Content uploaded by Zuzana Musilová
Author content
All content in this area was uploaded by Zuzana Musilová on Dec 28, 2017
Content may be subject to copyright.
Molecular phylogeny and biogeography of the Neotropical cichlid
fish tribe Cichlasomatini (Teleostei: Cichlidae: Cichlasomatinae)
Zuzana Musilova
´
a,b,*
, Oldr
ˇich R
ˇı
´c
ˇan
b,c
, Karel Janko
b
, Jindr
ˇich Nova
´k
c
a
Department of Zoology, Charles University in Prague, Vinic
ˇna
´7, 128 44 Praha, Czech Republic
b
Institute of Animal Physiology and Genetics of the Academy of Sciences of the Czech Republic, Rumburska
´89, 277 21 Libe
ˇchov, Czech Republic
c
Department of Zoology, Faculty of Sciences, University of South Bohemia, Branis
ˇovska
´31, 37005 C
ˇeske
´Bude
ˇjovice, Czech Republic
Received 31 May 2007; revised 5 October 2007; accepted 11 October 2007
Available online 22 October 2007
Abstract
We have conducted the first comprehensive molecular phylogeny of the tribe Cichlasomatini including all valid genera as well as
important species of questionable generic status. To recover the relationships among cichlasomatine genera and to test their monophyly
we analyzed sequences from two mitochondrial (16S rRNA, cytochrome b) and one nuclear marker (first intron of S7 ribosomal gene)
totalling 2236 bp. Our data suggest that all genera except Aequidens are monophyletic, but we found important disagreements between
the traditional morphological relationships and the phylogeny based on our molecular data. Our analyses support the following conclu-
sions: (a) Aequidens sensu stricto is paraphyletic, including also Cichlasoma (CA clade); (b) Krobia is not closely related to Bujurquina and
includes also the Guyanan Aequidens species A. potaroensis and probably A. paloemeuensis (KA clade). (c) Bujurquina and Tah-
uantinsuyoa are sister groups, closely related to an undescribed genus formed by the ‘Aequidens’pulcher–‘Aequidens’rivulatus groups
(BTA clade). (d) Nannacara (plus Ivanacara) and Cleithracara are found as sister groups (NIC clade). Acaronia is most probably the
sister group of the BTA clade, and Laetacara may be the sister group of this clade. Estimation of divergence times suggests that the diver-
gence of Cichlasomatini started around 44 Mya with the vicariance between coastal rivers of the Guyanas (KA and NIC clades) and
remaining cis-andean South America, followed by evolution of the Acaronia–Laetacara–BTA clade in Western Amazon, and the CA
clade in the Eastern Amazon. Vicariant divergence has played importantly in evolution of cichlasomatine genera, with dispersal limited
to later range extension of species within genera.
Ó2007 Elsevier Inc. All rights reserved.
Keywords: Molecular phylogeny; Cichlids; Cichlidae; Cichlasomatini; South America; Biogeography; Penalized likelihood; Partitioned Bremer support;
Cytochrome b; 16S rRNA; S7; DIVA; Vicariance; Dispersal
1. Introduction
Neotropical cichlids are extremely varied in morphol-
ogy, behaviour and ecology (Lowe-McConnell, 1991)
despite comprising fewer species than their relatives in Afri-
can lakes. Since the early 20th century, when Regan
(1905a,b,c, 1906a,b) completely revised the group, the
number of species and genera has increased considerably.
Kullander (2003) summarizes the cichlid diversity in the
Neotropics as including 54 genera and 407 species. These
numbers are likely to increase significantly, as there still
are many undescribed species and genera (e.g. Kullander
and Ferreira, 2006; Kullander and Lucena, 2006; Lo
´pez-
Ferna
´ndez et al., 2006; R
ˇı
´c
ˇan and Kullander, 2006).
The cichlid subfamily Cichlasomatinae has one of the
most complex taxonomic histories among cichlid groups
in the Neotropics and no other group of Neotropical cich-
lids has witnessed so many taxonomic changes and new
descriptions at the genus level. The subfamily Cichlasomat-
inae was formally diagnosed by Kullander (1998) and
today includes species placed during the 19th and 20th cen-
1055-7903/$ - see front matter Ó2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ympev.2007.10.011
*
Corresponding author. Address: Institute of Animal Physiology and
Genetics, Academy of Sciences of the Czech Republic, Rumburska
´89, 277
21 Libe
ˇchov, Czech Republic.
E-mail address: zuzmus@email.cz (Z. Musilova
´).
www.elsevier.com/locate/ympev
Available online at www.sciencedirect.com
Molecular Phylogenetics and Evolution 46 (2008) 659–672
tury mainly in the genera Cichlasoma and Aequidens.Kul-
lander (1998) also divided the subfamily into two tribes,
the Heroini and the Cichlasomatini. Since this seminal
work, both the monophyly and division of the subfamily
have been supported by independent data (Farias et al.,
1999, 2000, 2001; Sparks, 2004; Marescalchi, 2005). Cic-
hlasomatine species have mostly been associated with
the genus Aequidens, while most heroines have been
placed in the catchall genus Cichlasoma. Since Kullander
(1998) the genus Cichlasoma has been placed in the tribe
Cichlasomatini, while most of the species previously con-
stituting Cichlasoma are now in the tribe Heroini (type
genus Heros Heckel 1840). Kullander’s studies (1983a,
1986; Kullander and Nijssen, 1989) rejected the classifica-
tion of Regan (1905a,b) based predominantly on the num-
ber of anal fin spines. Regan (op. cit.) placed Neotropical
cichlids into two groups, one with more than three spines
in the anal fin (nearly all placed in former Cichlasoma and
most of them distributed in Central America), and the
other one with three anal fin spines, distributed in many
genera (and subfamilies according to current
classification).
At present, the Neotropical cichlid fish tribe Cichlaso-
matini Kullander, 1998 comprises 69 valid species placed
in 10 valid genera (Aequidens,Bujurquina,Cichlasoma,
Cleithracara,Ivanacara,Krobia,Laetacara,Nannacara,
Tahuantinsuyoa and Acaronia). The generic assignment of
several species is questionable (e.g. ‘Aequidens’potaroensis,
‘Aequidens’paloemeuensis,‘Aequidens’hoehnei;Kullander,
1998) and some species groups likely represent unnamed
genera (i.e. the ‘Aequidens’pulcher group and ‘Aequidens’
rivulatus group; Kullander, 1998). Species of the genera
Aequidens,Bujurquina,Krobia,Cleithracara and Laetacara
were formerly placed in the catchall genus Aequidens previ-
ous to the taxonomic revisions of Kullander (1986) and
Kullander and Nijssen (1989), which excluded them from
Aequidens. In addition to its current twelve species, Cichla-
soma, another catchall group, contained also the majority
of Mesoamerican cichlids according to Regan (1905b).
These have been later classified as the tribe Heroini (Kul-
lander, 1998). A revision of the Mesoamerican Heroini for-
merly placed in Cichlasoma is still pending, but recent
studies have contributed to that goal (Roe et al., 1997;
Martin and Bermingham, 1998; Hulsey et al., 2004; Cha-
krabarty, 2006; R
ˇı
´c
ˇan and Kullander, 2006; Concheiro
Pe
´rez et al., 2007). The South American ex-Cichlasoma
(now Heroini) have a stable genus-level taxonomy (Caque-
taia and Heroina—Kullander, 1996;Australoheros—R
ˇı
´c
ˇan
and Kullander, 2006). The genus Nannacara established by
Regan (1905a) remained without intrageneric changes until
2007, when Ro
¨mer and Hahn (2007) described and sepa-
rated the genus Ivanacara for two species of Nannacara.
The genus Acaronia has been variously assigned to Heroini
(Stiassny, 1991) or a separate tribe Acaroniini (Kullander,
1998). These placements based on morphology are refuted
by all molecular analyses, which clearly place Acaronia
among the Cichlasomatini (Farias et al., 1999, 2000,
2001; Sparks and Smith, 2004; Marescalchi, 2005; Conche-
iro Pe
´rez et al., 2007).
The distribution of Cichlasomatini covers most of cis-
andean South America and to a lesser extent also trans-
andean South America including lower Central America.
