Jean H. HUBER*



I. Further investigations on the systematic status of Panchax bualanus Ahl, Panchax escherichi Ahl, Haplochilus striatus Boulenger and Aphyosemion striatum microphtalmum Lambert & Géry.

Few cases of discussions have taken place, since Scheel's remarkable pioneering works (1968 et subseq.), about the systematics and the validity/synonymy status of old Rivulin taxa from the tropical west African forests genus Aphyosemion. However Scheel was unable to study the types of most taxa in the various institutions, including those from the Berlin Museum in the former East Germany. His conclusions are well known regarding Ahl's descriptions: most of Ahl's taxa were considered as synonyms of older taxa described by others and even by Ahl himself, e.g.:

1-  Panchax elberti Ahl, 1924 is a junior synonym of P. bualanus Ahl, 1924; the former taxon being named in a later issue of the same publication (Zool. Anz.) than the latter.

2-  Panchax escherichi Ahl, 1924 (in the same article as elberti) is a junior synonym of Haplochilus striatus Boulenger, 1911.

Radda who did study at the Berliner Museum on several occasions did not question Scheel's conclusions; we did not either, especially since we had the opportunity to survey "probably" paratypic material of P. escherichi from Z.M.B. (Berliner Zool. Museum) in the mid eighties for the latter case and because of distributional reasons for the former case.

The situation remained stable until Seegers's publications of 1986 and 1988: Seegers was, with the help and dedication of the present Z.M.B. curator Hans-Joachim Päpke, the first to stay long periods in Z.M.B. and find out most of Ahl's type series, study them and publish major shifts in their systematic status. Seegers's major contribution was well accepted because of his reputation of seriousness, his record of publications, especially regarding the east African genus Nothobranchius and because the proposed shifts did not hurt the then accepted validity of the studied taxa… except in 3 cases: the two stated above, plus the complicated problem involving fallax/ walkeri/ deltaense/ kribianum/schwoiseri/gulare, which we will not cover because Aphyosemion fallax Ahl, 1935 has been named from an aquarium population with a questionable origin and the types are said to be lost (nomen vanum).

It was not until recently that we were able to find the time to ask on loan Ahl's types through the courtesy of our old friendly connaissance H.-J. Päpke and, to avoid any confusion, we requested the holotype (or the lectotype) which is the single voucher of a species. The purpose of these notes is to report on the study of these two types and of related material for comparison.

1- Panchax bualanus Ahl (holotype and single type): ZMB 21947, "Buala, Kamerun, Elbert leg."

The subadult specimen is in a desperate condition: the fish is squashed laterally, which makes it higher than in reality, most fins are cut near their base, except the Pectorals (which position excludes a relationship with the Aplocheilichthyins), no trace of pigments (especially not a single bar on sides) can be found and many scales are lost, including in the lower post-orbital region; there is a lateral, somewhat darker, band on the lower mid sides from the Pectoral root to the level of half of the Anal base, but this may be seen as the standard lower median band, present in the Kathetys subgenus (and also other subgenera) in formalin preserved or frightened specimens; neither the scalation type, nor the predorsal scale count can be disclosed; due to the lateral squash, many scales are lost on sides: the lateral series count could only be hypothesized as 29 (+2 on peduncle), whereas the transversal series count could be evaluated as 8. Measurements: S.L.: 25.2 mm, head length: 6.8 mm (27% in S.L.), predorsal length: 18.2 mm (72% in S.L.), preanal length: 16.3 mm (65% in S.L.); all these data look reasonable for a non annual Aphyosemion i.e. for biogeographical reasons, belonging to the subgenera Aphyosemion, Mesoaphyosemion, Kathetys.

The radiophotograph was more informative: D= 10; A= probably 15; D/A= +8; vertebrae= 11+16; the skull is of the genus Aphyosemion type and the hypural plate is very little divided. The above mentioned data leave little probability of it belonging to another genus than Aphyosemion, i.e. either Epiplatys or Nothobranchius s.l.; the longer Dorsal base and the somewhat lower D/A deviation would favor for the subgenus Kathetys or Mesoaphyosemion more than Aphyosemion s.s.; for comparison, two series of specimens belonging to the Paris M.N.H.N. collections have been radiophotographed with the following results:

* MNHN 1978-387: Aphyosemion bualanum sensu Scheel; 3 specimens: D= 11, 10, 9; A= 16, 15, 16; D/A= +7, +8, +9; vertebrae= 11+16; 11+16; 12+15.

* MNHN 1981-1368: A. exiguum; 2 specimens: D= 9, 10; A= 15, 15; D/A= +7, +7; vertebrae= 12+15; 12+15.

All 5 specimens, with a very little divided plate, like Ahl's type.

In conclusion Ahl's type cannot be distinguished easily from the Cameroonian bualanum sensu Scheel: many aspects of morphology, meristics, osteology are similar; however it cannot be ruled out that Ahl's type belong to Mesoaphyosemion: the two known Mesoaphyosemion species from the neighboring regions are A. cameronense (D= 12; A= 16; D/A= +6) and A. wildekampi (D= 11; A= 16; D/A= +9), but both do not display the vertically barred color pattern disclosed by Ahl for his bualanus, and also present in the Cameroonian bualanum sensu Scheel; it cannot be ruled out too that following Seegers, bualanum is a fasciated, unknown live, Mesoaphyosemion species, which would be isolated in the highlands with distinctive morphomeristics from its neighboring counterpart, either cameronense or more probably wildekampi; it cannot be ruled out finally that Ahl's type is a distinctive Kathetys species related to the Cameroonian component, but we know of no geographical barrier in the area and bualanum sensu Scheel is widely distributed on the Plateau.

