全 文 ：Journal of Systematics and Evolution 46 (3): 424–438 (2008) doi: 10.3724/SP.J.1002.2008.08062
(formerly Acta Phytotaxonomica Sinica) http://www.plantsystematics.com
Inferring the Tree of Life of the order Cypriniformes, the earth’s most
diverse clade of freshwater fishes: Implications of
varied taxon and character sampling
1Richard L. MAYDEN* 1Kevin L. TANG 1Robert M. WOOD 1Wei-Jen CHEN 1Mary K. AGNEW
1Kevin W. CONWAY 1Lei YANG 2Andrew M. SIMONS 3Henry L. BART 4Phillip M. HARRIS
5Junbing LI 5Xuzhen WANG 6Kenji SAITOH 5Shunping HE 5Huanzhang LIU
5Yiyu CHEN 7Mutsumi NISHIDA 8Masaki MIYA
1(Department of Biology, Saint Louis University, St. Louis, MO 63103, USA)
2(Department of Fisheries, Wildlife, and Conservation Biology, Bell Museum of Natural History, University of Minnesota, St. Paul, MN 55108, USA)
3(Department of Ecology and Evolutionary Biology, Tulane University, New Orleans, LA 70118, USA)
4(Department of Biological Sciences, The University of Alabama, Tuscaloosa, AL 35487, USA)
5(Laboratory of Fish Phylogenetics and Biogeography, Institute of Hydrobiology, Chinese Academy of Sciences, Wuhan 430072, China)
6(Tohoku National Fisheries Research Institute, Fisheries Research Agency, Miyagi 985-0001, Japan)
7(Ocean Research Institute, University of Tokyo, Tokyo 164-8639, Japan)
8(Department of Zoology, Natural History Museum & Institute, Chiba 260-8682, Japan)
Abstract The phylogenetic relationships of species are fundamental to any biological investigation, including
all evolutionary studies. Accurate inferences of sister group relationships provide the researcher with an historical
framework within which the attributes or geographic origin of species (or supraspecific groups) evolved. Taken
out of this phylogenetic context, interpretations of evolutionary processes or origins, geographic distributions, or
speciation rates and mechanisms, are subject to nothing less than a biological experiment without controls.
Cypriniformes is the most diverse clade of freshwater fishes with estimates of diversity of nearly 3,500 species.
These fishes display an amazing array of morphological, ecological, behavioral, and geographic diversity and
offer a tremendous opportunity to enhance our understanding of the biotic and abiotic factors associated with
diversification and adaptation to environments. Given the nearly global distribution of these fishes, they serve as
an important model group for a plethora of biological investigations, including indicator species for future cli-
matic changes. The occurrence of the zebrafish, Danio rerio, in this order makes this clade a critical component in
understanding and predicting the relationship between mutagenesis and phenotypic expressions in vertebrates,
including humans. With the tremendous diversity in Cypriniformes, our understanding of their phylogenetic
relationships has not proceeded at an acceptable rate, despite a plethora of morphological and more recent mo-
lecular studies. Most studies are pre-Hennigian in origin or include relatively small numbers of taxa. Given that
analyses of small numbers of taxa for molecular characters can be compromised by peculiarities of long-branch
attraction and nodal-density effect, it is critical that significant progress in our understanding of the relationships
of these important fishes occurs with increasing sampling of species to mitigate these potential problems. The
recent Cypriniformes Tree of Life initiative is an effort to achieve this goal with morphological and molecular
(mitochondrial and nuclear) data. In this early synthesis of our understanding of the phylogenetic relationships of
these fishes, all types of data have contributed historically to improving our understanding, but not all analyses are
complementary in taxon sampling, thus precluding direct understanding of the impact of taxon sampling on
achieving accurate phylogenetic inferences. However, recent molecular studies do provide some insight and in
some instances taxon sampling can be implicated as a variable that can influence sister group relationships. Other
instances may also exist but without inclusion of more taxa for both mitochondrial and nuclear genes, one cannot
distinguish between inferences being dictated by taxon sampling or the origins of the molecular data.
Key words Cobitoidea, Cypriniformes, Cyprinoidea, taxon sampling, Tree of Life.

———————————
Received: 27 April 2008 Accepted: 9 May 2008
* Author for correspondence. E-mail: cypriniformes@gmail.com; Tel.: 1-314-977-3910; Fax: 1-314-977-3658.
MAYDEN et al.: Inferring the Tree of Life of the order Cypriniformes

425
Nothing in Biology Makes Sense Except in the Light of Evolution.
Theodosius Dobzhansky. 1973.
The American Biology Teacher

Phylogenetic reconstruction is fundamental to comparative biology research … as the phylogeneti-
cists’ conclusions (i.e., their phylogenetic inferences) become the comparative biologists’ assumptions.
Consequently, the generation of robust phylogenetic hypotheses and the understanding of the factors in-
fluencing accuracy in phylogenetic reconstruction are crucial to evolutionary hypothesis testing.
Antonis Rokas and Sean B. Carroll (2005: 1337).

