The Bathyclarias-Clarias species flock. A new model to understand rapid speciation in African Great Lakes

Abstract

Phylogenetic relationships between seven species of the catfish species flock from Lake Malawi (genus Bathyclarias) and other Clariid catfish have been investigated using cytochrome b partial sequences. Here we demonstrate that this species flock originated from a widespread, generalist species, Clarias gariepinus, still occurring in the lake. Bathyclarias species and their ancestor C. gariepinus form a simple model that can be used to understand the mechanisms of adaptation and rapid speciation in African Great lakes.

Full text

Génétique / Genetics
The Bathyclarias–Clarias species flock.
A new model to understand rapid speciation in
African Great lakes
Jean François Agnèse
a,b
*, Guy G. Teugels
c
a
Laboratoire genome et populations, IRD, CNRS UPR 9060, université Montpellier-II, place Eugène-Bataillon,
34095 Montpellier cedex 05, France
b
Molecular Genetics Department, National Museums of Kenya, P.O. Box 40658, Nairobi, Kenya
c
Ichthyology Laboratory, Africa Museum, B3080, Tervuren and Katholieke Universiteit Leuven, Laboratory of
Compared Anatomy and Biodiversity, B-3000 Leuven, Belgium
Received 6 October 2000; accepted 25 April 2001
Communicated by Jean Rosa
Abstract –Phylogenetic relationships between seven species of the catfish species flock
from Lake Malawi (genus Bathyclarias) and other Clariid catfish have been investigated
using cytochrome bpartial sequences. Here we demonstrate that this species flock
originated from a widespread, generalist species, Clarias gariepinus, still occurring in the
lake. Bathyclarias species and their ancestor C. gariepinus form a simple model that can
be used to understand the mechanisms of adaptation and rapid speciation in African
Great lakes. © 2001 Académie des sciences/Éditions scientifiques et médicales Elsevier
SAS
evolution / speciation / species flock / clariids
Résumé –L’essaim d’espèces Bathyclarias–Clarias, un nouveau modèle
pour comprendre les spéciations rapides dans les grands lacs africains. Les pois-
sons chats du genre Bathyclarias forment un essaim d’espèces (‘species flock’) dans le
lac Malawi. Les relations phylogénétiques entre sept espèces de Bathyclarias et d’autres
espèces de la famille ont été mises en évidence grâce à l’utilisation des séquences
partielles de cytochrome b. Nos observations montrent que cet essaim d’espèces a
évolué à partir d’une espèce toujours présente dans le lac, Clarias gariepinus. Cet essaim
d’espèces du genre Bathyclarias et leur ancêtre Clarias gariepinus forment un modèle
simple qui pourra être utilisé pour comprendre les mécanismes d’adaptation et de
spéciation rapide qui ont lieu dans les grands lacs africains. © 2001 Académie des
sciences/Éditions scientifiques et médicales Elsevier SAS
évolution / spéciation / species flock / Clariidae
. Version abrégée
Les grands lacs africains sont bien connus pour leurs
essaims d’espèces (‘species flocks’). Ces groupes
d’espèces ont pour caractéristiques principales d’être
monophylétiques, d’avoir évolué in situ, extrêmement
rapidement et abondamment. Pour cette raison, ils sont
très importants pour l’étude des mécanismes de
l’évolution et notamment en ce qui concerne la spécia-
tion sympatrique. On pense généralement que les
espèces à l’origine des essaims d’espèces étaient des
généralistes bien qu’aucune d’entre elles n’ai pu jusqu’à
*Correspondance and reprints.
E-mail address: agnese@crit.univ-montp2.fr, orstom@lion.meteo.go.ke (J.F. Agnèse).
683
C.R. Acad. Sci. Paris, Sciences de la vie / Life Sciences 324 (2001) 683–688
© 2001 Académie des sciences/Éditions scientifiques et médicales Elsevier SAS. Tous droits réservés
S0764446901013488/FLA

présent être identifiée. Dans la plupart des cas, ces
espèces ancestrales doivent être aujourd’hui éteintes.