Therefore, their distribution area covers biogeographically
crucial regions such as both Andean slopes, the Amazon
basin and the old geological formations (the Guyana and
Brazilian shields). Cichlasomatini are thus an ideally suited
model group for the study of historical biogeography and
evolutionary processes in the Neotropics, especially in
combination with their sister group, the predominantly
mesoamerican Heroini.
Kullander (1998) presented the first morphology-based
phylogeny which was not able to convincingly determine
relationships between and within the genera of Cichlaso-
matini. No robust tests of the relationships among genera
or their monophyly have been performed to date. Using
molecular markers and extensive taxonomical sampling
we evaluated for the first time the monophyly of cichlasom-
atine genera and their relationships. Our results contribute
to the ongoing investigation of the World’s richest (Neo-
tropical) freshwater biota, its diversity and biogeographic
history.
2. Materials and methods
2.1. Taxon sampling
Representative species of all valid genera of the tribe
Cichlasomatini as well as the genus Acaronia were included
in this study. We have strived to include multiple morpho-
logically and geographically distant species for each genus
to have as representative sampling as possible. The taxon
sampling includes 47 OTUs representing 41 species for all
three genes studied, i.e. the mitochondrial genes for cyto-
chrome b(cyt b) and 16S rRNA and the nuclear intron
in the ribosomal S7 gene. Specimens were wild-caught
and obtained from ornamental-fish importers with reliable
locality data, aquarium populations were used as well
(Table 1). Representative species of Heroini and Geopha-
gini were used as outgroup taxa.
2.2. DNA extraction, PCR amplification and sequencing
Genomic DNA was extracted from a fin clip (approx.
55 mm) using Dneasy
Ò
Tissue Kit (Qiagen), following
the manufacturer’s protocol. The polymerase chain reac-
tions (PCRs) used 1 ll of DNA as templates. Primers used
in the study are the following: for the 16S rRNA gene we
used the forward mtD-32 (50-CCG GTC TGA ACT
CAG ATC ACG T-30) and the reverse mtD-34 (50-CGC
CTG TTT AAC AAA AAC AT-30; both Marescalchi,
2005). The S7 intron primers S7RPE-X1F (50-TGG CCT
CTT CCT TGG CCG TC-30) and S7RPE-X2R (50-AAC
TCG TCT GGC TTT TCG CC-30) were from Chow and
Hazama (1998). The following combinations of cyt bprim-
660 Z. Musilova
´et al. / Molecular Phylogenetics and Evolution 46 (2008) 659–672
ers were used: forward Glu-DGL (50-TGA CTT GAA
RAA CCA YCG TTG-30;Palumbi et al., 1991), Fish-
CytB-F (50-ACC ACC GTT GTT ATT CAA CTA CAA
GAA C-30) and reverse H15149 (50-AAA CTG CAG
CCC CTC AGA ATG ATA TTT GTC CTC A-30;Kocher
et al., 1989), H15915 (50-AAC TGC CAG TCA TCT CCG
GGT TAC AAG AC-30;Irwin et al., 1991), Cb6b.H (GGA
ATT CAC CTC TCC-30;Martin and Bermingham, 1998)
and TrucCytB-R (50-CCG ACT TCC GGA TTA CAA
GAC CG-30). The PCR volume of 25 ll contained PPP
master mix (Top-Bio) including MgCl
2
(2.5 lM), both for-
ward and reverse primers (0.4 lM) and 1 ll of DNA sam-
ple. The reactions for all fragments consisted of an initial
denaturation step of 94 °C (2 min), followed by 35–40
cycles of denaturation at 94 °C for 1 min, annealing for
1 min at 46–50 °C for cyt b,48°C for 16S rDNA and
60 °C for the S7 intron and extension at 72 °C for 1 min.
The terminal extension was at 72 °C for 8 min. The reac-
Table 1
Species examined
Taxon name Source/locality Collection number Accession number
16S cyt bS7
Geophagus brasiliensis Aquarium stock ICCU 0700 AF049016 AF370659 EU199082
Thorichthys meeki Aquarium stock ICCU 0701 AY279669 U88860 EU199083
‘Aequidens’biseriatus Rio Atrato, Colombia ICCU 0702 EF432902 EF432958 EF432994
‘Aequidens’coeruleopunctatus Panama ICCU 0703 EF432905 EF432957 EF432997
‘Aequidens’pulcher AKV Aquarium stock ICCU 0704 EF432889 EF432944 EF432981
‘Aequidens’ cf. pulcher ‘‘Venezuela”Cunaviche, Los Llanos, Venezuela ICCU 0705 EF432888 EF432928 EF432980
‘Aequidens’ cf. pulcher ‘‘Rio Chirgua”Rio Chirgua, Carabobo, Venezuela ICCU 0706 EF432890 EF432934 EF432982
‘Aequidens’pulcher ‘‘Trinidad”Trinidad ICCU 0707 EF432887 EF432943 EF432979
‘Aequidens’rivulatus Rio Guayaquil, Ecuador ICCU 0708 EF432885 EF432935 EF432977
‘Aequidens’rivulatus AKV Aquarium stock ICCU 0709 EF432886 EF432935 EF432978
‘Aequidens’ sp. ‘‘Maracaibo”Maracaibo, Venezuela ICCU 0710 EF432904 EF432955 EF432996
Acaronia nassa Peru ICCU 0711 EF432897 EF432937 EF432989
Aequidens diadema Toboga
´n, Orinoco, Venezuela ICCU 0712 EF432880 EF432930 EF432972
Aequidens chimantanus Rio Caronı
´, Venezuela USBZC 1011 EF432884 EF432938 EF432976
Aequidens metae Pto Ayacucho, Venezuela ICCU 0713 EF432882 EF432927 EF432974
Aequidens patricki Aquarium stock ICCU 0714 EF432876 EF432913 EF432968
Aequidens potaroensis Las Claritas, Venezuela USBZC 1010 EF432870 EF432917 EF432962
Aequidens sp. ‘‘Atabapo”Rio Atabapo, Colombia ICCU 0715 EF432881 EF432924 EF432973
Aequidens sp. ‘‘Jenaro Herrera”Jenaro Herrera, Peru ICCU 0716 EF432912 EF432956 EF433004
Aequidens sp. Jaru
´Rio Jaru, Rondonia, Brazil ICCU 0717 EF432878 EF432926 EF432970
Aequidens tetramerus ‘‘Venezuela”Las Claritas, Venezuela ICCU 0718 EF432893 EF432929 EF432985
Aequidens tetramerus ‘‘Rio Negro”Rio Negro, Brazil ICCU 0719 EF432879 EF432936 EF432971
Aequidens tetramerus ‘‘Peru”Peru ICCU 0720 EF432911 EF432949 EF433003
Aequidens tubicen Rio Trombetas, Brazil ICCU 0724 EF432883 EF432945 EF432975
Bujurquina vittata Chaco, Argentina ICCU 0726 EF432892 EF432951 EF432984
Bujurquina peregrinabunda Aquarium stock ICCU 0727 EF432907 EF432954 EF432999
Bujurquina syspilus Aquarium stock ICCU 0728 EF432906 EF432952 EF432998
Bujurquina sp. ‘‘Maicuru
´”Maicuru
´ICCU 0729 EF432908 EF432953 EF433000
Cichlasoma amazonarum Iquitos, Peru ICCU 0730 EF432875 EF432914 EF432967
Cichlasoma bimaculatum Rio Tacutu, Bonfim, Brazil ICCU 0731 EF432874 EF432925 EF432966
Cichlasoma cf. araguaiense Rio Xingu, Altamira, Brazil ICCU 0732 EF432877 EF432923 EF432969
Cichlasoma dimerus Rio Paraguay, Paraguay ICCU 0733
*
EF432872 EF432941 EF432964
Cichlasoma orinocense Orinoco estuary, Venezuela ICCU 0734 EF432873 EF432922 EF432965
Cichlasoma cf. pusillum South America, exact locality unknown ICCU 0735 EF432871 EF432916 EF432963
Cleithracara maronii Aquarium stock ICCU 0736 EF432901 AY050614 EF432993
Krobia cf. guianensis Aquarium stock ICCU 0737
*
EF432910 EF432933 EF433002
Krobia sp. ‘‘Oyapock”Oyapock River, French Guyana ICCU 0738 EF432868 EF432932 EF432960
Krobia sp. ‘‘Xingu”Rio Xingu, Cachoeira Parati, Brazil ICCU 0739 EF432869 EF432931 EF432961
Laetacara curviceps Aquarium stock ICCU 0740 EF432895 EF432918 EF432987
Laetacara dorsigera Aquarium stock ICCU 0741 EF432894 EF432919 EF432986
Laetacara sp. ‘‘Buckelkopf”Aquarium stock ICCU 0742 EF432896 EF432940 EF432988
Laetacara thayeri Aquarium stock ICCU 0744 EF432909 AY050608 EF433001
Ivanacara adoketa Aquarium stock ICCU 0745 EF432903 EF432946 EF432995
Nannacara anomala Aquarium stock ICCU 0746 EF432898 AY050618 EF432990
Nannacara aureocephalus Aquarium stock ICCU 0747
*
EF432899 EF432939 EF432991
Nannacara taenia Aquarium stock ICCU 0749 EF432900 EF432921 EF432992
Tahuantinsuyoa macantzatza South America, exact locality unknown ICCU 0750 EF432891 EF432915 EF432983
The asterisk () in the collection number marks the sample, where only the fin tissue sample was obtained. The accession numbers beginning with EF and
EU mean original sequences from this study. Different letters are sequences downloaded from GenBank. ICCU, Ichthyological collections of Charles
University in Prague; USBZC, University of South Bohemia-Zoological Collections.