This famous quote from Dobzhansky emerged
over a century after Darwin’s transformational treatise
on descent with modification changed the face of
comparative biology. During the long hiatus following
Darwin’s hypothesis, much controversy emerged
about the idea of evolution and speciation. This
persisted until advances in genetics and population
biology made these disciplines which scientists could
use to provide first order, hypothesis-driven explana-
tions for evolution and speciation. While dominating
the field of evolutionary biology for decades, neither
this discipline nor the related disciplines of taxonomy
or systematics could offer a satisfactory theoretical
framework for reconstructing historical patterns of
speciation or evolutionary mechanisms underlying the
tree of life. This held true even after Hennig’s meth-
odology, now a fundamental part of phylogenetic
systematics, had already been published in
“Grundzüge einer Theorie der phylogenetischen
Systematik” (1950) (the translated English version
“Phylogenetic Systematics” that received a much
wider international distribution was not published
until 1966). Integral in Hennig’s work was the evalua-
tion of both population and species level evolution of
traits or attributes, speciation, and inheritance of these
traits throughout the tree of life. His theories were
quite contrary to the traditional view of systematic
biology of the time and, as a discipline dealing with
systematics, were not viewed as worthy of investiga-
tion by geneticists and population biologists, then
forging new ideas on evolution. As with any paradigm
shift (Kuhn, 1962), significant controversy and estab-
lished inertia delayed an unbiased assessment and the
eventual adoption of this now widely accepted phi-
losophical and methodological transformation, offer-
ing for the first time a mechanism for researchers to
reconstruct testable histories (species trees) of life.
Thus, at the time of Dobzhansky’s famous assessment
of biology, few really knew of or understood the
significance of Hennig’s work. This was much like the
theory of continental drift and the works of Alfred
Wegener (Wegener, 1915) which had almost no
impact on its field when originally published. Recog-
nition for both of these important interdisciplinary
scientists and their work did not come until after their
deaths.
In the wake of the transformation of the biologi-
cal community following the adoption of Hennig’s
theory and methods for reconstructing testable hy-
potheses of evolutionary relationships (i.e., phylogen-
ies), it is clear that Dobzhansky’s assessment of the
essential foundations of biology requires reconsidera-
tion. Rather, as indicated in the second quote by
Rokas and Carroll (2005), it is more appropriate to
keep in mind that nothing in biology makes sense
except in the light of the phylogenetic relationships of
species, including evolution. In fact, the conclusions
of any phylogenetic study of a group of organisms
serve as the beginning of any other biological investi-
gation of these same organisms. Only when a re-
searcher is equipped with a hypothesis of the sister
group relationships of targeted populations, species, or
supraspecific natural (monophyletic) taxa can one
address the variety of possible biological questions
about any of the organisms or species, including
evolutionary studies.
Given the obvious significance of hypotheses of
the genealogical relationships of species, it is critical
that these hypotheses reflect as closely as possible the
genealogical history of the species, including sister
group relationships, ancestral state reconstructions,
and branch lengths derived from meaningful optimi-
zations. Sister group relationships are critical for
appropriate comparisons, biogeographic investiga-
tions, and an eventual resolution of modes of speci-
ation. Ancestral character reconstructions and branch
lengths are essential to researchers investigating many
areas of evolution concerning ages of clades, rates of
anagenesis and cladogenesis, tracing the descent of
attributes, as well as other domains of comparative
biology.
Various methods have been advocated for the in-
ference of historical relationships of species that
involve both parametric and non-parametric algo-
rithms. In a perfect world, such a hypothesis of rela-
tionships would include all of the species within the
Journal of Systematics and Evolution Vol. 46 No. 3 2008 426
group and an abundance of data with straightforward
homology assessment, the resulting phylogeny can be
used by anyone to investigate all aspects of the species
and interpret its traits and its geographic distribution
within an historical framework. This perfect situation
rarely exists and several important considerations
must be given to any analysis, to allow a researcher to
most accurately infer relationships. At least three
issues must be considered for accurate phylogenetic
reconstructions; these are 1) accurate analytical
methods, 2) selection of appropriate and enough
character data (e.g., morphology, molecular, behav-
ioral) for reliable inference, and 3) appropriate selec-
tion of taxa for the question at hand (Swofford et al.,
1996).
The computational demands inherent in phy-
logenetic reconstructions, representing a type of
NP-complete problem (Graham & Foulds, 1982), are
astounding as the number of possible resolutions
increases exponentially as the number of taxa in-
volved also increases. Historically, this prevented
scientists from examining many species and a focus
was placed on an increase in the number of characters
for the analysis. Few morphological studies, however,
come close to or exceed 100 or more characters (see
Scotland et al., 2003). With the efficiency of generat-
ing molecular sequence data increasing and the cost
decreasing, the ability to produce hundreds and thou-
sands of potential characters quickly became the
standard approach to assemble systematic data sets
and soon the number of molecular phylogenies rela-
tive to those based on morphological data dramatically
increased. With this increase in the amount of data
from sequences and the character states being limited
to “A-T-G-C,” model-based analyses and a number of
sophisticated evolutionary models have changed the
face of phylogenetic systematics (Yang, 1996; Sulli-
van & Swofford, 1997; Stamatakis, 2006a).
The ease with which DNA sequence data may be
obtained, combined with the improved models and
model-based analyses, has led to rapid and widespread
adoption of these methods by the community. How-
ever, for a variety of reasons, many researchers none-
theless limited analyses to only a subset of taxa for a
proposed clade, often because of unacceptable com-
putation time, availability of specimens, and/or lim-
ited funds to collect large numbers of homologous
sequences. This has resulted in a culture of scientists
focusing on relatively few taxa with an abundance of
character data. Analyses of relationships among
relatively few taxa based on complete or nearly com-
plete genome data represent an excellent example of
one extreme (e.g., Inoue et al., 2001; Miya et al.,
2003; Mabuchi et al., 2007).
In recent years, algorithms have improved com-
putation time for data sets with large numbers of taxa
(e.g., Huelsenbeck & Ronquist, 2001; Stamatakis,
2006b). With this flexibility, one may then ask, when
given a choice, which of the two possible variables
should be increased, taxa or character data, to increase
the accuracy of the inferred phylogeny? This question
has been the focus of a number of studies and consid-
erable debate (Hillis et al., 2003; Rokas et al., 2003;
Rosenberg & Kumar, 2003; Cummings & Meyer,
2005; Rokas et al., 2005; Hedtke et al., 2006) and a
review of this controversy is provided in Heath et al.
(2008). The overwhelming evidence supports increas-
ing taxon sampling, even at the expense of great
quantities of character data, for improved accuracy of
topologies. In simulation studies, increased taxon
sampling appears to be more consequential than
increasing the number of characters for reaching the
“true” relationships in a group (Hillis, 1996). Several
authors have also agreed that the addition of species in
analyses results in more accurate estimates of rela-
tionships (Lecointre et al., 1993; Hillis, 1996, 1998;
Graybeal, 1998; Rannala et al., 1998; Zwickl & Hillis,
2002; Pollock et al., 2002; Poe, 2003; DeBry, 2005;
Hedtke et al., 2006). Furthermore, empirical studies
have attributed problematic reconstructions and poorly
resolved trees to researchers limiting analyses to an
inadequate number of taxa (Bremer et al., 1999;
Johnson, 2001; Lin et al., 2002; Braun & Kimball,
2002; Chen et al., 2003; Sorenson et al., 2003;
Albrecht et al., 2007).
The essential problems with focusing only on in-
creasing characters at the expense of taxa involves
complications with estimates of unobserved changes
or transformations in a tree—consequently poor
estimates of evolutionary models or a resulting matrix
that precludes parsimony from arriving at a correct
solution. If there are not enough taxa in an analysis
then it is difficult to accurately estimate parameters for
evolutionary models as there will be too many unob-
served changes inherent in a matrix (Felsenstein,
1978; Hendy & Penny, 1989; DeBry, 1992; Huelsen-
beck & Hillis, 1993; Yang, 1994; Huelsenbeck, 1995;
Gascuel et al., 2001; Huelsenbeck & Lander, 2003;
Susko et al., 2004). Serious complications include
either long-branch attraction (Felsenstein, 1978) or a
nodal-density effect (Gojobori et al., 1982; Fitch &
Bruschi, 1987; Fitch & Beintema, 1990; Bruno &
Halpern, 1999; Hugall & Lee, 2007), or both. Here, a
limited sampling of species results in an artificial
MAYDEN et al.: Inferring the Tree of Life of the order Cypriniformes