Therefore we feel the probability that Scheel's view is correct is high, but not 100% and both names for the Cameroonian populations are then possible: conservatively as bualanum (like in Killi-Data 1996 and according to the ethics of I.C.Z.N.) or following Seegers as elberti, until live topotypes from the mountainous village of Buala, alt. 1200m. (now Bouala: 6.42 N; 15.58 E) presently in the Central African Republic are collected by keen enthusiastic Killi-hobbyists or ichthyologists.

2- Panchax escherichi Ahl (lectotype), ZMB 21936; "Attogondema, Kamerun, Escherich leg."

The fully adult male specimen is in a rather satisfactory condition: few scales are missing, all fins are entire except the upper half of Caudal and the upper part of Dorsal; some pigments are indeed available, except the preopercular markings and the Caudal pattern. On the sides, there are 2-3 lateral rows of discontinuous, rather large light (i.e. red) spots, up to the level of the Dorsal insertion (the following ones being faded out) and in-between them, 3-4 dark contrasted lines, up to the Dorsal ending level (just like in Roman's photos -1971: fig. 66- of freshly preserved A. striatum from southern Ecuatorial Guinea); the largest number of red spots in the striation for a line was 16, the lowest was 3; a dark reticulation may be seen on the upper posterior sides; a typical marginal dark band is seen on Anal, not on Dorsal (because of the cut, at least at its beginning), both fins being pointed, the former having at its tip a streak like a flame and a faded mid-series of dots; similarly, because of the cut, no median band on Dorsal can be seen; equally important is the dark brown colors following the rays at the Dorsal base level and on the lower part of Caudal (red band?).

The morpho-meristic data are, after confirmation by a radiophotography for some of them:

frontal scalation= G-type; total length= 43.2 mm (124% in S.L.); S.L.= 34.9 mm; predorsal length= 23.4 mm (67% in S.L.); body height at Anal level= 6.0 mm (17% in S.L.); at peduncle level= 4.0 mm (11% in S.L.); head= 9.5 mm (27% in S.L.); interorbitar= 4.8 mm (14% in S.L.); eye diameter= 2.9 mm (8% in S.L.); scales, LL= 29 (+1 on peduncle), predorsal= 19, transversal at Anal= 7; fins, D= 11; A= 15; D/A= +6; vertebrae= 12+17; (relatively!) well divided hypural plate.

The key issue on the escherichi's case is not at the genus level (this is clearly an Aphyosemion species, belonging to the subgenus Mesoaphyosemion and to the striatum superspecies), but at the species level: is escherichi identical to striatum or to microphtalmum, both species being roughly isomorphs, having the same distribution North of the Ogooué River, being occasionally sympatric, having a roughly similar color pattern?

For comparison 2 series of specimens from the Paris MNHN, belonging to both species have been radiophotographed and studied with the following results (both series with a well divided hypural plate):

· MNHN 1981-873, A. striatum, 8 specimens: D= 11, 11, 10, 11, 11, 10, 11, 11 (mean 10.75); A= 14, 14, 14, 15, 16, 15, 16, 15 (mean 14.88); D/A= +6, +6, +6, +6, +6, +5, +7, +6 (mean +6.0); vertebrae= 12+16, 12+15, 12+15, 12+15, 12+15, 12+16, 11+18, 12+15

(mean 27.5).

· MNHN 1978-152, A. microphtalmum from N. Gabon (ex. simulans), 3 specimens: D= 11, 10, 10 (mean 10.33); A= 16, 16, 17 (mean 16.33); D/A= +7, +8, +8 (mean 7.67); vertebrae= 11+17, 11+17, 10+17 (mean 27.7).

Except for the Anal ray count and the D/A deviation of P. escherichi lectotype which are closer to striatum than to microphtalmum, two major factors, no statistical difference could be found.

The tendency of these results to be nearer to striatum have been confirmed on a larger number of specimens, the 10 remaining paralectotypes, requested on loan a few months later:

All paralectotypes are of lower interest than lectotype (Seegers's choice was then appropriate): most patterns are wiped out, except some faded brown striations on sides of three of them; many fins are cut and all, without markings; the overall morphology lies within Mesoaphyosemion (G-type frontal scalation, open neuromast pattern of head, average morphology); the radiophotographies were more informative: D= 11, 11, 11, 11, 11, 12, 11, 11, 11, 12 (mean with lectotype: 11.18); A= 14, 14, 14, 13, 14, 14, 14, 14, 14, 13 (mean with lectotype: 13.91); D/A= +6, +6, +6, +5, +6, +6, +6, +6, +5, +6 (mean with lectotype: 5.82); vertebrae: 13+16, 13+17, 12+17, 13+17, 13+17, 14+17, 13+17, 13+16, 13+16, 13+16 (mean with lectotype: 29.55); the meristic counts fit strongly with striatum and the abdominal vertebrae counts correspond better to striatum.

