Version classiqueVersion mobile

The inland water fishes of Africa

 | 
Didier Paugy
, 
Christian Levêque
, 
Olga Otero

Taxonomy and systematics

Didier Paugy et Christian Levêque

Résumé

Taxonomy is the science of naming and describing organisms. It is a highly formalized science whose rules are established and revised regularly by an International Commission of Zoological Nomenclature created in 1895. The International Code of Zoological Nomenclature contains all the rules for designating animal species which are the same for all zoological groups.
One of its basic principles is that the first name given to a species has priority over the others. It is not uncommon for taxonomic revisions of a group to reveal that some species had been described under different names, because the authors did not have access to all available information or had made an error. In these conditions, the so-called priority rule must be applied. This sometimes leads to a change in the name used for the species up to that time. The situation may prove even more complex, as illustrated by the case of Schilbe mystus (see box “Who is Schilbe mystus?”).
Another well-known case of ambiguity is that of the mounted holotype of Synodontis xiphias described in 1864 by Günther, which has a long and pointed snout similar to a swordfish. The species had never been recorded since then, for a perfectly good reason: an x-ray showed that the long snout was a fake nose, resulting from a metal frame used in mounting the animal (Poll, 1971) (figure 8.1). The holotype of Synodontis labeo Günther, 1865, described later, belongs to the same species, but the name S. xiphias remains valid in application of the principle of priority, and despite the deformity of the specimen retained as the type.
Systematics is the study of the diversity of organisms and the relationships between these organisms. Its goal is to classify species and detect phylogenies – a purpose that differs from the objectives of taxonomy. Classification consists of recognizing and defining groups or taxa (that is, a set of organisms that share a specific character) that taxonomists will then have to name. According to the hierarchical classification proposed by Linnaeus (1758), each level of the hierarchy corresponds to a taxon name. While ideas governing classification in particular have evolved significantly since the 18th century, the basic rules stated by Linnaeus remain solid. The discipline currently known as biosystematics is a modern approach to taxonomy and phylogeny that makes use of information from different sources: morphology, genetics, biology, parasite specificity, behaviour, and ecology. This type of approach is undoubtedly set to develop further.

Texte intégral

Who is Schilbe mystus?
After examining the presumed holotype of Schilbe mystus, De Vos & Skelton (1990) showed that the specimen, described in 1758 by Linnaeus, actually belonged to the same species as the one named S. (Eutropius) niloticus described later in 1829 by Rüppell.
If the two species are synonyms, application of the International Code of Zoological Nomenclature means that the older name S. mystus takes priority over S. niloticus. In keeping with this, all fishes that had been identified as S. niloticus up to that time need to be renamed S. mystus.
While the holotype of S. mystus is a synonym of S. niloticus, this is not the case for all the other fishes identified up to then as S. mystus but different from S. niloticus and the S. mystus type. What species do they belong to? Among the species erroneously considered a synonym of the S. mystus type, the oldest is S. intermedius described by Rüppell in 1832, and whose description matches that of the ex-mystus. As a result, fishes described before 1990 under the name S. mystus must be renamed S. intermedius.
The application of perfectly reasonable taxonomic rules can therefore lead to a particularly confusing situation for non-specialists. The name S. mystus is now used for a species that was widely cited in literature under the name S. niloticus, and the former S. mystus take the name S. intermedius.
However, this change in nomenclature is not always rigorously applied by everyone everywhere, to the point that one often wonders if the name being used is the former or the post-revision name.
In this specific case, the legendary formal discipline of the taxonomist did not take into account the confusion that would result for ichthyologists, biologists and ecologists alike.

FIGURE 8.1. Head radiography of the holotype of Synodontis xiphias (from Poll, 1971).

FIGURE 8.1. Head radiography of the holotype of Synodontis xiphias (from Poll, 1971).

What is a species?

1The binomial system originally proposed by Linnaeus (1758) is used by naturalists the world over to designate and identify species. But the concept itself of “species” has long been debated. To date, no definition is entirely satisfactory.

The typological species concept

2Until the mid-19th century, systematists viewed species as fixed. They were as God had created them, immutable and limited in number. The purpose of taxonomy then was to draw up an inventory of all existing life forms and describe their specific characters. In the words of Linnaeus, “(…) there are as many species as the infinite being created diverse forms (…)”, and he formalized this view by establishing the species through a type individual, the holotype (see box “International Code of Zoological Nomenclature”), to which a Latin binomial was assigned in order to identify and classify them.

Extract of the International Code of Zoological Nomenclature fourth edition
adopted by the XXth General Assembly of the International Union of the Biological Sciences December 1999
type, n. A term used alone, or forming part of a compound term, to denote a particular kind of specimen or taxon.
allotype, n. A term, not regulated by the Code, for a designated specimen of opposite sex to the holotype [Recommendation 72A].
cotype, n. A term not recognized by the Code, formerly used for either syntype or paratype, but that should not now be used in zoological nomenclature [Recommendation 73E].
genotype, n. A term not recognized by the Code, formerly used for type species, but that should not now be used in zoological nomenclature [Recommendation 67A].
hapantotype, n. One or more preparations consisting of directly related individuals representing distinct stages in the life cycle, which together form the name-bearing type in an extant species of protistan [Arts 72.5.4, 73.3].
A hapantotype, while a series of individuals, is a holotype that must not be restricted by lectotype selection; however, if a hapantotype is found to contain individuals of more than one species, components may be excluded until it contains individuals of only one species [Art. 73.3.2].
holotype, n. The single specimen (except in the case of a hapantotype, q.v.) designated or otherwise fixed as the name-bearing type of a nominal species or subspecies when the nominal taxon is established.
lectotype, n. A syntype designated as the single name-bearing type specimen subsequent to the establishment of a nominal species or subspecies [Art. 74].
neotype, n. The single specimen designated as the name-bearing type of a nominal species or subspecies when there is a need to define the nominal taxon objectively and no name-bearing type is believed to be extant. If stability and universality are threatened, because an existing name-bearing type is either taxonomically inadequate or not in accord with the prevailing usage of a name, the Commission may use its plenary power to set aside that type and designate a neotype.
paralectotype, n. Each specimen of a former syntype series remaining after the designation of a lectotype [Art. 72.1.3, Recommendation 74F].
paratype, n. Each specimen of a type series other than the holotype [Recommendation 73D].
syntype, n. Each specimen of a type series (q.v.) from which neither a holotype nor a lectotype has been designated [Arts. 72.1.2, 73.2, 74]. The syntypes collectively constitute the name-bearing type.
topotype, n.(topotypic, a.). A term, not regulated by the Code, for a specimen originating from the type locality of the species or subspecies to which it is thought to belong, whether or not the specimen is part of the type series.

3For practical reasons, the typological species concept is still used to this day: a species is a set of individuals that are identical to each other and to the “type” specimen, i. e., the specimen used to describe and characterize the species in morphological terms. This type is deposited in a museum where it serves as a reference or a sort of standard for comparisons or future revisions.

The biological species concept

4The typological species concept gradually gave way in the early 20th century to the concept of a biological species. In 1798, Cuvier had already used a biological criterion in defining a species as “the collection of all organized bodies, borne of each other or common parents, and of all those that resemble them as much as they resemble each other.” The definition given by Mayr (1942) a century and a half later is not fundamentally different but emphasizes the biological aspect: species are groups of populations whose members can interbreed, and who are reproductively isolated from other groups. One of the major criteria for differentiating neighbouring species was the fact that any eventual hybrids were not viable or were sterile. “The origin of species is therefore simply the evolution of some difference – any difference at all – that prevents the production of fertile hybrids between populations under natural conditions” (Wilson, 1992).

5On the genetic level, the concept of a biological species implies the existence of a gene pool that could recombine within the population during sexual reproduction, but that this gene pool is somehow “protected” from mixing with other pools by biological, physiological, or behavioural mechanisms.