427
accumulation of apomorphies possessed by species
(ancestral or descendant) because taxa are missing
from intervening nodes that would presumably “break
up” branches and more realistically disperse character
change in the phylogeny (apomorphies and homo-
plasy) (Wiens, 2005). Long-branch attraction results
from an accumulation of phylogenetic noise or homo-
plasy in two or more non-adjacent taxa that is inter-
preted as an accumulation of homologous characters
between two closely related species. In either case, a
restriction of taxon sampling, even with a limited
number of characters, can result in phylogenetic noise
(homoplasy via convergences, reversals or substitu-
tions) overwhelming the phylogenetic signal.
Cypriniformes is known as the most diverse
group of freshwater fishes with estimates of diversity
reaching close to 3,500 species (Nelson, 2006). The
family occurs in habitats ranging from lakes and rivers
to small springs and streams in Eurasia, North Amer-
ica, and Africa. Species of this order (particularly
those of the Cyprinidae) are usually perceived as
having very similar morphologies (Howes, 1991), an
attribute that has likely contributed to the paucity of
researchers investigating their phylogenetic relation-
ships because of suspected conserved or constrained
evolution limiting the number of phylogenetically
useful morphological characters. These fishes also
include many commercially important species (e.g.,
aquarium trade, fisheries), and as model organisms in
many areas of research ranging from community
ecology to developmental biology. The zebrafish or
zebra danio (Danio rerio), goldfish (Carassius aura-
tus), algae eater (Gyrinocheilus aymonieri), and many
carp species (Cyprinus, Hypophthalmichthys, Cteno-
pharyngodon) likely represent the most familiar
members of this diverse clade.
As species of Cypriniformes are diverse compo-
nents in most of the freshwater habitats around the
globe and serve as important model organisms in
comparative research, the phylogenetic relationships
of these fishes serve as a critical, historical framework
aiding directed research. Unfortunately, either consid-
erable uncertainty exists over the relationships of
these fishes, even at the higher levels, or there are no
data providing insight into the relationships of these
fishes. Furthermore, some of the species and clades
previously examined for phylogenetic relationships
have been problematic in their resolution, despite
efforts to increase character sampling to resolve their
sister group relationships. Most notable among these
are the subfamilial relationships within the families
Cyprinidae, Catostomidae, Cobitidae, and Balitoridae,
and the phylogenetic position of the common aquar-
ium species, the algae eater, and relatives in the genus
Gyrinocheilus (Gyrinocheilidae).
Herein, we examine the relationships of major
clades within Cypriniformes as they relate to the
impact of taxon sampling, while intentionally holding
character sampling constant. We focus particularly on
the historically problematic taxa identified above and
how increasing taxa from 49 to 110 species in the
analysis alters their phylogenetic placement and
support for their sister group relationships. These
comparisons are made relative to a previous analysis
of these same species based on a more limited sam-
pling of taxa (53 species) but with whole mitochon-
drial genomes (Saitoh et al., 2006) containing signifi-
cantly more character data (14,563 bp) than used in
this analysis (1497 bp). This analysis is based on
sequences of exon 3 of recombination activating gene
1 (RAG1), a commonly used gene in phylogenetic
relationships of gnathostome vertebrates (e.g., Groth
& Barrowclough, 1999; Waddell & Shelley, 2003;
San Mauro et al., 2004; Steppan et al., 2004; Krenz et
al., 2005), including ray-finned fishes (e.g., López et
al., 2004; Rüber et al., 2004; Holcroft, 2005; Sullivan
et al., 2006; Chen et al., 2007; Mayden et al., 2007).
This research is part of an ongoing international Tree
of Life initiative on Cypriniformes and is aimed at
furthering our understanding of not only the phyloge-
netic relationships of these species but also improving
our understanding of effective and accurate methods
for phylogeny reconstruction.
1 Methods
Taxon sampling attempted to match that of Sai-
toh et al. (2006). In cases where the same species was
not available, a congeneric representative was chosen
as a substitute, if possible. We conducted analyses on
two different sets of taxa. For the first analysis, we
compiled a data matrix of RAG1 sequence data to
match the species from Saitoh et al. (2006; fig. 2 &
table 2); this small data set included 7 outgroups and
49 cypriniform fishes. To explore the effects of taxon
sampling, our second analysis added an additional 61
cypriniform taxa; these included an additional two
species of catostomids, two cobitids, three botiids,
four balitorids, and 50 cyprinids (one acheilognathin,
one cultrin, one squaliobarbin, three gobionins, five
cyprinins, 17 rasborins, and 22 leuciscins). Of those,
23 RAG1 were sequences downloaded from Gen-
Bank. Please see Table 1 for a complete list of taxa
examined for this study.
Journal of Systematics and Evolution Vol. 46 No. 3 2008 428
RAG1 was chosen for this study as it is not part
of the mitochondrial genome, thus not overlapping
with the data presented in Saitoh et al. (2006), thereby
enabling an independent assessment of relationships.
In addition, RAG1 has been demonstrated to be
phylogenetically informative for this level of rela-
tionship (e.g., López et al., 2004; Rüber et al., 2004;
Holcroft, 2005; Sullivan et al., 2006; Chen et al.,
2007; Mayden et al., 2007). Methods for DNA data
collection followed standard procedures, as outlined in
Conway et al. (2008), a previous study that utilized
exon 3 of RAG1. The primers, RAG1F1 and
RAG1R1, published in López et al. (2004) and the
primer, R1-4061R, published in Chen et al. (2007)
were used to amplify and sequence approximately
1500 bp of this loci. Sequences were deposited in
GenBank (Table 1).
Bayesian analyses were conducted with the par-
allel version of MrBayes v.3.1.2 (Huelsenbeck &
Ronquist, 2001; Ronquist & Huelsenbeck, 2003;
Altekar et al., 2004). Seven non-cypriniform outgroup
taxa within the Ostariophysi were used as outgroups;
Chanos chanos was designated as the most distant
outgroup. Prior to the analysis, the sequence data were
partitioned by codon position and MrModelTest v.2.2
(Nylander, 2004) and PAUP* (Swofford, 2002) were
used to perform hierarchical likelihood ratio tests
(hLRT) on each partition to determine the most ap-
propriate model of nucleotide substitution. The best-fit
model for all three codon positions of RAG1, in both
the small and large data sets, was found to be
GTR+I+Γ, which was applied during the MrBayes
analyses. Two independent Bayesian searches were
conducted for each data set (four total), each search
ran for 1,000,000 generations, with 4 chains, sampling
every 1,000 generations. The distribution of log
likelihood scores was examined to determine station-
arity and burn-in time for each search. In both
searches involving the small data set, stationarity in
log likelihood scores was observed after approxi-
mately 30,000–40,000 generations. Trees from the
first 50,000 generations (51 trees) were discarded to
ensure all burn-in trees were excluded. This left 950
trees from each search, which were combined to form
a common pool of 1900 trees, these were then used to
construct the 50% majority-rule consensus. In the two
analyses of the large data set, stationarity was not
observed until after 50,000–60,000 generations; to
guarantee that all the trees examined were
post-burn-in, the first 101 trees from each search
(representing 100,000 generations) were discarded.
The remaining 1800 trees were used to generate the
50% majority-rule consensus tree. Branch support for
each clade was based on posterior probability values,
indicated by the frequency of occurrence of each clade
among the trees retained after the initial burn-in
topologies were discarded.
2 Results and Discussion
As with any extremely diverse group, the phy-
logenetic relationships of Cypriniformes has had a
troublesome history. In fact, some authors have
described the relationships among species, genera,
subfamilies and families as largely “chaotic” (Hubbs
& Miller, 1977; Mayden, 1989; Mayden et al., 2007).
Some taxa have been especially problematic in our
ability to confidently decipher their sister-group
relationships to develop a phylogenetically informa-
tive classification. The phylogenetic placement of
Gyrinocheilus and Tinca, two of the most phyloge-
netically problematic genera in the order, and the
naturalness of the traditionally recognized Balitoridae
and Cobitidae (now recognized as two families,
Botiidae and Cobitidae sensu Šlechtová et al., 2007),
as well as the naturalness and phylogenetic relation-
ships of the subfamilies of Catostomidae and Cypri-
nidae have all been difficult to seemingly intractable
problems. Much of the difficulty with these systematic
issues may, in part, owe its origin to the historic
difficulties in character assessment, obtaining taxa for
a global-wide Cypriniformes analysis, and the recent
emergence of molecular analyses for the order (see
Mayden et al., 2007). However, universal to all of the
molecular studies for the order has been the examina-
tion of a limited number of taxa, in most cases less
than 60–70 species and this degree of taxon sampling
has only occurred in the last few years. Historically,
even single morphological analyses of taxa in this
order have been limited in the number of taxa and
characters, with the notable exception of explicit
phylogenetic studies by Sawada (1982), Mayden
(1989), and Smith (1992). Consistency in the resolu-
tion of recent molecular trees for the traditionally
recognized Cobitidae led Šlechtová et al. (2007) to
elevate the subfamilies within Cobitidae (Botiinae and
Cobitinae) to the family level, thus resolving the
apparent polyphyly of the family. Their taxon sam-
pling of the Balitoridae did not provide the diversity to
identify the problems with this family identified
herein. No analyses have focused on resolving the
relationship of Gyrinocheilus, Tinca, or the subfamily
naturalness and relationships within Cyprinidae.
MAYDEN et al.: Inferring the Tree of Life of the order Cypriniformes

