Molecular Phylogenetics and Evolution 35 (2005) 624636

www.elsevier.com/locate/ympev

Utility of nuclear DNA intron markers at lower taxonomic levels:

Phylogenetic resolution among nine Tragelaphus spp.
Sandi Willows-Munro , Terence J. Robinson, Conrad A. Matthee
Evolutionary Genomics Group, Department of Zoology, Stellenbosch University, Private Bag X1, Matieland 7602, South Africa
Received 21 July 2004; revised 19 January 2005
Available online 2 March 2005

Abstract
Phylogenetic relationships among the nine spiral-horn antelope species of the African bovid tribe Tragelaphini are controversial.
In particular, mitochondrial DNA sequencing studies are not congruent with previous morphological investigations. To test the utility of nuclear DNA intron markers at lower taxonomic levels and to provide additional data pertinent to tragelaphid evolution, we
sequenced four nuclear DNA segments (MGF, PRKCI, SPTBN, and THY) and combined these data with mitochondrial DNA
sequences from three genes (cytochrome b, 12S rRNA, and 16S rRNA). Our molecular supermatrix comprised 4682 characters which
were analyzed independently and in combination. Parsimony and model based phylogenetic analyses of the combined nuclear DNA
data are congruent with those derived from the analysis of mitochondrial gene sequences. The corroboration between nuclear and
mtDNA gene trees reject the possibility that genetic processes such as lineage sorting, gene duplication/deletion and hybrid speciation account for the conXict evident in the previously published phylogenies. It suggests rather that the morphological characters
used to delimit the Tragelaphid species are subject to convergent evolution. Divergence times among species, calculated using a
relaxed Bayesian molecular clock, are consistent with hypotheses proposing that climatic oscillations and their impact on habitats
were the major forces driving speciation in the tribe Tragelaphini.
2005 Elsevier Inc. All rights reserved.
Keywords: Bovidae; Tragelaphini; Systematics; Phylogeny; Molecular clock

1. Introduction
Congruence among multiple data sets is arguably the
most reliable indicator of phylogenetic accuracy
(Cracraft and Helm-Bychowski, 1991; Miyamoto and
Cracraft, 1991; Miyamoto and Fitch, 1995; SwoVord,
1991). Current phylogenetic analyses thus often utilize a
multigene approach and analyses frequently include
morphological and/or palaeontological data (see
Flores-Villela et al., 2000; Gatesy et al., 2003; Nylander
et al., 2004 among others). One group providing a case in
point is the Cetartiodactyla. Initial investigations were
based mainly on morphological and palaeontological

Corresponding author. Fax: +27 021 808 2405.

E-mail address: sm2@sun.ac.za (S. Willows-Munro).

1055-7903/$ - see front matter 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ympev.2005.01.018

evidence (McKenna, 1975; Novacek, 1982; OLeary and

Geisler, 1999; Simpson, 1945; Van Valen, 1968) and
these phylogenetic hypotheses were subsequently tested
by making use of mitochondrial (mtDNA) analyses
(Graur and Higgins, 1994; Irwin and Arnason, 1994;
Montgelard et al., 1997; Ursing and Arnason, 1998).
ConXict between the mtDNA gene tree and morphological evidence (for example, the monophyly of ArtiodactylaLuckett and Hong, 1998; OLeary and Geisler,
1999) prompted the inclusion of independent nuclear
DNA markers (Gatesy et al., 1996; Gatesy et al., 1999,
2002; Matthee et al., 2001).
Although the phylogenetic utility of nuclear DNA
data are now well entrenched in the mammalian literature (Gatesy and Arctander, 2000; Madsen et al., 2001;
Matthee et al., 2001, 2004; Matthee and Davis, 2001;
Murphy et al., 2001; Springer et al., 2001), one of the