There are anyhow some other points in favor of the resemblance of males escherichi and striatum:

*           both have pointed vertical fins (rounded in microphtalmum);

*           both have the dark (red) streak at Anal tip;

*           both have the dark (red) margin at Anal (thin light edge in microphtalmum, dark submargin);

*           both have large red spots on sides (rather smaller in microphtalmum);

*           both have the red lines along the rays at Dorsal and Anal bases;

*           both are slenderer (microphtalmum being somewhat more robust).

Moreover, Ahl, in his description, mentions the strongly reflecting blue anterior sides and the conspicuous carmine markings: this is more so of striatum than of microphtalmum in life.

However this flow of converging characters towards striatum does not make a 100% proof, especially when for some of them Seegers's observations are opposite; and until live topotypes are made available by collectors (Attogondema, 0.85 N; 10.25 E, is now in N.W. Gabon at the foot of the Monts de Cristal) or until the lectotype may be genetically studied (actually recent techniques make now possible to undertake caryotype and DNA finger prints of preserved-in-alcohol-only specimens, even very old; however, the lectotype of escherichi has stayed in formaline for a while and thus we shall have to wait the improvement of the technique), both options are respectable: striatum and microphtalmum as valid with escherichi as a junior synonym of striatum, or, striatum and escherichi as valid with microphtalmum as a synonym of escherichi; but, alas, this is not even sure that live topotypes (both species may live sympatrically there, as it is often the case near the Monts) or caryotyping (both species are close: striatum, n= 20; A= 33; microphtalmum, n= 19-20; A= 35-38) rules out one of these two options; DNA fingerprinting techniques are more promising since they have demonstrated recently their capability of separating two heteromorphic species with similar caryotypes.

For the time being, like for bualanum, and because of the converging flow of evidence and due to our own experience of collecting both species in the Monts de Cristal foothills, we will stick personally to the conservative widely acknowledged option of keeping valid the well described striatum and microphtalmum, still waiting for new data or new techniques to strengthen or obviously to change our mind.

Seegers's works are nonetheless very valuable and full of merits: they pushed us to reconsider facts that may have been too quickly frozen; at that time, radiophotography techniques were not as currently used as to-day.

In the near future, hopefully DNA techniques will be made available to establish the identity (and the ancestry!) card, even for old type specimens and reveal their true identity (a further example would be the not 100% settled case of A. congicum and A. melanopteron); unfortunately this will not be possible with species for which types do not exist or were lost like in the A. fallax case (mentioned above) or the not 100% settled case of A. oeseri versus A. santaisabellae, or the undesired case of A. trilineatum, or the questionable case of A. tessmanni versus bualanum/elberti; in addition debates have arisen because some of the addressed taxa (fallax, oeseri, trilineatum) plus the ambiguous case of A. roloffi (versus etzeli) were based on confusing or vague type localities and on patterns figured out by hand drawings or poor black and white photos not surely identical with the so-called description: it is necessary to be cautious in drawing absolute conclusions from these patterns which are very standard, i.e. banal in Aphyosemion: as an example of difficulty it is appropriate to mention that recently a new phenotype of the Fundulopanchax series has been discovered in Ecuatorial Guinea which shows remarkable similarities but not identity with deltaense or fallax sensu Scheel (1990) but not Seegers (1988)… then who is who?

Let's wish and hope that luck will allow to find preserved old aquarium specimens still hidden in Museums that fit with these taxa and to apply rational powerful techniques.



II. Proposal of allocation of precise type localities for key taxa, still missing one.

Because of the frequent emergence of cryptic species in Cyprinodonts, the precise labeling of the type locality, enhanced by its geographical coordinates, has become overwhelmingly important and in fact a key issue in Systematics: as a prerequisite it is critical to fix these data for the named taxa of a given superspecies before starting a study of newer components.

Most type localities coordinates of Old World Cyprinodonts are now available, either because they were given by the describer or because they were easily deduced from the labeling of the type locality or else because they were inferred from the study of the bibliography or internal Museums reports. Wildekamp, Romand & Scheel working together have brought a large share of these data for C.L.O.F.F.A. in 1986 (by the way let's point out that the coordinates of Proc. loemensis are erroneous: in fact they plot in Western Zaïre and not Southwestern Congo: the confusion comes from the fact that there is no Lombo river in Congo, but one in Zaïre; actually, the collector was a civil administrator based in Loango, the most important place in the Loémé area at that time and not Pointe Noire; thus he collected in the northern tributaries of the Loémé river and one of them in the surroundings is named to-day Koulombo which crosses the old way from Loango to Brazza, at approximately 4.63 S; 11.98 E, which should be the proper coordinates!) and they must be thanked for this hard and obscure job! Anecdotically, and to illustrate the difficulty, they appeared in disagreement for one case (Scheel, 1990), i.e. the type locality of A. decorsei Pellegrin, 1904: it was only owing to luck and to the experience of two senior ichthyologists at the Paris Museum that the real type locality was found plotted on a map in a French geographical bulletin published by Chevalier in 1904, distinctively distant from their two proposals (Huber, 1994).