429
Table 1 Taxa used in this study, with GenBank accession numbers for RAG1 sequences
Taxon GenBank No. Taxon GenBank No.
Gonorynchiformes Gobioninae
Chanidae Abbottina rivularis EU711102
Chanos chanos AY430207 Coreoleuciscus splendidus EU711114
Gonorynchidae Gnathopogon elongatus EU711153
Gonorynchus greyi EU409606 Gobio gobio EU292689
Siluriformes Hemibarbus labeo EU711154
Callichthyidae Pseudorasbora pumila EU711155
Corydoras rabauti Chen unpublished Pungtungia herzi EU711156
Clariidae Romanogobio ciscaucasicus EU409624
Clarias batrachus DQ492521 Sarcocheilichthys variegatus EU711157
Heteropneustidae Leuciscinae
Heteropneustes fossilis DQ492522 Abramis brama EU711103
Characiformes Alburnoides bipunctatus EU711104
Alestidae Alburnus alburnus EU711143
Phenacogrammus interruptus Chen unpublished Aspius vorax EU711106
Characidae Blicca bjoerkna EU711108
Chalceus macrolepidotus EU409607 Campostoma anomalum EF452827
Cypriniformes Clinostomus elongatus EU711112
Balitoridae Couesius plumbeus EU711115
Barbatula barbatula EU711107 Cyprinella lutrensis EU711158
Barbatula toni EU711133 Erimystax dissimilis EU711116
Homaloptera leonardi EU711130 Exoglossum maxillingua EU711118
Homaloptera parclitella EU409610 Hemitremia flammea EF452828
Lefua echigonia EF458305 Hybognathus nuchalis EU711120
Nemacheilus longicaudus EU711124 Leucaspius delineatus EU711121
Schistura balteata EU711131 Luxilus chrysocephalus EF452829
Sewellia lineolata EU409609 Nocomis biguttatus EF452830
Botiidae Notemigonus crysoleucas EF452831
Botia striata EU711109 Notropis atherinoides EF452832
Chromobotia macracantha EU711137 Notropis baileyi EU292691
Leptobotia mantschurica EU711138 Opsopoeodus emiliae EF452833
Leptobotia pellegrini EU292683 Pelecus cultratus EU711144
Sinibotia superciliaris EU711110 Phoxinus percnurus EU409627
Catostomidae Pimephales promelas AY430210
Catostomus commersonii EU409612 Richardsonius balteatus EF452835
Cycleptus elongatus EU409613 Rutilus rutilus EU711126
Erimyzon oblongus EU711117 Scardinius erythrophthalmus EU409628
Hypentelium nigricans EU711134 Semotilus atromaculatus EU409629
Minytrema melanops EU711135 Tribolodon nakamurai EU711159
Myxocyprinus asiaticus EU711136 Rasborinae
Thoburnia rhothoeca EU711128 Aphyocypris chinensis EU292692
Cobitidae Aspidoparia morar EU711105
Acantopsis choirorhynchos EU711139 Barilius bendelisis EU292693
Cobitis striata Saitoh unpublished Boraras merah EF452838
Cobitis taenia EU711113 Chela dadiburjori EU292694
Misgurnus anguillicaudatus EU711122 Danio erythromicron EU292698
Misgurnus nikolskyi EU711140 Danio rerio U71093
Pangio oblonga EU711141 Danionella sp. EF452841
Cyprinidae Devario regina EU292701
Acheilognathinae Esomus metallicus EU292702
Acheilognathus typus EU292688 Horadandia atukorali EU292703
Rhodeus atremius EU711125 Inlecypris auropurpurea EU292708
Rhodeus ocellatus EU711142 Luciosoma setigerum EU292704
Cultrinae Microrasbora rubescens EU292706
Chanodichthys mongolicus EU711145 Nicholsicypris normalis EU711123
Hemiculter lucidus EU711119 Opsaridium sp. EF452846
Ischikauia steenackeri EU292687 Opsariichthys uncirostris EF452847
Cyprininae Rasbora bankanensis EU292709
Barbonymus gonionotus EU711146 Rasbora gracilis EU292710
Barbus barbus EU711147 Sundadanio axelrodi EU292711
Barbus trimaculatus EU711148 Trigonostigma heteromorpha EU711129
Capoeta capoeta EU711111 Zacco sieboldii EU292713
Carassius auratus DQ196520 Squaliobarbinae
Cyprinus carpio AY787040 Ctenopharyngodon idella EF178284
Garra orientalis EU292684 Tincinae
Gymnocypris przewalskii EU711149 Tinca tinca EU711162
Labeo batesii EU711150 Xenocyprinae
Labeo senegalensis EU711151 Xenocypris macrolepis EU711160
Puntius ticto EU711152 Gyrinocheilidae
Puntius titteya EU292685 Gyrinocheilus aymonieri EU292682
Sawbwa resplendens EU292686 Vaillantellidae
Schizopyge curvifrons EU711146 Vaillantella maassi EU711132

Journal of Systematics and Evolution Vol. 46 No. 3 2008 430
Unlike mitochondrial gene trees for the family,
previous phylogenetic analyses at a higher level
within Cypriniformes using nuclear genes are few and
include only those by Mayden et. al. (2007), Šlechtová
et al. (2007), Conway et al. (2008), and He et al.
(2008b). The first study examined not only the phy-
logenetic placement of the model species Danio rerio
but also evaluated some general relationships within
the order. Šlechtová et al. (2007) assessed supraspeci-
fic relationships in the Cobitoidea. He et al. (2008b)
focused on the subfamilies in Cyprinidae, exclusive of
Psilorhynchinae, using the first intron of S7. Conway
et al. (2008) reevaluated the classification of Ce-
lestichthys margaritatus, verifying it as a species of
Danio, and also examined some higher relationships
in the family Cyprinidae.
Phylogenetic relationships among the 53 targeted
taxa within Cypriniformes, based on whole mitochon-
drial genomes (Fig. 1; Saitoh et al., 2006: fig. 2),
identified a monophyletic Cypriniformes and two
monophyletic superfamilies, Cobitoidea (Gyrino-
cheilidae, Catostomidae, Botiidae, Cobitidae,
Balitoridae, Vaillantellidae) and Cyprinoidea (Cypri-
nidae with multiple natural and unnatural subfami-
lies). Nodal support for these relationships were very
good overall, with posterior probabilities of 95–100%.
Exceptions included relatively poor support for the
superfamily Cobitoidea (63%) and similar support for
one sister group hypothesis in Catotomidae (Catosto-
mus + Minytrema) and some sister group relationships
within the subfamily Gobioninae of Cyprinidae (Gna-
thopogon + (Puntungia + Pseudorasbora)). Within the
Cobitoidea, Gyrinocheilus forms the sister group to
Catostomidae; this Gyrinocheilidae + Catostomidae
clade is sister to a loach clade consisting of mono-
phyletic Botiidae, Cobitidae, and Balitoridae, along
with the monotypic Vaillantellidae. Vaillantella is
sister to a Cobitidae + Balitoridae clade, with Botiidae
as the basal member of this loach clade. The sister
group to the Cobitoidea is Cyprinidae, or the suborder
Cyprinoidea. Within the family, the subfamily Cy-
prininae is sister to the remaining cyprinids. Saitoh et
al. (2006) did not recover a monophyletic Rasborinae,
instead they find support for a monophyletic Rasbori-
nae sensu stricto, with the other putative rasborins
found elsewhere in the tree. A monophyletic subfam-
ily Acheilognathinae is recovered as the sister group
to all remaining cyprinids (excluding Cyprininae and
Rasborinae sensu stricto). A clade composed of
Xenocyprinae, Cultrinae, and the remaining “ras-
borins” was recovered by Saitoh et al. (2006). The
apical portion of the Cyprinidae is occupied by three
subfamilies, Gobioninae and Leuciscinae, which are
monophyletic, and the monotypic Tincinae. The
subfamily Tincinae is sister to Leuciscinae, and that
clade is sister to the Gobioninae.
With few exceptions, nodal support for both the
smaller (49 cypriniform species) and larger (110
cypriniform species) analyses of taxa using only
RAG1 sequences was generally high (90%–100%).
Phylogenetic resolution among the 49 species exam-
ined herein for RAG1 (Fig. 2) had some notable
differences from the topology of Saitoh et al. (2006)
(Fig. 1). In terms of higher-level relationships, the
RAG1-only tree (Fig. 2) did not find a monophyletic
Cobitoidea, instead catostomids are found in a basal
position as the sister group to a clade comprising all of
the other cypriniforms, consequently Gyrinocheilidae
is no longer the sister to Catostomidae, rather it is the
sister group to a loach clade similar to the one seen in
Fig. 1. Within the Catostomidae, relationships differed
from the mitochondrial tree in that Cycleptus and
Myxocyprinus did not form a clade. Rather, Cycleptus
is the basal sister group in the family, and Myxo-
cyprinus was sister to a clade wherein Hypentelium
was sister to Catostomus plus Minytrema. Overall, the
relationships within the loach clade (Botiidae, Vail-
lantellidae, Balitoridae, and Cobitidae) are roughly
comparable to those seen in the mitogenome tree. The
one major difference is that in the RAG1-only phy-
logeny, balitorids are not monophyletic, due to the
position of Homaloptera leonardi (Fig. 2). The other
major difference in the RAG1-only tree involves a
radical rearrangement of the subfamilial relationships
within the Cyprinidae. The subfamily Acheilognathi-
nae is still monophyletic but instead of being the sister
group to all cyprinids except Cyprininae and Rasbori-
nae sensu stricto, it is sister to Gobioninae. Associated
with that change is the relocation of Tinca, where it is
now sister to a Leuciscinae + (Acheilognathinae +
Gobioninae) clade. It should be noted that, although
the tree obtained from our RAG1 analysis is not
identical to the mitogenome topology presented by
Saitoh et al. (2006), it is congruent with an alternate
topology they obtained using the 12nRTn coding
scheme excluding third codon positions (not shown;
see Saitoh et al., 2006 for discussion of alternate
topologies they recovered).
The relationships recovered by the analysis of the
110-taxa, RAG1-only data matrix (Fig. 3) resolve
some of the conflicts between the two trees produced
from examination of fewer taxa (Figs. 1 & 2). Beyond