6While the concept of biological species is not debatable, it is nonetheless difficult to apply in fishes, in the sense that systematic cross-breeding trials in various natural populations are particularly difficult to carry out.

Concept of specific-mate recognition system

7In reaction to the concept of reproductive isolation that undergirds that of the biological species, Paterson (1985) considers that there is no obvious reason for species evolving independently in isolated geographic areas to find it advantageous to develop mechanisms that lead to reproductive isolation.

8The concept of a specific-mate recognition system, meanwhile, is based on the premise that conspecific sexual partners must share specific characters, such as a co-adapted set of signals and responses between males and females of the same species during their mating rituals, in order to meet and ensure fertilization of gametes (Ribbink, 1988). A species is thus made up of individuals possessing the same system of fertilization, that is, the set of adaptations that encourage and facilitate the encounter between sexual partners to ensure reproduction: visual, chemical, or sonic recognition signals; colouring; mating dances; synchronous reproductive periods; preference for the same types of habitat and laying sites; territoriality, etc.

9This concept is an interesting alternative. The existence of selective recognition systems for conspecific partners, rather than barriers to hybridization, serves as an important criterion for speciation. In other words, we are in the realm of ethology.

Sister species

10Sister species are biological species that have achieved reproductive isolation but are still difficult to discern based on morphology alone. These species probably result most frequently from recent speciation.

11Here are two examples that provide a good illustration of how some recent techniques can help separate species with similar phenotypes and that probably thus result from fairly recent speciation.

  • Labeo from Senegal and Niger. In the upper course of these two basins, we found numerous specimens that were intermediate between L. senegalensis and L. coubie, and initially thought they could be hybrids. While the shape is closer to L. senegalensis in terms of general morphology, the buccal anatomy was closer to L. coubie. Similarly, the number of gill rakers on the first gill arch was not conclusive and also suggested a hybrid form. Comparative analysis of gill parasite fauna (Monogeneans) of the three forms of Labeo made it possible to determine that the specimens were not hybrids but a separate valid species that had specific parasite fauna. Genetic analysis of the three forms using protein electrophoresis also lifted all doubt about the true specificity of the presumed hybrid. It was named L. roseopunctatus (Paugy et al., 1990) (figure 8.2) for its distinctive colour.
  • Petrocephalus from the Niger. Three Petrocephalus forms found in the upper Niger basin have identical metric and meristic characteristics and only differ in their colour pattern:
    • Form A: uniformly silver colouring.
    • Form B: first dorsal fin rays strongly black; black spot under and touching the base of the dorsal fin.
    • Form C: grey spot under and not touching the dorsal fin.

FIGURE 8.2. General morphology of three sympatric species of Labeo from Western Africa (from Paugy et al., 1994).

FIGURE 8.2. General morphology of three sympatric species of Labeo from Western Africa (from Paugy et al., 1994).

12Form A, long known to science, corresponds to P. bovei, a widely-distributed species but whose morphology – aside from colouring – does not differ from the other two forms. To draw a conclusion on the possible specificity of the three forms, the characteristics of the electrical discharges of these Mormyridae were studied. While forms A and C could not be distinguished, individuals of group B showed different characteristics. Also, while the type of discharge could not distinguish A and C, genetic separations confirmed the specific level of these three forms (Agnèse & Bigorne, 1992). In addition to the already identified P. bovei, two new species, P. soudanensis and P. pallidomaculatus were described with the help of new techniques that allowed recognition of the specific validity of these two other forms (Bigorne & Paugy, 1991) (figure 8.3).

FIGURE 8.3. General morphology of three sympatric species of Petrocephalus from Western Africa (from Paugy et al., 1994).

FIGURE 8.3. General morphology of three sympatric species of Petrocephalus from Western Africa (from Paugy et al., 1994).

Polytypic species

13In a species’ area of distribution, environmental factors are generally not uniform. Populations can react locally to changes in these factors through morphological changes. The concept of a polytypic species refers to a species that can present several forms and/or which is composed of several subspecies that may show differences on the morphological, physiological, ecological or behavioural level.

14Recognizing variability within a population can lead to two positions between which ichthylogists have long hesitated. Either “varieties” are local adaptations of the same species, or they are truly different species. Thus, for the species Brycinus nurse common in the Sahelian zone, there are two known lacustrine dwarf populations, each sympatric with riverine populations: B. nurse dageti in Lake Chad and B. nurse nana in Lake Turkana. The first subspecies was initially described as a different species, B. dageti (see Paugy, 1986). There are also numerous examples of polytypic species among the Cyprinidae (Lévêque, 1989a and b) and Mormyridae (Bigorne, 1987 and 1989; Bigorne & Paugy, 1991). In reality, given that hydrographic basins currently act as geographic islands, a species is most often split in several isolated populations with little or no contact among them (figure 8.4). In such conditions, it is unsurprising for isolated populations to begin developing certain adaptations in response to their environmental conditions, which would explain the variability observed.

FIGURE 8.4. Distribution of B. imberi, B. abeli, B. carolinae and B. nigricauda.

FIGURE 8.4. Distribution of B. imberi, B. abeli, B. carolinae and B. nigricauda.

15In tropical zones, several studies have shown the influence of food habits on the morphological variations of fishes (see chapter Diets and food webs).

Geographical clines

16In a terrestrial physical system, we have observed that when environmental factors vary gradually, there may also be gradual changes in certain morphological or morphometric characteristics in the populations of a species. These geographic variations result from the response of organisms to environmental conditions and/or the limitation of gene exchanges between isolated populations.

17Endler (1977) defined cline as follows: “a geographic gradient in a measurable character, or gradient in gene, genotype, or phenotype frequency.” Thus, phenotype variability or variability in morphological or meristic characters observed in a species can be considered a cline.

Regional scale

18In the upper Niger basin, the number of soft rays in the anal and dorsal fins of Petrocephalus bovei (Bigorne et al., unpublished) changes along an upstream-downstream gradient (figure 8.5). There is an average difference of two rays between extreme populations. No solid explanation has been given for this phenomenon.

FIGURE 8.5. Petrocephalus bovei: clinal changes in two meristic characters (average) along an upstream-downstream gradient (Upper Niger).

FIGURE 8.5. Petrocephalus bovei: clinal changes in two meristic characters (average) along an upstream-downstream gradient (Upper Niger).

West African scale: Schilbe mandibularis

19Schilbe mandibularis (Schilbeidae) is found in the Atlantic basins of Upper Guinea, from eastern Liberia to western Ghana (Lévêque & Paugy, 1999). The number of anal fin rays and vertebrae in this species increases from west toeast, as is also the case for body size, size at first maturity, and maximum size(figure 8.6) (Lévêque & Herbinet, 1982). From an ecological viewpoint, thewest-east gradient corresponds to the transition from forest rivers (Cess,Cavally) to savannah rivers (Comoé) with intermediate situations (Sassandra and Bandama). In this example, there is also a correlation between the maximum size of individuals and the number of vertebrae.

FIGURE 8.6. Schilbe mandibularis: clinal changes in two meristic characters (average) and maximum size observed along an east-west gradient in river catchments of Côte d’Ivoire.

FIGURE 8.6. Schilbe mandibularis: clinal changes in two meristic characters (average) and maximum size observed along an east-west gradient in river catchments of Côte d’Ivoire.

Subcontinental scale: Alestes baremoze, Brycinus macrolepidotus and Barbus bynni

20Alestes baremoze (Alestidae) is found in all the basins of the Nilo-Sudanian zone (Paugy et al., 1994; Lévêque & Paugy, 1999). Populations are not identical and, depending on the sub-regions, there are geographic variations in some meristic characters such as the number of scales along the lateral line, the number of vertebrae, and the number of soft rays of the anal fin (figure 8.7). The values of these three characters decrease along the cline from the Nile to the Côte d’Ivoire with intermediate values for the Sudanian basins. Three subspecies were described on the basis of the variations of these meristic characters. No detailed explanation has been given to explain this trend.

