REVIEW OF APLOCHEILICHTHYINS

Updates to the phylogeny and systematics of the african aplocheilichthyins, the lampeye schooling cyprinodonts (Aplocheilichthyinae, Cyprinodontiformes)

 

by   Jean H. HUBER (1)

 

(1) Muséum national d'histoire naturelle, Laboratoire d'Ichtyologie générale, 43,rue Cuvier, 75231 Paris Cedex 05, FRANCE.

 

ABSTRACT.
- Interrelationships of 19 predefined groups, belonging to the African Aplocheilichthyin genera, are proposed after the phylogenetic analysis of 74 external characters using PAUP 3.1.1. software. The resulting consensus tree is congruent with previous, not all, analyses and with the biogeographical distribution. The genus Micropanchax sensu Huber, 1981 is reported to be polyphyletic. The group of deep-bodied species from the Guinean Dorsale appears to be related to Procatopus and Plataplochilus and is conservatively described as a subgenus of the former under the name Rhexipanchax. The monophyly of the coastal forms is proposed, with Poropanchax rehabilitated as a full genus and Congopanchax becoming its subgenus from the Congo Basin. Micropanchax is restricted to the slender Sahelian forms, whereas the systematics of the groups of central and eastern Africa remain unclear, except for the two pelagic monotypic genera Lamprichthys and Laciris, which are basal.

 

RÉSUMÉ.
- Actualisation de la phylogénie et de la systématique des Aplocheilichthyines Africains, les Cyprinodontes grégaires aux yeux brillants (Aplocheilichthyinae, Cyprinodontiformes).

           Une nouvelle classification phylogénétique des genres Aplocheilichthyines africains est proposée à partir de 19 groupes prédéfinis et 74 caractères externes, en utilisant le logiciel PAUP 3.1.1. L'arbre de consensus résultant est congruent avec les analyses précédentes, mais pas avec toutes, et avec la distribution biogéographique. Le genre Micropanchax sensu Huber, 1981 apparaît polyphylétique. Le groupe d'espèces au corps haut et habitant la Dorsale Guinéenne apparaît lié à Procatopus et Plataplochilus et est décrit conservativement comme un sous-genre du premier, sous le nom de Rhexipanchax. La monophylie des formes côtières est proposée, avec Poropanchax réhabilité au niveau du genre, et Congopanchax, comme son sous-genre de la Cuvette Congolaise. Micropanchax est restreint aux formes allongées Sahéliennes, tandis que la systématique des groupes d'Afrique centrale et orientale reste instable, à l'exception des deux genres monotypiques et pélagiques, Lamprichthys et Laciris, branchés à la base de l'arbre.

 

Key-words.
- Aplocheilichthyines, Africa, Phylogeny, Taxinomy.

 

            The study of the schooling groups of tropical Cyprinodontiformes fish, also called lampeyes because of an iridescent spot on their eyes, has been more or less dormant for decades. One could argue that these fishes are "damned-for-research" since they are extremely fragile to collect and 5-10% only can be kept alive for a few days in the field; their bright electric blue reflecting colors need very special conditions and care in aquarium to remain present; and they often prefer somewhat larger bodies of waters than standard killifishes and thus can only represent a secundary objective of collections. All these reasons contribute to the poor interest shown in them by killi-hobbyists, which means collections are rare and live material available for study scarce. Moreover, none of the fragmentary past studies have succeeded in producing a complete phylogeny of that group; none have been capable of establishing solid diagnostic characters of the described taxa at the generic or subgeneric level; in addition to the above reasons, several factors can be added which make their study even more complicated than Aplocheilids:

- only a few display derived conspicuous color patterns for diagnosis: most species show uniform blue reflecting sides, with rare generalized markings which are all the more variable at the population level;

- no stable morphology is available: many species present an allometry in growth and/or in sex and/or according to population; and the variation in meristics is usually high, which has led to several erroneous descriptions;

- few biogeographical barriers have been disclosed because they are less than poor swimmers and can cover large ranges in small to big rivers, near their shores, and in lakes.

These features separate them from the Aplocheilids and make their study even more difficult, with their very small size being an additional difficulty.

            The purpose of the present study is to propose a phylogeny for this poorly known group of Cyprinodonts, by using the same methodology as the previous study (Huber, 1999a): determine trees of predefined phylogenetic groups, based exclusively on characters of the external phenotype, processed using PAUP software. Contrary to the Neotropical Rivulins (Huber, 1999a), the other two modern approaches of systematics are not available, but they will be covered soon: Costa, (in prep.) for the osteology; Ghedotti (doctoral dissertation), also for the osteology; and maybe Sonnenberg for the molecular biology. Their work will complement and contrast the results of this study.

            Finally, it is important to mention that this study is limited to the Old World lampeyes. The single Neotropical Aplocheilichthyin-like genus, Fluviphylax Whitley, 1965, has not been included because: first, no live material was easily available to us; second, Costa (1996) has published a complete monography on it, both on external and internal characters; and third, Meyer and Lydeard (1993) have shown, within a more general molecular analysis, that Fluviphylax was distantly related to two Old World species, Aplocheilichthys spilauchen and Micropanchax kassenjiensis.

 

History of the genera and subgenera belonging to the African Aplocheilichthyins

 

            Sixteen genera or subgenera have been described in total, which belong to the subfamily Aplocheilichthyinae Myers, 1928 (sensu Parenti, 1981); these are the following, in historical order, with their original or subsequent diagnosis:

1- Aplocheilichthys Bleeker, 1863; type species: typus (a junior synonym of spilauchen); not diagnosed versus other Aplocheilichthyins; amended without any justification as Haplochilichthys by Garman, 1895; restricted (Huber, 1981) to a monotypic genus, specialized in estuarian biotopes and diagnosed from the other Aplocheichthyins by its brackish coastal distribution, by its relatively lower-set pectorals, a deep caudal peduncle, a slightly compressed body and the unique fasciated color pattern.

2- Procatopus Boulenger, 1904; type species: nototaenia; diagnosed by the far-forward ventral fins, almost below the pectoral fins, which are inserted very high on the sides.

3- Lamprichthys Regan, 1911; type species: tanganicanus; diagnosed from Aplocheilichthys s.l. by the very long anal fin and by the more numerous vertebrae (41 instead of 28).

4- Mohanga Boulenger, 1911; type species: tanganicana; diagnosed by the ctenoid scales and by the conical unequally sized teeth forming a band which is exposed when the mouth is closed. An obvious synonym of Lamprichthys, with the same type species, was described in the same publication of the same year, one volume and four months later (Myers, 1924).

5- Hypsopanchax Myers, 1924; type species: platysternus; diagnosed from Aplocheilichthys s.l. by the lack of enlarged teeth at the upper jaw, by the narrower head, by the very compressed deep body and the slender caudal peduncle. It also shares the somewhat advanced position of ventrals with Procatopus. Parenti (1981) has added another diagnostic character; expanded pleural ribs.

6- Micropanchax Myers, 1924; type species: schoelleri (a synonym of loati, fide Sandon, 1950); diagnosed from Aplocheilichthys s.l. by its smaller size, by the very widely united gill membrane, and by a single series of teeth on each jaw.

7- Lacustricola Myers, 1924; type species: pumilus; diagnosed as a subgenus of Micropanchax, by the somewhat heavier and wider head, by the elongate pointed dorsal and anal fins and by the deeper body form.

8- Cynopanchax Ahl, 1928; type species: bukobanus; diagnosed from Aplocheilichthys s.l. by three rows of teeth, the first outer having fewer, less regular, bigger, canine-like teeth than the two interior rows, and by the lack of joint teeth.

9- Platypanchax Ahl, 1928; type species: modestus; diagnosed by the teeth of the upper jaw (probably the canine-like teeth of the outer row). Since its single species, P. modestus, is close to Hypsopanchax deprimozi, Parenti (1981) has synonymized the up-to-then neglected genus, Platypanchax, with Hypsopanchax.

10- Plataplochilus Ahl, 1928; type species: ngaensis; with the teeth of the upper jaw in a wide band, the outer row only a little enlarged (not "slightly less" as per Myers, 1931); not diagnosed from other Aplocheilichthyins; diagnosed by difference with the genera described in the same publication: no scale on peduncle (contrary to Cynopanchax), dorsal fin not short (contrary to Platypanchax). Lambert and Clausen (1967), after having proven that the caudal scalation character was erroneous and that the teeth character was correct but not stable, presented a rediagnosis based on the supraorbital channel with 4 pores (like Lamprichthys) and on the conspicuous ctenoid scales in both sexes (unlike Procatopus). However, we have found (Huber, 1981) some specimens of Procatopus with ctenoid scales and of Plataplochilus with an open sensory system in grooves; thus these characters appear unstable too.

