specimens, in some specimens this margin is much wider. Ventral white marginalways wider than dorsal margin.Centralfin with 15 to 40 small dark red spots or short stripes.

Females (Fig. 6b). Dorsum and flanks brownish, dorsally darker, ventrally light grey or brownish. Scales on dorsum and sides with irregular narrow dark margins, forming a reticulate pattern. All fins hyaline; unpaired and pelvic fins with many red spots, often arranged in rows. All fins, except posterior caudal fin with narrow white margins.

Colour after 3 months in 96% ethanol

Males (Fig. 7a). Dorsum completely black. Flank scales with many small melanophores and black edges. Black bars on caudal peduncle. Operculum scales also with many small black spots, but without black margins. Throat and ventrum light brown, scales with black edges. Unpaired fins as in live specimens, but light blue colour now dark grey and dark red colour black. Pelvic fins black, pectoral fins translucent dark grey. Females (Fig. 7b). Dorsum dark brown, flanks and ventrum light brown. All dorsal and flank scales with prominent black margins, resulting in a reticulated pattern. Ventral side not reticulated. Unpaired fins greyish with dark spots. Pectoral and pelvic fins dark grey hyaline.

Distribution and habitat

The new species is restricted to a small area just southeast of Monte Alén in central Equatorial Guinea (Fig. 1). All collections were made 2–3 kilometers southeast of this mountain along the road from Niefang to Bicurga. It is not unlikely, however, that it also occurs east and west of the present distribution area, since no collections were made there. North and south of the distribution area was frequently sampled and no further occurrence of M. montealenense was found. Habitat: slowly running shallow clear water; pH 6.5; conductivity 3 Microsiemens, water temperature 22.5ºC.

Etymology

The specific epithet is derived from Monte Alén, the highest mountain in Equatorial Guinea (1250 m). The new species occurs at the southeastern foot of this mountain at an altitude of ca. 690 meter.

Aphyosemion mitemelense sp. nov.

urn:lsid:zoobank.org:act:B1C83E77-5B71-4373-9C83-DA769AE6A3A4

Figs 8–10, Tables 6–7

Aphyosemion aff. herzogi – ‍Ott, 2019: 43–45.

Type material

Holotype. MRAC 2019.008.P.0017, male 34.9 mm SL, Equatorial Guinea, Province Litoral, Mitimele River basin, Nki-Besua creek, 1º06´00.6´´N, 10º08´25.9´´E, collected on 4 December 2018 by F. Malumbres, J. G. Povez & D. González, Field code GEMPG18-14.

Paratypes. MRAC 2019.008.P.018–0021, 3 males and 1 female, 23.8–30.2 mm SL. — MNCN _ICTIO 296.762, male, 35.8 mm SL. — MNCN_ICTIO 296.763, female, 24.4mm SL. — ZSM 47593, 1 male 24.9 mm SL and 1 female, 23.0 mm SL. — AMNH 275004, 1 male, 30.2 mm SL and 1 female, 28.4 mm SL. All collected with the holotype.

Additional material. Aphyosemion mitemelense, Equatorial Guinea, Komo River basin, Rio Mbé, Yama, F. Malumbres, J. Sanjuán & G-J. van Huijgevoort, 1 July 2000, field code GEMHS00-10 (JZ collection). — Equatorial Guinea, Wele River basin, Rio Chiguo drainage, A. González, C. Vizcaíno, F. Portal & H. Ott, 18 August 2018, field code GEGVPO18-8 (RS collection). — Equatorial Guinea, Mitemele River basin, unnamed tributary between Rio Nkiém and Rio Mba, Esakora (Acurnam), A. González, C. Vizcaíno, F. Portal & H. Ott, 19 August 2018, field code GEGVPO18-9 (JZ collection). — Equatorial Guinea, Mitemele River basin, Rio Nkién, Ayem II, A. González, C. Vizcaíno, F. Portal & H. Ott, 19 August 2018, field code GEGVPO18-10 (JZ collection). — Equatorial Guinea, Mitemele River basin, unnamed left bank tributary, A. González, C. Vizcaíno, F. Portal & H. Ott, 19 August 2018, field code GEGVPO18-15 (RS collection). — Equatorial Guinea, Mitemele River basin, Rio Be drainage, A. González, C. Vizcaíno, F. Portal & H. Ott, 22 August 2018, field code GEGVPO18-25 (DNA specimen).