MAYDEN et al.: Inferring the Tree of Life of the order Cypriniformes

431


Fig. 1. A 50% majority rule consensus tree of 10,800 trees generated from Bayesian analysis of whole mitogenome sequence data from 53 cyprini-
form taxa, redrawn from Saitoh et al. (2006; fig. 2). Classification and family names are modified to reflect the taxonomy used herein (Figs. 2 & 3).
Species names were drawn from information provided in Saitoh et al. (2006; table 2), with updates to reflect current nomenclature, following
Eschmeyer (Catalog of Fishes, online version, updated 23 April 2008).

Journal of Systematics and Evolution Vol. 46 No. 3 2008 432


Fig. 2. A 50% majority rule consensus tree of 1,900 trees generated from Bayesian analysis of RAG1 sequence data collected from 49 cypriniform
taxa, representing the same species as illustrated in Fig. 1 (Saitoh et al., 2006; fig. 2); when some species were not available, a congeneric species was
used where possible. Bayesian posterior probabilities are displayed above each node.
MAYDEN et al.: Inferring the Tree of Life of the order Cypriniformes

433
the addition of taxa, the overall structure of the larger
phylogeny is closely congruent with that observed in
the 49-taxa, RAG1-only tree, in terms of familial and
subfamilial relationships. In looking at the higher
level relationships, the larger tree (Fig. 3) is congruent
with the RAG1-only tree shown in Fig. 2. At the base
of the tree, the superfamily Cobitoidea is not mono-
phyletic, with Catostomidae as the most basal member
of Cypriniformes, and Gyrinocheilidae is sister to a
loach group. Catostomid relationships were com-
pletely consistent with those observed in the RAG1
phylogeny with 49 taxa except that Thoburnia and
Erimyzon were included in the analysis; the former
forms the sister group to Hypentelium and the latter is
sister to Minytrema. Within Cyprinidae, the large tree
agrees with the topology in Fig. 2: Acheilognathinae
and Gobioninae as sister taxa, with Tincinae as the
sister to a Leuciscinae + (Acheilognathinae + Gobi-
oninae) clade. In addition, the monophyly of the
Cyprininae, the Rasborinae sensu stricto, the Gobion-
inae, and the Leuciscinae is not challenged when more
species from those groups are added. However, the
Cultrinae becomes paraphyletic with respect to Xeno-
cypris macrolepis in the larger analysis. The one
major conflict between the mitogenome phylogeny
and that using 49 taxa was the status of Balitoridae; a
conflict that the incorporation of additional taxa
resolved in favor of the Saitoh et al. (2006) tree. The
inclusion of more balitorid species appears to solidify
the monophyly of the Balitoridae (98% nodal sup-
port).
2.1 Relationships within Cobitoidea
Gyrinocheilus has been resolved as sister to a
clade inclusive of Botiidae, Vaillantellidae, Balitori-
dae, and Cobitidae, sister to Catostomidae, or the
basal sister group to all other Cypriniformes (He et al.,
2008a). The former relationship is observed in both
analyses herein for RAG1 and by Šlechtová et al.
(2007), also based on analysis of RAG1 sequences.
The sister relationship of Catostomidae and Gyrino-
cheilidae is observed in Saitoh et al. (2006) for 53 taxa
and He et al. (2008a) for 17 ingroup taxa. One analy-
sis in the latter study resolved Gyrinocheilus as the
sister group to all other Cypriniformes. The variable
relationships observed for the Cobitoidea cannot be
resolved herein with the increase in taxon sampling as
the results of all analyses appear to be partitioned on
the basis of whether the character base is mitochon-
drial or nuclear. Future analyses of many more taxa
for both mitochondrial and nuclear genes, and with an
eye towards the relationships at this basal portion of
the evolution of Cypriniformes, will be important in
resolving this early diversification of the group.
2.2 Relationships within Catostomidae
Several hypotheses have been presented for rela-
tionships in Catostomidae, but all of the studies have
had limited taxon sampling within the family and
order. The mitochondrial phylogeny of Saitoh et al.
(2006) identifies Cycleptus and Myxocyprinus as a
basal monophyletic group, while that of Harris and
Mayden (2001) identifies Myxocyprinus as the basal
sister group in the family and Cycleptus sister to other
taxa (excluding Carpiodes and Ictiobus, more basal in
the tree). These studies are not, however, consistent in
the resolution of Carpiodes and Ictiobus. Both RAG1
analyses identify Cycleptus as the basal sister group
and Myxocyprinus as more closely related to other
taxa (Carpiodes and Ictiobus not included in these
analyses). The sister relationship of Hypentelium and
Thoburnia and between Minytrema and Erimyzon is
consistent with that of Harris and Mayden (2001) and
Saitoh et al. (2006). However, the placement of the
Hypentelium plus Thoburnia clade in the nuclear gene
phylogeny is inconsistent with those analyses using
fewer taxa (Saitoh et al., 2006; Harris & Mayden,
2001). While these differences could be the result of
different resolutions based on alternative gene trees,
we hypothesize that the placement of this clade is
likely novel due to increasing taxon sampling, a
hypothesis that should be tested using greater taxon
sampling with the mitochondrial genes.
2.3 Relationships within Balitoridae
This group, as we recognize it today, was first
proposed by Sawada (1982) based on morphological
characters. In both instances, depending on the analy-
sis, it has been resolved as either an unnatural or
natural group. Previously, Tang et al. (2006) included
a high number of species of the then recognized
Cobitinae, Botiinae, Nemacheilinae, and Balitorinae
and the relatively rapidly evolving mitochondrial
genes cytochrome b and control region. In their study,
the general relationships among these major groups
were as observed herein for the larger sample of taxa
(Fig. 3), except Tang et al. (2006) did not include the
Vaillantellidae. The Cobitidae and Botiidae are sepa-
rate monophyletic groups and Balitoridae is mono-
phyletic and sister to the Cobitidae.
In other analyses with fewer taxa, the family
Balitoridae is consistently resolved as a paraphyletic
grade relative to the Cobitidae. This is not only ob-
served in the present study but was also found by
Tang et al. (2006), using cyt b sequence data, and later
by He et al. (2008a) using whole mitochondrial gene
sequences for only 17 species of Cypriniformes.
Journal of Systematics and Evolution Vol. 46 No. 3 2008 434


Fig. 3. A 50% majority rule consensus tree of 1,800 trees generated from Bayesian analysis of RAG1 sequence data collected from 110 cypriniform
taxa, including all of the species shown in Fig. 2. Bayesian posterior probabilities are displayed above each node.
MAYDEN et al.: Inferring the Tree of Life of the order Cypriniformes