Z. Musilova
´et al. / Molecular Phylogenetics and Evolution 46 (2008) 659–672 661
tions were performed on thermo-cyclers PTC-200 (MJ
Research) and iCyclerTM Thermal Cycler (Bio-Rad).
PCR products were purified using the QIAquick
Ò
PCR
Purification Kit (Qiagen) and directly used as a template
for the sequencing reaction using the ABI PRISM
Ò
Big-
Dye
TM
Terminator v3.1 Ready Reaction Cycle Sequencing
Kit (Applied Biosystems). Sequences were read on the
3130 Genetic Analyser (Applied Biosystems, Hittachi)
automatic sequencer.
2.3. Phylogenetic analyses
Phylogenetic analyses included 47 taxa sequenced for
two mitochondrial genes (16S rRNA gene, cyt bgene)
and one nuclear marker (first intron in the S7 gene). All
obtained sequences were submitted to GenBank (Accession
Nos. EF432868–EF433004).
Sequences were aligned using the Clustal W program
with the default settings as implemented in the BioEdit
software package (Hall, 1999) and then were manually edi-
ted. The fragment of 16S rRNA was aligned with respect to
its secondary structure (Moyer et al., 2004). The amino
acid translation of the cyt bsequences was examined for
stop codons. Each gene was analyzed separately before a
combined analysis was performed in order to explore topo-
logical differences and possible sources of conflict between
the signals of individual genes. Additionally we performed
saturation test in cytochrome bdataset using PAUP
*
4.0b10 (Swofford, 2002).
Maximum parsimony (MP) analyses were performed in
PAUP under the heuristic search with 500 random addi-
tions, holding 10 trees in memory in each step. In the next
step we ran a search on the saved trees to find all the short-
est trees (command: hsearch addseq = random hold = 10
nchuck = 5 chuckscore = 1 nreps = 1000; hsearch start =
current nchuck = 0 chuckscore = 0). Branch support was
estimated with 1000 bootstrap replicates and also using
Partitioned Bremer Support (PBS; Baker and DeSalle,
1997) performed in PAUP
*
using constraint topologies
and the converse constraints command. The PBS reveals
the support for each node from every gene separately and
was used to pinpoint possible sources of conflict between
datasets at individual nodes, since it is a commonly used
measure of congruence in the genetic signal (e.g. Wahlberg
et al., 2005).
Maximum likelihood (ML) analyses were conducted in
PAUP
*
under the evolutionary model that best fit the ana-
lyzed sequence dataset. The model was selected using Mod-
eltest version 3.7 and the hierarchical likelihood (hLRT)
criterion (Posada and Crandall, 1998). Bootstrap analysis
with 1000 replications was performed in the software
PhyML (Guindon and Gascuel, 2003).
Bayesian analyses were performed in MrBayes, version
3.1 (Ronquist and Huelsenbeck, 2003). The Bayesian tree
was inferred using the models suggested by Modeltest.
We analyzed every gene separately, as well as in the com-
bined dataset. The analyses were performed with
1,000,000 (genes separately) to 2,500,000 (combined data-
set) generations, sampling every 100 trees using two paral-
lel runs with four chains. The final burnin was 40% of
sampled trees (i.e. 40,000 from the run of 1,000,000 gener-
ations). The combined dataset was partitioned into loops
and stems in the case of the 16S rRNA gene and codon
positions in the cytochrome bgene. Each analysis was rep-
licated to ensure that convergence was reached.
Monophyly of genera and higher clades was addition-
ally tested by the approximately unbiased test (AU test;
Shimodaira, 2002) as implemented in the software Consel
(Shimodaira and Hasegawa, 2001). This test assesses the
significance of differences in likelihood scores between the
best maximum likelihood (ML) tree and the tree con-
structed under the user-defined constraints. We thus con-
structed the tree with the contrained monophyly of
selected clades and tested whether there is significant differ-
ence against the best ML tree. The likelihood scores of con-
strained and unconstrained trees were then compared by
AU test.
We also applied the incongruence length difference par-
tition homogeneity test (ILD; Farris et al., 1995) as imple-
mented in PAUP to assess whether there are significant
topological differences between individual gene-trees before
combining all loci into single phylogenetic analysis. This
test was run under 1000 replicates with the randomly
selected characters from dataset (Farris et al., 1995).
2.4. Dating of divergences and biogeography
Since our dataset deviated from clocklike behaviour, we
could not use standard molecular clock estimation.
Instead, we have performed the Penalized likelihood anal-
ysis (Sanderson, 2002) in R8S software (Sanderson, 2003)
which allows to estimate absolute substitution rates and
divergence times under a relaxed molecular clock using
truncated Newton (TN) algorithm.
We used ML topology with branch length as input tree
for the analyses. The penalty function was set to additive
and cross validation was used to assess the most appropri-
ate value of smoothing. We have used geological dating to
calibrate the penalized likelihood tree. As a point of the
geological calibration we used the final separation of
Parana
´–Amazon drainages (10–11.8 Mya; Lundberg
et al., 1998; Montoya-Burgos, 2003) evident in the phylog-
eny of the genera Cichlasoma and Bujurquina.
The geographic distribution of individual species was
obtained from the Neodat II and Fishbase (www.fish-
base.org) databases (Froese and Pauly, 2006) and from
Stawikowski and Werner (1998). The studied samples are
coded as present in eight ichthyological provinces of South
America (Hubert and Renno, 2006). Dispersal–vicariance
analysis as implemented in DIVA version 1.1 (Ronquist,
1997) was used to explain the biogeographic history of cic-
hlasomatine cichlid fishes in South America in terms of
vicariance, dispersal and extinctions. This method places
one or more ancestral areas (depending upon the con-
662 Z. Musilova
´et al. / Molecular Phylogenetics and Evolution 46 (2008) 659–672
straints imposed) at each internal node so as to minimize
the costs associated with dispersal and extinction events
(Ronquist, 1997).
3. Results
3.1. Phylogenetic analyses
The final alignment of 47 taxa consisted of a 615 bp
fragment of the 16S rRNA, the complete cyt bgene
(1111 bp) and 510 bp of the first intron in the S7 gene
(2236 bp in total). The ILD test allowed combining all data
partitions in a single dataset (p= 0.09). We found slight
tendency of saturation for cytochrome bgene in the third
codon position. Nevertheless, when the third position was
excluded from analyses, the tree topology remained gener-
ally unchanged. The only difference was lower resolution
for some nodes.