Diagnosis

Aphyosemion mitemelense belongs to the ‘Aphyosemion’ herzogi group, based on the oval shaped caudal fin, broad flat head, low number of anal fin rays and females that are less dull coloured and show often parts of male colour patterns than in other Aphyosemion s.l. It can be distinguished from all other members of the species group by a unique combination of colour pattern characters.

Aphyosemion mitemelense can be distinguished from both described species of the ‘A.’ herzogi species group by the male colour pattern of body and unpaired fins. In A. mitemelense the caudal fin has several more or less horizontal relatively wide dark red bands on centre and lower part of fin vs. whole fin with many red interradial lines on blue centre and lower part of fin in ‘A.’ bochtleri and two red lines edging a broad yellow band on fin centre in ‘A.’ herzogi. Background of anal fin green to blue green in A. mitemelense vs. yellow to orange, at least basally, in the other two species.

Aphyosemion mitemelense can be distinguished from neighboring undescribed species from the ‘A.’ herzogi group by having a greenish appearance and a green anal fin with irregular red blotches and/or stripes and a large dark red spot at the base of pectoral fin vs. a reddish appearance and an orange anal fin with regular interradial lines and lacking a red spot at the base of pectoral fin in populations on the right bank of the Mitemele River in southern Equatorial Guinea; by lacking a yellow horizontal band at the base of anal fin vs. having such a band in a population that occur sympatric near the village of Mveayong in the extreme southeast of Equatorial Guinea. Aphyosemion mitemelense differs from populations between Ngolensok and Acurenam in southeastern Equatorial Guinea by absence of more or less regular red interradial lines in caudal fin centre, lower half of caudal greenish versus yellow and upper part of fin with many irregular red spots versus most part of upper half of caudal fin with red interradial lines; it differs also by the presence of broad red blotches, stripes or bands on greenish or blue-greenish anal fin versus more narrow, short red interradial lines on yellow or orange anal fin.

Description

See Figs 8 and 9 for general appearance and colour pattern, see Table 6 for morphometric and Table 7 for meristic data. Medium sized Aphyosemion s.l. with strong sexual dimorphism, males being colourful and females rather dull but showing faint traces of male colour pattern. Dorsal profile convex from nape to halfway dorsal fin base. Ventral profile straight convex from operculum to end of anal fin base. Caudal peduncle straight from end of both anal and dorsal fin base to base of caudal fin. Snout slightly pointed, mouth directed upwards; lower jaw longer than upper; posterior end of mouth at approximately same level as anterior edge of eye. On both jaws unicuspid, slightly posteriorly directed teeth, some larger teeth on outer row. Open frontal cephalic sensory system with two separate pits accompanies by 2 small lobes, large distance between pits (more than twice the length of a pit). Preopercular system tubular with 6 small pores, anterior most pore reduced and sometimes hardly visible. Supra-orbital system open groove with 3 neuromasts on both sides, with 2 proximal lobes and 3 very small distal lobes, pre- and post-orbital system short tube with 2 large pores.

All fins rounded. Dorsal and anal fins located posterior to mid-body. Dorsal fin with 9–11 rays; anal fin with 12–14 rays, first ray of dorsal fin above anal fin ray 6–7. Pectoral fin reaching origin of pelvic fin in males, not in females; pelvic fin reaching origin of anal fin in males, not in females. Frontal squamation of G-type, with medium sized round to rather square G scale, not much smaller than E-scales. Scales on mid-longitudinal series 29–31. Transverse rows of scales above pelvic fin 8; circumpeduncular scale row 12.