It is with this example in mind that we propose the allocation of precise geographical coordinates for some remaining taxa, after detailed studies; they have been already published in Killi-Data 1996 without comments (Huber, 1996): with them there are still remaining 3 imprecise type localities for old world Cyprinodonts:

*           Pachypanchax nuchimaculatus (Guichenot, 1866) from "Madagascar" for which a preserved type exists (MNHN -2937, 162-2-38-1); some hope has been raised recently by Loiselle et al. (De Rham, pers. comm.) due to their discovery of another Pachypanchax species in Southern Madagascar, which might be linked with Goudot's neck-marked fish (he actually stayed long periods of his chaotic life in Southern Madagascar, whereas the two hitherto known species -omalonotus and sakaramyi- are from the North). I have not studied sakaramyi, but I have compared four syntypes of omalonotus and the single type of nuchimaculatus: these two lots cannot be assigned to the same species, as it has been done by Scheel (1968) and subsequent authors; nuchimaculatus (S.L.= 52.3 mm; T.L.= 64.5 mm, 123% of S.L., predorsal length= 75 % of S.L.) is slenderer, with D= 11, A= 15, D/A= +9, vertebrae= 14+18; the hypural plate is well divided and large, much more than omalonotus (not divided at all, like in playfairii); although the specimen is in poor condition, it seems that the frontal scalation and neuromasts patterns are also different (small upper scales; pores); I take this opportunity to hereby designate a lectotype for omalonotus (S.L.= 47.8 mm; T.L.= 54.7 mm, 114%; predorsal length= 74%; LL= 29+6, TRAV.= 12), in good condition but with a slightly cut Caudal; the other fifteen specimens being paralectotypes with the new number MNHN 1997-655; the four studied types have the following counts on radiophotographies: D= 12, 12, 13, 12; A= 19, 19, 18, 19; D/A= +11, +10, +10, +10; vertebrae: 14+17, 14+16, 15+16, 13+17; pending further collections and field observations, nuchimaculatus should be considered as distinctive from omalonotus; it should be compared with Pachy. sakaramyi (Holly, 1928), described later, for which proper meristic data are lacking. Guichenot's description is clear and precise, although his meristics are underestimated as usual. Not surprisingly from above, Guichenot compares his nuchimaculata to spilargyreia and not to omalonota, the other Old World Poecilia sp. known at that time. Finally, let's add that Scheel (1990) mentions that Myers has reported that nuchimaculatus is a brackish species: no data in M.N.H.N. can confirm this hypothesis, yet.

*           Aphyosemion spoorenbergi Berkenkamp, 1976 from "between Calabar and Mamfé, in Biafra, Southeastern Nigeria", in fact from an aquarium trade import: this case is dependent on aquarists's expeditions in the area to re-discover the fish and bring back data on the gardneri superspecies; understandably, these expeditions have been rare due to the political situation, but it is only a question of time.

*           Epiplatys bifasciatus and Ep. marnoi (Steindachner, 1881) from "Bahr-el-Seraf, Nile basin, central Soudan", approx. 9 N; 30-31 E: this case is somewhat similar to that of Rivulus micropus (Steindachner, 1863) for which the type locality is "Rio Negro", a large river 2200 km long (Huber, 1992), but this is simpler -the above quoted river, a by-pass of the White Nile, is only about 250km long!- although hazardous since it has become mainly dry due to Sahel extension, but also marshy; the fact that a Sahelian population of one of these species (Ep. bifasciatus?) has been disclosed at Malakal (central Soudan: 9.52 N; 31.67 E) by Guma'a (1982) to enjoy internal fertilization may be a spur for killi-hobbyists to undertake collecting trips along this river and thus allow to propose an actual type locality… and, why not, find a Nothobranchius species and Micropanchax loati in addition (note that in 1919, Werner has reported also as viviparous Ep. marnoi from Gonga and Reik, two localities of central Soudan which are not mentioned in the present Gazetteer).


In Killi-Data 1996, the following proposals have been brought forward ; they are detailed below:

*           Micropanchax pfaffi (Daget, 1954) from several localities in Mali: this case has been fairly easy; on our request, the describer decided to choose "Diafarabé, upper Niger river, S.W. Mali" as the most appropriate type locality (14.15 N; 5.02 W); a lectotype is hereby designated with the old number MNHN 1960-417 (S.L.= 16.4 mm; T.L.= 21.6 mm or 132% of S.L.), the remaining six syntypes (paralectotypes) from the same collection granted with the new number 1997-55.

*           Aphyosemion guineense Daget, 1954 from several localities in highlands of Guinée: this case has been equally easy; on the same request, the describer decided to choose "Dabola, S.E. Guinée" as the most appropriate type locality (10.81 N; 11.03 W); he also selected for us the lectotype hereby designated, 43.3 mm T.L. and 35.8 mm S.L., with the old number MNHN 1960-475, the remaining seven syntypes (paralectotypes) from the same collection granted with the new number MNHN 1997-33; the lectotype has 12+16 vertebrae.