Results of maximum parsimony (MP), maximum likeli-
hood (ML) and Bayesian analyses (BI) of the combined
dataset are shown in Fig. 1. The analyses resulted in very
similar topologies, different only in the position of the
genus Krobia and of the Nannacara–Ivanacara–Cleithra-
cara clade. The ML and BI topologies are almost identical
including estimated branch lengths, differing only in the
position of B. vittata–B. peregrinabunda and A. patricki–
C. cf. araguaiense.
Analyses of individual gene partitions also resulted in
very similar topologies (not shown), with conflicts limited
to nodes with low support. Possible sources of conflict
between the data partitions were further studied by analyz-
ing the PBS results, which found five points of conflict (see
Fig. 1). All of these are nodes having low bootstrap sup-
port and four nodes also differ between the MP and ML
(BI) topologies: (1) in the MP topology, Krobia sp.
‘‘Xingu”is the basal most Krobia, while in ML (BI) it is
Aequidens potaroensis; (2) in the MP topology Cichlasoma
is monophyletic, while in BI Aequidens patricki is nested
as a sister species to Cichlasoma, (in this case, ML shows
the same result as MP, see Fig. 1); (3) MP topology groups
Fig. 1. Results of phylogenetic analyses of the 47 taxon dataset including 2236 bp of three genetic markers (16S rRNA, cyt b, first intron in S7 gene). The
maximum parsimony (MP) analysis (A) is a strict consensus of six trees (L = 3930; CI = 0.3143; RI = 0.6356). The maximum likelihood (ML) analysis (B)
has a log-likelihood score of 20249.37366. (C) Bayesian inference tree. Numbers below nodes represent bootstrap support (1000 replications) in the MP
and ML analyses and posterior probabilities in the BI analysis. Numbers above nodes in the MP tree show Bremer support values and partitioned Bremer
Support in the following order: BS: 16S, cyt b, S7. The roman numbers in the MP tree link to the table above the tree. There are the BS and PBS values of
several nodes because of limited space in the tree figure. The ML and BI topologies differ only in two terminal nodes (position of B. vittata–B.
peregrinabunda and position of A. patricki–C. cf. araguaiense).
Z. Musilova
´et al. / Molecular Phylogenetics and Evolution 46 (2008) 659–672 663
the CA and BTA–Acaronia–Laetacara clades to the exclu-
sion of the NIC clade, while the ML (BI) topology groups
the NIC clade with the BTA–Acaronia–Laetacara clade to
the exclusion of the CA (and KA) clade; (4) the MP and
ML (BI) topologies differ in the position of the KA clade
(see below for explanation of clade acronyms).
3.1.1. Monophyly of individual genera
With the exception of the genera Tahuantinsuyoa,Ivana-
cara,Acaronia and the monotypic Cleithracara, our taxon
sampling includes multiple species per genus and thus
enables us to test their monophyly. Monophyly of all gen-
era except Aequidens is supported. The monophyly of
Aequidens was rejected by the AU test (p= 0.000005).
The genus Cichlasoma is nested within Aequidens. Further-
more, Aequidens potaroensis is not an Aequidens, but a
member of the genus Krobia (Fig. 1). Even after exclusion
of A. potaroensis is monophyly of Aequidens still rejected
by our data (AU test p= 0.005). ML analysis (Fig. 1B)
additionally did not recover a monophyletic Cichlasoma
since Aequidens patricki was placed as a basal Cichlasoma
to the exclusion of Cichlasoma cf. araguaiense.
Both the ‘Aequidens’pulcher group and the ‘Aequidens’
rivulatus group were monophyletic and strongly supported
sister taxa, unrelated to Aequidens. These two species
groups thus clearly represent an unnamed genus, justified
also on morphological grounds (OR
ˇunpublished results).
3.1.2. Relationships of cichlasomatine genera
We recovered four suprageneric clades among cichla-
somatine cichlids with strong support (Fig. 1). One clade
includes the genera Bujurquina,Tahuantinsuyoa and the
‘Aequidens’ pulcher–‘Aequidens’rivulatus group (termed
here the BTA clade). The second clade, referred to as the
CA clade, includes Aequidens (excluding A. potaroensis)
and Cichlasoma. The third clade (NIC clade) includes Nan-
Fig. 1 (continued)
Fig. 2. Dispersal–vicariance analysis (DIVA) of the biogeography of cichlasomatine cichlids in South America. (A) DIVA of the maximum parsimony
(MP) tree (see Fig. 1A). (B) DIVA of the maximum likelihood (ML) tree using the same dataset. Three samples of Aequidens tetramerus are shown as one
branch. ML and BI topologies (see Fig. 1) are almost identical, resulting in very similar mapping of biogeography. The map shows the Neotropical
biogeographic provinces based on Hubert and Renno (2006). Representatives of eight provinces were included in the analyses. Two provinces marked with
white color in the map (Magdalena River basin in Colombia and Atlantic coast of Brazil and Uruguay) show South American ichthyological provinces
from which we have no samples analyzed. Branches with more than three ancestral areas are marked by white color.
"
664 Z. Musilova
´et al. / Molecular Phylogenetics and Evolution 46 (2008) 659–672
nacara,Ivanacara and Cleithracara. Since Krobia also
includes Aequidens potaroensis, we refer to it as the KA
clade (Krobia–Aequidens potaroensis clade). The genera
Acaronia and Laetacara form successive sister groups to
the BTA clade, but the relationships are only weakly sup-
ported, except in the Bayesian analysis (Fig. 1C). Further-
Z. Musilova
´et al. / Molecular Phylogenetics and Evolution 46 (2008) 659–672 665
Fig. 2 (continued)
666 Z. Musilova
´et al. / Molecular Phylogenetics and Evolution 46 (2008) 659–672
more, the combined dataset analyzed with ML and BI
methods places the CA and KA clades as sister groups
(Fig. 1B and C), while the MP analysis places the KA clade
as the basal most cichlasomatine clade (Fig. 1A). Again
only the BI analysis found statistical support for the rela-
tionships of the KA clade.
3.2. Biogeography and dating of cichlasomatine
diversification
Biogeography of cichlasomatine cichlids has been stud-
ied using dispersal–vicariance analysis (DIVA; Ronquist,
1997) applied to the MP and ML (BI) topologies. The
ML and BI topologies differ in only two terminal nodes,
and thus the ML topology only was used for the DIVA
(the two terminal nodes in the BI analysis being in agree-
ment with those in the MP analysis). Due to differences at
basal nodes between the MP and ML (BI) topologies the
DIVA resulted in slightly different ancestral areas and bio-
geographical interpretations. Both topologies show a high
degree of vicariant events at basal nodes, with dispersal
events largely limited to terminal nodes. The MP topology
shows cis-andean South America excluding the La Plata
drainage as ancestral for the Cichlasomatini, while the
ML (BI) topologies include the La Plata drainage.
In the 45 nodes of the fully resolved phylogeny, DIVA
of the MP topology shows 17 vicariant events, while on
the ML (BI) topology there is one less (16). See Fig. 2
for the DIVA results and Fig. 3 for dating of cichlasoma-
tine phylogeny and of biogeographical events.
Cichlasomatine divergence started at around 44 Mya
(penalized likelihood estimation = 44.74 Mya) with a
vicariance between the Guyana rivers and cis-andean
South America (Orinoco, Western and Eastern Amazon,
and in the ML (BI) also the Tocantins–Xingu drainages
and the La Plata). This is the oldest vicariant event. Both
MP and ML (BI) topologies show this vicariance. In the
MP topology it is a single event, while in the ML (BI) there
are two parallel vicariances between these areas (Figs. 2
and 3). The Guyana rivers fauna is thus supported as very
old with a long independent history, which is also demon-
strated by the highest number of endemic genera per bio-
geographic province (see below).
The MP topology shows as a second event the vicariance
between Western Amazon and Orinoco–Eastern Amazon,
while the ML (BI) topology shows an extinction in Ori-
noco–Eastern Amazon plus the Tocantins–Xingu and La
Plata. Leaving aside the mechanism Western Amazon is
in both MP and ML (BI) topologies reconstructed as the
ancestral area of the Acaronia–Laetacara–BTA clade (at
ca 38 Mya). Both MP and ML (BI) topologies also show
colonization of trans-andean South America from the Wes-
tern Amazon at around 30 Mya, followed by vicariance
between the two areas (and within the ancestor of Western
Amazonian Bujurquina–Tahuantinsuyoa and trans-andean
‘Aequidens’). Both topologies also reconstruct a dispersal
of ‘Aequidens’ into the Orinoco and the Maracaibo from
trans-andean South America between 7 and 11 Mya, as
well as a dispersal of Bujurquina into the La Plata. Laeta-
cara has in both reconstructions colonized the Eastern
Amazon–Tocantins–Xingu areas, later diverging by vicari-
ance. Acaronia has similar to Laetacara extended its distri-
bution from the Western Amazon.