11- Pantanodon Myers, 1955; type species: podoxys (a synonym of stuhlmanni, fide Seegers, 1996); diagnosed from all other Aplocheilichthyins by the lack of external teeth, by the presence of tricuspid inner teeth, by three spines at ventral fins in male and several osteological characters.

12- Andreasenius Clausen, 1959; type species: aberrans; diagnosed as a subgenus of Procatopus by the less advanced position of the ventral fins; by a lower number of dorsal and anal fin rays; synonymised with Procatopus by Parenti (1981) because of no osteological difference between the two taxa; and no clear cut distinction in the morphomeristics, but a gradation.

13- Hylopanchax Poll & Lambert, 1965; type species: silvestris (a synonym of stictopleuron, fide Huber, 1982); diagnosed from Hypsopanchax by a less deep bodied shape, by the much larger size of post-opercular scales compared to the peduncle scales, by the lower number of scales in longitudinal series and by the presence in male, like in Procatopus, of branchiostegal appendages.

14- Poropanchax Clausen, 1967; type species: macrophthalmus; diagnosed by the presence of a supraorbital tubular channel with 3 pores only (4, in Lamprichthys and Plataplochilus), and not a groove and exposed neuromasts like in all other then studied Aplocheilichthyins.

15- Congopanchax Poll, 1971; type species: myersi; diagnosed from other Aplocheilichthyins by the presence of a supraorbital tubular channel with 2 pores only (and not 3 or 4), by the advanced origin of the dorsal fin nearly superimposed with the anal, and, except for Poropanchax, by the dimorphism in vertical fin shapes (with filaments in male). However, Wildekamp (1995) has challenged the specificity of the pore number: juveniles present a third middle pore like in Poropanchax, which is closed at the adult stage. Moreover, we have studied material of "Micropanchax" scheeli from Gabon with 2 pores in the adult, like for Congopanchax (Huber, 1998b).

16- Laciris Huber, 1981; type species: pelagicus; diagnosed from the other Aplocheilichthyins by the pelagic ecology, by the relatively lower insertion of the pectoral fins, by the "very vestigial neuromast frontal pattern (only two narrow grooves, without pores or sensory buttons)", and by the broad dark vertical margin at the caudal fin.

            In total, these sixteen taxa are not well diagnosed from each other after their first and subsequent diagnosis, except for the three exhibiting isolated characters: Lamprichthys (with much higher fin rays and vertebrae counts); Pantanodon (with no teeth); and Laciris (a deep water pelagic species with a vestigial sensory system). All other taxa were only diagnosed by characters which show a continuum and which overlap each other.

            The history of general reviews of Aplocheilichthyins is much shorter. It is held in nine papers across three periods:

- during the earlier period, the diagnoses are based mainly on the teeth structure and on general morphology. Two papers cover this: Ahl (1928) and Myers (1938), the latter author being highly suspicious about the value of the teeth as a diagnostic tool in that group (I agree, after the study of many specimens, except for Pantanodon, obviously; besides, most species have the same diet). Both authors concede that Myers's Micropanchax and Lacustricola are based on an erroneous observation (a single series of teeth) and lack a sufficient diagnosis in comparison to Aplocheilichthys s.l. Myers suspects the same for two of the three genera raised by Ahl, without naming them (Platypanchax and Plataplochilus?).

- during the second period, a strong emphasis is given to micromorphological characters, like ctenii on scales and fins, and even more, the frontal sensory system, following remarkable observations by Clausen (1967), which are in line with Gosline's for Fundulin genera (1949). This latter character has also been used by Lambert and Clausen (1967) and by Poll (1971). However, Huber (1981) has shown that there is a continuum of intraspecific and interspecific variations in the structure and position of the cephalic sensory system: indeed, a supraorbital system is basically the same whether it is a tubular channel with 3 pores at one end or an open groove with 3 neuromasts at the other end, with intermediates like fleshy lobes partly covering the groove. Moreover, if the channel reaches the preorbital sensory pair, then 4 pores are available; or if the middle pore is covered, then only 2 pores are available. The result, in terms of water contact with the sensory organs, is the same.

- during the third period, in recent modern times, the systematic analysis is focused on two approaches other than that of the external phenotype, but it concerns very few taxa. First approach, the detailed osteology: Pantanodon (with podoxys) by Whitehead (1962); and Aplocheilichthys (with spilauchen), Poropanchax (with macrophthalmus), Lamprichthys (with tanganicanus), "Aplocheilichthys" (with johnstoni), Procatopus (with nototaenia in Procatopus s.s. and gracilis in Andreasenius), Hylopanchax (with silvestris), Hypsopanchax (with platysternus and zebra) and Pantanodon (with madagascariensis) by Parenti (1981); i.e., 10 species in total. Moreover, the latter author shows that Aplocheichthyins are closer to viviparous Poeciliids than to Fundulins as per Myers (1931). The second approach, using molecular biology, concerns only two distinctive species to-date (Meyer and Lydeard, 1993).

            As emphasized for Rivulus (Huber, 1992), Aplocheilichthyins show a continuous range of variations and an overlap in most of their morphomeristic characters. Therefore, not a single character is clear cut enough to be used for absolute identification without a precise geographical origin. More specifically, we had reached a dead-end after having shown that the supraorbital system could not be used. With the availability of software for systematics, like PAUP, this situation has changed considerably; since it is now possible to use numerous characters without a hierarchy, provided that the continuity of each is numerized. Besides, software allows valuable descriptive characters, such as Clausen's with their statistically dominant state in each group, to reconsidered.

 

 

MATERIAL AND METHODS

 

            The material used in the present study is listed in appendix 1 (one star indicates the series was already studied for our 1981 publication). A species to species comparison between the studied material and the valid list, in table I, highlights the fact that all species groups have been surveyed, resulting in 51 presumably valid species out of a total of 63, i.e. 81%.

            The methodology is the same as for Rivulus and its allies (Huber, 1999a). All character states have been treated as unordered when several states are disclosed, i.e. the lengths of 0-1, 0-2, 0-3, 0-4 are equal. The hypothesis concerning  the numerization of continuous, pseudo-discontinuous (such as micromorphological characters, e.g. the 4 supraorbital sensory system patterns) and discontinuous characters, and the assumptions concerning the options of "no data" or "question marks" are also the same. The note of warning concerning the lack of sufficient and live material is again stressed here, and further emphasized, for the reasons mentioned in the introduction. No character known with a potential diagnostic value has been discarded. All characters are considered primitive (coded: 0) in the outgroup.

            The selected outgroup is distantly related to Aplocheilichthyins: a monotypic subgenus (or genus, depending on authors) of Fundulus with supposed primitive characteristics, Plancterus from the Mississipi basin, in central U.S.A. Other Funduline species belonging to different phylogenetic lines, like Fundulus (Fundulus) heteroclitus or F. (Fontinus) diaphanus, also from U.S.A., and Valencia hispanica could have been added to the outgroup, with the same results. One could argue that it would have been better to select a Poeciliid because it could be a sister group to Aplocheilichthyins; however, their relationships have not been yet robustly supported by molecular biology and this decision would not have proven useful in terms of external phenotype characterization. Poeciliids are all -except for one species- viviparous, a derived character and all show relatively derived external characters (not only the gonopodium in the male!). Fundulus zebrinus comes closer to the "ideal profile" of a primitive Cyprinodont (Huber, 1999a) and it exhibits a tubular supraorbital system, like the presumed primitive Aplocheilichthyins, and an A-type frontal scalation (i.e. the scale above the primitive pineal organ is covering all others).

 

 

The matrix of 20 groups and 74 characters

 

            A total of 19 presumed monophyletic groups have been defined and used as terminals, in addition to the single outgroup. All sixteen described taxa at the generic level have been retained, except for three well-acknowledged synonyms, Mohanga, Platypanchax and Andreasenius.

            Four new homogeneous groups are created from the split of Micropanchax into five subunits (then Micropanchax s.s. becomes Micropanchax1_loa; loati et al.), and two new groups from Poropanchax into 3 subunits (then Poropanchax s.s. becomes Poropanchax1_mac; macrophthalmus et al.):

- Micropanchax2_kin (kingii et al.), with a somewhat less slender body than Micropanchax1_loa, is mainly separated from other species of Micropanchax by the presence of seven preopercular cephalic pores, as suggested by Wildekamp (1995) and confirmed by us on a small separate series of specimens.

- Micropanchax3_hut (hutereaui et al.) with a rather deep body, and a marked sexual dichromatism. This group, notably the species camerunensis, shares some external characters with normani, the monotypic component of Poropanchax2_nor, which inhabits both the interior savannah and the coastal forest plain, west of the Bénin/Dahomey gap.

- Micropanchax4_joh (johnstoni et al.) with a strong sexual dichromatism and a unique pattern of independent blue spots on the sides, instead of the standard blue shiny background. Parenti (1981) has separated johnstoni from all other studied species by the distal arm of the premaxillary curved (not straight).