*           Epiplatys sexfasciatus Gill, 1862: this case is very difficult, because the type locality is given as lower Ogooué River, N.W. Gabon; the collector is Du Chaillu in 1860 (for which a mountainous region was besides named in Southern Gabon East and South of Lambaréné!); however Gill worked at the same time with fishes from very distant locations in Africa, including marine ones, and therefore no further information may be obtained from the Philadelphia Museum (E. Boehlke, pers. comm.). The study of types has confirmed us the diagnosis (six bars, open frontal neuromasts system): according to our present knowledge, such a fish may not live downstream from Lambaréné to the Ogooué mouth into the Atlantic Ocean, but is quite frequent North of the city, i.e. on the right side of the river (we have collected it 9 km S.W. Bifoun and an uncontrolled local collection reports it from not far from the Northern suburbs of Lambaréné) and it is replaced by Ep. ansorgii (also named berkenkampi) with a closed frontal neuromast system in the South, about 30 km away, according to our present knowledge (but, the limit of distribution of both species is unclear); the search of the fishing and shipping books of the mid nineteenth century has clearly disclosed that, to the contrary of the present time, Lambaréné was the most important harbor in the area, and neither it was the present Libreville, nor Port Gentil, at Ogooué mouth; therefore we feel reasonable and safe to propose as for the type locality, close to the Northern parts of Lambaréné (approx. 0.52 N; 10.22 E). Etzel (1995) has conducted interesting crossing experiments between components of the Ep. sexfasciatus superspecies which has led him to restrict the distribution of Ep. sexfasciatus s.s. (which has no more subspecies) to N.W. Gabon and S.W. Ecuatorial Guinea.

*           Aplocheilichthys spilauchen (Duméril, 1861) from 3 type specimens collected by M. Aubry-Lecomte: this case is very similar and is apparently contradictory to the preceding one; Duméril gave "Gabon River" (now the Ogooué) only as the type locality and it is known that the fish prefers brackish biotopes, which explains its gigantic distribution along coast from Sénégal to Angola; then it should be close to the sea, i.e. far from Lambaréné, 150 km distant! This apparent contradiction was levied when we disclosed in the Tervuren Museum a collection of this fish by the renowned U.S. ichthyologist, Tyson Roberts, in 1978 from a lake only 35 km S.W. Lambaréné (precisely at the Northern end of Lake Ezanga); therefore we feel reasonable to propose the type locality close to the Southern parts of Lambaréné (approx. 0.72 N; 10.11 E) at the beginning of the delta marshes, all the more that an uncontrolled local aquarist report has confirmed its occurrence there.

*           Aphyosemion ogoense ogoense (Pellegrin, 1930): this case is again linked to the Ogooué River, but in the upper part of its stream and precisely from one of its tributary, La Passa or Mpassa River, which crosses the Ogooué in the Eastern surroundings of Franceville; it is simpler but anecdotically attractive; in-depth studies of the Baudon's trips in 1929 disclosed that he let sample in the surroundings of Franceville (Pellegrin, 1930: 196 and MNHN registers), where now A. ogoense ogoense and A. pyrophore, two components of the same superspecies with distinct color patterns, are frequently collected; the anecdote is that, unaware of this fact, Radda and ourselves in 1976, discovered a fish further East of Franceville with a similar color pattern but with a distinctive morphology (in fact a member of the elegans superspecies), and we were convinced that the type locality given for ogoense was erroneous and decided to name that fish rectogoense (i.e. the correct ogoense)! To be fully complete, we must add that Pellegrin had 2 series of syntypes, the other from Léconi or Lécéni, but we designated (Huber, 1981) a lectotype from La Passa. Anyhow it is clear now that the type locality of A. ogoense ogoense should be close to East of Franceville (approx. 1.70 S; 13.78 E) near the confluence of the Ogooué and Mpassa Rivers, and lately it has been collected very close: 6 km East Binomi from Franceville to Onkoua, 1.72 S; 13.72 E (Eberl, pers. comm.).

*           Procatopus terveri Huber, 1981: this case is identical to that of ogoense, since it was also collected by Baudon at Mpassa, i.e. probably sympatrically (further West, Proc. terveri is sympatric with another Aphyosemion species: A. coeleste) or close to each other ; we therefore propose the same coordinates for the type locality (approx. 1.70 S; 13.78 E); further South it is replaced by another Procatopodin with a restricted distribution (maybe a frontier species, as per Huber, 1981) : Hypsopanchax catenatus.

*           Aphyosemion splendidum (Pellegrin, 1930): this case is parallel to the two preceding ones; Baudon (1929) gave 2 syntypes series to Pellegrin, one from the "Upper Sangha basin", the other from "between Souanké and Garabinzam", Ivindo basin, in Northern Congo. To avoid (if at all possible) its identity with Radda's taxon, A. kunzi, we redescribed (1982) splendidum and fixed a lectotype from the former series; in fact, in-depth studies showed that Baudon had collected in the neighborhood of Ouesso, near the Sangha river (Pellegrin, 1925: 98 and MNHN registers) and without knowing this fact we "rediscovered" that fish 40 km South Ouesso in a marshy area along the river (where we got dangerously lost!) during our 1978 trip; therefore we feel reasonable to propose Ouesso as the type locality, near the Eastern marshy banks of the river (1.62 N; 16.07 E).