FIGURE 8.7. Alestes baremoze: clinal changes in two meristic characters (average).

FIGURE 8.7. Alestes baremoze: clinal changes in two meristic characters (average).

21Prior to the morphometric analysis of B. macroplepidotus, two species had been identified throughout its area of distribution in tropical north Africa: Brycinus rutilus in the forest sector and B. macrolepidotus in the savannah zone. On the basis of two main morphometric characteristics (body height and position of the dorsal fin in relation to the pelvic fin), we were able to highlight the existence of a geographic cline (figure 8.8), and only B. macrolepidotus is now considered a valid species (Paugy, 1982a). There are in fact two extreme morphs and several intermediate forms. A look at the forest-savanna gradient shows that the bodies of individuals lengthen and become more slender, and that the position of the anal fin retreats in relation to the insertion of the pelvic fins.

22Within the Barbus bynni complex, three allopatric species have been identified: B. bynni (Nile basin), B. occidentalis (Sahelian zone: Chad, Niger, Ouémé, Volta, and Senegal basins) and B. waldroni (coastal basins of Côte d’Ivoire and Ghana). In fact, a parasitology study (fauna of gill parasites, Monogenean) (Lévêque & Guégan, 1990) show that the two West African species B. waldroni and B. occidentalis had the same Monogenean fauna as B. bynni. The strict specificity of this type of parasite thus proves that these three forms belong to the same species and must each be considered a subspecies of that group. The parasitology study showed that there was but one species whose meristic characters change according to a geographic cline. In West Africa, the two subspecies (respectively B. b. occidentalis and B. b. waldroni) are distinguishable through the number of scales (figure 8.9).

FIGURE 8.8. Brycinus macrolepidotus: clinal changes in two morphological characters (average) across the whole distribution area of the species.

FIGURE 8.8. Brycinus macrolepidotus: clinal changes in two morphological characters (average) across the whole distribution area of the species.

FIGURE 8.9. Barbus bynni: clinal changes in two meristic characters (average).

FIGURE 8.9. Barbus bynni: clinal changes in two meristic characters (average).

Question of hybrids

23Natural or artificial hybridization is a way of testing the concept of biological species. Hybridization in the natural world has been observed, but it remains a rare phenomenon to the extent that there has been little systematic research in the matter and that hybrids are not always identifiable to an untrained eye. Daget (1963) was able to identify young specimens from the Benue as hybrids of Citharidium ansorgii and Citharinus distichodoides. In Lake Itasy in Madagascar, Oreochromis macrochir and O. niloticus that were introduced in the 1960s hybridized to produce a form called tilapia 3/4 which co-existed for many years with the parental species (Daget & Moreau, 1981). A similar phenomenon was observed in Lake Naivasha in Kenya between the introduced species Oreochromis spilurus niger and O. leucostictus (Elder et al., 1971).

24More recently, natural hybridization between Coptodon zillii and Tilapia guineensis was observed in the lake of the Ayamé dam (Côte d’Ivoire). These two species are usually parapatric, with C. zillii found exclusively in fresh water and T. guineensis in brackish water. It is probable that with the building of the dam in 1958, populations of the two species were trapped, forcing them to cohabit (Pouyaud, 1994). Another example of hybridization that involved three Cichlidae species (Coptodon zillii, T. guineensis and T. dageti) was also found in the Comoé River in Côte d’Ivoire and confirmed by protein analysis, as was the previous case (Pouyaud, 1994).

25The last two cases are probably an illustration of what is called a hybrid zone, that is, a contact zone where parapatric populations may hybridize. It is likely that the phenomenon was underestimated up to now owing to limited technical means and a lack of interest.

26Generally speaking, hybridizations in the natural environment among Cichlidae occur when environmental changes remove the barriers that ensure reproductive isolation between species, or when a species colonizes or is introduced to a physical system where similar species already exist.

27Another example is the case of an Alestidae hybrid between Alestes baremoze and a Brycinus which was observed in Côte d’Ivoire with the closure of the Kossou storage lake. Finally, there is the case of a Hydrocynus somonorum species described in the Niger River by Daget (1954) which was then found to be a hybrid of H. forskalii and H. brevis. In all these cases, we see the rather elusive nature of the appearance of a hybrid form.

28Artificial hybridization was much used by aquariologists studying Cyprinodontiformes in an attempt to identify the biological species among the numerous, sometimes highly similar forms, some of which possess different karyotypes (Scheel, 1968).

Genetics and species concept

29Given the difficulty of recognizing the existence of species using morphology alone, ichthyologists sought to use other data, notably from genetics (see box “Genetic differentiation of Sarotherodon melanotheron populations” in chapter Diversity of African fish: heritage of evolution).

Number and shape of chromosomes

30Each living cell contains several chromosome pairs. Each pair contains one chromosome from the mother’s egg cell and another from the father’s sperm. The number of chromosomes is constant in a given species, and 2n (n being a set of chromosomes, either paternal or maternal) is used to define the number of chromosomes in somatic cells, which are diploid. Gametes or sex cells (eggs, sperm) contain only one set of chromosomes (n) and are called haploid. After fertilization, the gametes fuse and yield diploid cells. In some cases when the number of chromosomes in a cell is higher than 2n, such cells are called polyploid.

31The number, shape, and size of chromosomes vary from one species to another. These characters can be useful for taxonomic or phylogenetic research, and are increasingly used given the advances in karyotype preparation techniques (Ozouf-Costaz & Foresti, 1992). In particular, we can expect significant progress to be made in chromosome marking techniques which would make them easier to identify.

32The number of chromosomes for several African fish species has been determined (table 8.I). For many of them, this number ranges between 2n = 48 and 2n = 52, but in some cases the number is different owing to diverse reasons:

  • In some Cyprinidae species, 2n can be as high as 148-150. In this case it is a hexaploid number resulting from a polyploid phenomenon that consists of a mutation that leads to an increase in the number of chromosomes within a population. This number may be double (tetraploid individuals), triple (hexaploid individuals), or any other multiple of the number of chromosomes usually observed. The recent discovery of large hexaploid Barbus in South, East, and West Africa suggests that the origin of Cyprinidae should be reconsidered.
  • In the Siluriformes, 2n is between 70 and 72 in species of the genera Clarotes and Chrysichthys. This higher number, compared for instance with the genera Bagrus and Auchenoglanis, could trace its origin to centromere fissions in some chromosomes (Agnèse, 1989).
  • The highly variable karyotype of Cyprinodontiformes (2n between 9 and 24) appears to be unique among fish, if not in the animal world altogether. The mechanism behind the reduction in the number of chromosomes apparently results from the fusion of centromeres in some ancestral chromosomes (referred to as Robertsonian fusion).

Genetic distances and morphological differences: two separate phenomena

33Speciation and morphological evolution are two independent phenomena. In particular, there is not necessarily a correlation between the genetic distance between species and the morphological divergence observed. Thus, Cichlidae in the Great Lakes that evolved rapidly through adaptive radiation are close from a genetic perspective, even though many of the species are well-differentiated morphologically (figure 8.10). For ten Haplochromis species in Lake Victoria, for instance, there is little divergence in terms of protein analysis but the same species show significant anatomical divergence (Sage et al., 1984).

FIGURE 8.10. Genetics as marker of phylogeny: example of Cichlidae (from Meyer, 1993).

FIGURE 8.10. Genetics as marker of phylogeny: example of Cichlidae (from Meyer, 1993).

TABLE 8.I Number of chromosomes for different African species (2n: diploid number of chromosomes).

Families / species

2n Authors

Families / species

2n Authors

Protopteridae

Cyprinodontiformes (cont.)