- Micropanchax5_nim (nimbaensis et al.) with a group of dwarf, very deep-bodied species from west of the Bénin-gap, described as isolated by Huber (1981) and Wildekamp (1995).

- Poropanchax2_nor (normani), a single species which is presently intermediate between Poropanchax s.s. and Micropanchax3_hut: it shows the tubular supraorbital system of macrophthalmus (though not always: numerous specimens have fully open grooves) but it is less slender, and exhibits a continuous blue hue on the sides.

- Poropanchax3_sch (scheeli), is also one species, brackish like Aplocheilichthys spilauchen and Pantanodon sp. It exhibits also a variable supraorbital system: open in northern populations (Wildekamp, 1995),  and tubular with two pores in one southern population (Huber, 1999b), like in Congopanchax. As stated in Huber (1999b), this species may represent a phylogenetic bridge between the coastal corridor distributed group Poropanchax s.s. and the Congo Basin relict group Congopanchax.

            All presumably valid species are listed in table I, together with their major morphomeristic data. The validity status of the eastern African taxa follows Wildekamp (1995) and that of the western species, Huber (1981; 1982): 63 valid or presumably valid species are counted out of a total of 104 described taxa (Huber, 1996), i.e. 61%.

Each species has been assigned to one of the 19 groups in table I, based on published evidence and/or our own observations, with in a few cases an uncertain positioning due to contradictory or scarce data; such is labelled with a question mark.

            A total of 74 characters has been defined, according to the 5 types given in Huber (1998a). The list of the 74 characters, with their up to 5 states, is given in table II; many are common with the ones used for Rivulus et al. (Huber, 1999a).

All but the following three characters do not necessitate further specific comment:

- the position of the pectoral fin (high) and of the ventral fin (advanced) is variable, although for Procatopus they are at one extreme (high and advanced); however, we have seen specimens of Hypsopanchax with very highly inserted pectorals and of Plataplochilus with very far forward ventrals!

- the relatively bigger size of the post-opercular scales compared to the prehypural scales is variable too, although for Hylopanchax it lies at one extreme; however, some specimens of Hypsopanchax do show enlarged post-opercular scales.

- the frontal scalation pattern is also variable within a population of a single species, like for Rivulus, but not enough material has been seen to determine the dominant state for all species with certainty; for example, we have fixed the frontal scalation of Aplocheilichthys spilauchen to the E-type after having studied 40 selected, large, well-preserved specimens from all over its whole range (9 "E" including the lectotype, 7 "G", 5 mixed "G-H", 4 mixed "E-G", 4 "H" and 11 uncertain), whereas Whitehead (1962) states E-type and Wildekamp (1995) G-type. In fact, among Aplocheilichthyins, there are only two major distinctive types, "G" and "H", contrary to the Rivulins which show a greater diversity; and the allocation to the "G" or "H" type is uncertain because these two scales are neighbours on the middle line of the front and it is not easy to determine which one overlaps the other in many specimens. For example, we have disclosed different frontal scalations from Wildekamp's (1995) for the following species: pfaffi, myersi and camerunensis. This is probably due to the limited number of specimens that were studied, but anyhow this criterion should be cautiously handled.

            The matrix of 20 groups and 74 characters was processed using PAUP version 3.1.1 (Swofford, 1993) to produce the most parsimonious tree. This was done only through the heuristic search method, because the matrix was too big for a branch-and-bound search. The matrix is presented in table III.

The following analyses were undertaken:

- subminimal trees;

- for each tree, computations of classical homoplasy measurements as CI, RI and HI, the effect of excluding individual characters, the effect of weighting characters (by the RI index) and the effect of ordering characters;

- finally the strict consensus tree, bootstrap computation (Felsenstein, 1985), and minimal and maximal branch lengths according to optimizations of homoplasy.

Branch lengths in trees are proportional to the number of changes occurring along the branches and shown under ACCTRAN optimization. In the strict consensus tree, numbers above the branch lengths are bootstrap proportions obtained from 1000 replicates using PAUP. A scale is given to associate branch length and the number of changes.

Some phylogenetic alternatives were tested using the Mac Clade software (Maddison and Maddison, 1990) to compute the corresponding necessary extra-steps.

 

RESULTS

 

Two most parsimonious trees have been found with the following characteristics: tree length= 236 steps; consistency index (CI)= 0.555; retention index (RI)= 0.585; and homoplasy index (HI)= 0.445. Both trees are very similar with only one difference: the relative reciprocal position of Micropanchax4_joh, Lacustricola and the block "Cynopanchax­- Micropanchax3_hut". In tree 1, Micropanchax4_joh is basal to the three other lineages, whereas in tree 2, Micropanchax4_joh and Lacustricola are sister groups, like Cynopanchax and Micropanchax3_hut. This results in a single unresolved tritomy in the consensus tree (Fig. 1).

The subminimal test confirms the unequivocal nature of the disclosed configuration: for 237 steps, 34 subminimal trees are already found, and for 238 steps, over 100. The reweighting of all characters by the maximum value of the RI index produces a single tree which is identical to tree 1 and is then congruent with the consensus tree.

Changing all characters (or only the morphomeristic ones) from unordered to ordered produces only one parsimonious tree of 287 steps. A comparison with the consensus tree obtained with all unordered characters reveals that:

- the two trees are rather different in the position of groups of lineages (e.g., Aplocheilichthys s.s., basal to the block Poropanchax-Congopanchax; the 4 sister lineages allied to Lacustricola, placed basal to all Aplocheilichthyins, except Lamprichthys and Laciris);

- however, all assemblages comprise the same lineages, except Aplocheilichthys.

The bootstrap analysis of the unordered consensus tree gives an idea of the robustness of the various branches, complementing the lengths of the branches themselves. For example, a bootstrap value of over 70% and an assigned branch length of over 8 steps indicate a strong support by the data; values between 50% and 70%, and 4 to 8 steps indicate good support; and values below these limits reflect uncertainties in the data.

For the Aplocheilichthyins, with the external phenotype only, some nodes are supported as:

- the distinction between the primitive taxa (Lamprichthys and Laciris) and the more derived taxa is well supported (82%);

- two monophyletic assemblages are also strongly supported: Micropanchax1_loa and 2_kin (79%), and Poropanchax+Congopanchax (93%);

- the branches of the deep-bodied forms (minus Pantanodon) and of Lacustricola+Mic4_joh and Cynopanchax+Mic3_hut are also supported (between 54% and 70%);

On the other hand, there are some poorly supported nodes like those of Aplocheilichthys and Pantanodon which correspond to monotypic, very specialized taxa or alike, with a lot of autapomorphies and few shared derived characters (between 16 and 20 steps in their terminal branches). The sensitivity analysis, conducted by excluding individual characters one by one, does not change significantly the number and shape of the minimal trees and the general framework of the consensus tree. The exclusion of characters, with RIs equal to zero, i.e. numbers 2, 11, 21, 23, 26, 31, 41, 48, 51, 55, 58, 64, 68 and 74, has been processed for a heuristic search: the two resulting trees (196 steps) are similar to the two minimal trees of the unordered analysis, only different by the position of Micropanchax4_joh and in a few branch lengths, and the consensus tree is identical to the consensus tree resulting from the unordered analysis.

 

 

DISCUSSION

 

            Understandably, the general configuration of the consensus tree fits with the basic morphology, but it provides much more information.

1. Supporting evidence:

- the present results, as anticipated in Huber (1981), do not agree with the previous older classifications based on teeth (Ahl, Myers, op. cit.), on frontal neuromasts systems or ctenoidy on scales and fins (Clausen, Lambert, Poll, op. cit.): notably, the various genera diagnosed by a tubular sensory system are scattered across the tree. It is probable that this character is labile, especially because it differs according to population. For ctenoidy, it is the same: Procatopus, as per Huber (1981), comes close to Plataplochilus (respectively without and with ctenoidy).

- the terminal position of the block "Procatopus- Hypsopanchax" (Plataplochilus, not studied; Hylopanchax, synonymized with Procatopus) and the nested position of Pantanodon, as per Parenti (1981) is confirmed here.

- the hypothetical correspondance of the two primitive genera Lamprichthys and Laciris as per Huber (1981) is corroborated,  despite the important differences in detailed morphology.

- the well-separated position of the monotypic taxon, Aplocheilichthys spilauchen, is also retained, as per Huber (1981), despite its unstability. The consensus places it away from its ecological counterparts, likely deep-bodied or not (e.g. Porop3_sch), but this requires further consideration.

 

2. Novelties:

- the major finding of the present study is that the large genus, Micropanchax, is polyphyletic and its various groups (5 groups in a narrow sense, plus the 3 Poropanchax groups) are scattered throughout the tree (to the contrary to the homogeneity of Rivulus in the preceding study).