*           Epiplatys boulengeri (Pellegrin, 1926): this case is an easy one, since Schouteden (1925) collected 3 syntypes series, one from Mongendé, W. Zaïre, with 1 specimen, the other from Kidada with two specimens (MRAC 18371-72) and the third from Tshimbu with three specimens (MRAC 17414-415; MNHN 1926-242); out of the three, only Mongendé is given in the Gazetteer and therefore this is the proposed type locality (2.15 S; 16.17 E), close to the Congo river which makes the frontier between Zaïre and Congo; a lectotype is hereby designated accordingly; because the types of this species have been split between Tervuren and Paris, the hereby designated lectotype happens to be in Belgium (MRAC 18370): a specimen of 41.5 mm S.L. and 51.5 mm T.L. (kindly measured for us by Guy Teugels), with the paralectotypes being stated above. The two other localities are probably not far: there is a village named Tshimbi, to-day, also along the banks of the Congo River, further upstream. The radiophotography of the lectotype gives the following meristics: D= 11, A= 16, D/A= +9, vertebrae= 13+14. It should be stressed that this taxon is well defined by populations not far from the type locality, in Eastern Congo; on the contrary the oldest taxon of that superspecies, Ep. multifasciatus, is unknown live from its type locality, although it is precisely quoted: Kondoué, Kasaï, S.E. Zaïre, 4.98 S; 23.30 E; we encourage aquarists to target it as a collecting trip when the political situation permits: it was sympatric with Aphyosemion ferranti and A. lujae, all with the same type locality (and several Nothobranchius species are reported not too far away).



On top of those proposals reported in Killi-Data 1996, we would like to put forward suggestions of type localities of 3 important taxa belonging to two Aphyosemion superspecies, with components having variable caryotypes and thus emerging cryptic species; we feel they are reasonable and desirable:

*           Aphyosemion bitaeniatum (Ahl, 1924): here the systematic move by Seegers (1988, op. cit.), i.e. multicolor, although widely used, as a junior synonym of bitaeniatum, has been readily accepted, including by us: bitaeniatum is a well defined taxon, lacking only a type locality information; multicolor is a confusing taxon because this is an aquarium based "poor" description claimed by two authors, Brüning (1929) and Meinken (1930); however, multicolor has the advantage of having been allocated a posteriori a precise type locality (Ladiges, 1951), near Lagos, where the fish is very common: this is to-day Apapa and we suggest to allocate it to bitaeniatum (6.48 N; 3.37 E), along a principle of parsimony.

*           Aphyosemion calliurum (Boulenger, 1911): in this case the type locality, Liberia, is erroneous (or is it a misunderstanding between Boulenger and Arnold, regarding the origin of this trade import?); anyway, Clausen has demonstrated that the types were having an open frontal neuromasts system and were originating from Nigeria; it is known that in those days most aquarium fishes were collected and exported from nearby big cities and harbors: this is especially the case with Lagos in Nigeria where the type locality of Ep. grahami falls and which was described also by Boulenger (1911) in the same paper; besides A. calliurum is very common in this part of Nigeria; we therefore suggest to consider "inland Lagos" as the type locality for calliurum, like for Ep. grahami and Micropanchax macrophthalmus macrophthalmus: 6.45 N; 3.38 E.

*           Aphyosemion ahli (Myers, 1933): this is the most complicated case of all; it was "described" by Meinken (1932) as a variety of calliurum, Panchax calliurus var. caeruleus, and renamed as ahli since the name caeruleus was preoccupied by caeruleus (1) as per Boulenger (1915); this so called description was based on imported aquarium material of unknown origin and no type has been deposited and none is to be found in Berlin (Päpke & Seegers, 1986); additionally, no other fish has been reported from the same import; there is then absolutely no rational for a type locality; however, we tentatively suggest to allocate it to "Longji, approx. 10 km N.W. Kribi" for the following reasons: first, Longji is close to Kribi, an harbor that was used in exporting fishes in the twenties and thirties; second, this place is in the middle of the distribution of ahli s.l., i.e. far from the type localities of calliurum and australe, its two most related species; third, a population of ahli has been discovered there by Scheel and Clausen, on January 9., 1966 (in fact, it is most probably one of the first aquarium strain since the original trade import): this population has a well known caryotype (Scheel, 1990) and a standard one for ahli, i.e. different from the populations with variable caryotypes collected between the Sanaga and the Nyong rivers; fourth, the females there (according to Radda, 1973) of ahli are easily set apart because they show the rare fasciated pattern that has been described by Meinken in his drawing; fifth, this option leaves open for comparative study two recent taxa of the superspecies, Aphyosemion heinemanni Berkenkamp, 1983 and A. edeanum Amiet, 1987, and also for a further study and eventually a separate naming the populations having a different caryotype (from in between the two above mentioned rivers) and those, poorly known from Ecuatorial Guinea.

Finally, this option does not take a position on the pending validity of ahli and pascheni: either, following Radda (1973), ahli and pascheni, both as valid species and then types of both species must be restudied in depth, or following Amiet (1987) who may well be right, pascheni as a senior synonym of ahli and festivum as a valid species, both fishes being sympatric at Longji and pascheni sensu Radda to be renamed; the geographical coordinates of the type locality of A. ahli would then be identical to those of pascheni, i.e. 3.07 N; 9.98 E and a neotype is hereby designated from Scheel's collection in the Copenhagen Museum of 21 specimens: a female of 34.2 mm S.L. and 43.2 mm T.L., with vertical markings (registration number P 35 4023), the remaining 20 specimens, males, females and juveniles, with registration numbers P 35 4220-39 being paraneotypes. The exact location name is Nziou, which lies in the immediate southern surroundings of Longji. The following meristics have been counted for respectively the neotype and an adult male paraneotype (n°35 4238): D= 11, 11; A= 14, 15; D/A= +8, +8; LL= 30, 29; both show a G-type frontal scalation.