Protopterus annectens

34 Wickbom, 1945

Aphyosemion christyi

18 Scheel, 1968

Polypteridae

Aphyosemion exiguum

36 Scheel, 1968

Erpetoichthys calabaricus

36 Denton & Howell, 1973

Aphyosemion filamentosum

36 Scheel, 1968

Polypterus palmas

36 Denton & Howell, 1973

Aphyosemion franzwerneri

22 Scheel, 1972

Polypterus senegalensis

36 Urishido et al., 1977

Aphyosemion gardneri

40 Scheel, 1968

Phractolaemidae

Aphyosemion guineense

38 Scheel, 1968

Phractolaemus ansorgii

28 Vervoort, 1979

Aphyosemion gulare

32 Scheel, 1968

Pantodontidae

Aphyosemion labarrei

28 Scheel, 1972

Pantodon buchholzi

48 Uyeno, 1973

Aphyosemion louessense

20 Scheel, 1972

Notopteridae

Aphyosemion lujae

40 Scheel, 1968

Notpterus afer

34 Uyeno, 1973

Aphyosemion mirabile

32 Scheel, 1972

Cyprinidae

Aphyosemion obscurum

34 Scheel, 1968

Barbus bynni

150 Golub. & Krys., 1993

Aphyosemion roloffi

42 Scheel, 1968

Barbus bynni occidentalis

148 Guégan et al., 1995

Aphyosemion scheeli

40 Scheel, 1972

Barbus capensis

150 Oellerm. & Skelt., 1990

Aphyosemion sjoestedti

40 Scheel, 1968

Barbus ethiopicus

150 Golub. & Krys., 1993

Aphyosemion walkeri

36 Scheel, 1968

Barbus intermedius

150 Golub. & Krys., 1993

Poropanchax luxophthalmus

48 Scheel, 1972

Barbus natalensis

150 Oellerm. & Skelt., 1990

Poropanchax normani

48 Scheel, 1972

Barbus petitjeani

150 Guégan et al., 1995

Epiplatys annulatus

50 Scheel, 1972

Barbus wurtzi

148 Guégan et al., 1995

Epiplatys barmoiensis

34 Scheel, 1972

Barbus ablabes

48 Rab et al., 1996

Epipltys bifasciatus

40 Scheel, 1968

Barbus anema

50 Golub. & Krys., 1993

Epiplatys chaperi

50 Scheel, 1972

Barbus bigornei

50 Rab et al., 1996

Epiplatys dageti

50 Scheel, 1968

Barbus holotaenia

50 Rab, 1981

Epiplatys duboisi

48 Scheel, 1968

Barbus macrops

50 Rab et al., 1996

Epiplatys fasciolatus

38 Scheel, 1968

Garra dembeensis

50 Krys. & Golub., 1993

Epiplatys sexfasciatus

48 Scheel, 1968

Garra quadrimacukata

50 Krys. & Golub., 1993

Epiplatys spilargyreius

34 Scheel, 1968

Labeo senegalensis

50 Paugy et al., 1990

Nothobranchius guentheri

35 Scheel, 1981;

Labeo coubie

50 Paugy et al., 1990

Nothobranchius guentheri

36 Ewulonu et al., 1985

Labeo roseopunctatus

50 Paugy et al., 1990

Nothobranchius kirki

36 Scheel, 1972

Raiamas steindachneri

58 Rab et al., 2000

Nothobranchius melanospilus

36 Ewulonu et al., 1985

Bagridae/Claroteidae

Nothobranchius palmquisti

34 Scheel, 1968

Auchenoglanis occidentalis

56 Agnèse, 1989

Nothobranchius palmquisti

36 Ewulonu et al., 1985

Bagrus docmak

54 Agnèse, 1989

Nothobranchius patrizii

36 Ewulonu et al., 1985

Chrysichthys auratus

72 Agnèse, 1989

Nothobranchius rachovii

18 Scheel, 1981

Chrysichthys maurus

70 Agnèse, 1989

Nothobranchius rachovii

16 Ewulonu et al., 1985

Clarotes laticeps

70 Agnèse, 1989

Cichlidae

Clariidae

Astatotilapia burtoni

40 Thompson, 1981

Clarias anguillaris

56 Agnèse, 1989

Aulonocara kornelia

44 Foerster & Schartl, 1987

Clarias gariepinus

56 Teugels et al., 1992

Aulonocara huesheri

44 Foerster & Schartl, 1987

Heterobranchus longifilis

52 Teugels et al., 1992

Aulonocara stuartgranti

44 Foerster & Schartl, 1987

hybrid C. gariepinus x H. longifilis

54 Teugels et al., 1992

Hemichromis bimaculatus

44 Zahner, 1977

Mochokidae

Heterotilapia multispinosa

48 Zahner, 1977

Synodontis bastiani

54 Agnèse et al., 1990

Melanochromis auratus

46 Thompson, 1981

Synodontis budgetti

54 Agnèse et al., 1990

Alcolapia alcalica

48 Park, 1974

Synodontis courteti

54 Agnèse et al., 1990

Oreochromis aureus

44 Kornfield et al., 1979

Synodontis filamentosus

56 Agnèse et al., 1990

Oreochromis karongae

38 Harvey et al., 2002

Synodontis membranaceus

54 Agnèse et al., 1990

Oreochromis macrochir

44 Jalabert et al., 1971

Synodontis ocellifer

54 Agnèse et al., 1990

Oreochromis mossambicus

44 Fukoka & Muram., 1975

Synodontis schall

54 Agnèse et al., 1990

Oreochromis niloticus

44 Jalabert et al., 1971

Synodontis sorex

54 Agnèse et al., 1990

Sarotherodon galilaeus

44 Kornfield et al., 1979

Synodontis violaceus

54 Agnèse et al., 1990

Tilapia guineensis

44 Vervoort, 1980

Cyprinodontiformes

Oreochromis macrochir

44 Vervoort, 1980

Aphyosemion ahli

36 Scheel, 1968

Tilapia mariae

40 Vervoort, 1980

Aphyosemion arnoldi

38 Scheel, 1968

Coptodon rendalli

44 Michele & Takah., 1977

Aphyosemion bivittatum

40 Scheel, 1968

Tilapia sparmanni

42 Vervoort, 1980

Aphyosemion, bualanum

40 Scheel, 1968

Coptodon zillii

44 Kornfield et al., 1979

Aphyosemion calliurum

26 Scheel, 1972

Channidae

Aphyosemion cameronense

34 Scheel, 1972

Parachanna obscura

34 Nayyar, 1966

34Meanwhile, significant genetic differences have been observed for populations of the genus Tropheus, a group of Cichlidae in Lake Tanganyika that dates back around 1.25 million years, whereas these populations are only distinguishable through minor morphological variations and some marked differences in colouring. In Mormyridae, genetic distances measured using enzymatic polymorphism are greater between Petrocephalus populations from different West African basins than between the genera Mormyrops, Pollimyrus and Marcusenius which are very well-differentiated morphologically (Agnèse & Bigorne, 1992) (figure 8.11). This result shows that there is no correlation between genetic distance and morphological similarity in fish.

FIGURE 8.11. Phylogenetic network of some West African mormyrids (from Agnèse & Bigorne, 1992).

FIGURE 8.11. Phylogenetic network of some West African mormyrids (from Agnèse & Bigorne, 1992).

Petrocephalus
PBb: P. bovei (Bandama)
PBf: P. bovei (Federougouba)
PBn: P. bovei (Niger)
PBp: P. bovei (Pampana)
PBr: P. bovei (Rokel)
PBs: P. bovei (Sassandra)
PLk: P. levequei (Konkouré)
PLl: P. levequei (Kaba)
PLp: P. levequei (Pampana)
PLr: P. levequei (Rokel)
PP: P. pallidomaculatus
PPc: P. pellegrini (Cavally)
PPn: P. pellegrini (Cess)
PS: P. sudanensis
Other mormyrids
HP: Hippopotamyrus paugyi
MM: Marcusenius mento
MPS: Mormyrops caballus
MS: Marcusenius senegalensis
MT: Marcusenius thomasi
HP: Hippopotamyrus paugyi
POL: Pollimyrus petricolus

35Some Cyprinodontiformes species are also similar in appearance while having very different karyotypes.