- the second major finding is that the tree is in agreement with the biogeography of the various Aplocheilichthyin groups. Not a single lineage is in contradiction with biogeography, although a gap exists between the two distributions of Pantanodon and Hypsopanchax, due to the lack of collections in Mozambique.

 

3. Discordances:

- Pantanodon, closest to Hypsopanchax. This relationship is robust and the two taxa are not very distant in Parenti's cladogram (only separated by johnstoni) and show a general external resemblance. As an alternative, branching off Pantanodon at the base of the tree (in line with its biogeography) costs 12 extra-steps i.e. 5% of the number of steps. According to the present data, a basal branching of Pantanodon is therefore rather unlikely.

- Micropanchax4_joh, closest to Lacustricola. This relationship is not coherent with Parenti's observations (1981). However, she did not analyse the osteology of any component of Lacustricola.

- Labile placement of Aplocheilichthys in the tree. Following the change from one state to another of characters that may be rather ambiguous, the otherwise well-defined Aplocheilichthys genus moves in the tree and is branched in the "deep-bodied" lineage. As an alternative, branching off Aplocheilichthys as a sister group of Hypsopanchax only costs 3 extra-steps (1.2%). Therefore, we refrain from considering the exact placement of Aplocheilichthys to be robust in the tree, pending further studies.

 

In total, only further studies with new material and with the two other approaches can confirm or contradict these results.

 

SYNOPSIS OF CHARACTERS USED FOR IDENTIFICATION

 

            Ideally, groups are best defined when the diagnosis matches the synapomorphies, i.e. defining a group with its uniquely derived character (or several of them). Such characters exhibit a state that occurred only once during the course of evolution. This is the case of 11 character states (among 8 characters: number 9, 33, 47, 53, 56, 63, 65, 67). Autapomorphies apart, all other character states are homoplastic. Therefore, the diagnoses proposed here are mostly based on combinations of homoplastic character states. Biogeographical data are also given.

            In total, the African Aplocheilichthyins are brought into 7 major lineages in the present study:

- The "ancestral" lineage. Two long-since separated genera, Lamprichthys and Laciris (8 steps). Both are very large sized, with a short head. High dorsal and anal fin ray counts, combined with high lateral and transversal scale counts, produce a unique condition. No post-orbital colored plaque. Both, when preserved, are like sardines, with a broad longitudinal dark zone on the upper sides of both sexes. Both are strictly lacustrine, living in mid-waters of very old lakes of the Rift valley (Tanganyika and Edouard) and may have been trapped independently there and, as single Cyprinodont entities, may have been forced to evolve only by other fish constraints. No other Aplocheilichthyin is reported to be strictly endemic to Rift valley lakes; the lacustrine Aplocheilichthyins are also available in inflows or outflows of these lakes.

- Lamprichthys. Autapomorphies: in male, a crescent curved caudal fin, which is the shortest in Aplocheilichthyins (17% of standard length); striations of golden and blue spots on sides and in parallel on the fins of the male; and a longer egg incubation period than all other known Aplocheilichthyins. Other characteristics: by far the largest Aplocheichthyin; body compressed with a marked sexual dimorphism in fin shapes and dichromatism; a short predorsal length, like Congopanchax and Pantanodon (independently acquired); and the highest anal fin count, circa 27. Shares with Plataplochilus, the tubular supraorbital system with 4 pores. Shares with Procatopus a very high insertion of the pectoral fins. Shares with Procatopus and Plataplochilus the choice of breeding substratum: eggs are deposited in rock holes. Lamprichthys is also diagnosed by a series of osteological characters (Parenti, 1981), including a very high number of vertebrae (40-43).

- Laciris. Autapomorphies: a vestigial supraorbital sensory system with shallow continuous grooves and invisible buttons; a broad dark border at male caudal; a posteriorily cut, short caudal fin without extensions; and a pelagic form collected in mid-waters where lake Edouard was 38 m deep. Other characteristics include: a more cylindrical and slender body, with a rounded transversal profile of mouth, like Lacustricola et al.; male and female, with silvery sides, no markings; and with no dimorphism and no allometry in growth (like Mic1_loa and Mic2_kin). Shares with Hylopanchax, the posterior insertion of the dorsal fin at the vertical of the anal ending. This species of a monotypic genus is thought probably extinct (Wildekamp, pers. comm.). It feeds on aquatic prey.

 

- The "deep-bodied" lineage with a strongly compressed body shape and a rather pointed transversal profile of mouth. Three genera may be retained: Procatopus (Plataplochilus and the new western taxon, Rhexipanchax i.e. Mic5_nim, being subgenera, pending further studies), Hypsopanchax and Hylopanchax. This lineage dwells in the lowlands and adjacent foothills of the plateau, preferably in brooks, of the primary or derived forest of western Africa from Senegal to northern Angola, including the Congo Basin. Members of this lineage replace each other allopatrically: Mic5_nim in the peri-Guinean region, Procatopus in Nigeria to Cameroun, Plataplochilus in Gabon to Cabinda, Hylopanchax in the Congo Basin and the Ivindo Basin in Gabon, Hypsopanchax in the surrounding belts of the Congo Basin and to southern Africa, in the Zambezi Basin.

- Procatopus+Plataplochilus+Mic5_nim+Hylopanchax. Synapomorphies: flat predorsal shape; discontinuous internal markings on male fins.

- Procatopus+Plataplochilus+Mic5_nim+Hypsopanchax. Smooth break in the body outline at the vertical fin level. Except for Plataplochilus, shares a supraorbital system made of continuous zig-zag grooves, like also Aplocheilichthys and Mic4_joh.

- Procatopus+Plataplochilus+Mic5_nim. Synapomorphies: red markings on male body and fins (unlike Hylopanchax and Hypsopanchax); upper and lower caudal fin extensions in dominant male. Other characteristics: deep-bodied; higher than Hylopanchax but less so than Hypsopanchax, with no break at the frontal level (unlike Hypsopanchax); with a truncated caudal fin with extensions in male; acuminate dorsal and anal fins in male; and a pointed transversal profile of mouth.

- Procatopus+Plataplochilus. Synapomorphy: the upper extension of caudal fin is longer than the lower in old, dominant males. Shares with Lamprichthys the choice of breeding substratum; eggs are deposited in rock holes. Shares also the presence of branchiostegal appendages in male with Hylopanchax. These 2 lineages are separated from Mic5_nim by the smoother body outline and the larger size, and by the longer caudal fin in dominant male (up to 50% of standard length).

- Procatopus. Autapomorphies: ventral fins, closer to the pectoral fin insertion level; and anal fin with posterior filaments in dominant male. Other characteristics: somewhat larger than Plataplochilus (60 mm T.L. versus 50 mm), with a distinct cephalic sensory system in grooves. Shares with Lamprichthys a very high insertion of the pectoral fins.

- Plataplochilus. Autapomorphy: a wide postero-inferior dark band on sides (irregular). Other characteristics:  all sensory organs below pores and pits; more frequent ctenoidy on male sides and fins (like Lamprichthys); the dorsal insertion is placed more anteriorily (circa 65% of standard length) than in Procatopus (circa 70%). Shares with Lamprichthys the tubular supraorbital system with 4 pores. Shares with Poropanchax (independently acquired) the occurrence of two, separate, longitudinal blue shiny zones on the sides of the male (instead of a continuous hue), but this character is not easy to distinguish and variable, as in terveri ( Plataplochilus), scheeli (Porop3_sch) and normani (Porop2_nor).

- Micropanchax5_nim. See the description of Rhexipanchax below.

- Hypsopanchax+Hylopanchax. Distinctive from Procatopus+Plataplochilus+Mic5_nim by the caudal fin, without extensions; by the conspicuous widened dark reticulation on mid to lower scales of the sides; by the more or less enlarged post-opercular scales in both sexes; and by the rounded caudal fin in the female, different from the shape of the male.

- Hypsopanchax. Autapomorphies: a strong break in the frontal outline of old male and in the lower body outline behind ventrals, which forms a keel; and frequent dark vertical markings (wide-angled chevrons). Other characteristics: the deepest and most compressed lineage, with somewhat enlarged post-opercular scales; no or little body dichromatism; and a rounded caudal fin in male, distinct from Hylopanchax.

- Hylopanchax. Autapomorphies: among non-slender forms, the single lineage with a very low transversal scale count (5 scales); post-opercular scales, strongly enlarged, compared to those at the peduncle level (less strongly in Hypsopanchax). Other characteristics: smaller and less deep-bodied monotypic group, with a truncated caudal fin without extension in the male; and a dark mid-line in anal fin of the male. Shares the presence of branchiostegal appendages in the male with Procatopus+ Plataplochilus. Shares with Laciris, the posterior margin of the dorsal fin, located directly above the anal fin ending.