Live colouration

Males (Fig. 8a). Top of head and dorsum brownish with dark green shining. Flank metallic green. Caudal peduncle with yellow to golden shine except western-most populations with blueish green caudal peduncle. Operculum with three dark red stripes, top stripe very short and sometimes hardly visible, lower stripe longest and most prominent and posteriorly often connected to the red spot in pectoral fin (Fig. 9a–c). This spot varies from small and rounded to large and triangle shaped. The lower opercular stripe anteriorly forming a suborbital stripe. Flank with 3 (very rarely 4) longitudinal lines with red spots. Top line reaching caudal fin base, other lines getting shorter from dorsum to ventrum. Caudal peduncle in western-most populations with broad longitudinal ventral red line, in other populations with irregular narrow red bars or blotches and often yellow to orange ventral edge. Pectoral fins transparent grayish to yellow with a dark red spot at the base, pelvic fins yellow to green with some larger red spots and/or streaks. Dorsal and anal fins green, sometimes with a narrow white margin. Dorsal fin with many dark red spots. Anal fin with dark red submarginal band and narrow blue or whitish edge and variable median band of elongated blotches, sometimes forming a band. Dorsal part of caudal fin with same colour pattern as dorsal fin, separated from rest of this fin by a straight narrowred line. Mid section of caudal fin light green to yellow-green with 2 or 3redlines, mid section separated from ventral edge of fin by a broader red line. Ventral margin of caudal fin blue. Pelvic fin very variable, with or without red spots and/or stripes, generally green, but also populations with red and yellow are known. Pectoral fin hyaline, yellow or green with diagnostic red marking at base.

Females (Fig. 8b). Relatively colourful compared to other Aphyosemion s.l. females as all females of the ‘Aphyosemion’ herzogi group. Dorsum and flanks brownish, dorsally darker, ventrally light grey or whitish. Sides with 3 or 4 relatively regular rows of small red spots. Unpaired fins light grey and yellowish with narrow light grey edge. Anal fin and caudal fin with red inter-radial red stripes at the base (not well visible in Fig. 8). Dorsal fin and dorsal part of caudal fin with many red spots. Pectoral fin hyaline often with orange or yellow spot at the base.

Colour after 3 months in 96% ethanol

Males (Fig. 10a). Dorsum black, flanks dark grey, flanks and caudal peduncle ventrally light brown, ventrum light brown. Flanks scales with dark grey posterior margin. Third, fourth and fifth scale row with rounded red spots. Third row complete red, fourth row to onset dorsal fin, fifth row shorter and sometimes incomplete. Ventral scale row caudal peduncle with irregular dark red band. All fins black, red pigmentation hardly visible. Head dark grey, ventrally light brown. Space between pectoral fin and operculum marked light brown. Lower lip anteriorly black.

Females (Fig. 10b). Dorsum and dorsal part of flanks dark grey. Ventrum, ventral part of flanks and caudal peduncle light brown. Three to four irregular rows of red spots on anterior flank, some posterior flank scales with narrow crescent shaped red markings. Unpaired fins grey. Dorsal fin with many red spots, anal and caudal fin with red spots and/or interradial red stripes. Pelvic and pectoral fin hyaline. Space between pectoral fin and operculum marked light brown. Lower lip anteriorly black.

Colour after 1 month in formalin and then transferred to 70% ethanol

Males. Dorsum grey-brown, flanks and caudal peduncle ventrally light grey. Flank scales without dark posterior margin. Third, fourth, and fifth scale row with rounded red spots. Third row complete red, fourth row to onset dorsal fin, fifth row shorter and sometimes incomplete. Ventral scale row caudal peduncle with irregular red band. All fins dark grey to black. Anal with dark red median band. Caudal fin ventrally with red interradial stripes, dorsally with red spots. Dorsal fin with elongated red spots. Pectoral fin with dark red spot at base. Head grey. Space between pectoral fin and operculum white. Lower lip anteriorly black.

Females. As preserved directly in 96% ethanol.