The rules of classification

36Classification of the living world is hierarchical, with groups included in larger sets that do not overlap. But this hierarchy can be based on different principles. Phenetic hierarchy is based on the similarity of appearance of the classified groups: number and position of fins, presence or absence of barbels, number of fin rays, etc. Phylogenetic hierarchy, meanwhile, is based on evolutionary relationship: groups are defined according to the closeness of their relationships and the age of their common ancestors. These two classification systems, which gave rise to different schools of thought, can provide concordant results but may also differ in their conclusions. The numerical taxonomy school fosters phenetic classification, whereas phylogenetic classification is supported by the so-called cladistic school.

37Supporters of numerical taxonomy believe that organisms sharing common characteristics have a similar evolutionary history, but do not make hypotheses on genealogy. By using a large number of characters in a large number of individuals, and assigning them equal weights, statistical methods are expected to identify homogeneous sets. However, morphological convergences during evolution occasionally led to artificial groupings, and the biases that could be generated by statistical methods were also criticized (Ridley, 1989). This school peaked in the 1960s but lost much of its relevance since. An example of its application in African fishes was proposed by Daget (1966); he was able to establish a phenogram of Citharinidae (figure 8.12) that confirmed the empirical classification used up to that time.

FIGURE 8.12. Phenetic and cladistic classifications of the citharinids (from Daget, 1966 and 1980).

FIGURE 8.12. Phenetic and cladistic classifications of the citharinids (from Daget, 1966 and 1980).

38Cladistic classification (sometimes called Hennigean classification) is based on the principle that during evolution, an ancestral species gave rise to two daughter species. For each species, it is therefore necessary to determine with which other species it shares the most recent common ancestor, because this will form the basis for the first group. A monophyletic group is derived from a single common ancestor, whereas a polyphyletic group includes species that have similarities but are not direct descendants of a common ancestor.

39Unlike phenetic relationships, phylogenetic relationships cannot be observed directly. How, then, can they be identified? The method proposed by Hennig (1950) consists of looking for characters that can be qualified as evolutionary innovations. We can thus distinguish evolved or apomorph characters in opposition to ancestral characters, referred to as primitive or plesiomorph. This method, which often used comparisons with distant groups to determine if characters were evolved or primitive, also shows a bias – it considers that evolution always proceeds from a plesiomorph state to an apomorph state, and never the reverse (Daget, 1980).

40This method was also applied to the Citharinidae (Daget, 1980). We see (figure 8.12) that it would be possible to include the species C. distichodoides and C. ansorgii in the genus Citharidium, a sister group of the genus Citharinus comprising six species. A conservation measure would be to consider that the species C. distichodoides is perhaps not the sister species of C. ansorgii, and create a genus Citharinops. We would then have two monotypic genera which are also sister groups, and this set would be a sister group of the genus Citharinus. The phylogenetic proximity between Citharidium and Citharinops does not contradict the possibility of hybridization of these two species in a natural physical system as signalled by Daget (1963). Among the Citharinus, the group of three Nilotic species is a sister group of three other Congolese species. Phenetic and cladistic classifications lead to similar conclusions on the systematic level, which means that in the present case, the morphological similarity reflects the phylogenetic relationship. The only difference with regard to the phenetic analysis is that C. congicus is closer to C. gibbosus than to C. macrolepis.

41Using 15 anatomical and morphological characters, Daget & Desoutter (1983) also proposed a phylogenetic classification of Polypteridae from a common ancestor. The group Polypterus bichir-endlicheri probably diverged very early on, whereas the monospecific genus Calamoichthys, with a distinct morphology compared to the others, is curiously classified with the other Polypterus, which could indicate that it is of recent origin (figure 8.13).

42Using her own work as well as results obtained by other ichthyologists, Stiassny (1991) proposed a cladogram summarizing the current state of knowledge on phylogenetic relationships within the Cichlidae family. In Africa and South America, this family forms two sister groups. There are nonetheless indications that the genus Tylochromis could be the sister group of other African Cichlidae, whereas the Congolese genus Heterochromis, phylogenetically different from the other African species, could be a sister group of Asian Cichlidae (Etroplines). Malagasy Cichlidae (Ptychochromines) form the sister group of all other Cichlidae (figure 8.14).

FIGURE 8.13. Cladogram illustrating the phylogenetic relationships between the African Polypteridae (from Daget & Desoutter, 1983).

FIGURE 8.13. Cladogram illustrating the phylogenetic relationships between the African Polypteridae (from Daget & Desoutter, 1983).

43For many scientists, phylogenetic classification is currently preferred to phenetic classification, even if methods are not quite perfected. While anatomical studies have revealed both their relevance and their limitations, molecular tools currently in development are hoped to yield spectacular technical progress in coming years, thereby allowing a better understanding of the relationships between species.

FIGURE 8.14. Summary cladogram of cichlid intrafamily relationships (from Stiassny, 1991).

FIGURE 8.14. Summary cladogram of cichlid intrafamily relationships (from Stiassny, 1991).

44Thus, based on enzymatic polymorphism and chromosome studies, a phylogeny of West African Bagridae was proposed by Agnèse (1989) (figure 8.15). For the eight species investigated, two sister groups can be distinguished: on one hand Bagrus and perhaps Auchenoglanis which possess 54 and 56 chromosomes, and Chrysichthys-Clarotes on the other hand with 70-72 chromosomes. Within Chrysichthys we can also observe two sister groups, the sub-genera Chrysichthys (C. maurus, C. auratus) and Melanodactylus (C. nigrodigitatus and C. johnelsi) as Risch (1986) proposed on the basis of anatomical and morphological studies. Clarotes is a sister group of the Chrysichthys and they were both separated, along with Auchenoglanis, in the Claroteidae family by Mo (1991).

FIGURE 8.15. Phylogenetic relationships among West African Bagridae and Claroteidae (from Agnèse, 1989).

FIGURE 8.15. Phylogenetic relationships among West African Bagridae and Claroteidae (from Agnèse, 1989).

45The use of genetic markers such as mitochondrial DNA, which are more precise than protein markers, is also becoming more widespread. Kornfield (1991) managed to propose a phylogenetic classification of Oreochromis niloticus subspecies in East Africa, and the phylogeny of Tanganyika Cichlidae of the genus Tropheus was established by Sturmbauer & Meyer (1992). For other groups, phylogeny was determined thanks to molecular biology techniques. This is the case for the Mormyridae whose familial phylogeny was defined (Lavoué, 2001), and for the Alestidae, for which group monophyly was underscored, confirming the clearly individualized status of this strictly African family (Calcagnotto et al., 2005; Hubert et al., 2005a and 2005b). In the same way, some families with an indeterminate status could be grouped together afterwards. For instance, within the Gonorynchiformes, the monophyly of the Cromeriidae, Grasseichthyidae, and Kneriidae has been proven (Lavoué et al., 2005).

46A fairly recent study on the phylogeny of the Protopteriformes highlights the monophyly of the African forms of the genus Protopterus (Tokita et al., 2005). Within this genus, three clades are identified: P. amphibius (Eastern province), P. dolloi (Congolese and Lower Guinea provinces) and P. aethiopicus/P. annectens (P. aethiopicus: lakes Victoria, Edouard, Tanganyika and Albert; P. annectens: Nilo-Sudanian province). However, the differentiation of each of these three units does not enjoy strong support.

47It has long been thought that the divergence of the three Protopteriforme genera followed the fragmentation of Gondwana (see chapter General characteristics of ichthyological fauna). Protopterus and Lepidosiren (Amazon) have always been considered sister groups, and analysis of all the species in this study confirms it. The authors’hypothesis on the closeness of relationships among the species is based on the more or less sustained presence of a neotenic character, the external gills. According to them, this character is not generated during the ontogeny of Neoceratodus (Australia), disappears rapidly in the juveniles of P. dolloi and P. aethiopicus, but can persist longer in P. annectens and even remain in the adult P. amphibius.