 

- The "angel" lineage. Synapomorphy: circumcaudal, coloured, wide margin in the male (white to red). Other characteristics: strongly chromatic forms, having all more or less extended fins, especially the ventrals, and exhibiting broken blue metallic zones on sides. The shape of the body is cylindrical to slightly compressed and slender (except Porop2_nor); although the shape of body is similar to the "Sahelian" lineage, a slight allometry in growth occurs (never in the "Sahelian" lineage). Small-sized without any body dimorphism. A narrow transversal profile of mouth. A tubular supraorbital system with 2-3 pores. The caudal fin is oval and elongated in the male with a complete light blue-to-white or red border (sometimes dark-edged, depending on populations). One genus may be retained, but not fully in Clausen's sense: Poropanchax. We propose here to raise Poropanchax to the genus status again, with Congopanchax as a subgenus, pending further studies. This is a presently disjunct distribution. The 4 allopatric sublineages dwell in brackish lagoons and estuaries (Porop3_sch) as well as nearby coastal freshwater creeks not far from the sea (Porop1_mac), and also the Congo-Zaïre Basin (Congopanchax). The single opportunistic species of Porop2_nor, P. normani, has been able to invade the Sahelian lowlands, probably when more humid climatic conditions existed.

- Porop1_mac+Porop2_nor+Porop3_sch (Poropanchax s.s.). These 3 lineages share, with Micropanchax3_hut (independently acquired?), a posteriorily cut anal fin in the female.

- Porop1_mac+Porop2_nor. Long (but not reaching the caudal peduncle) extension in ventral fins of the dominant male, like Aplocheilichthys (independently acquired). Internal part of caudal fin, with light blue blotches.

- Porop1_mac. Autapomorphy: upper and lower margins of the male caudal are acuminate. Shares, with Plataplochilus, the occurrence of two independent longitudinal blue shiny zones on sides of the male (instead of a continuous hue), but this character is not easy to distinguish and may be variable like in Porop3_sch and Porop2_nor.

- Porop2_nor. Distinguished from the 3 other Poropanchax lineages by its somewhat higher depth and by the shape of the vertical fins; continuous (not broken) blue hue on sides. Dark and light markings near the posterior borders of dorsal and anal fins in the male. In most populations, a small, relatively light triangular lobe in the lower part of the male caudal fin (Wildekamp, 1995).

- Porop3_sch+ Congopanchax. Very long (beyond the caudal peduncle) extension in ventral fins of the dominant male; also in the female, but shorter. The caudal fin of the male is rounded or oval, without acuminate margins.

- Porop3_sch. Body shape, similar to the other Poropanchax lineages. Rather continuous, more greenish shiny hue on the sides, which are more prominent on the posterior half. Restricted to brackish coastal biotopes, like Aplocheilichthys, with which it is often sympatric.

- Congopanchax. Five autapomorphies: a low D/A deviation (the lowest in Aplocheilichthyins, with a ratio lower than 4); bilobate dorsal and anal fins in the male, with extensions in anterior and posterior rays; a dwarf size; and a specific food (crustaceans, not insects).

 

-The "vertically patterned" lineage. Autapomorphies: a very pointed transversal profile of the mouth, with an unequal mouth opening and a pin-pointed upper maxillary; a unique pattern of dark grey and golden bars on sides and caudal fin in the male; golden markings on fins, but not in lines across rays like in Lamprichthys; and the mid anal fin rays are extended in the male, whereas the posterior rays are truncated. A single strongly dichromatic component, Aplocheilichthys spilauchen. Other characteristics: large sized, deep-bodied, but not strongly compressed, unlike the "deep-bodied" lineage; a truncate caudal fin shape, without extensions in the male; a marked sexual dimorphism in body shape, the female being less deep, like the "deep-bodied" lineage. Unique to Aplocheilichthyins is the deep peduncle, even in juveniles. Shares with Pantanodon, another brackish monotypic genus, a presumably dominant E-type frontal scalation. Shares, with Lacustricola and Cynopanchax, the lack of a second thin dark line along the lower sides. Dwells in brackish waters in lagoons and rivers along the coast of West Africa, except in the case of large river deltas where it can opportunistically swim much further upstream; often sympatric with Poropanchax scheeli (Porop3_sch); and the pattern of distribution is parallely similar to that of Pantanodon, in East Africa.

- The "East-African" lineage. Moderately dichromatic, rather robust, slightly compressed forms, with a rounded transversal shape of mouth, and rounded, deeply coloured inner fins (unlike other Aplocheilichthyins). A moderate dimorphism in fin shapes. This major lineage comprises Lacustricola, Cynopanchax, and the unnamed Micropanchax3_hut and Micropanchax4_joh. With the external phenotype approach and the present limited knowledge, it is not possible to analyse this more deeply and fix the status of these taxa-groups; if this phylogeny is confirmed, then the oldest taxon Lacustricola will be given in priority. The key colour differentiating characteristics are each groups autapomorphies. Lacustricola and Cynopanchax share, with Aplocheilichthys, the lack of a second thin dark line along the lower sides and exhibit golden-green colours on sides, like Mic3_hut, and which are absent in Mic4_joh. This major lineage dwells in East Africa in tributaries of the Rift Valley lakes, the lakes shores themselves and coastal rivers from Central to South Africa, eastwards from Angola, westwards from Tanzania; also in the Congo belts, but seemingly not the Basin itself.

- Lacustricola. Autapomorphy: a broad longitudinal mid-side dark band. Other characteristics: larger size, with a moderate upward break in the front profile of the male (similar to Hypsopanchax, but much less conspicuous).  Angola and dominantly in western parts of the Rift.

- Cynopanchax. Autapomorphy: a large vertically oblong post-opercular dark blotch. Other characteristics: smaller size, with a short fan-shaped rounded caudal fin. Eastern coastal Africa and neighboring plateau, up to and including the Rift.

- Micropanchax3_hut. Autapomorphy: dark transversal markings on vertical fins, on the mid-posterior region (unlike Poropanchax2_nor and Aplocheilichthys, for which it is confined near to the fin border). Other characteristics: higher depth than the 3 other groups, with a somewhat greater dimorphism in body shape; and mozaic markings on the male sides. Belts of the Congo Basin and the Rift Valley.

- Micropanchax4_joh. Autapomorphy: well separated blue spots in longitudinal series on sides of the male, and also the female, but less numerous (Lamprichthys has also blue spots but they are associated with golden spots and are in the male only). Other characteristics: similar size to Lacustricola, but more slender. South Africa and the southern parts of the Rift.

 

- The "Sahelian" lineage. Synapomorphy: back and front are more melanistic. Other characteristics: slender, not dichromatic and not dimorphic forms exhibit rounded short vertical fins; little colour on the males, excluding the often silvery shiny on sides. There is a tendency for the male to be slightly smaller than the female in Mic_loa, like in Poeciliids, but this should be confirmed on larger samples. No allometry in growth is observed. Only one genus may be retained; this is Micropanchax sensu restricto (Mic.1_loa and Mic.2_kin). The supraorbital sensory system is made of discontinuous zig-zag grooves with buttons. This lineage dwells in the Sahelian savannahs in northern Africa up to the northern Rift Valley lakes and eastern Africa. This is the most widely distributed group, often collected sympatrically with members of the "East-African" lineage.

- Micropanchax1_loa: a short head length, like the primitive lacustrine groups Lamprichthys and Laciris; and the caudal fin is slightly longer in the male.

- Micropanchax2_kin: only distinguished from Mic1_loa by the somewhat deeper body shape and the seven preopercular pores (not a narrow groove); but this is not an autapomorphy, since this state of the sensory system is also seen in other Aplocheilichthyins.

 

- The "Madagascan" lineage with compressed, deep-bodied, brackish fish of the genus Pantanodon. Autapomorphies: uniquely combines a short base dorsal and the largest base anal fin (D= 7-8 and A= 19-20); a marked break in body outline at the dorsal fin level, much stronger than in the "deep-bodied" line; ventral fins with spiny hooks; wing-shaped dorsal fin in the male, posteriorily pointed; anal fin shape distinctive from that of dorsal; pre-orbital and post-orbital systems in a deep groove with embedded buttons; external teeth, absent; a non-equal mouth, with a nearly closed opening; and a filtering food behavior. Shares with Aplocheilichthys, the other brackish monotypic genus, a dominant E-type frontal scalation. Shares with Lamprichthys and Congopanchax, a shorter predorsal length (independently acquired). A deeper anal fin in female, like Mic5_nim. A pronounced difference between the greatest depth and the depth at the peduncle (much stronger than the "deep-bodied" line). This lineage is distributed along the sea-coast of Tanzania, Kenya and eastern Madagascar (Huber, 1998b), with a similar pattern as the monotypic brackish Aplocheilichthys and Porop3_sch from western Africa. Little is known about the two species and live specimens have never been studied and compared: several external characters appear distinctive between the two species, based on the descriptions and the study of preserved material.