Finally, a last case may be forwarded, because of its high importance in Epiplatys phylogeny and of its huge distribution from Sénégal to central Soudan (see above its possible junior synonym Ep. marnoi) and to lower Zaïre (Kinshasa):

*           Epiplatys spilargyreius (Duméril, 1861) has been described from "les eaux douces de la Côte des Mandingues" (freshwaters of the Coast of Mandingues), collected by M. Schill, Captain of the Russian merchant navy which may imply a collection not far from the sea coast and not far from a major harbor of that time; according to Rochebrune (1883), also quoted in Pellegrin (1923), the type locality is "obviously situated either on the Falémé river (i.e. at its mouth with the Sénégal river: 14.02 N; 11.10 W) or near the Gouina falls, on the Sénégal river (i.e. in Mali, upstream from Kayes and from the first falls where the ships contents had to be transferred at that time: 14.69 N; 12.78 W)"; however, both localities are far inland in the Sénégal basin; according to Prof. Daget, Rochebrune's "obvious" statements have been proven erroneous in several cases and these data must be taken with care, since moreover, Duméril has given "the coast" as a clue. We are then facing three solutions, which are all possible, since the fish is reported all over that region; for the sake of simplicity, because of Duméril's remark and Daget's consideration, we therefore suggest that the type locality of spilargyreius is recorded as the surroundings of Saint-Louis, the main harbor at the mouth of Sénégal into the Atlantic ocean (16.03 N; 16.50 W): most merchant ships stopped there for their business in the mid nineteenth century, the fish has been collected there and Epiplatys senegalensis, its junior synonym, is having a type locality not too far upstream (Dagana: 16.65 N; 15.52 W).



In conclusion we would like to add that we feel confident the above proposals and suggestions for type localities are adequate; this is because not only of course the various species are now known from these places, but also because it is now experienced that, in those old days, collections for the aquarium trade were mostly undertaken from large harbors surroundings in order not to face too many logistic problems to keep the fishes live during shipment; however the future may disclose some new data which may contradict some of them: we must always stay humble with our own reasoning and not precipitate on an apparently attractive solution and we may expect a lot of advances when DNA techniques will be efficient for old formalin-preserved types or topotypes, so that an identity card of all taxa may be prepared.


This article exemplifies also the very important role the aquarists play in disclosing solutions simply by visiting the places still unknown: several cases are brought here as a challenge for them; may we add a last one : in 1995 a probably new species of the Aphyosemion wildekampi superspecies has been discovered by a French cooperant in Southern Central African Republic, about 60 km Southeast Mbaïki (3.61 N; 18.42 E) which is said to show a postopercular ocellus (we only have poorly preserved material); in fact the whole region, between Nola and Mongoumba, is very interesting because it is a primary forest where the non fish fauna (reptiles, insects) and the flora have been shown to be endemic.

Above all, a quick look to the maps published in Killi-Data 1996 strikes how many collecting trips in virgin regions remain to be undertaken! Good luck to the aquarists explorators and many anticipated thanks to them.



This hard work would not have been possible without the generous help of the following persons: Dr. M.L. Bauchot (Paris, MNHN), E. Böhlke (Philadelphia, ANSP), Dr B. Herzig (Vienna, NMW), Dr J. Nielsen (Copenhagen, ZMUC), Dr H.J. Päpke (Berlin, ZMB) and Dr G. Teugels (Tervueren, M.R.A.C.).

Because we felt it is of utmost importance that the above proposals and suggestions be reviewed by the experts on Old World Cyprinodonts, we have sent the manuscript to them including to Lothar Seegers by courtesy : those who took the time to comment receive also kindest thanks: Prof. J. Daget (Paris MNHN), Prof. J.L. Amiet (Nyons, France), Univ.-Prof. Dr A.C. Radda (Vienna, Austria).

This article has been given for publication to the Journal of the American Killifish Association (A.K.A.) in recognition to being acknowledged the honorary membership some time ago. Richard Martino receives our personal gratitude for his action in this context and for having informed us by phone across the Atlantic.



Ahl,E. 1924a. Zur Systematik der altweltlichen Zahnkarpfen der Unterfamilie Fundulinae. Zool. Anz. , 60 (1-2): 49-55.

Ahl, E. 1924b. Über neue Afrikanische Zahnkarpfen der Gattung Panchax. Zool. Anz. , 60 (3-4): 303-313.

Amiet, J.-L. 1987. Le genre Aphyosemion Myers (Pisces, Teleostei, Cyprinodontiformes). Sciences Nat , Compiègne: 262 pp., 76 pl.

Boulenger, G.A. 1911. Description of new African Cyprinodont Fishes. Ann. & Mag. Nat. Hist.  ser. 8, vol. 8 (44): 260-268.

Guichenot, M. 1866. Catalogue des Poissons de Madagascar du Musée de Paris. Mém. Soc. Imp. Sc. Nat. Cherbourg, XII: 129-148.

Guma'a, S.A. 1982. On the Biology of Female Epiplatys bifasciatus (Cyprinodontidae) from Southern Sudan. Hydrobiologia  89: 285-300, 5 figs., 8 plates, 3 tables.