Le genre Bathyclaras, avec douze espèces décrites, est
endémique du lac Malawi. C’est le seul essaim d’espèces
de poisson des grands lac africains qui ne soit pas
composéde Cichlidae. Afin de statuer sur l’origine
monophylétique ou polyphylétique des espèces du
genre Bathyclarias ainsi que pour établir leurs liens
phylogénétiques avec les autres Clariidae africains, une
étude des séquences partielles du cytochrome baété
entreprise. Pour cela, sept espèces du genre Bathycla-
rias ont étééchantillonnées dans le lac Malawi : B.
euryodon, B. gigas, B. ilesi B. longibarbis, B. nyaensis,
B. rotundifrons et B. worthingtoni. Des spécimens de
Clarias gariepinus ont étécollectés dans le lac Malawi,
les marais de Luapula (Zambie), la rivière Oubangui
(Centre Afrique) et le lac Manzalla (Égypte). Les autres
Clariidae étudiés comprennent : Clarias ngamensis (bas
Zambèze, Mosambique), C. stappersii,Heterobranchus
boulengeri (tous deux des marais Luapula en Zambie),
C. buthupogon (de la rivière Oubangui en Centre
Afrique), C. agboyiensis (de la rivière Ouéméau Benin)
et H. longifilis (de la lagune Ébriéen Côte d’Ivoire).
Clarotes laticeps, une espèce d’une famille proche des
Clariidae, les Claroteidae, a étéutilisée pour raciner le
réseau phylogénétique. Tous les spécimens étudiés ont
étédéposésaumusée royal de l’Afrique Centrale de
Tervuren (Belgique). Un fragment de 660 paires de
bases représentant une partie du cytochrome bde
l’ADN mitochondrial a étéamplifié grâce au couple
d’amorces L15267 et H15891. Les séquences obtenues
ont étédéposées dans la bibliothèque genBank/EMBL
sous les références AF126820 àAF126829 et AF235922
àAF235934.
Chaque espèce ou échantillon est caractérisépar un
haplotype propre àl’exception des espèces du genre
Bathyclarias pour lesquelles quatre haplotypes seule-
ment ont pu être observés : haplotype A (B. gigas,B.
longibarbis,B. euryodon,B. rotundifrons); haplotype
B(B. nyaensis); haplotype C (B. nyaensis et B. ilesi);
haplotype D (B. worthingtoni). Les divergences entre
séquences (p distance) vont de 0,2 à17 %. Comme
attendu, les divergences les plus grandes sont observées
entre la racine Clarotes laticeps et les autres espèces
(16 % à17 %). Les divergences les plus faibles ont été
observées entre d’une part les haplotypes C et D de
Bathyclarias et d’autre part entre l’haplotype A de
Bathyclarias et l’haplotype de C. gariepinus de Zam-
bie. Sur l’arbre consensus obtenu àpartir de la méth-
ode de ‘neighbor joining’, on peut observer plusieurs
groupes d’espèces : C. buthupogon et C. agboyiensis
sont placées ensemble comme le sont les deux espèces
du genre Heterobranchus.C. gariepinus de Centre
Afrique et d’Égypte sont regroupées avec un autre
ensemble composéde toutes les espèces du genre
Bathyclarias et C. gariepinus de Zambie et du lac
Malawi.
Tous les haplotypes de Bathyclarias et ceux de C.
gariepinus de Zambie et du lac Malawi forment un
groupe monophylétique supportépar une très forte
valeur de bootstrap (100 %). Ce groupe forme avec les
autres haplotypes de C. gariepinus un second groupe
monophylétique (bootstrap = 95 %). Les trois arbres
obtenus par la méthode de parsimonie sont similaires à
l’arbre consensus obtenu par la méthode précédente.
Tous trois ne diffèrent entre eux que par la position
respective des haplotypes de Clarias gariepinus du lac
Malawi, Bathyclarias C et D. Bien que les arbres
phylogénétiques ne permettent généralement pas de
reconstruire les relations d’ancêtre àdescendant, les
résultats observés sont ici particuliers. Les différentes
populations de Clarias gariepinus et les différentes
espèces de Bathyclarias forment un ensemble mono-
phylétique au sein duquel deux populations de C.
gariepinus sont groupe frère d’un ensemble constitué
des espèces du genre Bathyclarias et de deux autres
populations de C. gariepinus (lac Malawi et Zambie).