 

DESCRIPTION OF RHEXIPANCHAX, N. SUBGEN., A NEW PERIGUINEAN TAXON WITH A BROKEN OUTLINE OF THE LOWER BODY SHAPE

 

- Type species: Haplochilichthys nimbaensis Daget, 1948.

- Synonyms: "Micropanchax" Huber, 1981; unnamed subgenus of Aplocheilichthys s.l. (Wildekamp, 1995).

- Derivatio nominis: from the Greek "Rhexis" (broken) and Panchax, a generic name of related fish, in reference to the body outline. Gender: masculine.

- Components: nimbaensis; lamberti (Daget, 1962); kabae (Daget, 1962); schioetzi (Scheel, 1968); monikae (Berkenkamp & Etzel, 1976), a synonym of schioetzi, fide Huber (1982); and terofali (Berkenkamp & Etzel, 1981), a synonym of schioetzi, fide Wildekamp (1995).

- Diagnosis: a member of the genus Procatopus as defined here within the "deep-bodied" lineage (see supra). Autapomorphy: an additional break in its body outline, at the lower opercle level (hence the name). Rhexipanchax is also defined by a combination of primitive and derived characters. Its smaller size (less than 45 mm in T.L.) in the genus, the unique combination of a low D/A deviation with a high number of anal fin rays, the post-orbital sensory system as a conspicuous groove and, especially in the type-species, a deeper anal fin in the female, like Pantanodon.

- Distribution: the Guinean Dorsale in Senegal, Guinea, Mali, Sierra Leone, Liberia, Ivory Coast, Burkina Faso and Ghana, i.e. west of the Dahomey/Bénin gap; restricted to primary and derived forests: absent in the Sahelian savannah, and rare near the coast where it is replaced by Poropanchax. East of the Dahomey gap, Procatopus takes over allopatrically.

 

 

CONCLUSION

 

This study, based only on external phenotypic characters, is the first global analysis of the African Aplocheilichthyins. It allowed:

- the confirmation of the isolated characteristics of the two pelagic forms and of the specialized, distinctive taxa, Aplocheilichthys and Pantanodon;

- the proposition of the monophyly of the "angel" forms (Poropanchax, Congopanchax), the monophyly of the deep-bodied forms (Procatopus, Hylopanchax, Plataplochilus, Hypsopanchax and Rhexipanchax n. subgen.);

- the reorganisation of the heterogeneous Micropanchax sensu Huber (1981), by restricting it only to two lineages of closely related species, allied to loati and kingii;

- the re-establishment of taxa that were abandoned, like Lacustricola and Cynopanchax, because they correspond to well-separated lineages of evolution; Lacustricola being eventually the available older taxon to be used as a generic name, if these results are verified.

- the proposition of a consensus tree of 7 lineages in accordance with the biogeography;

- the proposition of future exploration of three insufficiently understood positions; those of Pantanodon, Aplocheilichthys and the species johnstoni and its allies.

            Like in the preceding twin study for the genus Rivulus and its allies, the processing of all data using PAUP results in a consensus tree of two very similar minimal trees, with some strong branches. Like for Rivulus, it proposes for the first time a relationship among isomorphic or heteromorphic superspecies that conventional previous methods could only at best separate. However, contrary to the preceding study, no other in-depth comprehensive study of the Aplocheilichthyins is available and thus the inferred phylogenetic relationships need to be corroborated by the two other approaches, namely the internal phenotype and the genotype. Corroboration is likely, since the present findings do not contradict the known biogeography of these fishes, at least in the general framework of the tree. If this is not the case, then the matrix will be at least useful for identification purposes and new characters will have to be established and processed in order to reach reconciliation.

 

 

Acknowledgements. - This study would not have been possible without the generous help of the Paris Museum staff, familiar with the details of PAUP: Véronique Barriel during its first stage, Guillaume Lecointre for the initial spur and during the final and detailed processing and analysis. Both are warmly thanked.

Out of courtesy, the manuscript has been sent to experts in the group who currently work on the two other basic aspects: W.J.E.M. Costa (Rio de Janeiro) and M.J. Ghedotti (Lawrence) in osteology and G.C. Collier (Tulsa) in molecular biology, and also to R.H. Wildekamp (Gemert) and L. Seegers (Dienslaken), who pursue a morphological study of each species. I wish here to thank them for their encouragements and remarks. The manuscript has benefitted from the remarks of two anonymous reviewers, who are also thanked.

 

 

REFERENCES

 

Ahl E., 1928. - Beiträge zur Systematik der afrikanische Zahnkarpfen. Zool. Anz., 79(3-4): 113-123.

Clausen H.S, 1967. - Tropical Old World Cyprinodonts. 64 p. Akademisk Forlag, Copenhagen, Denmark.

Costa W.J.E.M., 1996. - Relationships, monophyly and three new species of the neotropical miniature poeciliid genus Fluviphylax (Cyprinodontiformes: Cyprinodontidei). Ichthyol. Explor. Freshw., 7(2): 111-130.

Huber J.H., 1981. - A review of the cyprinodont fauna of the coastal plain in Rio Muni, Gabon, Congo, Cabinda and Zaïre, with taxonomic shifts in Aphyosemion, Epiplatys and West African Procatopodins. 46 p. B.K.A. Publ.

Huber J.H., 1982. - Cyprinodontidés récoltés en Côte d'Ivoire (1974-1978). Cybium, 6(2): 49-74.

Huber J.H., 1992. - Review of Rivulus. Ecobiogeography - Relationships. 586 p. Soc. fr. Ichtyologie Ed., Paris.

Huber J.H., 1996. - Killi-Data 1996. Updated checklist of taxonomic names, collecting localities and bibliographic references of oviparous Cyprinodont fishes (Atherinomorpha). 400 p; in French, English, and German. Soc. fr. Ichtyologie, Ed., Paris.

Huber J.H., 1999a. - Updates to the phylogeny and systematics of the neotropical cyprinodont genus Rivulus and its allies (Cyprinodontiformes, Rivulinae). Cybium, 23(1).

Huber J.H., 1999b. - The estuarian cyprinodont angel, Poropanchax scheeli (Romand, 1970) (Aplocheilichthyinae, Cyprinodontiformes): its first discovery from Gabon, with further insights on African lampeyes systematics and on specialized cyprinodonts in brackish waters (F.A.M.A., in press).

Lambert J.G. & H.S. Clausen, 1967. - The genus Plataplochilus Ahl, 1928 redefined (Pisces, Cyprinodontidae). Rev. Zool. Bot. Afr., 76: 392-396.

Meyer A. & C. Lydeard, 1993. - The evolution of the copulatory organs, internal fertilization, placentae and viviparity in killifishes (Cyprinodontiformes), inferred from a DNA phylogeny of the thyrosine kinase gene X-src. Proc. R. Soc. London, B, 254: 153-162.

Myers G.S., 1931. - The primary groups of oviparous cyprinodont fishes. Stanford Univ. Publ. (Biol. Sci.), 6(3): 245-254.

Myers G.S., 1938. - Studies on the Genera of Aplocheilichthys and its Allied. Copeia: 136-143.

Parenti L.R., 1981. - A phylogenetic analysis of cyprinodontiformes fishes (Teleostei; Atherinomorpha). Bull. Am. Mus. Nat. Hist., 168: 335-557.

Poll M., 1971. - Un genre nouveau et une espèce nouvelle de Cyprinodontidae congolais. Rev. Zool. Bot. Afr., 83: 302-308.

Wildekamp R.H., 1995. - A world of killies. Atlas of the oviparous cyprinodontiformes fishes of the world. Vol. 2. 384 p. Amer. Killifish Ass. Publ.

 

Appendix 1.- List of studied material with their older generic names (corresponding new names are given in table I). * denotes material already studied by Huber, 1981.