S. Willows-Munro et al. / Molecular Phylogenetics and Evolution 35 (2005) 624636

caveats for employing these markers is the slow rate at

which nucleotide changes accumulate over time. For
example, comparisons among sequences revealed that
the rates of nucleotide change among members of the
family Bovidae are on average 3.2 times higher for coding mtDNA comparisons than for non-coding nuclear
DNA introns (Matthee and Davis, 2001). In many
instances this slow rate of change in the nuclear DNA
data seems to suggest that inferences at lower taxonomic
levels (for example, among mammalian species belonging to the same genus) will not be resolved with high
bootstrap and or statistical support. Limited empirical
examples exist to support this hypothesis (but see
Moore, 1995; Springer et al., 2001).
In the present study, we ampliWed a set of four
nuclear intron markers from the spiral-horned antelope
of the tribe Tragelaphini (family Bovidae). This group
forms a monophyletic lineage within the subfamily Bovinae (Essop et al., 1997; Gentry, 1992; Georgiadis et al.,
1990; Hassanin and Douzery, 1999a; Matthee and Robinson, 1999) and is endemic to the African continent.
Nine extant species are recognized, belonging to three
genera. These genera were conventionally identiWed
based on morphology and limited fossil evidence
(Ansell, 1971): Tragelaphus (containing T. strepsiceros,
T. imberbis, T. angasi, T. buxtoni, T. spekei, and T. scriptus), Taurotragus (containing T. oryx and T. derbianus)
and Boocercus (for T. euryceros). Cranial similarity,
however, suggests an sister taxon relationship between
T. spekei and T. angasi (Kingdon, 1982; Roberts, 1952),
and T. imberbis and T. strepsiceros (Alden et al., 1995;
Kingdon, 1982), while the presence of horns in both
sexes of T. euryceros, T. oryx, and T. derbianus has been
taken to indicate that these three lineages form a derived
monophyletic entity (Gentry, 1990). Likewise, general
body conformation and pelt coloration unite the eland
(T. derbianus and T. oryx) and kudu (T. strepsiceros and
T. imberbis) lineages (Walker, 1964). A recent mtDNA
investigation based on cytochrome b data (Matthee and
Robinson, 1999) provided an alternative working
hypothesis for the evolution of the tribe that did not support the recognition of the three genera (see also Essop
et al., 1997; Gatesy et al., 1997; Georgiadis et al., 1990).
In addition, the mtDNA data contradicted most of the
abovementioned morphological associations with
the contrasting evolutionary associations suggested by
the mitochondrial and morphological phylogenies being
attributed to convergent evolution at the morphological
level (Matthee and Robinson, 1999; Hassanin and
Douzery, 1999b). It is equally likely, however, that the
group experienced a fairly rapid radiation during which
independent lineage sorting could have resulted in
increased homoplasy obscuring the correct phylogenetic aYnities in the mtDNA gene tree. In other words,
the conXict between the gene tree and the morphological
tree could equally be explained by shared mtDNA char-

625

acters becoming Wxed in only a portion of the descendents of a polymorphic ancestor (Hillis, 1999).
The aims of this study were Wrst to test the utility of
nuclear DNA markers in a group that experienced a
fairly rapid radiation during the last 15 My (Hassanin
and Douzery, 2003; Matthee and Robinson, 1999; Vrba,
1985). Second, resolution at the nuclear DNA level could
clarify the apparent conXict that exists between the
mtDNA and morphological topologies. Additionally it
was anticipated that a more comprehensive molecular
data set would allow for greater accuracy in determining
of the pattern and timing of episodes of speciation
within the group, as well as providing further insights on
the factors thought to be driving speciation in these African antelope. This is particularly pertinent since suggestions that species inhabiting moist forest habitats (T.
euryceros, T. spekei, and T. scriptus) are derived, and
those adapted to more arid environments (T. imberbis,
T. angasi, T. strepsiceros, T. oryx, and T. derbianus) are
more basal, rest on a single, maternally inherited locus
(Matthee and Robinson, 1999).