Huber, J.H. 1980. Rapport sur la deuxième Expédition au Gabon (Aôut 79). Etude des Cyprinodontidés récoltés. Rev. fr. Aquariol.  7: 37-42, fig., 8 photos, table, map.

Huber, J.H. 1981. Die Überart Aphyosemion ogoense. Eine systematische Übersicht, sowie die Darstellung von hypothetischen Artbildungsmodellen. Deutsche Killifisch Gem. J. , 13 (9): 133-148, 16 figs., 3 maps.

Huber, J.H. 1992. Review of Rivulus. Ecobiogeography - Relationships. Cybium Suppl., Société Française d'Ichtyologie Publ.: 586 pp., 40 pls., 85 figs, 8 tabs, 13 maps

Huber, J.H. 1994. Aphyosemion passaroi, espèce inédite du Gabon Sud-oriental, au patron de coloration unique et Description complémentaire de Aphyosemion decorsei Pellegrin. Revue fr. Aquariol. , 20 (1993): 77-79, 1 fig.

Huber, J.H. 1996. Killi-Data 1996. Updated checklist of taxonomic names, collecting localities and bibliographic references of oviparous Cyprinodont fishes (Atherinomorpha); in french, english, and german. Cybium, Soc. fr. Ichtyologie, Ed. , Paris: 400 pp., 20 maps.

Ladiges, W. 1951. Aphyosemion multicolor Meinken. Der Fisch in der Landschaft : 97-98, fig.

Pellegrin, J. 1923. Les Poissons des Eaux douces de l'Afrique Ocidentale Française (du Sénégal au Niger). Gouverment Général de l'Afrique Occidentale Française. Publications du Comité d'Etudes Hist. et Scient., Emil La Rose Ed.  373 pp., 75 figs.

Pellegrin, J. 1930. Poissons de l'Ogooué, du Kouilou, de l'Alima et de la Sangha recueillis par M. A. Baudon. Description de cinq Espèces et cinq Variétés nouvelles. Bull. Soc. Zool. France  55: 196-210, fig.

Radda, A.C. 1973. Notes on the Distribution, Ecology and Variability of the Aphyosemion gardneri Superspecies in West Cameroon- Aphyosemion pascheni and A. ahli. Brit. Killifish Assoc. Sep. , 94: 12 pp., 10 figs., 2 maps.            

Rochebrune, A.T. de. 1882. Faune de la Sénégambie. Actes Soc. Linn. Bordeaux  36: 23-190, 6 pls.

Scheel, J.J. 1968. Rivulins of the Old World. Tropical Fish Hobbyist Publication, New Jersey . 480 pp., figs., photos, tables, maps.

Scheel, J.J. 1990. Atlas of Killifishes of the Old World. T.F.H. Publications, Neptune City, New Jersey  (1986): 1-448, fig.

Seegers, L. 1986. Bemerkungen über die Sammlung der Cyprinodontiformes des Zoologischen Museums Berlin. 1. Die Gattungen Aphyosemion Myers, 1924 und Fundulosoma Ahl, 1924, Teil 1. Mitteilungen zool. Mus. Berl.  62 (2): 303-321, figs.

Seegers, L. 1988. Bemerkungen über die Sammlung der Cyprinodontiformes (Pisces: Teleosta) des Zoologischen Museums Berlin. 1. Die Gattungen Aphyosemion Myers, 1924 und Fundulosoma Ahl, 1924. Teil 2. Mitt. Zool. Mus. Berl. , 64 (1): 3-70, figs.

Werner, F. 1919. Fische aus dem Blauen und Weissen Nil. In: Wissenschaftliche Ergebnisse der unternommenen zoologischen Expedition nach dem anglo-ägyptischen Sudan (Kordofan). Denk. Akad. Wiss. Wien mathem.-naturwiss. Klasse  96: 439-452, 10 figs.

Wildekamp, R.H., R. Romand & J.J. Scheel. 1986. Cyprinodontidae. in: Daget J., J.P. Gosse & D.F.E. Thys van den Audenaerde (Eds.). Checklist of the Freshwater Fishes of Africa (C.L.O.F.F.A. 2). Office de la Recherche Scientifique d'Outre Mer (ORSTOM), Paris.  vol. 2: 165-276.


Paris, December 1994-September 1997.


* author's address: Laboratoire d'ichtyologie, M.N.H.N., 43 rue Cuvier, 75231 Paris Cedex 05, France.


Post-finalisation note (dedication):

Although this article ends up with an optimistic flavor, one must not forget the risks still facing those expeditions; the renown enthusiastic US aquarist and a good friend of us, Dale Weber, was killed in Brasil, early 1997 during a collecting trip; this article is dedicated to his memory: I am proud to have named after him an atypical species in 1992: Rivulus weberi.



(1) Boulenger described the taxon as a variety of Fundulus gularis, the spellings caeruleus or coeruleus have been equally used by many authors, all the more that another preoccupied name Fundulopanchax bivittatus var. coerulea had been described by Meinken (1930). In the original Boulenger's paper, it is not easy to disclose which of the two spellings is the right one: Prof. Daget, through his vast experience, found out that "in the index, the taxon was placed just after cabrae (Tilapia) which proves unequivocally that for Boulenger, it was with an a, and not an o".

Additional reading of this author's publications ? Select any ONLINE article.