 

Fundulus zebrinus, MNHN 1985-109, 5 specimens, Arikaree Riv., Kansas, USA.; Aplocheilichthys spilauchen (types), MNHN 2936 and 1997-4083, 3 spms, near Lambaréné, Gabon; Aplocheilichthys spilauchen, MNHN 1967-237, 5 spms, Zambi, Congo; Aplocheilichthys spilauchen, MNHN 1967-240, 1 spm., Sounda, Congo; Aplocheilichthys spilauchen*, MNHN 1963-445, 14 spms, Lamto, Côte d'Ivoire; Aplocheilichthys spilauchen*, MNHN 1967-234, 49 spms, Kayes, Congo; Aplocheilichthys spilauchen, MNHN 1967-236 & 239, 27 spms, Lac Noumbi, Congo; Aplocheilichthys spilauchen, MNHN 1979-234, 2 spms, Lokundjé, Cameroun; Aplocheilichthys spilauchen, MNHN 1982-10, 2 spms, Mayumba, Gabon; Aplocheilichthys spilauchen, MNHN 1961-690, 18 spms, Kassagna, Guinée; Aplocheilichthys spilauchen*, MRAC 73-39-P-2187-88, 2 spms, Lagos, Nigeria; Congopanchax brichardi* (types), MRAC 168566-73, 8 spms, Boendé, Zaïre; Congopanchax myersi* (types), MRAC 86801-802, 77394, 3 spm., Stanley Pool, Zaïre; Congopanchax myersi, MNHN 1958-125, 1 spm., Stanley Pool, Zaïre; Hypsopanchax catenatus (types), MRAC 80-54-P-1503-35, 4 spms, Franceville-Boumango, Gabon; Hypsopanchax deprimozi*, MRAC 64305-524, 15 spms, Rutchuru, Lac Edouard, Zaïre; Hypsopanchax deprimozi (types), MNHN 1927-339, Rutchuru, Lac Edouard, Zaïre; Hypsopanchax jubbi (types), MRAC 142614, 142640-46, 4 spms, Upper Zambezi, Lula, Zimbabwe; Hypsopanchax platysternus*, MRAC 129931-991, 10 spms, Kitutu, Zaïre; Hypsopanchax zebra*, MNHN 1930-162 & 163, 14 spms, Djouéli, Congo; Hypsopanchax zebra, MNHN 1982-830, 4 spms, Zanaga, Congo; Hypsopanchax zebra (types), MNHN 1929-228, 6 spms, Zanaga, Congo; Laciris pelagicus*, MRAC 142810-14, 4 spms, Lac Edouard, Zaïre; Lamprichthys tanganicanus*, MRAC 90488-493, 6 spms, Albertville, Tanganyika, Zaïre; Lamprichthys tanganicanus, MNHN 1985-697, 3 spms, Tanganyika, Burundi; Micropanchax antinorii*, MNHN 1904-243, spms, British Eastern Africa; Micropanchax bracheti (types), MRAC 73-11-P-847-850, 4 spms, Namoundjoka, Togo; Micropanchax camerunensis*, MRAC 75-38-P-1179-212, 20 spms, near Sangmelima, Cameroun; Micropanchax camerunensis*, MRAC 75-38-P-2672-703, 30 spms, near Ayos, Cameroun; Micropanchax dispar (types), MNHN 1960-393, 10 spms, Sakété, Bénin; Micropanchax eduardensis, MRAC 38836-40, 4 spms, Vitshumbi, Lac Edouard, Zaïre; Micropanchax hutereaui, MNHN 1936-108, 1 spm., Chiumbe, Angola; Micropanchax hutereaui, MNHN 1979-562, 4 spms, Yalouké, Centrafrique; Micropanchax jeanneli (types), MNHN 1933-112, 5 spms, Lac Rodolphe, Ethiopie; Micropanchax johnstoni, MRAC 73-39-P-1954-1969, 16 spms Kampala-Masindi, Uganda; Micropanchax kabae (types), MNHN 1959-116, 6 spms, Mamou, Guinée; Micropanchax katangae*, MRAC 150140-180, 4 spms, Katanga, Zaïre; Micropanchax kingii (types), MNHN 1959-352, 4 spms,Nganatir, Lac Tchad; Micropanchax kongoranensis, MNHN 1983-259, 10 spms, Kibiti, Tanzania; Micropanchax lamberti* (types), MNHN 1960-464, 77 spms, Ballay, Guinée; Micropanchax lamberti (types), MNHN 1959-118, 4 spms, Guinée; Micropanchax lualabaensis (types), MRAC 49991-98, 4 spms, Lualaba, Zaïre; Micropanchax macrophthalmus, MNHN 1985-1096, 13 spms, Mouanko, Cameroun; Micropanchax mahagiensis (type), MRAC 21030, 1 spm., Mahagi, Lac Albert, Zaïre; Micropanchax mediolateralis (types), MRAC 162489-91, 3 spms, Muquehe, Angola; Micropanchax nigrolateralis (type), MRAC 162484, 1 spm., Riv. Tchimenji, Angola; Micropanchax nimbaensis*, MNHN 1949-57 (type), 1 spm., Mount Nimba, Guinée; Micropanchax nimbaensis*, MNHN 1957-39, 40, 41, 42, 1960-207, 7 spms, Mount Nimba, Guinée; Micropanchax nimbaensis, MNHN 1990-1383, 6 spms, Mount Nimba, Guinée; Micropanchax nimbaensis, MNHN 1991-523, 12 spms, Mount Nimba, Guinée; Micropanchax normani*, MNHN 1960-126, 26 spms, Burkina; Micropanchax normani, MNHN 1981-672, 10 spms, Béré, Côte d'Ivoire; Micropanchax normani*, MNHN 1961-674, 28 spms, Banharé, Sénégal; Micropanchax omoculatus (types), MRAC 77-16-P-60-63, 4 spms, Mbeya-Morogoro, Tanzania; Micropanchax pfaffi*, MNHN 1963-444, 12 spms, Korhogo, Côte d'Ivoire; Micropanchax pfaffi (types), MNHN 1960-417 & 1997-55, 7 spms, Diafarabé, Niger; Micropanchax pfaffi (types), MNHN 1960-422, 10 spms, Kokry, Soudan; Micropanchax pfaffi, MNHN 1981-690, 31 spms, Madji, Côte d'Ivoire; Micropanchax pfaffi, MNHN 1981-683, 25 spms Mbingué, Côte d'Ivoire; Micropanchax pumilus*, MRAC 10944-82, 10 spms, Tanganyika, Zaïre; Micropanchax rancureli (types), MNHN 1964-277, 14 spms, Azaguié, Côte d'Ivoire; Micropanchax scheeli, MNHN 1997-181 & 182, 2 spms and MNHN 1998-483, 30 spm., Animba River, Gabon; Micropanchax schioetzi*, MNHN 1963-441, 68 spms, Bouaké, Côte d'Ivoire; Micropanchax schioetzi, MNHN 1981-655, 4 spms, Riv. M'pédo, Côte d'Ivoire; Micropanchax schioetzi, MNHN 1981-657, 22 spms, Bouaké, Côte d'Ivoire; Micropanchax schioetzi* (types), MRAC 164478-81, 4 spms, Kumasi, Ghana; Micropanchax silvestris*, MRAC 79121-145, 10 spms, Ibembo, Zaïre; Micropanchax silvestris*, MRAC 73-23-P-12868-14275, 40 spms, Boendé, Zaïre; Micropanchax silvestris* (types), MRAC 124066-075, 10 spms, Yangambi, Zaïre; Micropanchax stictopleuron, MNHN 1981-1412, 7 spms, Débrouillé, Congo; Micropanchax stictopleuron, MNHN 1972-72, 11 spms, Makokou, Gabon; Micropanchax usanguensis (types), MRAC 77-16-P-64-78, 3 spms, Ilongo, Tanzania; Pantanodon madagascariensis, MNHN 1963-166 (types), 5 spms, Mahambo, Madagascar; Procatopus abbreviatus, MNHN 1929-88 (types), 7 spms, Yabassi, Cameroun; Procatopus cabindae*, MNHN 1967-228, 63 spms, Tombo, road to Sounda, Congo; Procatopus cabindae*, MNHN 1967-230, 18 spms, Trib. to Kouilou, road to Sounda, Congo; Procatopus cabindae*, MNHN 1967-231, 5 spms, Djembo, Congo; Procatopus cabindae*, MNHN 1967-232, 34 spms, Zibati, Congo; Procatopus cabindae, MNHN 1967-225, 16 spms, Les Saras, Congo; Procatopus cabindae, MNHN 1967-226, 51 spms, Kondé, Congo; Procatopus chalcopyrus* (types), MRAC 140290-311, 20 spms, road Lambaréné-Libreville, Gabon; Procatopus loemensis* (types), MNHN 1924-87, 2 spms, Lombo, Congo; Procatopus miltotaenia*, MRAC 80-27-P-234-237, 4 spms, Lambaréné-Mouila, Gabon; Procatopus miltotaenia (types), MRAC 140312-42, 3 spms, Lambaréné-Mouila, Gabon; Procatopus mimus* (types), MRAC 153265-78, 4 spms, Libreville, Gabon; Procatopus ngaensis* (type), MNHN 1936-36, 1 spm., Attogondema, Gabon; Procatopus ngaensis, MNHN 1982-13, 22 spms, Ntoum, Gabon; Procatopus nototaenia*, MNHN 1979-349, 2 spms, Lokundjé, Cameroun; Procatopus pulcher* (types), MRAC 164734-742, 5 spms, Libreville-Kango, Gabon; Procatopus similis, MNHN 1985-597, 2 spms, Somakek, Cameroun; Procatopus terveri* (types), MNHN 1930-50 & 51, 9 spms, riv. La Passa, Gabon; Procatopus terveri*, MRAC 20203-204, 2 spms, riv. La Passa, Gabon.