Cela implique que les espèces du genre Bathyclarias et
les populations de C. gariepinus du lac Malawi et de
Zambie ont un ancêtre commun qui leur est propre. Le
seul scénario évolutif qui puisse rendre compte de cet
arbre consiste àconsidérer qu’une population de C.
gariepinus a donnénaissance aux espèces du genre
Bathyclarias sans que l’espèce C. gariepinus ne cesse
d’exister.
Dans la lac Malawi, C. gariepinus est présent en eaux
peu profondes alors que les espèces du genre Bathy-
clarias sont rencontrées seulement en eau profonde,
jusqu’à 70 m. Certaines d’entre elles ne possèdent plus
l’organe suprabranchial caractéristique des Clariidae et
qui leur permet de respirer l’air atmosphérique.
L’âge de cet essaim d’espèces peut être évaluéà
l’aide des divergences observées entre les différents
haplotypes de Bathyclarias et ceux de C. gariepinus de
Zambie et du lac Malawi. Ces différents haplotypes
reflètent partiellement le polymorphisme ancestral des
populations fondatrices. Si l’on considère que la plus
faible différenciation entre un haplotype de Bathycla-
rias et un haplotype de Clarias gariepinus reflète la
différenciation maximum depuis l’événement de radia-
tion, on peut alors considérer que cet essaim d’espèces
est âgéde 140 à151 000 ans.
Ànotre connaissance, c’est le seul essaim d’espèces
pour lequel l’espèce ancestrale a étéidentifiéeetest
toujours présente sur les lieux de la radiation. L’étude
de ce modèle simple ancêtre–essaim d’espèces pourra
apporter beaucoup d’informations sur les mécanismes
d’adaptation et de spéciation dans les grands lacs
d’Afrique.
684
J.F. Agnèse, G.G. Teugels / C.R. Acad. Sci. Paris, Sciences de la vie / Life Sciences 324 (2001) 683–688

1. Introduction
East African lakes are generally known for their unique
large cichlid species flocks [1–5]. The latter are considered
as extremely important in the study of evolution [5, 6].
Although it is assumed that the ancestor of each of these
species flocks is a generalist species [6], it has not been
able so far to properly identify it. Most of these ancestral
species very likely became extinct, which is unfortunate
because they could have been informative in the under-
standing of mechanisms of adaptation and rapid specia-
tion in these lakes. Bathyclarias is a clariid catfish genus
(Siluriformes) endemic to Lake Malawi [7]. With its twelve
species it is the only non-cichlid species flock known from
the East African Rift valley lakes.
2. Materiel and Methods
Sequences of part of the cytochrome bwere used to
differentiate between a polyphyletic or monophyletic ori-
gin of the Bathyclarias species flock and to determine their
phyletic affinities with other clariid species, including
those present in the same region. Bathyclarias species [B.
euryodon, B. gigas, B. ilesi B. longibarbis, B. nyaensis, B.
rotundifrons and B. worthingtoni] were sampled in Lake
Malawi (Malawi); Clarias gariepinus specimens were col-
lected from Lake Malawi (Malawi), the Luapula Swamps
(Zambia), the Ubanghi River (Central African Republic)
and Lake Manzalla (Egypt). Other clariid material studied
included Clarias ngamensis (Lower Zambezi, Moçam-
bique), C. stappersii,Heterobranchus boulengeri (all from
Luapula Swamps, Zambia), C. buthupogon (Ubanghi River,
Central African Republic), C. agboyiensis (Oueme River,
Benin) and H. longifilis (Ebrie Lagoon, Ivory Coast). Clar-
otes laticeps, a Claroteidae species has been used to root
the networks. Specimens studied were deposited in the
Musée Royal de l’Afrique Centrale, Tervuren (Belgium).