The Clariidae: an example where morphological differentiation does not necessarily reflect evolutionary history
Jean-François Agnèse
The genus Clarias Scopoli 1777 is an interesting example of a fairly well-defined group in terms of appearance based on a set of morphological criteria, whereas they actually represent a discontinuous evolutionary unit. Günther (1864) defined the genus Clarias based on about ten characters including the shape of the gill accessory organ, the presence of a spine in the pectoral fin, and the absence of an adipose fin.
Although Teugels (1986) had recognized six sub-genera within Clarias, this genus appeared to be composed of a set of closely-related species from an evolutionary viewpoint.
Recently, a number of phylogenetic reconstitutions carried out using mitochondrial markers clearly showed that the genus Clarias was in fact extremely paraphyletic (Agnèse & Teugels, 2000, 2001, 2005) (figure 8.16). Some species that are currently classified in other genera of the Clariidae family (because they are highly differentiated morphologically), are in fact very closely related
genetically to species of the genus Clarias. Among the Clariidae, it thus seems that there is no link between morphological differentiation and genetic differentiation.
As a result, morphological criteria are inadequate for classifying species in one genus or another, and a revision of the family that takes genetic affinities into account is now needed.

FIGURE 8.16. Phylogenetic relationships among Clariidae species.

FIGURE 8.16. Phylogenetic relationships among Clariidae species.

Morphological and osteological evolution (in this example “anguilliform body”) in Clariidae did not follow an orthogenetic series (from Agnèse & Teugels, 2005). Group A represent the “big head and numerous fins” species, eel-like species are only found in group B. Sketches of species with particular body shapes are represented on the right side, arrows indicated their position in the tree.

48Moreover, genetic differentiation is not necessarily linked to morphological differentiation, as seen in the Clariidae (Agnèse & Teugels, 2005). In this case, it was shown that the ‘anguilliform’ criterion of some species was a character independent of the phyletic groups to which they belong. It clearly seems to be an adaptive response that appeared independently in several groups (see box “The Clariidae: an example where morphological differentiation does not necessarily reflect evolutionary history”).

The genus concept

49While the notion of the species concept is relatively clear, despite some difficulties related to the concept’s implementation, how about higher taxonomic levels? Are they conventional sets, or can they be defined on a similar basis as that used to distinguish species?

50According to the principles of taxonomy, each species must belong to a genus, which is a category comprising at least one species or according to the principles of phylogenetic classification, a monophyletic group of species. Obviously, the principle of interbreeding cannot be used to define higher taxonomic levels.

51On the other hand, we can distinguish morphologic sets, with the attendant uncertainties (bias resulting from convergences, for instance) regarding their affiliation to the same evolutionary line.

52The improvement of ideas about classification and the lack of standardization in defining taxa higher than species gave rise to a certain anarchy, and led to many nomenclature changes as a result of advances in knowledge of phylogenetic relationships. It was suggested (Dubois, 1988) that the genus should constitute a discontinuous evolutionary unit that could be defined through a set of genetic, phylogenetic, and ecological characters.

53An interesting example is that of the genus Tilapia sensu lato that Trewavas (1983) split into three genera based on their reproductive behaviour:

  • Tilapia sensu stricto which lays eggs on substrate;
  • Sarotherodon, which includes paternal and biparental mouth breeders;
  • Oreochromis which includes only maternal mouth breeders.

54This classification, based on behavioural aspects, sets itself apart from the conventional approach based on morphology and anatomy. It aroused many reactions, and Thys van den Audenaerde (see Teugels & Thys van den Audenaerde, 1992) long argued that it was preferable to not split the genus Tilapia according to these criteria. This debate attracted the interest of molecular biologists, and results of genetic studies showed on the whole that Trewavas’ division was quite right from the phylogenetic viewpoint, with a few exceptions for species such as Sarotherodon galilaeus that is close to Oreochromis (Sodsuk & McAndrew, 1991; Pouyaud & Agnèse, 1995) and S. melanotheron, the sole paternal incubator among the Tilapia (Seyoum, 1990; Pouyaud, 1994) and whose position is indeterminate.

Higher classification of African fishes

55In this chapter, the classification schema above genus level is limited to the ranks of class, order and family. This has the advantage of simplicity, but ignores some important groups in fish phylogeny such as Chondrichthyes, Teleostei, Acanthomorpha, and Percomorpha.

56In general, it follows Nelson’s “Fishes of the World” (2006) with the notable exception of Elasmobranchii and Holocephali being elevated to class rank. But at family level, we may follow Eschmeyer’s “Catalog of Fishes”, as it is updated more frequently to integrate new stable groupings.

57Wiley & Johnson published a new classification for Teleostei (2010). They proposed the stabilization of several recent or (much) older published hypotheses by listing known evidence (synapomorphies). We will follow this new arrangement as soon as the ichthyological community endorses it, e. g., through publication in a subsequent edition of Fishes of the World. In the meantime, to avoid confusion for non-specialists, we shall continue to use the current one.

58We adopted the rules followed by FishBase, as it is currently the better updated database.

59Note that the numbers of species given below do not take subspecies into account.

60For the world as a whole, total count of genera is 5,056 and total count of species is 33,148

61The origin and evolution of species could be reconstituted through the study of fossils. Like all palaeontological investigations, conclusions from such studies can be called into question upon the discovery of new deposits and new fossils. Current hypotheses must thus be viewed as conclusions based on the current state of knowledge rather than statements of fact.

62We have also seen that for existing fishes, systematics has far from resolved all issues, and the introduction of new molecular techniques in phylogenetic classifications will probably lead to changes in these classifications, most of which were established using anatomical and morphological criteria. Moreover, ichthyologists are not always in agreement regarding the proposed novel phylogenies, which complicate the situation. There is nevertheless a sort of consensus for the broad lines of the fish classification that we present here.

63The oldest known fish remains date to the Cambrian (over 500 million years ago). A form of lamprey, Hardistiella montanensis, from the Upper Carboniferous was found in Montana, USA (-325 M.A.), but a fossil that may have been related to lampreys, Haikouichthys ercaicunensis, was found in Lower Cambrian layers in Yunan, China (-530 M.A.) (Lecointre & Leguyader, 2001). Jawless fishes (Agnatha, still represented today by lampreys) and Placodermi initially dominated but gave way to bony and cartilaginous fishes in the late Devonian and Carboniferous (-400 to-350 million years) eras, during which they diversified greatly in seas and freshwaters. New groups, the holostei, then radiated during the Triassic (-250 to-210 million years) and Jurassic (-210 to-150 million years), followed by teleosts during the Jurassic and Cretaceous. The radiation of teleosts is behind most of the marine and freshwater species that we know today.

FIGURE 8.17. Cladogram representing the phylogeny of the major living group of current craniata (from Lecointre & Le Guyader, 2001).

FIGURE 8.17. Cladogram representing the phylogeny of the major living group of current craniata (from Lecointre & Le Guyader, 2001).

64We shall only mention the Myxini (hagfish: 1 family, 6 genera, 78 species) and the Cephalaspidomorphi (lampreys: 3 families, 10 genera, 47 species), two jawless Craniata that do not exist in the inter-tropical zone. Forms likely to be encountered in Africa belong to the Gnathostomata, that is, jawed vertebrates.

65The Elasmobranchii contain the craniata with cartilaginous skeletons, commonly known as rays or sharks. From a phylogenetic viewpoint, these “fishes” are very distant from the bony fishes (see box “Systematic ichthyology”; figure 8.17).