Distribution and habitat

‘Aphyosemion’ mitemelense was thought to be restricted to small left bank tributaries of the southeastern Mitemele River basin in Equatorial Guinea (Fig. 1). One population, however, is known from just outside the basin 8 km southeast of Evinayong and recently the species was collected in north-west Gabon, where it was found in a small creek 13 km north of the village of Song in the Noya River basin (personal communication L. Chirio, August 2020). This species is restricted to a small area at the southwestern range of the species group. The currently described species are found at the southern (‘A.’ herzogi within the Okano River drainage) or southeastern (‘A.’ bochtleri, Mvoung River drainage) range of the species groups distribution area. Water conditions at the type locality (altitude 354 meter): pH 6.6; conductivity 11 µS; water temperature 24ºC. Sympatric occurring Cyprinodontiformes: Mesoaphyosemion mimbon and Plataplochilus sp. aff. pulcher.

Etymology

The specific epithet is derived from the Mitemele River (Equatorial Guinea), to which its distribution is restricted, except for one population occurring in adjacent northern Gabon.

Discussion[Up]

The freshwater fish fauna of Equatorial Guinea is still incompletely known as shown by the descriptions of killifish species in the last years (‍Legros & Zentz, 2007; ‍Malumbres & Castelo, 2001; ‍Sonnenberg, 2008; this paper) and the results of recent surveys which still find new reports or undescribed species (‍Roman, 1971; ‍Castelo, 1994; ‍Lasso et al., 1998; ‍Schmidt & Barrientos, 2019; R. Schmidt, pers. comm., Sep. 2019). The recently described killifish species are often endemic to Equatorial Guinea or shared with usually only one of the neighboring countries (‍Malumbres & Castelo, 2001; ‍Legros & Zentz, 2007; ‍Sonnenberg, 2008; this paper). All are usually known from a relative small distribution area.

Mesoaphyosemion is a large and complex group of species living in small rainforest streams in southern Cameroon, northern Gabon, eastern Equatorial Guinea, and northern Republic of Congo (‍Amiet, 1987; ‍Wildekamp, 1993; ‍Huber, 2007; Stiassny et al., 2007). Currently eight species are accepted as valid, but actually nine more or less distinct male colour pattern phenotypes are recognized. In combination with the here published genetic data this indicates the occurrence of a large number undescribed species. One obstacle is the repeated recurrence of certain male colour pattern characters within the genus, like the marbled pattern of M. maculatum, M. losantosi, M. montealenense, and M. mimbon, or a yellow caudal peduncle in M. amoenum, M. halleri, and M. sp. aff. cameronense phenotypes 3, 4, 5, 9, and 10 (see also Sonnenberg et al., submitted). However, the genetic data do not support a single, monophyletic origin of these characters, but a repeated independent occurrence in different genetic lineages (Fig. 2). Despite their similarity, M. losantosi, M. montealenense, M. maculatum, andM. mimbon do not represent a single species, nor a distinct lineagewithin Mesoaphyosemion. The genetic data also indicate a quite high genetic divergence between them of more than 5% observed pairwise distance, which is in the range found in other nothobranchiid species groups for cytochrome b or COI mtDNA datasets between different species. Similar colour pattern characters are not uncommon in unrelated nothobranchiid species, e.g. a pattern of narrow bars on the caudal peduncle in different genera (Aphyosemion s.s., ‘A.’ ogoense species group, Diapteron, Fundulopanchax, Kathetys, Scheelsemion). Huber (‍1980) coined the term “frontier species” for similarities in colour pattern of neighboring Aphyosemion s.l. populations. Probably it will be possible in future to use Quantative Trait Locus (QTL) mapping and next generation sequencing to find out if these similar colour pattern characters in different species are based on the same genes and ‘switched’ on and off due to small mutations independently.