Reconstructions in the Aequidens–Cichlasoma clade (CA
clade) are more different between the MP and ML (BI)
topologies due to the different position of the Orinoco–
Guyanan endemic A. chimantanus and the interchanging
positions of A. patricki and C. araguaiense. The MP topol-
ogy shows a vicariance between A. chimantanus and all
remaining species, while the ML (BI) reconstruction is more
complex, with exctinction in the Eastern Amazon, Tocan-
tins–Xingu and La Plata. Both reconstructions however
agree on extensive dispersal and vicariance between the
Eastern Amazon, Western Amazon and Orinoco–upper
Rio Negro (in the A. tubicen–A. metae clade). A better
taxon sampling in this clade is needed to resolve the incon-
gruences between the MP and ML (BI) reconstructions. The
widespread A. tetramerus is under the current taxon sam-
pling reconstructed in both scenarios as ancestral in the
Eastern Amazon, later having dispersed into all cis-andean
South America except the La Plata. A. tetramerus popula-
tions have a long independent history, the oldest split being
ca 7 Mya old, and it is thus quite probable that several inde-
pendent species are hidden under the name. Cichlasoma
plus A. patricki have diverged from A. tetramerus and A.
sp. ‘‘Jaru”through vicariance between the Eastern Amazon
and rest of cis-andean South America (or alternatively
between East and West Amazon in the MP reconstruction).
As in the case of the A. tubicen–A. metae clade dispersal and
vicariance within cis-andean South America have been
extensive in the Cichlasoma plus A. patricki clade. The old-
est vicariance is the Xingu–Tocantins (C. araguainse;
around 16 Mya), followed by the La Plata (C. pusillum,C.
dimerus and C. portalegrense). The rest of Cichlasoma is
Western Amazonian–Orinocoan, from where it also sec-
ondarily colonized the Guyana rivers (C. bimaculatum).
In general the biogeographic analysis shows that vicari-
ance has been the main factor responsible for the generic
phylogeny of the Cichlasomatini and that genera have sec-
ondarily extended their distributions through dispersal.
Very few exctinctions are required to reconcile the dis-
persal–vicariance reconstruction. Biogeographic interpre-
tations at more inclusive levels within genera have to
await a better taxon sampling. However at the generic level
our taxon sampling is extensive, and except in the case of
Aequidens the ancestral areas of genera will probably
remain as reconstructed in this study.
4. Discussion
4.1. Phylogeny of Cichlasomatini
The systematics of cichlasomatine cichlids at the genus
level has received much attention (see Kullander, 1998
Z. Musilova
´et al. / Molecular Phylogenetics and Evolution 46 (2008) 659–672 667
for a review) and our results support the monophyly of all
but one genus. According to Kullander (1998),Cichlasoma
and Aequidens are very similar and well-diagnosed genera.
While our data support their close relationship, the genus
Aequidens appears to be paraphyletic (even after excluding
A. potaroensis), which is in agreement with indications
already present in Marescalchi (2005) or R
ˇı
´c
ˇan and Kul-
lander (2006). The paraphyly of Aequidens relative to Cic-
hlasoma has never been postulated on morphological
grounds. Some previous studies (Kullander, 1983b) sug-
gested the inclusion of two Guyanan Aequidens species
(A. potaroensis and A. paloemeuensis) into Krobia.Kul-
lander (op. cit.) already hypothesized the relation of these
species to Krobia, although he did not include them into
this genus (Kullander and Nijssen, 1989). He subsequently
omitted them from his phylogenetic analysis (Kullander,
1998). Our data clearly suggest the inclusion of the Guya-
nan river species (represented herein by A.potaroensis) into
Fig. 3. Dating of cichlasomatine phylogeny (ML topology) using the Penalized likelihood analysis as implemented in R8S software (Sanderson, 2003)
calibrated with geological dating (see Section 2).
668 Z. Musilova
´et al. / Molecular Phylogenetics and Evolution 46 (2008) 659–672
the genus Krobia.Aequidens paloemeuensis was not
included in our analysis, but according to Kullander and
Nijssen (1989),A. potaroensis and A. paloemeuensis are sis-
ter species. Our data did neither support nor reject the jus-
tification of the newly established genus Ivanacara.
Although the species I. adoketa (formerly Nannacara adok-
eta) resulted as a sister species to other Nannacara species
with the genetic uncorrected p-distance of about 20%
(19.4–20.8%) from the rest of species belonging to the
NIC clade, we lacked the other member of the genus, I.
bimaculata, to test the monophyly of Ivanacara and Nanna-
cara in our analyses (Fig. 4).
Three suprageneric clades are supported by all our analy-
ses. These are the Aequidens (Cichlasoma) clade (CA), the
Bujurquina–Tahuantinsuyoa–‘Aequidens’ clade (BTA) and
the Cleithracara–Nannacara–Ivanacara clade (NIC). Acaro-
nia is in all analyses the sister group of the BTA clade, and
Laetacara is the sister group of this clade (Fig. 4). Statistical
support for this grouping is however much lower than in the
previous cases. The genus Krobia and the NIC clade are basal
lineages in the cichlasomatine phylogeny. Krobia is either the
sister group of the CA clade (ML and BI analyses) or the
basal most lineage of cichlasomatines (MP analysis). The
NIC clade is similarly either the sister group of the BTA–
Acaronia–Laetacara clade (ML and BI), or of all clades
except the KA clade (MP). As can be seen, support decreases
and disagreements increase towards the more inclusive
nodes. Examination of our data suggests that the weak sup-
port and differences in basal nodes between MP and ML (BI)
analyses are a combination of lack of phylogenetic signal and
character conflict (the latter not being significant and may be
the result of the former).
The above-described suprageneric relationships of cic-
hlasomatine cichlids are very different from those reported
by Kullander (1998) based on phylogenetic analysis of
morphological characters. Our results agree with the mor-
phology-based phylogeny of Cichlasomatini of Kullander,
1998; see Fig. 5) in only one point, i.e. the close relationship
of Nannacara +Ivanacara and Cleithracara (herein the
NIC clade). All other suprageneric relationships are differ-
ent. There is however very good agreement with other stud-
ies based on molecular markers (Farias et al., 1999, 2000,
2001; Marescalchi, 2005; Concheiro Pe
´rez et al., 2007). In
agreement with these studies and contrary to Kullander
(1998) also the genus Acaronia is a part of the Cichlasoma-
tini, while Kullander (op. cit.) placed it as intermediate
between Cichlasomatini and Heroini. While Kullander
(op. cit.) placed the genera Krobia and Bujurquina as sister
groups, our analyses find them only distantly related.
Bujurquina is together with Tahuantinsuyoa and the ‘Aequi-
dens’pulcher–‘Aequidens’rivulatus group part of the
strongly supported BTA clade. This topology agrees well
with life history traits and biogeography, since Tah-
uantinsuyoa and Bujurquina are strongly supported sister
groups, with sympatric distribution and a mouthbrooding
behaviour, which is unique among cichlasomatine cichlids
(Stawikowski and Werner, 1998). The sister clade of these
two Western Amazonian genera is the ‘Aequidens’pul-
cher–‘Aequidens’rivulatus group, distributed both in cis-
Fig. 4. Topologies observed in the particular markers. Each clade is
shown if at least one of the Bayesian or the MP analyses supported its
monophyly. Resolved nodes are marked only if they appeared in both MP
and Bayesian analyses or in one of them with the bootstrap value >50 or
posterior probability value >80. Numbers above branch correspond to
bootstrap values in MP analyses, numbers below branches to Bayesian
posterior probabilities (100). Clade abbreviations are as follows: BTA
(Bujurquina,Tahuantinsuyoa,‘Aequidens’), CA (Cichlasoma, Aequidens),
KA (Krobia, Aequidens potaroensis), NIC (Nannacara,Ivanacara and
Cleithracara), La (Laetacara genus), Ac (Acaronia nassa). The BTA + Ac
grouping in the 16S analysis depicts a position of Acaronia within the BTA
clade. White clades in the 16S analysis mark their unresolved topology.