66The Osteichthyes include other vertebrates including bony fishes that are unevenly distributed in two groups: the Sarcopterygii (lungfish and cœlacanth, but also mammals, birds, squamata, etc.) and the Actinopterygii (or ray-finned fishes).

living fossils
Groups of archaic fish can still be found on the African continent. The best-known of these is certainly the cœlacanth which, in the sixties, was the star of a media frenzy. We shall not devote much space to this species, which remains oceanic and therefore distant from our continental concerns. It is worth noting that this very ancient group, which appeared in the Devonian (400-350 106 BP) and disappeared at the same time as the dinosaurs (around 70 106 BP), was once found throughout the planet and thought to have lived in fresh waters.
Polypterus or bichir: this group is endemic in Africa. The only known fossils were also found in Africa (135-65 106 BP) in the same area of distribution as modern species. These fishes have a swim bladder whose highly vascularized wall serves as a “lung” that allows the animal to survive in oxygen-poor waters.
Note as well that juveniles have external gills (see box “Polypterids”, chapter Geographical distribution and affinities of African freshwater fishes).
Lungfishes currently exist in the Amazon (Lepidosiren), Australia (Neoceratodus), and Africa (Protopterus).
African Protopterus can survive during dry periods by entering a state of aestivation.
To do so, P. annectens burrows a tunnel into the mud when the water level begins to go down but before the water dries out completely.
Once the water goes below the level of the tunnel opening, the animal blocks the entrance with mud then goes to the end of the tunnel, where it covers itself with a layer of integumentary mucus that hardens and forms a cocoon to keep itself moist.
Thus protected, the lungfishes can live for up to four years in the cocoon, using their own muscle fibres as food.
Like their South American cousins, African lungfish larvae have external gills whose level of development depends on the water’s oxygen content.
As the juveniles age, the external gills are generally resorbed and the fish breathe using their lungs and internal gills.

67Among the Sarcopterygii, the lungfishes or Dipnoi are the oldest freshwater bony fishes, and their origins can be traced to the Devonian (Rosen et al., 1981) (see box “Living fossils”). The swim bladder is connected to the oesophagus and can serve as a lung. They were represented by many forms in the primitive era, but only a few species survive, including those belonging to the genus Protopterus in Africa.

68Recent studies have shown that the mitochondrial DNA of Protopteridae is closer to an amphibian’s DNA than to a coelacanth’s, which tends to support the hypothesis that terrestrial vertebrates could come from a split in a line that gave rise to lungfishes (Meyer & Wilson, 1990). Dipnoi would thus be the sister group of tetrapods (figure 8.17).

69The Polypteridae are currently considered to be the earliest line of Actinopterygii. Their bodies are covered with bony scales. The swim bladder can serve as an accessory respiratory organ. Juveniles possess arborescent external gills that disappear in adults, and which are perhaps morphological characteristics of primitive Actinopterygii. In the past, this primitive character misled zoologists into classifying them away from Actinopterygii, and even placing them among Sarcopterygii.

70In continental Africa, the Actinopterygii comprise the majority of other families and species grouped among the teleosts, whose most ancient known fossil, Pholidophorus, dates to the lower Triassic (-195 Myrs). This extremely diverse group in fact represents the overwhelming majority of modern fishes. There are several major subdivisions, some of which are represented in continental Africa (Wiley & Johnson, 2010):

  • Osteoglossomorpha (bony-tongued fishes) including the most primitive of modern teleosts. They have been known since the late Jurassic and are currently represented by five families in Africa (figure 8.18);
  • Elopomorpha comprises different families and species that are almost exclusively marine, but which can occasionally penetrate into freshwater. Eels belong to this group;
  • Otomorpha is composed of 4 large monophyletic groups:
    • Clupeiformes, which appeared in the lower Cretaceous. They include primarily marine species, but some have adapted to freshwater;
    • the order Gonorynchiformes which has 3 families in Africa, including the Kneriidae (which now covers the former Cromeriidae and Grasseichthyidae families) and two other families represented by a single species (Phractolemidae and Chanidae);
    • the order Cypriniformes whose origin is under debate;
    • the order Characiformes represented by numerous species. The origin of the Characiformes and the Cypriniformes could date back to the end of the Cretaceous or the Palaeocene (Cavender, 1991);
    • the order Siluriformes (catfishes) forms a highly characteristic group of fishes with scale-free bodies and bony spines in their dorsal and pectoral fins.
  • the Euteleosteomorpha including the majority of species, in particular:
    • the Osmeriformes represented by a single family (Galaxiidae) in Africa;
    • the Cyprinodontiformes including numerous fresh and brackish water species.
    • the Synbranchiformes represented by two families: the Synbranchidae of marine origin, and the Mastacembelidae (spiny eels) that were formerly classified among the Perciformes (Travers, 1984).
    • the Perciformes which is poorly defined and whose classification is unclear and problematic, as it is probably not a monophyletic grouping. In total it includes 150 mostly marine families, some of which have adapted to continental waters. This is the case in particular for the Cichlidae, Latidae, Nandidae, and Anabantidae. The Cichlidae probably appeared very early in the Cretaceous and their evolution was already well on its way prior to the separation of Gondwana (Stiassny, 1987; 1991).
    • The Pleuronectiformes and the Tetraodontiformes including families of marine origin of which certain species have adapted to freshwater. The Tetraodontiformes and Pleuronectiformes appeared in the lower Eocene and mostly include marine forms.

Systematic ichthyology
The term “fishes” is not scientific and in any case refers to a grade, that is, a group without its own history. In fact, the relationship is not stronger between a teleost and a shark (figure 8.17) than between birds and bats, or even some flying or gliding reptiles or fishes. We could even note that they are in truth more distant. In concrete terms, “fishes” comprise a vast heterogeneous set traditionally defined as legless craniata. This definition is fitting, if, like Nelson (1994) or more recently Lecointre & Le Guyader (2001), we agree that this current fauna includes:
• Myxini or hagfish (without vertebrae or jaws)
• Cephalaspidomorphi = lampreys (jawless)
• Elasmobranchii = cartilaginous fishes (sharks, rays)
• Actinopterygii = ray-finned fishes
• Sarcopterygii = vertebrates with fleshy members (cœlacanth and lungfishes, closer to tetrapods than to the other “fishes”)
Until around the 1960s, this “fishes superclass” was considered a valid taxon, owing to lack of recognition or use of phylogenetic criteria. The shift from a purely utilitarian classification to a phylogenetic one only took place later, even though the Darwinian theory of evolution (Darwin, 1859) had already been long accepted. It appears that the situation only moved forward after the translation of Hennig’s method (1966).
ladistics – which is what we are discussing now – has the advantage “(…) of highlighting the feasibility – or refutability – of a theory of phylogenetic relationships” (Janvier, 1986).
Following the palaeontologists, palaeoichthyologists and ichthyologists rapidly began using cladistics and, in doing so, gained an advantage over other disciplines involving vertebrates.
The classification of “fishes”, by becoming phylogenetic, was obviously going to be overthrown. As such, compared with previous classifications (Bertin & Arambourg, 1958) that still served as the reference until recently, things changed considerably, particularly under the initial impetus of Greenwood et al. (1966) which was a sort of pre-cladistics classification that was not informed by the works of Hennig.
Even if the classifications can be debated, it is mentally easier to arrange taxa in categories that include groups or species that have the most similarities.
As such, the notions of species, genus, family, or even order are still generally accepted or tolerated by the cladistic school.
We have summarized data on current forms following the structure orders/families/genera/species in table 8.II which is necessarily incomplete, as not everything is known, and reflects only the specific interest focused on the different taxa.
Thus, we have cause to believe that the number of mammal and bird species is close to being accurate, which is certainly not the case for amphibians and fishes, for which the number of species is by all indications underestimated.

TABLE 8.II. Number of Craniata known at the present time.

TABLE 8.II. Number of Craniata known at the present time.

Advances in African freshwater ichthyology

71Currently 3,360 species of fresh and brackish water fish have been described from Africa, belonging to 95 families (Lévêque et al., 2008). Numerous other species collected in the East African lakes are awaiting description.