‘Aphyosemion’ herzogi and ‘A.’ bochtleri were the first described species of the ‘A.’ herzogi group. Although both species differ considerably in colour pattern, several authors considered ‘A.’ bochtleri as a subspecies of ‘A.’ herzogi, since populations showing sometimes mixed combinations of colour patterns as in the above species where found (‍Seegers, 1997; ‍Wildekamp, 1993; ‍Wildekamp et al., 1986). These discussions were summarized by Eberl (‍2016). The phylogenetic tree in this paper (Fig. 3) shows that similar to the results of Mesoaphyosemion the ‘A.’ herzogi group most probably consists of a potentially larger number of distinct species, which is in agreement with the different male colour phenotypes found (‍Eberl, 2016; ‍Seegers, 1997; own observations).

‘Aphyosemion’ mitemelense is genetically quite distinct from the remaining samples of the group. Within this species, the DNA sample from the collection site GEGVPO18-8 (DNA sample MIT002), which is outside the Mitemele basin, is genetically quite divergent from the other samples of the species. It differs up to 4.76% from populations inside the basin which is considerably high. Since it cannot be distinguished from the nearest population inside the basin, based on external morphology and colour pattern, we consider it yet as conspecific with ‘A.’ mitemelense. A similar case is e.g. Chromaphyosemion riggenbachi (‍Ahl, 1924) in Cameroon, which contains several distinct lineages with quite high genetic distances between them (haplotype groups I-V with uncorrected mean distances between 3.99%-7.24%) (‍Völker et al., 2006), which are all very similar by male colour pattern, but have quite different chromosome numbers.

Inside the Mitemele basin two colour patterns occur in ‘Aphyosemion’ mitemelense (Fig. 9). The majority of known populations have a colour pattern as seen in Fig.9b and 9c: a barred caudal peduncle with a more or less yellow to golden colour. Two populations at the extreme south-west of the distribution area lack the bars and have a blue-greenish caudal peduncle (Fig. 9a): the type locality (DNA samples MIT020 and 021) and a population just 3 km north-east of it (DNA sample MIT009). Yet they are not closely related with a genetic distance of almost 4%. In the other (golden and barred) populations there is also a genetic gap, although they all look very similar. The two southern-most populations (DNA samples 004 and 006) group with the type specimens (MIT020 and MIT021) and the more northern populations group with the MIT009 population (Fig.1). Maybe the species developed the two different phenotypes in allopatric populations, that came together in secondary contact and between the phenotypes some exchange of genetic material took place. Actually the mitochondrial genetic data and the male colour pattern of the two different phenotypes do not show a congruent pattern. Probably population genetic data sets based on highly variable nuclear marker might be able to determine if mitochondrialintrogression has occurred across the two ‘A.’ mitemelense phenotypes.

Small scale species distribution in nothobranchiids is quite common. This is usually also reflected in genetic data. Even within species quite some genetic divergence can occur between populations (‍Völker et al., 2006; own unpublished data). Possible causes for low genetic exchange between populations might be low dispersal capabilities, local adaptations (‍McKenzie et al., 2013), and strong mate choice preference (see references in ‍Sonnenberg, 2007).

Different views exist about the relationships of the ‘A.’ herzogi species group. Amiet (‍1987) assumes a closer relationship to Scheelsemion franzwerneri (Scheel, 1971), based on: a) less sexual colour dimorphism as in other Aphyosemion s.l., b) a racket shaped caudal, c) dark post-cephalic mark, d) few dorsal and anal fin rays, dorsal inserted above 6-‍7 anal fin ray, e) dark crossbars as chevrons or Xs in stressed specimens. Both species share similar habitat preferences (wells, shallow water in rain forest), behavior (mainly bottom orientated, less free swimming) and very faint similarity in colour pattern (‍Amiet, 1987; ‍Huber, 2013, own observations, pers. comm. T. Blum). However, these characters are not unique to these species and genetic data include S. franzwerneri into the ‘A.’ calliurum species group (‍Agnèse et al., 2009; ‍Collier, 2006). Recently, Huber (‍2013: fig. 2) presented a morphological study and places the ‘A.’ herzogi group in a clade with ‘A.’ tirbaki, which in turn is the sister group to the ‘A.’ calliurum group (in a basal position), Chromaphyosemion, Diapteron, and Kathetys [((‘A.’ tirbaki, ‘A.’ herzogi) (Scheelsemion, (Chromaphyosemion, (Diapteron, Kathetys))))] This is not supported by the phylogenetic tree based on mitochondrial DNA sequences as published in Collier (‍2006: fig. 5), in which the nodes of ‘A.’ bochtleri, A.’ herzogi, and ‘A.’ tirbaki are printed close to the samples of Scheelsemion, but a relationship of these species groups has no statistical significant support. Huber’s (‍2013) inclusion of the ‘A.’ herzogi species group and ‘A.’ tirbaki into a subgenus Scheelsemion in his sense makes it a para- or probably polyphyletic taxon.