Techniques used for tissues preservation, DNA extrac-
tion, DNA amplification and sequence analysis have been
described earlier [8]. Primers used (L15267 and H15891)
[9] enabled the amplification of a 660-base-pair (bp) frag-
ment of the mitochondrial DNA that comprised a small
part of the Glutamic acid transfer RNA (tRNA-Glu) and
nearly half of the cytochrome bgene. Sequences were
deposited in the genBank/EMBL data libraries under the
accession numbers AF126820 to AF126829 and
AF235922 to AF235934. Distances were estimated using
Kimura’s two-parameters method [10]. The phylogenetic
trees were constructed using two different methods: the
Maximum Parsimony technique (DNAPENNY program
[11]) and the neighbor joining method (NJ) using Kimura2
distances [12] (using SEQBOOT, DNADIST, NEIGHBOR
and CONSENSE programs [11]). For the NJ method, boot-
strap values have been assessed as support for internal
nodes present in the data matrix. A total of 500 replicates
of the sequences matrix was generated using the SEQ-
BOOT program [11] then 500 matrices of Kimura2 dis-
tances were generated using DNADIST program. Using
NEIGHBOR program, these matrices were transformed
into 500 trees and finally summarized in one consensus
tree using CONSENSE program. Number of specimen
studied are: 1 for Bathyclarias ilesi,Clarias buthupogon, C.
agboyiensis,Clarotes laticeps and Heterobranchus bou-
lengeri; 2 for B. euryodon, B.gigas , B. worthingtoni, C.
gariepinus from Central African Republic, Egypt, Malawi
and Zambia, C. ngamensis and C. stappersii ; 3 for B.
longibarbis, B. nyaensis,B. rotundifrons and H. longifilis.
3. Results
The mean base composition of the cytochrome b
sequences are similar to those previously observed in
actinopterygian fishes [13]: a low G content (14.9 %) and
almost equal A, T and C contents (respectively, 28.1, 27.7,
29.2 %). On the 553 bp segment analysed, 148 (26.7 %)
nucleotide positions were variable. Most variable sites
(116; 78.4 % of the total variable sites) were found at third
codon positions. Twenty five (16.9 %) were found at first
codon positions, and 7 (4.7 %) were at second codon
positions. Each species or sample was characterised by
one private haplotype, except for the Bathyclarias species
in which four different haplotypes were observed: haplo-
type A (B. gigas,B. longibarbis,B. euryodon,B. rotundi-
frons); haplotype B (B. nyaensis); haplotype C (B. nyaensis
and B. ilesi); haplotype D (B. worthingtoni). Uncorrected
sequence divergences (pdistance) among different species
ranged from 0.2 to 17 %. As expected, the highest
sequence divergences were observed between the out-
group (Clarias batrachus) and the other species (16 to
17 %). The lowest divergence (0.2 %) was observed
between Bathyclarias haplotypes C and D on one hand
and between Bathyclarias haplotype A and C. gariepinus
from Zambia on the other hand. 83 (56 %) of the 148
variable sites were informative for parsimony analysis (i.e.
showing at least two kinds of nucleotides, each present at
least twice). Of these phylogenetically informative sites,
74 (89.2 % of all informative sites) were at third codon
position, 8 (9.6 %) at first codon position, and 1 (1.2 %) at
second codon position. Empirical attempts were made to
detect putative saturation. The observed number of transi-
tions were plotted against the observed number of trans-
versions for all pairs of sequences and all codon positions.
Transitions did not increase when transversions were
higher than 16. Transition type differences were saturated
mainly when comparing the rooting species Clarotes lati-
ceps with other species. Following these results all substi-
tutions were considered. If we consider that a node is
significant when it is supported by a bootstrap value of
70 % and more [14–16], the following clusters can be
identified with the NJ consensus tree (figure 1): C. buth-
upogon and C. agboyiensis are grouped together as are the
685
J.F. Agnèse, G.G. Teugels / C.R. Acad. Sci. Paris, Sciences de la vie / Life Sciences 324 (2001) 683–688

two Heterobranchus species. C. gariepinus from Central
African Republic and Egypt are grouped with a cluster
composed of all the Bathyclarias haplotypes and C. gar-
iepinus from Zambia and Lake Malawi.