 

Fig. 1.- Strict consensus of two equiparsimonious trees of 236 steps (C.I.= 0.555, R.I.= 0.585) obtained using PAUP 3.1.1 and based on 74 externals characters. Numbers above branches are bootstrap proportions obtained from 1000 replicates.

Table I.- List of the presumably valid species with their old names, with their assignment to each superspecies in the matrix, with the number of studied specimens and the following major morphomeristic data: maximum length in mm, mean of dorsal fin rays, anal fin rays, D/A deviation, lateral and transversal scales, predorsal length, head length, depth at anal level and maximum caudal fin length, as a % of S.L. The index is computed by the formula "(D+A+D/A)*LL". The proposed new names are given with an asterisk when this is conditioned by the confirmation of additional studies.

 

 

Table II.- List of the 74 characters and their up to 5 states, for Aplocheilichthyins.

1- Max. size: large, over 65 mm=0; medium sized, between 65 and 45 mm=1; small, between 45 and 25 mm=2; dwarf, under 25 mm=3

2- Female larger than male: yes=0; no=1

3- Body shape: cylindrical to slightly compressed=0; moderately compressed=1; strongly compressed=2

4- Body shape at lower opercle level: forming a continuous line=0; showing a disruption=1

5- Front shape of old males behind eyes: without a discontinuity=0; with a small upwards break=1; with a strong disruption=2

6- Body shape behind Ventrals level: slightly curved=0; forming a pronounced keel=1

7- Body outline at vertical fins insertion: continuous=0; forming a break=1

8- Body depth at Anal level of fully adult males: relatively slender (under 23% standard length)=0; deep-bodied (23 to 30%)=1; overheightened (over 30%)=2

9- Predorsal shape: oval, convex=0; flat=1

10- Body shape at peduncle level: nearly as high as at Anal level=0; with a moderate difference=1; with a strong difference=2

11- Strong peduncle, even in juvenile specimens: yes=0; no=1

12- Mouth transversal shape: downward, rather rounded=0; upward, rather rounded=1; upward, narrow=2; upward, rather pointed=3; upward, very pointed=4

13- Mouth opening: equal, slightly oblique=0; subequal, with a pointed upper part=1; non equal, the upper part being shortened=2; non equal, with a nearly closed opening=3

14- Eye diameter: small=0; large=1

15- Long preorbital length, equal to eye: no=0; yes=1

16- Branchiostegal appendages in adult male: no=0; yes=1

17- Frontal sensory supraorbital structure: in tubular channels=0; in two continuous straight grooves=1; in two continuous zig-zag grooves=2; in two discontinuous zig-zag grooves=3

18- Frontal sensory system: hidden beneath 4 pores=0; hidden beneath 3 pores=1; hidden beneath 2 pores=2; exposed neuromasts=3; strongly reduced to buttons=4

19- Pre-opercular sensory system: tubular with pores=0; in a groove with exposed neuromasts or buttons=1; vestigial or not visible=2

20- Pre-orbital sensory system: tubular with three pores=0; with a groove embedding reduced buttons=1

21- Post-orbital (dermosphenotic) sensory system: reduced to a channel with two pores=0; with a conspicuous groove=1

22- Frontal scalation pattern (major tendency): A or E-type=0; G-type=1; H-type (i.e. no type)=2

23- Post-operculum side scales: with similar size than other mid-sides scales=0; somewhat enlarged=1; strongly enlarged=2

24- External teeth: present=0; absent=1

25- Pectoral fins relative level of insertion: low=0; rather median=1; high=2; very high=3

26- Head length in % S.L.: relatively long, over 24%=0; relatively short, under 24%=1

27- Predorsal length in % S.L.: relatively short, under 64%=0; average, between 64% and 70%=1; relatively long, over 70%=2

28- Average Dorsal fin count: over 10=0; between 10 and 7=1; under 7=2

29- Average Anal fin count: relatively high, over 15=0; average, between 15 and 11=1; relatively low, under 11=2

30- Average D/A deviation: under 0=0; between 1 and 3=1; between 4 and 7=2; between 7 and 10=3; over 10=4

31- Ventral fins relative position to Pectoral level and Anal insertion: closer to Anal fin=0; near to the middle=1; closer to Pectoral level=2

32- Ventral fins extremity with spiny hooks in male: no=0; yes=1

33- Average LL scales count: over 35=0; under 35=1

34- Average circumpeduncular scales count: over 17=0; between 13 and 17=1; under 13=2

35- Additional longitudinal scales behind hypural plate: over 4=0; under 3=1

36- Regular hexagonal drawing of exposed edges of side scales: no=0; yes=1

37- Average transversal (TRAV.) scales count in an oblique row below Dorsal: over 10=0; between 7 and 10=1; under 7=2

38- Sexual dimorphism in body shape: weak=0; strong=1

39- Sexual dimorphism in Dorsal, Anal and Ventral fins shape: weak=0; moderate=1; strong=2

40- Ventral Fin extensions in male: no or limited=0; yes, beyond Anal insertion=1; yes, very long to Caudal peduncle=2

41- Anal Fin extensions of mid-posterior rays in male: no=0; yes, short=1; yes, filamentous=2

42- Dorsal fin shape tip in male: rounded or slightly pointed=0; straight and acuminate=1; with streamers=2; posteriorily truncate=3; wing-shaped, asymetrical to Anal=4

43- Dorsal to Anal comparison in male: similarly shaped=0; distinct with bilobate Dorsal= 1; distinct with deep rounded Anal=2

44- Anal fin border in male: rounded=0; slightly curved=1; rather straight, comb-tipped=2; parallely cut=3

45- Anal fin shape in female: posteriorily rounded=0; posteriorily cut=1; straight and posteriorily pointed=2; deep=3

46- Caudal Fin shape in male: rounded, short=0; truncate=1; oval, elongate=2; oval and pointed in the middle=3

47- Caudal fin extension in male: none=0; with short upper and lower extensions=1; with a longer upper streamer=2; crescent-shaped=3

48- Caudal fin posterior margin in female: like male=0; distinct (rounded)=1

49- Iris with a definite iridescence (lampeye) in both sexes: no=0; yes=1

50- A thin black line on mid-sides (even after preservation): no=0; yes=1

51- A thin black line on lower sides from Anal insertion (even after preservation): no=0; yes=1

52- Sexual dichromatism (excluding fins): strong (distinctive)=0; weak, subdued=1; none=2

53- Back and front, with melanistic markings: no=0; yes=1

54- Fine dark reticulation, more or less regularly in chevrons, on scales of sides: no=0; yes=1

55- Widened dark reticulation on mid to lower anterior sides: no=0; yes=1

56- Post-orbital lightly colored plaque: no=0; yes, silvery=1

57- Post-opercular dark upper vertical marking: no=0; yes=1

58- Dark longitudinal broad band on sides: none=0; yes, faint on upper mid sides=1; yes, median and gray=2; yes, median and conspicuous in both sexes=3

59- Vertical dark bars for the male: prominent on sides only=0; gray, on sides and fins=1; none=2

60- Blue hue on the male sides: no=0; yes, rather grey to silver=1; rather yellow to green=2; overall blue=3; blue, with 2-3 separate longitudinal bands=4

61- Series of blue longitudinal spots on scales center of the male sides: no=0; yes, continuous=1; yes, discontinuous=2

62- Golden markings on the male sides and fins: no=0; yes, scattered mainly on lower sides=1; yes, mainly on fins in parallel series=2

63- Prominent red markings on the male sides and fins: no=0; yes= 1

64- Cross-rays dark reticulations in the male posterior Dorsal and Anal fins: no=0; yes=1

65- Coloured markings in mid-vertical fins: no=0; yes, spotted lines=1; yes, wide band=2

66- Broad black border on Caudal of Male: no=0; yes=1

67- Circumcaudal coloured margin in male (light blue to orange): no=0; yes=1

68- Ecology: strictly in freshwaters=0; also in brackish waters=1

69- Ecology preference: large rivers=0; creeks, small bodies of water=1; endemic to lakes=2

70- Swimming: open water=0; not far from surface=1

71- Biotope: in savannah=0; in primary or derived forest=1

72- Behaviour: forms small groups=0; gregarious in large schools=1

73- Behaviour, breeding: among plants=0; in rock holes=1; pelagic=2

74- Food preference: insects and their larves=0; crustaceans/aquatic prey=1

 

Table III.- Matrix of 20 groups (the outgroup and the 19 aplocheilichthyin groups) and 74 characters.

---

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