Fig. 5. Comparison of the morphological phylogeny of cichlasomatines
(Kullander, 1998) and our molecular study. Gray circles mark congruence
in topology between the two studies, white crossed circles mark conflicts.
All tested genera except Aequidens are supported as monophyletic in both
studies, but relationships between genera are mostly different and in
conflict. The monophyly of the genera Tahuantinsuyoa,Krobia and
Cleithracara was not tested since only one valid species per genus
analyzed. ‘Aequidens’hoehnei was not included in our molecular analysis.
The genus Acaronia is not shown as it was considered to a separate tribe
Acaroniini by Kullander (1998). Our results place it firmly among
cichlasomatines. Tree modified from Kullander (1998).
Z. Musilova
´et al. / Molecular Phylogenetics and Evolution 46 (2008) 659–672 669
and trans-andean northern South America. Krobia is either
related to Aequidens–Cichlasoma (ML and BI analyses), or
the basal most genus among Cichlasomatini (MP analysis).
Laetacara is in all our analyses the sister group of the BTA
clade. None of our analyses thus supports its relationship
to the NIC clade (Nannacara,Ivanacara,Cleithracara)as
found in Kullander (1998; see Fig. 5).
4.2. Biogeography and dating of divergence
Cichlasomatini and Heroini have wide and more or less
complementary distributions. While the latter are distrib-
uted throughout the Neotropics from Argentina up to Mex-
ico, being mostly represented in Mesoamerica and covering
also trans-andean South America, western Amazonia and
the Parana
´–Paraguay–Sa
˜o Francisco drainages, the former
group appears strictly in South America, with only one spe-
cies (‘Aequidens’coeruleopunctatus) present in lower Central
America. The distribution of both tribes is to some extent
complementary also in South America as heroines are
mostly restricted to northern and western Amazonia,
trans-andean South America and the Parana
´–Paraguay–
Sa
˜o Francisco drainages (Brasilian shield) (Concheiro Pe
´rez
et al., 2007), while cichlasomatines are well represented in
eastern Amazonia and especially on and around the Guy-
ana shield (eastern Amazonia, Guiana rivers and Orinoco)
where heroines are almost completely absent.
The divergence between the Guyana rivers and rest of
South America is well reflected in our biogeographic anal-
yses, being the oldest vicariant event in the phylogeny of
Cichlasomatini. The origin of the genera Krobia and the
NIC clade (Nannacara,Ivanacara,Cleithracara) dates back
to this old vicariance event dated at around 44 Mya. The
area was later additionally colonized by species of several
other genera (e.g. Cichlasoma bimaculatum,Acaronia nassa,
Aequidens tetramerus). Dating of these secondary invasions
is hard to ascertain from our data and taxon sampling, but
source areas for the colonizations can be pinpointed. In
case of Cichlasoma it was Western Amazon–Orinoco plus
Upper Rio Negro (C. taenia from Trinidad and Orinoco
delta further supporting this route was not present in our
analysis), in the case of Acaronia it was the Western Ama-
zon (probably also including Orinoco plus Upper Rio
Negro, as A. vultuosa present there was not included in
our analysis), and in case of Aequidens tetramerus it was
the Eastern Amazon based on the present taxon sampling.
However a much better locality sampling of this wide-
spread species group (see above) is needed to reconcile its
biogeography. The first two species thus colonized the
Guyana rivers probably through river anastomosis in the
Orinoco delta area during a period of lower sea levels,
while Aequidens might have dispersed through the low
lying Rupununi area between Guyana and Brazil or
through river anastomosis in the present day Amazon delta
following the change of the Amazon outlet to the east
between the Guyana and Brazilian shields (Lundberg
et al., 1998), or even through the same route as Cichlasoma
and Acaronia. It is not possible to rule out even indepen-
dent invasions into the Guyana rivers through several
routes by the A. tetramerus superspecies.
The Guyana shield, an area divided into three areas of
endemism (Orinoco– Upper Rio Negro, Eastern Amazon
and the coastal Guyana rivers), harbours the highest level
of endemism and diversity among cichlasomatine cichlids
(eight genera, i.e. Cleithracara,Krobia, Nannacara, Ivana-
cara, Acaronia,Aequidens,Cichlasoma and marginally Lae-
tacara). The former four are endemic or have their centre
of distribution in the Guyana shield. On the contrary, only
five genera, none of them endemic (Cichlasoma,Aequidens,
Laetacara,Bujurquina,Krobia; latter three only margin-
ally), are found in the much larger Brazilian shield. Whole
Western Amazonia possesses five genera (Bujurquina,
Aequidens,Cichlasoma,Laetacara,Tahuantinsuyoa) and
only one of them is endemic (Tahuantinsuyoa;Bujurquina
is nearly endemic). Western Amazonia has however based
on our results played importantly in the evolution of Cic-
hlasomatini, as the whole BTA clade (Bujurquina–Tah-
uantinsuyoa–‘Aequidens’) plus Acaronia and Laetacara
have their ancestral area in Western Amazonia. From Wes-
tern Amazonia, this clade has colonized trans-andean
South America (30 Mya) and also the La Plata (10–
12 Mya). In contrast to the Western Amazonian clade,
the Aequidens–Cichlasoma clade is better represented in
eastern South America (in Eastern Amazon and Orinoco
plus Upper Rio Negro), which is also its reconstructed
ancestral area in the DIVA based on the MP topology
(Fig. 2A). The oldest species of this clade in the Orinoco
province is A. chimantanus, in MP topology DIVA recon-
struction the basal most species of the whole clade.
The two largest clades of Cichlasomatini thus have com-
plementary and largely exclusive distributions in South
America, one (the BTA clade plus Acaronia and Laetacara)
in the Western Amazon, the other (CA clade) in Eastern
Amazon. Orinoco plus Upper Rio Negro is an intermediate
area, today still connected to the Eastern Amazon through
several points (the best known being the Casiquiare), while
previously being the northern portion of the Western Ama-
zon previous to the final rise of the Andes (Lundberg et al.,
1998).
The cichlasomatine fauna of the Xingu–Tocantins prov-
ince is supported as being old and independent for a long
time (e.g. Hubert and Renno, 2006). Within Krobia and
Nannacara it has vicariant relationships to the Guyana
Rivers province (ca. 15 Mya in Krobia and ca. 10 Mya in
Nannacara), while in Laetacara and Cichlasoma with
Amazonia in the former (around 22 Mya) and most of
cis-andean South America in the latter (around 16 Mya).
The dating of age and initial divergence within Cichlaso-
matini is placed around 44 Mya, which is in good agree-
ment with the independently estimated age of heroines,
the cichlasomatine sister group (Concheiro Pe
´rez et al.,
2007). Heroines very likely colonized Mesoamerica at least
16–24 Mya (at least 10 Mya prior to the final emergence of
the Isthmus of Panama) in agreement with other secondary
670 Z. Musilova
´et al. / Molecular Phylogenetics and Evolution 46 (2008) 659–672
freshwater fishes (Concheiro Pe
´rez et al., 2007; Hrbek
et al., 2007). Although the colonization of Mesoamerica
by the only cichlasomatine species (‘Aequidens’coeruleo-
punctatus) likely followed the same route from trans-
andean South America (Figs. 2 and 3), it happened much
more recently, after the separation from the remaining spe-
cies at around 10 Mya (Fig. 3). This is in agreement with
Bussing’s (1976, 1985) postulate that cichlasomatines
belong to the New Southern Element and heroines to the
Old Southern Element of the Central American freshwater
ichthyofauna. The Heroini have radiated in at least two lin-
eages in Mesoamerica already at around 16–24 Mya
(Concheiro Pe
´rez et al., 2007). Based on our biogeographic
analyses, ‘Aequidens’ has been present in trans-andean
South America as long as the Heroini (ca. 28 Myr; Figs.
2and3), and the time discrepancy between the heroine
and cichlasomatine colonization of Central America thus
cannot be explained using the available studies and is wor-
thy of further study.