All Bathyclarias haplotypes and those of C. gariepinus
from Lake Malawi and Zambia formed a monophyletic
group supported by a very high bootstrap value (100 %).
This group and the two other C. gariepinus haplotypes
(Central African Republic and Egypt) also formed a mono-
phyletic group (bootstrap = 95 %). Each sample of C. gar-
iepinus is characterised by a private haplotype, reflecting
its widespread distribution (South, Central, West, East and
North Africa, Israel and Syria).
The three MP trees obtained only differ by the relative
positions of the different haplotypes of Clarias gariepinus
from Lake Malawi, Bathyclarias C and D and can be easily
represented by one single tree (figure 2). This tree is similar
to the NJ consensus tree previously described.
4. Discussion
A phylogenetic tree usually does not allow to define
ancestral–descendant relationships between different spe-
cies. It is not because two species are sister species that
one descents from the other. Generally, when one species
splits into two new species, this ancestor species disap-
pears in the same time. Nevertheless, in the present case,
it seems that speciation events have not used this usual
dichotomic way. If this would have been the case, we
should have observed a monophyletic group composed of
Bathyclarias species and C. gariepinus populations, in
which all Bathyclarias species would have been sister
species of all C. gariepinus populations. Our results clearly
indicated that all Clarias gariepinus populations do not
represent a monophyletic assemblage but rather a para-
phyletic one.
Species from the genus Bathyclarias together with C.
gariepinus from lake Malawi and C. gariepinus from Zam-
bia form a monophyletic group. This implies that they
share a common ancestor (figure 3). Then the only sce-
nario which can explain these observations is to consider
that Bathyclarias species originated from a C. gariepinus
population.
In Lake Malawi, C. gariepinus is present in shallow
waters; most Bathyclarias species only occur in deep
waters down to as much as 70 m [17]. The basis for the
occurrence of this species flock is adaptation to deep
water conditions. C. gariepinus, like most clariids, pos-
sesses a suprabranchial organ, formed by folds of the
second and the fourth branchial arches, which enables
aerial respiration. They are capable of walking on land for
distances of several hundred metres, breathing atmo-
spheric air and using their pectoral spines as support [18].
They are obligate air breathers and therefore C. gariepinus
is only found in shallow waters. In Bathyclarias species,
the suprabranchial chamber is reduced as are the arbores-
cent organs, which are even absent in some species. Living
at depth precludes air breathing and, based on the large
relative gill size in Bathyclarias species compared to C.
gariepinus, indicates that most Bathyclarias have returned
to a greater dependence upon aquatic respiration [19].
Clarias gariepinus is a generalist species: studies on its
feeding biology described it as an omnivorous scavenger
Figure 1. Consensus tree showing the phylogenetic relationships
between Bathyclarias species and other clariids. Clarotes laticeps,a
Claroteidae species, was used to root the network. Numbers repre-
sent bootstrap values (500 replicates). Scale bar = 0.007D.
Figure 2. One of the three Maximum Parsimony trees showing the
phylogenetic relationships between Bathyclarias species and other
clariids. Black dot indicates the only node where differences appeared
between trees. Each tree represented one of the possible arrange-
ments between Bathyclarias C, D and C. gariepinus from Lake
Malawi. Clarotes laticeps, a Claroteidae species, was used to root the
network.
686
J.F. Agnèse, G.G. Teugels / C.R. Acad. Sci. Paris, Sciences de la vie / Life Sciences 324 (2001) 683–688

[20] and its reproduction is characterised by a high fecun-
dity and no parental care [21].