2. Materials and methods

2.1. Taxonomic representation
All nine recognized tragelaphid species were
included in our study (Table 1). Moreover, wherever
possible, multiple representatives drawn from diVerent
geographic areas were included. Unfortunately in some
cases no exact locality data were available and only a
broad geographical region from where the specimen
was collected could be obtained. The Indian nilgai
(Boselaphus tragocamelus), domestic cow (Bos taurus),
and African buValo (Syncerus caVer) were used as
closely related outgroup taxa in conjunction with the
impala (Aepyceros melampus), a member of Antilopinae (Kingdon, 1982, 1997; Bronner et al., 2003), as a
more distantly related lineage (Matthee and Davis,
2001).
2.2. DNA ampliWcation, sequencing, and alignment
DNA was extracted using standard laboratory protocols (Sambrook et al., 1989). Four nuclear DNA markers
(thyrotropinTHY, protein-kinase CIPRKCI, Bspectrin non-erythrocyticSPTBN, and stem cell
factorMGF) were ampliWed and sequenced using published primers and protocols (Matthee et al., 2001). An
additional 24 mtDNA sequences were obtained from
three mtDNA genescytochrome b, 12S rRNA, and
16S rRNA (for mtDNA PCR and primer details see
Matthee and Davis, 2001). The remainder of the
mtDNA data were obtained from public databases (33
sequences in total; Table 1).

626

S. Willows-Munro et al. / Molecular Phylogenetics and Evolution 35 (2005) 624636

Table 1
Taxonomic sampling used in the study
Species name (Vernacular name)

DNA region
Mitochondrial DNA

Taurotragus derbianus (Giant Eland)

Taurotragus oryx (Common Eland)
Taurotragus oryx (Common Eland)
Tragelaphus strepsiceros (Greater Kudu)
Tragelaphus strepsiceros (Greater Kudu)
Tragelaphus imberbis (Lesser Kudu)
Tragelaphus angasi (Common Nyala)
Tragelaphus angasi (Common Nyala)
Tragelaphus buxtoni (Mountain Nyala)
Tragelaphus buxtoni (Mountain Nyala)
Tragelaphus euryceros (Bongo)
Tragelaphus euryceros (Bongo)
Tragelaphus spekei (Sitatunga)
Tragelaphus scriptus (Bushbuck)
Tragelaphus scriptus (Bushbuck)
Boselaphus tragocamelus (Indian Nilgai)
Bos taurus (Domestic Cow)
Syncerus caVer (African BuValo)
Aepyceros melampus (Impala)

Nuclear DNA

12S rRNA

16S rRNA

Cyt b

MGF

PRKCI

SPTBN

THY

AY667206*
AF091710d
AY667207*
AY667208*
AF091696d
AY667209*
AF091698d
AY667210*
AY667211*
AY667212*
AF091691d
AY667213*
AF091692d
AY667214*
AF091693d
U86963c
NC001567a
AF091688d
U86964c

AY667193*
AY667194*
U87060c
AY667195*
AY667196*
M86493b
AY667197*
AY667198*
AY667199*
AY667200*
AY667201*
AY667202*
AY667203*
AY667204*
AY667205*
U87013c
NC001567a
U87061c
U87014c

AF022062e
AF036278d
AF022057e
AF022063e
AF036280d
AF036279d
AF091633d
AF022066e
AY667215*
AY667216*
AF036276d
AF022065e
AJ222680d
AF036277d
AF022067e
AJ222679d
NC001567a
AF036275d
AF036289d

AY667177*
AY667178*
AY667179*
AY667180*
AY667181*
AY667182*
AY667183*
AY667184*
AY667185*
AY667186*
AY667187*
AY667188*
AY667189*
AY667190*
AY667191*
AF165724g
AF165716g
AY667192*
AF165780g

AY667162*
AY667163*
AY667164*
AY667165*
AY667166*
AY667167*
AY667168*
AY667169*
AY667170*
AY667171*
AY667172*
AY667173*
AY667174*
AY667175*
AY667176*
AF165725g
AY029293f
AF210175h
AF165781g

AY667232*
AY667233*
AY667234*
AY667235*
AY667236*
AY667237*
AY667238*
AY667239*
AY667240*
AY667241*
AY667242*
AY667243*
AY667244*
AY667245*
AY667246*
AF165726g
AF165718g
AF210197h
AF165782g

AY667217*
AY667218*
AY667219*
AY667220*
AY667221*
AY667222*
AY667223*
AY667224*
AY667225*
AY667226*
AY667227*
AY667228*
AY667229*
AY667230*
AY667231*
AF165729g
AF65721g
AF210219h
AF165785g

Accession numbers with * indicates specimens sequenced in the present study.

a
Anderson et al. (1982).
b
Allard et al. (1992).
c
Gatesy et al. (1997).
d
Hassanin and Douzery (1999a, b).
e
Matthee and Robinson (1999).
f
De Donto et al. (2001).
g
Matthee et al. (2001).
h
Matthee and Davis (2001).