435
Interestingly, as observed for taxon sampling with the
nuclear gene RAG1, with an increase in taxon sam-
pling from these 17 species to 53 species in Saitoh et
al. (2006) the family goes from being paraphyletic to
monophyletic.
2.4 Monophyly of “Cobitidae” or Cobitidae and
Botiidae
The unnaturalness of the historically conceived
family Cobitidae was identified by Tang et al. (2006)
and Saitoh et al. (2006), but taxonomic changes were
not advocated in these analyses as caution was exer-
cised in anticipation of more taxa and character data to
provide further support that this resolution was not a
result of long branch attraction in these highly mor-
phologically divergent fishes. Šlechtová et al. (2007),
however, recognized the Cobitidae and Botiidae
(sensu Nalbant, 2002) and elevated the Nemacheilidae
and Vaillantellidae based on RAG1 sequences. In this
analysis the Cobitidae and Botiidae both resolve as
monophyletic groups with either the smaller or larger
taxon base. Further, both nuclear and mitochondrial
gene analyses support these families as monophyletic
groups. Although Šlechtová et al.’s (2007) Balitoridae
and Nemacheilidae are recovered as reciprocally
monophyletic groups in the large data set tree (Fig. 3),
they also are recovered as sister groups, therefore we
continue to recognize them as subfamilies of a mono-
phyletic Balitoridae.
2.5 Relationships within Cyprinidae
The purported chaos regarding placement and
relationships of species in the various subfamilies of
Cyprinidae (Conway et al., in press) is not surprising.
As currently conceived, this family is monophyletic
(Cavender & Coburn, 1992; Saitoh et al., 2006; He et
al., 2008b) but the vast majority of investigations of
these species are pre-Hennigian revisionary and
systematic studies, faunal works, and comparative
taxonomic studies. There remains much diversity to
be described and the current subfamilies’ taxonomy
rest largely on non-phylogenetic statements of inclu-
siveness that are essentially derived from phenetic
similarity. Thus, it is expected that many of the forth-
coming systematic studies of this family will result in
changes in the taxonomy of the group but this is only
because very few explicitly phylogenetic studies exist.
He et al. (2004) identified multiple, separate
lineages for species referred to the Rasborinae, con-
sistent with the polyphyletic origin of the subfamily
observed by Saitoh et al. (2006), Mayden et al. (2007),
Rüber et al. (2007), Conway et al. (2008) and herein
for both analyses, regardless of taxon sample size
(Figs. 2 & 3). The monophyly of Acheilognathinae
and its sister relationship to the remaining apical
cyprinids (excluding Cyprininae and Rasborinae sensu
stricto), agrees with the results of Conway et al.
(2008). The location of Tinca in the cypriniform tree
of life is troublesome. However, the resolution of this
lineage as the sister to the Leuciscinae + (Acheilog-
nathinae + Gobioninae) clade is congruent with the
Bayesian analyses derived from the nuclear intron S7
(He et al., 2008b: fig. 2C). The variable placement of
Tinca in this instance may be related to the gene
origin (mitochondrial versus nuclear), as well as the
number and composition of taxa in the ingroup. The
relative sister group relationships of taxa within the
subfamilies Cyprininae, Gobioninae, and Leuciscinae
are dependent on taxon sampling. In all three groups,
the relationships observed in the larger taxon base are
more consistent with previous studies (Cavender &
Coburn, 1992; Simons & Mayden, 1997, 1998, 1999;
Simons et al., 2003; He et al., 2008a, b) based on
either morphological or molecular data.
Acknowledgements We thank Yin-Long Qiu and
Zhiduan Chen and the Institute of Botany, Chinese
Academy of Sciences, for extending an invitation to
RLM for presenting this research at the Tree of Life
symposium, June 2007 in Beijing, China. RLM
wishes to thank David M. Hillis for great conversa-
tions regarding systematics and biodiversity, as well
as the topic of focus in this paper, and urging us to
publish this study. Finally, we thank the USA Na-
tional Science Foundation for financial support for the
Cypriniformes Tree of Life research (EF-0431326).
References
Albrecht C, Kuhn K, Streit B. 2007. A molecular phylogeny of
Planorboidea (Gastropoda, Pulmonata): insights from
enhanced taxon sampling. Zoologica Scripta 36: 27–39.
Altekar G, Dwarkadas S, Huelsenbeck JP, Ronquist F. 2004.
Parallel Metropolis-coupled Markov chain Monte Carlo
for Bayesian phylogenetic inference. Bioinformatics 20:
407–415.
Braun EL, Kimball RT. 2002. Examining basal avian
divergences with mitochondrial sequences: Model
complexity, taxon sampling, and sequence length.
Systematic Biology 51: 614–625.
Bremer B, Jansen RK, Oxelman B, Backlund M, Lantz H, Kim
K-J. 1999. More characters or more taxa for a robust
phylogeny—case study from the coffee family
(Rubiaceae). Systematic Biology 48: 413–435.
Bruno WJ, Halpern AL. 1999. Topological bias and
inconsistency of maximum likelihood using wrong
models. Molecular Biology and Evolution 16: 564–566.
Cavender TM, Coburn MM. 1992. Phylogenetic relationships
of North American Cyprinidae. In: Mayden RL ed.
Journal of Systematics and Evolution Vol. 46 No. 3 2008 436
Systematics, historical ecology, and North American
freshwater fishes. Palo Alto: Stanford University Press.
293–327.
Chen W-J, Bonillo C, Lecointre G. 2003. Repeatability of
clades as a criterion of reliability: A case study for
molecular phylogeny of Acanthomorpha (Teleostei) with