What is the age of this species flock? African cichlid
species flocks are of different estimated ages ranging from
12 million years in Lake Tanganyika [22] to 12 400 years
in Lake Victoria [23] and not more than about 7 000 years
in Lake Natron [24]. Many of the Lake Malawi cichlid
species have radiated within the past 25 000 years and
some as recently as 200 years [25–26] in congruence with
the lake level variations. The age of the Bathyclarias
species flock can be evaluated using the sequence diver-
gences observed between the different Bathyclarias hap-
lotypes (A, B, C and D) on one hand and those of C.
gariepinus from Zambia and Lake Malawi on the other
hand. The different haplotypes observed may partly reflect
some intraspecific ancestral polymorphism because some
species possess two or more unrelated haplotypes [B.
nyaensis with haplotypes B and C and C. gariepinus for
which each sample possess a private haplotype]. Then the
lowest sequence divergence between a Bathyclarias and a
Clarias haplotype may represent the maximum differentia-
tion from the radiation event. If we use the rate for mito-
chondrial gene evolution estimated for other fishes
(1.2–1.3 % sequence divergence per million years [27]),
the maximum time divergence between Bathyclarias hap-
lotype A and the haplotype of C. gariepinus from Zambia
(0.18 %) which could correspond to the maximum age of
this species flock, is 140 000–151 000 years.
To our knowledge, this is the first species flock for which
the ancestral species has clearly been demonstrated and
for which the ancestor is still extant. This simple ancestor–
species flock model can provide useful information on
adaptation and speciation mechanisms in African Great
lakes.
Acknowledgements. We thank M. Banda, R. Bills, P.C.
Goudswaard and J. Sullivan for their assistance in col-
lecting the material. The contribution of GGT forms
part of the INCO.DC project on demersal fish commu-
nities of Lake Malawi financed by the European Com-
mission (IC18-CT97-0195).
References
[1] Avise J.C., Flocks of African fishes, Nature 347 (1990) 512–513.
[2] Fryer G., Iles T.D., The Cichlid Fishes of the Great Lakes of Africa,
Oliver & Boyd, Edinburgh, 1972.
[3] Meyer A., Kocher T.D., Basasibwaki P., Wilson A.C., Monophyletic
origin of Lake Victoria cichlid fishes suggested by mitochondrial DNA
sequences, Nature 347 (1990) 550–553.
[4] Sturmbauer C., Meyer A., Genetic divergence, speciation and
morphological stasis in a lineage of African cichlid fishes, Nature 358
(1992) 578–581.
[5] Lévêque C., Biodiversity dynamics and conservation; The freshwa-
ter fish of tropical Africa, Cambridge University Press, 1997.
[6] Greenwood P.H., African cichlids and evolutionary theories, in:
Echelle A.A., Kornfield I. (Eds.), Evolution of Fish species flocks, Univer-
sity of Maine Press, Orono, 1984, pp. 141–154.
[7] Anseaume L., Teugels G.G., On the rehabilitation of the clariid
catfish genus Bathyclarias endemic to the East African Rift Lake Malawi,
J. Fish Biol. 55 (1999) 405–419.
[8] Agnèse J.F., Teugels G.G., Monophyly of the genus Heterobranchus
(Siluriformes, Clariidae) and paraphyly of the genus Clarias: morphologi-
cal and genetical evidence, Copeia, in press, 2000.
[9] Briolay J., Galtier N., Brito R.M., Bouvet Y., Molecular phylogeny of
Cyprinidae Inferred from cytochrome bDNA Sequences, Mol. Phyl. Evol.
9 (1998) 100–108.
[10] Kimura M., A simple method for estimating evolutionary rates of
base substitutions through comparative studies of nucleotide sequences,
J. Mol. Evol. 16 (1980) 111–120.
[11] Felsenstein J., PHYLIP (Phylogeny Inference Package) Version
3,5c. (University of Washington, Seattle) (1993).
[12] Kimura M., A simple model for estimating evolutionary rates of
base substitutions through comparative studies of nucleotide sequences,
J. Mol. Evol. 16 (1980) 111–120.
[13] Lydeard C., Roe K.J., in Molecular systematics of fishes (eds.
Kocher T.D. & Stepien, C.A.) (Academic Press, San Diego) (1997).
[14] Zharkikh A., Li W.H., Statistical properties of bootstrap estimation
of phylogenetic variability from nucleotids sequences: II. Four taxa with-
out a molecular clock, Mol. Biol. Evol. 9 (1992) 1119–11147.