Collier (‍2006) coined the term ‘orphan’ taxa for those species and species groups that do not show any closer relationship to any other group. Interestingly even the phylogenetic tree published by Huber (‍2013: fig. 2), based on morphological characters does not allow the inclusion of ‘A.’ tirbaki and the ‘A.’ herzogi group into a monophyletic Scheelsemion (see also above). A preliminary analysis of Huber’s (‍2013) published dataset indicates that of the 91 morphological characters only 50 are informative. A simple Maximum Parsimony analysis of the dataset results in 10 equally short trees, the 50% consensus tree of a bootstrap analyses with 1000 replicates did not show any significant support for a taxon Scheelsemion as proposed in Huber (2013). Of all presented characters only one is informative for a grouping of the ‘A.’ calliurum group with ‘A.’ herzogi and ‘A.’ tirbaki (character 65, isolated yellow spot surrounded by red, postopercular, above pectoral fin insertion), two additional (characters 31 and 35, dorsal longitudinal fin shape and anal posterior fin shape in females) are also found either in Chromaphyosemion or Diapteron and one (character 3, Body-outline (longitudinal) inmale:higher at about ventral fin level) supports a grouping of ‘A.’ tirbaki, ‘A.’ herzogi, Scheelsemion, Chromaphyosemion, Diapteron, and Kathetys (‍Huber, 2013). At the moment it seems that character 65, the occurrence of a yellow spot is not diagnostic, because this is not found in several populations and species of the ‘A.’ calliurum group and in ‘A.’ tirbaki, but present in other species, as e.g. ‘A.’ coeleste or ‘A.’ wachtersi, the same might be the case for character 3, but needs further study, as characters 31 and 35.

Huber (‍2013) diagnoses Scheelsemion (including the ‘A.’ herzogi group and ‘A.’ tirbaki) with several characters, of which some are only diagnostic for the ‘A.’ calliurum species group, which includes the type species of Scheelsemion, like a dark blotch under the chin (see Fig. 11) and long extensions of unpaired fins in males or only diagnostic for the ‘A.’ herzogi group like red lower lip and the three-part colour pattern of the caudal fin. Some characters are not really useful as diagnostic characters (e.g. incomplete rows of red dots or blotches, fin extensions of unpaired fins in males, red lower lip) as they occur in several other species groups, which can’t be distinguished by the given characters. The problem with the diagnosis of Scheelsemion is, that it has to include three different species groups (‘A.’ calliurum, ‘A.’ herzogi, ‘A.’ tirbaki) that do not share some unique diagnostic features and, by DNA studies, are most probably not closely related.

Currently, all published molecular genetic studies of the relationships within Aphyosemion s.l. that include members of both species groups do not support a closer relationship of the ‘A.’ herzogi species group with or a placement within Scheelsemion (‍Agnèse et al., 2009; ‍Collier, 2006). If Scheelsemion is restricted to the ‘A.’ calliurum group, this is supported by the available DNA studies (‍Agnèse et al., 2009; ‍Collier, 2006) and it is diagnosed by the dark blotch on chin and a unique pattern of the frontal neuromast system in all species. In turn, the ‘A.’ herzogi complex seems to be a separate species groups that needs further studies.