larger number of taxa. Molecular Phylogenetics and
Evolution 26: 262–288.
Chen W-J, Ruiz-Carus R, Ortí G. 2007. Relationships among
four genera of mojarras (Teleostei: Perciformes:
Gerreidae) from the western Atlantic and their tentative
placement among percomorph fishes. J. Fish Biol. 70
(Supplement B): 202–218.
Conway KW, Chen W-J, Mayden RL. 2008. The “Celestial
Pearl danio” is a miniature Danio (s.s) (Ostariophysi:
Cyprinidae): evidence from morphology and molecules.
Zootaxa 1686: 1–28.
Conway KW, Hirt MV, Yang L, Mayden RL, Simons AM.
Cypriniformes: Systematics and paleontology. In: Schultz
H-P, Wilson MVH, Nelson JS eds. Origin and
phylogenetic interrelationships of teleosts. Munich: Verlag
Dr. Freidrich Feil. (in press)
Cummings MP, Meyer A. 2005. Magic bullets and golden
rules: Data sampling in molecular phylogenetics. Zoology
108: 329–336.
DeBry RW. 1992. The consistency of several phylogeny-
inference methods under varying evolutionary rates.
Molecular Biology and Evolution 9: 537–551.
DeBry RW. 2005. The systematic component of phylogenetic
error as a function of taxonomic sampling under
parsimony. Systematic Biology 54: 432–440.
Dobzhansky T. 1973. Nothing in biology makes sense except in
the light of evolution. American Biology Teacher 35: 125–
129.
Eschmeyer WN, Ferraris CJ, Hoang MD, Long DJ. 1998.
Species of fishes. In: Eschmeyer WN ed. Catalog of
fishes. San Francisco: California Academy of Sciences.
25–1820. Online version, updated 23 April 2008.
Felsenstein J. 1978. Cases in which parsimony or compatibility
methods will be positively misleading. Systematic
Zoology 27: 401–410.
Fitch WM, Beintema JJ. 1990. Correcting parsimonious trees
for unseen nucleotide substitutions: The effect of dense
branching as exemplified by ribonuclease. Molecular
Biology and Evolution 7: 438–443.
Fitch WM, Bruschi M. 1987. The evolution of prokaryotic
ferredoxins—with a general method correcting for
unobserved substitutions in less branched lineages.
Molecular Biology and Evolution 4: 381–394.
Gascuel O, Bryant D, Denis F. 2001. Strengths and limitations
of the minimum evolution principle. Systematic Biology
50: 621–627.
Gojobori T, Ishii K, Nei M. 1982. Estimation of average
number of nucleotide substitutions when the rate of
substitution varies with nucleotide. Journal of Molecular
Evolution 18: 414–423.
Graham RL, Foulds LR. 1982. Unlikelihood that minimal
phylogenies for a realistic biological study can be
constructed in reasonable computational time. Mathemat-
ical Bioscience 60: 133–142.
Graybeal A. 1998. Is it better to add taxa or characters to a
difficult phylogenetic problem? Systematic Biology 47:
9–17.
Groth JG, Barrowclough GF. 1999. Basal divergences in birds
and the phylogenetic utility of the nuclear RAG-1 gene.
Molecular Phylogenetics and Evolution 12: 115–123.
Harris PH, Mayden RL. 2001. Phylogenetic relationships of
major clades of Catostomidae (Teleostei: Cypriniformes)
as inferred from mitochondrial SSU and LSU rDNA
sequences. Molecular Phylogenetics and Evolution 20:
225–237.
He S, Liu H, Chen Y, Kuwahara M, Nakajima T, Yang Z. 2004.
Molecular phylogenetic relationships of Eastern Asian
Cyprinidae (Pisces: Cypriniformes) inferred from
cytochrome b sequences. Science in China (Series C Life
Sciences) 47: 130–138.
He S, Gu X, Mayden RL, Chen W-J, Conway KW, Chen Y.
2008a. Phylogenetic position of the enigmatic genus
Psilorhynchus (Ostariophysi: Cypriniformes): Evidence
from the mitochondrial genome. Molecular Phylogenetics
and Evolution 47: 419–425.
He S, Mayden RL, Wang X, Wang W, Tang KL, Chen W-J,
Chen Y. 2008b. Molecular phylogenetics of the family
Cyprinidae (Actinopterygii: Cypriniformes) as evidenced
by sequence variation in the first intron of S7 ribosomal
protein-coding gene: Further evidence from a nuclear gene
of the systematic chaos in the family. Molecular
Phylogenetics and Evolution 46: 818–829.
Heath TA, Hedtke SM, Hillis DM. 2008. Taxon sampling and
the accuracy of phylogenetic analyses. Journal of
Systematics and Evolution 46: 239–257.
Hedtke SM, Townsend TM, Hillis DM. 2006. Resolution of
phylogenetic conflict in large data sets by increased taxon
sampling. Systematic Biology 55: 522–529.
Hendy MD, Penny D. 1989. A framework for the quantitative
study of evolutionary trees. Systematic Zoology 38:
297–309.
Hennig W. 1950. Grundzüge einer Theorie der
phylogenetischen Systematik. Berlin: Deutscher
Zentralverlag.
Hennig W. 1966. Phylogenetic systematics (Davis DD, Zangerl
R translation). Urbana: University of Illinois Press.
Hillis DM. 1996. Inferring complex phylogenies. Nature 383:
130–131.
Hillis DM. 1998. Taxonomic sampling, phylogenetic accuracy,
and investigator bias. Systematic Biology 47: 3–8.
Hillis DM, Pollock DD, McGuire JA, Zwickl DJ. 2003. Is
sparse taxon sampling a problem for phylogenetic
inference? Systematic Biology 52: 124–126.
Holcroft NI. 2005. A molecular analysis of the
interrelationships of tetraodontiform fishes (Acanthomor-
pha: Tetraodontiformes). Molecular Phylogenetics and
Evolution 34: 525–544.
Howes GJ. 1991. Systematics and biogeography: An overview.
In: Winfield IJ, Nelson JS eds. Cyprinid fishes:
systematics, biology and exploitation. London: Chapman
and Hall.
Hubbs CL, Miller RR. 1977. Six distinctive cyprinid fish
species referred to Dionda inhabiting segments of the
Tampico embayment drainage of Mexico. Transactions of
the San Diego Society of Natural History 28: 267–335.
Huelsenbeck JP. 1995. Performance of phylogenetic methods in
MAYDEN et al.: Inferring the Tree of Life of the order Cypriniformes