All six ‘Aequidens’ species are primarily trans-andean,
including the putative sister group of ‘A.’pulcher (A. lati-
frons), which was not available for analyses. ‘Aequidens’
pulcher is thus the only species of the group that is second-
arily found also in cis-andean South America (Orinoco
river Basin in Venezuelan llanos and on Trinidad). Based
on our analyses this secondary dispersal into cis-andean
South America and vicariance from trans-andean relatives
is dated at ca. 10 Mya (Figs. 2 and 3). This dating is well in
agreement with the final rise of the easternmost chain of the
Andes in Venezuela, isolating the Maracaibo drainage
from surrounding coastal areas and from the Llanos (dated
by Lundberg et al., 1998 at ca. 10–11.8 Mya).
The internal phylogenetic structure of cichlasomatine
cichlids contrasts with the Heroini, where the genera are
vaguely defined and except for monotypic ones they are
not supported as monophyletic (Concheiro Pe
´rez et al.,
2007). Cichlasomatine genera, on the other hand, are well
differentiated as suggested by phylogenetic support for all
but one (Aequidens) genus. Internodes are much longer
among cichlasomatine cichlids compared to heroines (cf.
Concheiro Pe
´rez et al., 2007) suggesting completely different
cladogenetic histories. While at least the Mesoamerican her-
oine diversity is probably the result of repeated radiations
(Concheiro Pe
´rez et al., 2007;OR
ˇunpublished results), cic-
hlasomatines seem to be shaped primarily by interspaced
vicariant events. Studies comparing the macroevolution of
the two tribes are thus potentially extremely informative
on Neotropical biogeography and faunal evolution.
We have provided the first detailed molecular phyloge-
netic study of Cichlasomatini cichlids and found patterns
of diversification that are very different from those in their
sister group—the Heroini cichlids (Concheiro Pe
´rez et al.,
2007). Our data allowed us to solve the systematic position
of all genera belonging to this tribe and their large scale
biogeography in South America. Furthermore, we offer a
comparison with the morphology-based phylogenetic
hypothesis of this cichlid assemblage.
Acknowledgments
We thank Rainer Stawikowski, Lutz Krahnefeld, Ing-
oma Kratz, Antonı
´n Prouza, Aqua Ferrytale and Hobby
Zoo Tillmann for help with obtaining several samples.
Thanks are also to Vendula S
ˇlechtova
´, Ingo Schindler, Petr
Ra
´b, Ivan C
ˇepic
ˇka, J. Andres Lo
´pez and two anonymous
reviewers for the interesting discussion and for all com-
ments significantly improving the manuscript. Daniel Fry-
nta encouraged the initial stage of the study. Thanks to
Petr Jans
ˇta, Luka
´s
ˇKubic
ˇka, Zuzana Starostova
´, Milan
Kaftan and Miroslav S
ˇva
´tora for their help with the fish
manipulation. We also thank M. Noren (Stockholm) for
providing the TrucCytB-R primer. We would like to thank
the Biodiversity Research group for providing us labora-
tory background.
This work was supported by the Grant Agency of
Charles University in Prague (GAUK 182/2004/B-BIO/
PrF and GAUK 139407), Hla
´vka foundation and the foun-
dation of the Centre for Biodiversity LC06073 (MSMT) to
Z.M., and by the MSM6007665801 grant to OR
ˇ. The Insti-
tute of Animal Physiology and Genetics receives continu-
ous support from IRP IAPG (No. AV0Z50450515).
References
Baker, R.H., DeSalle, R., 1997. Multiple sources of character information
and the phylogeny of Hawaiian Drosophilids. Syst. Biol. 46 (4), 654–673.
Bussing, W.A., 1976. Geographical distribution of the San Juan ichthy-
ofauna of Central America with remarks on its origin and ecology. In:
Thorson, T.B. (Ed.), Investigations of the Ichthyofauna of Nicaraguan
Lakes. University of Nebraska, Lincoln, pp. 157–175.
Bussing, W.A., 1985. Patterns of distribution of the Central American
ichthyofauna. In: Stehli, F.G., Webb, S.D. (Eds.), The Great American
Biotic Interchange. Plenum Publishing, pp. 453–473.
Chakrabarty, P., 2006. Systematics and historical biogeography of
Greater Antillean Cichlidae. Mol. Phylogenet. Evol. 39 (3), 619–627.
Chow, S., Hazama, K., 1998. Universal PCR primers for S7 ribosomal
protein gene introns in fish. Mol. Ecol. 7, 1255–1256.
Concheiro Pe
´rez, G.A., R
ˇı
´c
ˇan, O., Ortı
´, G., Bermingham, E., Doadrio, I.,
Zardoya, R., 2007. Phylogeny and biogeography of 91 species of
heroine cichlids (Teleostei: Cichlidae) based on sequences of the
cytochrome bgene. Mol. Phylogenet. Evol. 43, 91–110.
Farias, I.P., Ortı
´, G., Sampaio, I., Schneider, H., Meyer, A., 1999.
Mitochondrial DNA phylogeny of the family Cichlidae: monophyly
and fast molecular evolution of the Neotropical assemblage. J. Mol.
Evol. 48, 703–711.
Farias, I.P., Ortı
´, G., Meyer, A., 2000. Total evidence: molecules,
morphology, and the phylogenetics of cichlid fishes. J. Exp. Zool. 288,
76–92.
Farias, I.P., Ortı
´, G., Sampaio, I., Schneider, H., Meyer, A., 2001. The
Cytochrom bgene as a phylogenetic marker: the limits of resolution for
analyzing relationships among Cichlid fishes. J. Mol. Evol. 53, 89–103.
Farris, J.S., Kallersjo, M., Kluge, A.G., 1995. Constructing a significance
test for incongruence. Syst. Biol. 44, 570–572.
Froese, R., Pauly, D., 2006. FishBase. Published by author. <http://
fishbase.com/search.php/>.
Guindon, S., Gascuel, O., 2003. PhyML - A simple, fast, and accurate
algorithm to estimate large phylogenies by maximum likelihood. Syst.
Biol. 52, 696–704.
Hall, T.A., 1999. Bio Edit: a user-friendly biological sequence alignment
editor and analysis program for Windows 95/98/NT. Nucleic Acids
Symp. 41, 95–98.
Z. Musilova
´et al. / Molecular Phylogenetics and Evolution 46 (2008) 659–672 671
Hrbek, T., Seckinger, J., Meyer, A., 2007. A phylogenetic and biogeo-
graphic perspective on the evolution of poeciliid fishes. Mol. Phylo-
genet. Evol. 43, 986–998.
Hubert, N., Renno, J.F., 2006. Historical biogeography of South
American freshwater fishes. J. Biogeogr. 33, 1414–1436.
Hulsey, C.D., de Leon, F.J.G., Johnson, Y.S., Hendrickson, D.A., Near,
T.J., 2004. Temporal diversification of Mesoamerican cichlid fishes
across a major biogeographic boundary. Mol. Phylogenet. Evol. 31 (2),
754–764.
Irwin, D.M., Kocher, T.D., Wilson, A.C., 1991. Evolution of the
cytochrome b gene in mammals. J. Mol. Evol. 32, 128–144.
Kocher, T.D., Thomas, W.K., Meyer, A., Edwards, S.V., Paabo, S.,
Villablanca, F.X., Wilson, A.C., 1989. Dynamics of mitochondrial—
DNA evolution in animals—amplification and sequencing with con-
served primers. Proc. Natl. Acad. Sci. USA 86, 6196–6200.
Kullander, S.O., 1983a. A Revision of the South American Cichlid genus
Cichlasoma. Swedish Museum of Natural History, Stockholm, Sweden.
Kullander, S.O., 1983b. Taxonomic Studies on the Percoid Freshwater
Fish Family Cichlidae in South America. PhD. thesis, Stockholm.
Kullander, S.O., 1986. Cichlid fishes of the Amazon River Drainage of
Peru. Swedish Museum of Natural History, Stockholm, Sweden.
Kullander, S.O., 1996. Heroina isonycterin a, a new genus and species of
cichlid fish from western Amazonia, with comments on cichlasomine
systematics. Ichthyol. Explor. Freshwaters 7, 149–172.