[15] Hillis D.M., Bull J.J., An empirical test of bootstrapping as a
method for assessing confidence in phylogenetic analysis, Systematic
Biology 42 (2) (1993) 182–192.
[16] Lecointre G., Philippe H., LêH.L.V., Le Guyader H., How many
nucleotides are required to resolve a phylogenetic problem? The use of a
new statistical method applicable to available sequences, Mol. Phyl. Evol.
3 (4) (1994) 292–309.
[17] Jackson P.B.N., Revision of the clariid catfishes of Nyasaland with
a description of a new genus and seven new species, Proc. Zool. Soc.
London 132 (1959) 109–128.
[18] Teugels G.G., A systematic revision of the African species of the
genus Clarias (pisces; Clariidae), Ann. Mus. R. Afr. Centr 247 (1) (1986)
1–199.
[19] Greenwood P.H.A., revision of the genus Dinotopterus Blgr.
(Pisces, Clariidae) with notes on the comparative anatomy of the
suprabranchial organs in the Clariidae, Bull. Br. Mus. (Nat. Hist.) 7 (1961)
215–241.
Figure 3. Reconstitution of the speciation events which occurred
during and before the origin of Bathyclarias species. All Bathyclarias
species and Clarias gariepinus from Lake Malawi and Zambia have a
common ancestor. This ancestor is very likely a population of Clarias
gariepinus.
687
J.F. Agnèse, G.G. Teugels / C.R. Acad. Sci. Paris, Sciences de la vie / Life Sciences 324 (2001) 683–688

[20] Bruton M.N., The food and feeding behavior of Clarias gariepinus
(Pisces: Clariidae) in Lake Sibaya, SouthAfrica,withempha-
sisonitsroleasapredatorofcichlids, Trans. Zool. Soc. Lond. 35 (1979)
47–114.
[21] Bruton M.N., The breeding biology and early development of
Clarias gariepinus (Pisces: Clariidae) in Lake Sibaya, South Africa with a
review of breeding in species of the subgenus Clarias (Clarias), Trans.
Zool. Soc. Lond. 35 (1979) 1–45.
[22] Sturmbauer C., Verheyn E., Rüber L., Meyer A., phylogenetic
patterns in populations of Cichlid Fishes from Rocky habitats in Lake
Tanganyika, in: Kocher Stepien, Stepien C.A. (Eds.), Molecular Systemat-
ics of Fishes, Academic Press, San Diego, 1997, pp. 97–111.
[23] Johnson T.C., Scholz C.A., Talbot M.R., Kelts K., Ricketts R.D.,
Ngobi G., Beuning K., Ssemmanda I., McGill J.W., Late pleistocene
desiccation of Lake Victoria and rapid evolution of cichlid fishes, Science
273 (1996) 1091–1093.
[24] Tichy H., Seegers L., The Oreochromis alcalicus flock (Teleostei:
Cichlidae) from lakes Natron and Magadi, Tanzania and Kenya: a model
for the evolution of “new”species flocks in historical times? Ichthyol.
Explor. Freshwaters 10 (2) (1999) 147–174.
[25] Moran P., Kornfield I., Retention of an ancestral polymorphism in
the mbuna species flock (Teleostei: Cichlidae) of Lake Malawi, Mol. Biol.
Evol. 10 (5) (1993) 1015–1029.
[26] Owen R.B., Crossley R., Johnson T.C., Tweddle D., Davidson S.,
Kornfield I., Eccles D.H., Engstrom D.E., Major low levels of lake Malawi
and their implications for speciation rates in cichlid fishes, Proc. R.. Soc.
Lond. [Biol.] 240 (1990) 519–533.
[27] Berminghan E., McCafferty S.S., Martin A.P., Fish Biogeography
and Molecular clocks: perspectives from the Panamanian Isthmus, in:
Kocher T.D., Stepien C.A. (Eds.), Molecular Systematics of Fishes, Aca-
demic Press, San Diego, 1997, pp. 113–128.
688
J.F. Agnèse, G.G. Teugels / C.R. Acad. Sci. Paris, Sciences de la vie / Life Sciences 324 (2001) 683–688