437
simulation. Systematic Biology 44: 17–48.
Huelsenbeck JP, Hillis DM. 1993. Success of phylogenetic
methods in the 4-taxon case. Systematic Biology 42:
247–264.
Huelsenbeck JP, Lander KM. 2003. Frequent inconsistency of
parsimony under a simple model of cladogenesis.
Systematic Biology 52: 641–648.
Huelsenbeck JP, Ronquist F. 2001. MRBAYES: Bayesian
inference of phylogeny. Bioinformatics 17: 754–755.
Hugall AF, Lee MSY. 2007. The likelihood node density effect
and consequences for evolutionary studies of molecular
rates. Evolution 61: 2293–2307.
Inoue JG, Miya M, Tsukamoto K, Nishida M. 2001. A
mitogenomic perspective on the basal teleostean
phylogeny: Higher-level relationships with longer DNA
sequences. Molecular Phylogenetics and Evolution 20:
275–285.
Johnson KP. 2001. Taxon sampling and the phylogenetic
position of passeriformes: Evidence from 916 avian
cytochrome b sequences. Systematic Biology 50:
128–136.
Krenz JG, Naylor GJ, Shaffer, HB, Janzen FJ. 2005. Molecular
phylogenetics and evolution of turtles. Molecular
Phylogenetics and Evolution 37: 178–191.
Kuhn T. 1962. The structure of scientific revolutions. Chicago:
University of Chicago Press.
Lecointre G, Philippe H, Vân Lê HL, Le Guyader H. 1993.
Species sampling has a major impact on phylogenetic
inference. Molecular Phylogenetics and Evolution 2:
205–224.
Lin Y-H, McLenachan PA, Gore AR, Phillips MJ, Ota R,
Hendy MD, Penny D. 2002. Four new mitochondrial
genomes and the increased stability of evolutionary trees
of mammals from improved taxon sampling. Molecular
Biology and Evolution 19: 2060–2070.
López JA, Chen W-J, Ortí G. 2004. Esociform phylogeny.
Copeia 2004: 449–464.
Mabuchi K, Miya M, Azuma Y, Nishida M. 2007. Independent
evolution of the specialized pharyngeal jaw apparatus in
cichlid and labrid fishes. BMC Evol. Biol. 7: 10.
Mayden RL. 1989. Phylogenetic studies of North American
minnows, with emphasis on the genus Cyprinella
(Teleostei: Cypriniformes). Miscellaneous Publication
(University of Kansas Museum of Natural History) 80:
1–189.
Mayden RL, Tang KL, Conway KW, Freyhof J, Chamberlain
SJ, Haskins M, Schneider L, Sudkamp M, Wood RM,
Agnew M, Bufalino A, Sulaiman Z, Miya M, Saitoh K, He
S. 2007. Phylogenetic relationships of Danio within the
order Cypriniformes: A framework for comparative and
evolutionary studies of a model species. Journal of
Experimental Zoology B: Molecular and Developmental
Evolution 308B: 642–654.
Miya M, Takeshima H, Endo H, Ishiguro NB, Inoue JG, Mukai
T, Satoh TP, Yamaguchi M, Kawaguchi A, Mabuchi K.,
Shirai SM, Nishida M. 2003. Major patterns of higher
teleostean phylogenies: A new perspective based on 100
complete mitochondrial DNA sequences. Molecular
Phylogenetics and Evolution 26: 121–138.
Nalbant TT. 2002. Sixty million years of evolution. Part one:
Family Botiidae (Pisces: Ostariophysi: Cobitoidea).
Travaux du Muséum National d’Histoire Naturelle
“Grigore Antipa” 44: 309–344.
Nelson JS. 2006. Fishes of the world. 4th ed. Hoboken: John
Wiley and Sons, Inc.
Nylander JAA. 2004. MrModeltest v2.2. Uppsala: Evolutionary
Biology Centre, Uppsala University.
Poe S. 2003. Evaluation of the strategy of long-branch
subdivision to improve the accuracy of phylogenetic
methods. Systematic Biology 52: 423–428.
Pollock DD, Zwickl DJ, McGuire JA, Hillis DM. 2002.
Increased taxon sampling is advantageous for
phylogenetic inference. Systematic Biology 51: 664–671.
Rannala B, Huelsenbeck JP, Yang Z, Nielsen R. 1998. Taxon
sampling and the accuracy of large phylogenies.
Systematic Biology 47: 702–710.
Rokas A, Carroll SB. 2005. More genes or more taxa? The
relative contribution of gene number and taxon number to
phylogenetic accuracy. Molecular Biology and Evolution
22: 1337–1344.
Rokas A, Krüger D, Carroll SB. 2005. Animal evolution and
the molecular signature of radiations compressed in time.
Science 310: 1933–1938.
Rokas A, Williams BL, King N, Carroll SB. 2003.
Genome-scale approaches to resolving incongruence in
molecular phylogenies. Nature 425: 798–804.
Ronquist F, Huelsenbeck JP. 2003. MRBAYES 3: Bayesian
phylogenetic inference under mixed models.
Bioinformatics 19: 1572–1574.
Rosenberg MS, Kumar S. 2003. Taxon sampling,
bioinformatics, and phylogenomics. Systematic Biology
52: 119–124.
Rüber L, Britz R, Kullander SO, Zardoya R. 2004.
Evolutionary and biogeographic patterns of the Badidae
(Teleostei: Perciformes) inferred from mitochondrial and
nuclear DNA sequence data. Molecular Phylogenetics and
Evolution 32: 1010–1022
Rüber L, Kottelat M, Tan HH, Ng PKL, Britz R. 2007.
Evolution of miniaturization and the phylogenetic position
of Paedocypris, comprising the worlds smallest
vertebrate. BMC Evolutionary Biology 7: 38.
Saitoh K, Sado T, Mayden RL, Hanzawa N, Nakamura K,
Nishida M, Miya M. 2006. Mitogenomic evolution and
interrelationships of the Cypriniformes (Actinopterygii:
Ostariophysi): The first evidence toward resolution of
higher-level relationships of the world’s largest freshwater
fish clade based on 59 whole mitogenome sequences.
Journal of Molecular Evolution 63: 826–841.
San Mauro D, Gower DJ, Oommen OV, Wilkinson M, Zardoya
R. 2004. Phylogeny of caecilian amphibians
(Gymnophiona) based on complete mitochondrial
genomes and nuclear RAG1. Molecular Phylogenetics and
Evolution 33: 413–427.
Sawada Y. 1982. Phylogeny and zoogeography of the
superfamily Cobitoidea (Cyprinoidei, Cypriniformes).
Memoirs of the Faculty of Fisheries, Hokkaido University
28: 65–223.
Scotland RW, Olmstead RG, Bennett JR. 2003. Phylogeny
reconstruction: The role of morphology. Systematic
Biology 52: 539–548.
Simons AM, Mayden RL. 1997. Phylogenetic relationships of
the creek chubs and the spine-fins: An enigmatic group of
Journal of Systematics and Evolution Vol. 46 No. 3 2008 438
North American cyprinid fishes (Actinopterygii:
Cyprinidae). Cladistics 13: 187–205
Simons AM, Mayden RL. 1998. Phylogenetic relationships of
the western North American phoxinins (Actinopterygii:
Cyprinidae) as inferred from mitochondrial 12S and 16S
RNA sequences. Molecular Phylogenetics and Evolution
9: 308–329.
Simons AM, Mayden RL. 1999. Phylogenetic relationships of
North American cyprinids and assessment of homology of
the open posterior myodome. Copeia 1999: 13–21.
Simons AM, Berendzen PB, Mayden RL. 2003. Molecular
systematics of North American phoxinin genera
(Actinopterygii: Cyprinidae) inferred from mitochondrial
12S and 16S ribosomal RNA sequences. Zoological
Journal of the Linnean Society 139: 63–80.
Šlechtová V, Bohlen J, Tan HH. 2007. Families of Cobitoidea
(Teleostei; Cypriniformes) as revealed from nuclear
genetic data and the position of the mysterious genera
Barbucca, Psilorhynchus, Serpenticobitis and Vaillantella.
Molecular Phylogenetics and Evolution 44: 1358–1365.
Smith GR. 1992. Phylogeny and biogeography of the
Catostomidae, freshwater fishes of North America and
Asia. In: Mayden RL ed. Systematics, historical ecology,
and North American freshwater fishes. Palo Alto: Stanford
University Press. 778–826.
Sorenson MD, Oneal E, García-Moreno J, Mindell DP. 2003.
More taxa, more characters: The hoatzin problem is still
unresolved. Molecular Biology and Evolution 20:
1484–1498.
Stamatakis A. 2006a. Phylogenetic models of rate
heterogeneity: A high performance computing perspective.
In: Parallel and Distributed Processing Symposium, 2006.
IPDPS 2006, Rhodos, Greece, 25–29 April 2006.
Stamatakis A. 2006b. RAxML-VI-HPC: Maximum likelihood-
based phylogenetic analyses with thousands of taxa and
mixed models. Bioinformatics 22: 2688–2690.
Steppan SJ, Storz BL, Hoffmann RS. 2004. Nuclear DNA
phylogeny of the squirrels (Mammalia: Rodentia) and the
evolution of arboreality from c-myc and RAG1. Molecular
Phylogenetics and Evolution 30: 703–719.
Sullivan JM, Swofford DL. 1997. Are guinea pigs rodents? The
importance of adequate models in molecular
phylogenetics. Journal of Mammalian Evolution 4: 77–86.
Sullivan JP, Lundberg JG, Hardman M. 2006. A phylogenetic
analysis of the major groups of catfishes (Teleostei:
Siluriformes) using rag1 and rag2 nuclear gene
sequences. Molecular Phylogenetics and Evolution 41:
636–662.
Susko E, Inagaki Y, Roger AJ. 2004. On inconsistency of the
Neighbor-Joining, least squares, and minimum evolution
estimation when substitution processes are incorrectly
modeled. Molecular Biology and Evolution 21:
1629–1642.
Swofford DL. 2002. PAUP*. Phylogenetic analysis using
parsimony (*and other methods). Version 4. Sunderland:
Sinauer Associates.
Swofford DL, Olsen GJ, Waddell PJ, Hillis DM. 1996.
Phylogenetic inference. In: Hillis DM, Moritz C, Mable
BK eds. Molecular systematics, 2nd ed. Sunderland:
Sinauer Associates. 407–514.
Tang Q, Liu H, Mayden RL, and Xiong B. 2006. Comparison
of evolutionary rates in the mitochondrial DNA
cytochrome b gene and control region and their
implications for phylogeny of the Cobitoidea (Teleostei:
Cypriniformes). Molecular Phylogenetics and Evolution
39: 347–357.
Waddell PJ, Shelley S. 2003. Evaluating placental inter-ordinal
phylogenies with novel sequences including RAG1,
γ-fibrinogen, ND6, and mt-tRNA, plus MCMC-driven
nucleotide, amino acid, and codon models. Molecular
Phylogenetics and Evolution 28: 197–224.
Wegener A. 1915. Die Entstehung der Kontinente und Ozeane.
Braunschweig: F. Vieweg and Sohn.
Wiens JJ. 2005. Can incomplete taxa rescue phylogenetic
analyses from long-branch attraction? Systematic Biology
54: 731–742.
Yang Z. 1994. Statistical properties of the maximum likelihood
method of phylogenetic estimation and comparison with
distance matrix methods. Systematic Biology 43:
329–342.
Yang Z. 1996. Among-site rate variation and its impact on
phylogenetic analyses. Trends in Ecology and Evolution
11: 367–372.
Zwickl DJ, Hillis DM. 2002. Increased taxon sampling greatly
reduces phylogenetic error. Systematic Biology 51:
588–598.