72The knowledge of the ichthyological brackish and freshwater fauna in Africa is a very long story (Paugy, 2010a). From the time of the Ancient Egyptians to the present, more than 3,300 species (95 families and 542 genera) have been discovered, drawn and described. Michel Adanson (mid-18th century) initiated the first material collections during the eighteenth century. During the 19th century, the work of travelling scientists (Étienne Geoffroy Saint-Hilaire, Andrew Smith) and explorers (including Mungo Park, Pierre Savorgnan de Brazza and Henry Morton Stanley) added substantially to developing zoological collections from their field trips. At that time, many species descriptions were based on fish preserved in these collections. Towards the end of the 19th and into the early part of the 20th century, knowledge of African fishes was greatly enhanced, especially through the work of Georges A. Boulenger, Albert C.L.G. Günther and Franz Steindachner who, respectively, described 640, 119 and 53 species. Boulenger did visit Africa once during his life, in 1905 when he went to Cape Town for a conference. With this one exception, none of these scientists ever travelled to Africa themselves. In contrast to those of the previous century, the majority of naturalists of the 20th century who were interested in African fishes took part in collecting expeditions. The majority of the naturalists working in Africa during the middle and later parts of the 20th century tended to specialize in particular groups.

FIGURE 8.18. Cladogram representing the phylogeny of the major living group of teleosts (adapted from Wiley & Johnson, 2010). In the right column, taxa existing in inland Africa (fresh and brackish waters) are underlined.

FIGURE 8.18. Cladogram representing the phylogeny of the major living group of teleosts (adapted from Wiley & Johnson, 2010). In the right column, taxa existing in inland Africa (fresh and brackish waters) are underlined.

73Thanks to these naturalists the number of known African freshwater fish species reached 1,900 before the Second World War, 2,150 at the end of the 1950s and finally more than 3,300 at present (figure 8.19). In addition to conventional systematic studies, there was a steady rise in the number of contributions dealing with genetics, specific parasites, and electrophysiology, amplifying fish identification using criteria other than morphology. These methods have proved helpful in finding explanations for the radiation of cichlids in the Rift Valley Lakes of East Africa. Blending all these methods, descriptions of hitherto unknown species continue to be published (see also chapter General characteristics of ichthyological fauna).

FIGURE 8.19. Number of African freshwater fish species currently described (updated from Paugy, 2010a).

FIGURE 8.19. Number of African freshwater fish species currently described (updated from Paugy, 2010a).

74Most of the African continent has remained above sea level since more than 600 Myrs ago (Precambrian). Such a long period of land emergence may explain the diversity of the freshwater fish fauna and its unparalleled assemblage of so-called archaic families of which most are endemic.

Table des illustrations

Titre FIGURE 8.1. Head radiography of the holotype of Synodontis xiphias (from Poll, 1971).
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-1.jpg
Fichier image/jpeg, 117k
Titre FIGURE 8.2. General morphology of three sympatric species of Labeo from Western Africa (from Paugy et al., 1994).
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-2.jpg
Fichier image/jpeg, 54k
Titre FIGURE 8.3. General morphology of three sympatric species of Petrocephalus from Western Africa (from Paugy et al., 1994).
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-3.jpg
Fichier image/jpeg, 47k
Titre FIGURE 8.4. Distribution of B. imberi, B. abeli, B. carolinae and B. nigricauda.
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-4.jpg
Fichier image/jpeg, 283k
Titre FIGURE 8.5. Petrocephalus bovei: clinal changes in two meristic characters (average) along an upstream-downstream gradient (Upper Niger).
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-5.jpg
Fichier image/jpeg, 196k
Titre FIGURE 8.6. Schilbe mandibularis: clinal changes in two meristic characters (average) and maximum size observed along an east-west gradient in river catchments of Côte d’Ivoire.
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-6.jpg
Fichier image/jpeg, 317k
Titre FIGURE 8.7. Alestes baremoze: clinal changes in two meristic characters (average).
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-7.jpg
Fichier image/jpeg, 141k
Titre FIGURE 8.8. Brycinus macrolepidotus: clinal changes in two morphological characters (average) across the whole distribution area of the species.
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-8.jpg
Fichier image/jpeg, 283k
Titre FIGURE 8.9. Barbus bynni: clinal changes in two meristic characters (average).
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-9.jpg
Fichier image/jpeg, 213k
Titre FIGURE 8.10. Genetics as marker of phylogeny: example of Cichlidae (from Meyer, 1993).
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-10.jpg
Fichier image/jpeg, 92k
Titre FIGURE 8.11. Phylogenetic network of some West African mormyrids (from Agnèse & Bigorne, 1992).
Légende PetrocephalusPBb: P. bovei (Bandama)PBf: P. bovei (Federougouba)PBn: P. bovei (Niger)PBp: P. bovei (Pampana)PBr: P. bovei (Rokel)PBs: P. bovei (Sassandra)PLk: P. levequei (Konkouré)PLl: P. levequei (Kaba)PLp: P. levequei (Pampana)PLr: P. levequei (Rokel)PP: P. pallidomaculatusPPc: P. pellegrini (Cavally)PPn: P. pellegrini (Cess)PS: P. sudanensisOther mormyridsHP: Hippopotamyrus paugyiMM: Marcusenius mentoMPS: Mormyrops caballusMS: Marcusenius senegalensisMT: Marcusenius thomasiHP: Hippopotamyrus paugyiPOL: Pollimyrus petricolus
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-11.jpg
Fichier image/jpeg, 115k
Titre FIGURE 8.12. Phenetic and cladistic classifications of the citharinids (from Daget, 1966 and 1980).
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-12.jpg
Fichier image/jpeg, 161k
Titre FIGURE 8.13. Cladogram illustrating the phylogenetic relationships between the African Polypteridae (from Daget & Desoutter, 1983).
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-13.jpg
Fichier image/jpeg, 127k
Titre FIGURE 8.14. Summary cladogram of cichlid intrafamily relationships (from Stiassny, 1991).
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-14.jpg
Fichier image/jpeg, 139k
Titre FIGURE 8.15. Phylogenetic relationships among West African Bagridae and Claroteidae (from Agnèse, 1989).
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-15.jpg
Fichier image/jpeg, 45k
Titre FIGURE 8.16. Phylogenetic relationships among Clariidae species.
Légende Morphological and osteological evolution (in this example “anguilliform body”) in Clariidae did not follow an orthogenetic series (from Agnèse & Teugels, 2005). Group A represent the “big head and numerous fins” species, eel-like species are only found in group B. Sketches of species with particular body shapes are represented on the right side, arrows indicated their position in the tree.
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-16.jpg
Fichier image/jpeg, 312k
Titre FIGURE 8.17. Cladogram representing the phylogeny of the major living group of current craniata (from Lecointre & Le Guyader, 2001).
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-17.jpg
Fichier image/jpeg, 257k
Titre TABLE 8.II. Number of Craniata known at the present time.
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-18.jpg
Fichier image/jpeg, 141k
Titre FIGURE 8.18. Cladogram representing the phylogeny of the major living group of teleosts (adapted from Wiley & Johnson, 2010). In the right column, taxa existing in inland Africa (fresh and brackish waters) are underlined.
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-19.jpg
Fichier image/jpeg, 136k
Titre FIGURE 8.19. Number of African freshwater fish species currently described (updated from Paugy, 2010a).
URL http://books.openedition.org/irdeditions/docannexe/image/25202/img-20.jpg
Fichier image/jpeg, 151k

Auteurs

Research director at IRD. He has published numerous works on the systematics, distribution, and ecology of West African fresh water fishes, and expanded the collections of the MNHN in Paris.

Research director at IRD. He has published numerous works on the systematics, distribution, and ecology of West African fresh water fishes, and expanded the collections of the MNHN in Paris.

Le texte et les autres éléments (illustrations, fichiers annexes importés) sont sous Licence OpenEdition Books, sauf mention contraire.

Acheter

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search