Kullander, S.O., 1998. A Phylogeny and Classification of the South
American Cichlidae (Teleostei: Perciformes). In: Malabarba, L.R., Reis,
R.P., Lucena, Z.M., Lucena, C.A.S. (Eds.), Phylogeny and Classifica-
tion of Neotropical Fishes. Edipucrs, Porto Alegre, Brazil, pp. 461–498.
Kullander, S.O., 2003. Family Cichlidae (Cichlids). In: Reis, R.E.,
Kullander, S.O., Ferraris, C.J., Jr. (Eds.), Check List of the Freshwater
Fishes of South and Central America. Edipucrs, Porto Alegre, pp.
605–654.
Kullander, S.O., Nijssen, H., 1989. The Cichlids of Surinam. Teleostei:
Labroidei. E.J. Brill, Leiden, The Netherlands.
Kullander, S.O., Ferreira, E.J.G., 2006. A review of the South American
cichlid genus Cichla, with descriptions of nine new species (Teleostei:
Cichlidae). Ichthyol. Explor. Freshwaters 17, 289–398.
Kullander, S.O., Lucena, C.A.S., 2006. A review of the species of
Crenicichla (Teleostei: Cichlidae) from the Atlantic coastal rivers of
southeastern Brazil from Bahia to Rio Grande do Sul States, with
descriptions of three new species. Neotrop. Ichthyol. 4, 127–146.
Lo
´pez-Ferna
´ndez, H., Taphorn, D.C., Kullander, S.O., 2006. Two new
species of Guianacara from the Guiana Shield of Eastern Venezuela
(Perciformes: Cichlidae). Copeia 2006 (3), 384–395.
Lowe-McConnell, R.H., 1991. Ecology of cichlids in South American and
African waters, excluding the African Great Lakes. In: Keensleyside,
M.H.A. (Ed.), Cichlid Fishes: Behaviour, Ecology and Evolution.
Chapman Hall, London, Great Britain, pp. 60–85.
Lundberg, J.G., Marshal, L.G., Guerrero, J., Horton, B., Malabarba,
M.C.S.L., Wesselingh, F., 1998. The stage for Neotropical fish diver-
sification: a history of tropical South American rivers. In: Malabarba,
L.R., Reis, R.P., Lucena, Z.M., Lucena, C.A.S. (Eds.), Phylogeny and
Classification of Neotropical Fishes. Edipucrs, Porto Alegre, Brazil, pp.
13–48.
Marescalchi, O., 2005. Karyotype and mitochondrial 16S gene character-
izations in seven South American Cichlasomatini species (Perciformes,
Cichlidae). J. Zool. Syst. Evol. Res. 43, 22–28.
Martin, A.P., Bermingham, E., 1998. Systematics and evolution of lower
central American Cichlids inferred from analysis of cytochrome bgene
sequences. Mol. Phylogenet. Evol. 9, 192–203.
MontoyaBurgos, J.I., 2003. Historical biogeography of the catfish genus
Hypostomus (Siluriformes: Loricariidae), with implications on the
diversification of Neotropical ichthyofauna. Mol. Ecol. 12, 1855–1867.
Moyer, G.R., Burr, B.M., Krajewski, C., 2004. Phylogenetic relationships
of thorny catfishes (Siluriformes: Doradidae) inferred from molecular
and morphological data. Zool. J. Linn. Soc. 140, 551–575.
Palumbi, S., Martin, A.P., Romano, S., McMillan, W.O., Stice, L.,
Grabowski, G., 1991. The Simple Fool’s Guide to PCR, Version 2.0.
Univ. of Hawai i Press, Honolulu.
Posada, D., Crandall, C.A., 1998. MODELTEST: testing the model of
DNA substitution. Bioinformatics 14, 817–818.
Regan, C.T., 1905a. A revision of the fishes of the American cichlid genus
Acara,Nannacara,Acaropsis and Astronotus. Ann. Mag. Nat. Hist. 15,
329–347.
Regan, C.T., 1905b. A revision of the fishes of the American cichlid genus
Cichlosoma and of the allied genera. Ann. Mag. Nat. Hist. 7 (15), 60–
67, 225–243, 316–340, 433–445.
Regan, C.T., 1905c. A revision of the fishes of the South-American cichlid
genera Crenacara,Batrachops, and Crenicichla. Proc. zool. Soc. Lond.
1905, 152–168.
Regan, C.T., 1906a. A revision of the South American cichlid genera
Retroculus,Geophagus,Heterogramma, and Biotoecus. Ann. Mag. Nat.
Hist. 17 (7), 49–66.
Regan, C.T., 1906b. A revision of the fishes of the South American
cichlid genera Cichla,Chaetobranchus, and Chaetobranchopsis, with
notes on the genera of American Cichlidae. Ann. Mag. Nat. Hist.
17 (7), 230–239.
R
ˇı
´c
ˇan, O., Kullander, S.O., 2006. Character and tree-based delimitation of
species in the ‘Cichlasoma’ facetum group (Teleostei, Cichlidae) with
the description of a new genus. J. Zool. Syst. Evol. Res. 44, 136–152.
Roe, K.J., Conkel, D., Lydeard, C., 1997. Molecular systematics of middle
American cichlid fishes and the evolution of trophic-types in ‘Cichla-
soma (Amphilophus)’ and ‘C-(Thorichthys)’. Mol. Phylogenet. Evol. 7,
366–376.
Ro
¨mer, U., Hahn, I., 2007. Ivanacara gen. n. (Teleostei: Perciformes,
Cichlasomatini)—a new genus of cichlids from the Neotropis. In:
Ro
¨mer, U. (Ed.), Cichlid Atlas, volume 2, Natural History of South
American Dwarf Cichlids, Part 2. Mergus Verlag GmbH, Melle,
Germany, pp. 1190–1197.
Ronquist, F., 1997. Dispersal–vicariance analysis: a new approach to
the quantification of historical biogeography. Syst. Biol. 46 (1),
195–203.
Ronquist, F., Huelsenbeck, J.P., 2003. MrBayes 3: Bayesian phylogenetic
inference under mixed models. Bioinformatics 19, 1572–1574.
Sanderson, M.J., 2002. Estimating absolute rates of molecular evolution
and divergence times: a penalized likelihood approach. Mol. Biol.
Evol. 19, 101–109.
Sanderson, M.J., 2003. R8S: inferring absolute rates of molecular
evolution and divergence times in the absence of a molecular clock.
Bioinformatics 19, 301–302.
Shimodaira, H., 2002. An approximately unbiased test of phylogenetic
tree selection. Syst. Biol. 51, 492–508.
Shimodaira, H., Hasegawa, M., 2001. CONSEL: for assessing the
confidence of phylogenetic tree selection. Bioinformatics 17, 1246–1247.
Sparks, J.S., 2004. Molecular phylogeny and biogeography of the
Malagasy and South Asian cichlids (Teleostei: Perciformes: Cichlidae).
Mol. Phylogenet. Evol. 30, 599–614.
Sparks, J.S., Smith, W.L., 2004. Phylogeny and biogeography of cichlid
fishes (Teleostei: Perciformes: Cichlidae). Cladistics 20, 501–517.
Stawikowski, R., Werner, U., 1998. Die Buntbarsche Amerikas. Band 1.
Eugen Ulmer GmbH and Co., Stuttgart, Germany.
Stiassny, M.L.J., 1991. Phylogenetic intrarelationships of the family
Cichlidae: an overview. In: Keensleyside, M.H.A. (Ed.), Cichlid
Fishes: Behaviour, Ecology and Evolution. Chapman Hall, London,
Great Britain, pp. 1–35.
Swofford, D.L., 2002. PAUP
*
. Phylogenetic analysis using parsimony (and
other methods), version 4.0b10. Sinauer Associates, Sunderland,
Massachusetts.
Wahlberg, N., Brower, A.V.Z., Nylin, S., 2005. Phylogenetic relationships
and historical biogeography of tribes and genera in the subfamily
Nymphalinae (Lepidoptera: Nymphalidae). Biol. J. Linn. Soc. 86, 227–
251.
672 Z. Musilova
´et al. / Molecular Phylogenetics and Evolution 46 (2008) 659–672