Cycle sequencing was performed using Big Dye chemistry (Version 3.1, Applied Biosystems) with the resulting
products analyzed on a 3100 ABI automated sequencer.
Sequences were aligned using the default settings of
CLUSTAL X (Thompson et al., 1997) and optimized
manually to ensure accuracy. Areas of ambiguously
aligned sites were limited to a single gene and this region
was excluded from all analysis (positions 149159 in
aligned MGF data set). Heterozygous sites were coded
using the IUB codes. The exon and intron boundaries
were deWned from previously published sequence data
for each DNA region (Matthee et al., 2001; Matthee and
Davis, 2001). To further ensure the homology of the
DNA sequences, exon regions were translated into
amino acids and screened for functionality. No stop
codons or insertions were present in the exonic regions.
All new sequences were deposited in GenBank, accession
numbers are shown in Table 1.
2.3. Data partitioning
Each DNA region was analyzed separately. Additionally, data sets were partitioned and analyzed according
to origin of marker (i.e., mitochondrial or nuclear), and
all characters and taxa were merged into a single data

matrix or molecular supermatrix. To explore inconsistencies among data partitions, pair-wise incongruence
length diVerences (ILD) were calculated in PAUP*
4.0b10 (SwoVord, 2003) excluding uninformative characters (Hipp et al., 2004). To further explore the homogeneity of signal among data partitions the crossed
ShimodairaHasegawa tests (crossed SH tests, Shimodaira and Hasegawa, 1999) were performed in PAUP*
4.0b10. In these tests the most likely topology obtained
by the mitochondrial, nuclear partitions and the supermatrix were compared in pair-wise fashion.
2.4. Data analysis
Parsimony and maximum likelihood searches were
performed in PAUP* 4.0b10 using the heuristic search
option with 100 replicates of random taxon addition and
TBR branch swapping. For the maximum likelihood
analyses the optimal evolutionary model and model
parameters were estimated applying the Akaike information criteria (AIC, Akaike, 1974) using MODELTEST
3.06 (Posada and Crandall, 1998). All character transformations were equally weighted and alignment gaps
treated as missing characters in parsimony. Nodal support for parsimony and maximum likelihood trees was

S. Willows-Munro et al. / Molecular Phylogenetics and Evolution 35 (2005) 624636

assessed by 1000 bootstrap iterations as well as Bremer

support values (Bremer, 1988, 1994) using TreeRot v.2
(Sorenson, 1999).
Bayesian analyses were performed using MrBayes
3.0b3 (Huelsenbeck and Ronquest, 2001) with each
Bayes run launched from a random starting tree and run
for 1 106 generations. To conWrm that the analysis was
not trapped at a local optima (Leach and Reeder, 2002)
all runs were performed three times. Trees were sampled
every 100th generation and burnin was determined using
the sumt command in MrBayes 3.0b3. The priors were
set to the GTR model of evolution and the combined
data (mitochondrial, nuclear, and supermatrix partitions) were examined using the partition option. MrBayes 3.0b3 was also used to estimate the 95% posterior
credibility intervals for each DNA region and combined
partition.
Recent studies have suggested that the posterior
probability values generated by Bayesian inference may
be inXated (Cummings et al., 2003; Suzuki et al., 2002;
Waddell et al., 2001) but in some instances may be better
estimates of true phylogeny (Wilcox et al., 2002). In
the present study, robust associations were identiWed by
bootstrap values 770% and posterior probability values
795%.
2.5. Distribution of support among the diVerent data
partitions
An important aspect of any multigene analysis is the
examination of support contributed by the diVerent data
sets in the supermatrix topology. First, following
Kluges (1989) suggestion, we determined the number of
unambiguous synapomorphies for selected nodes using
MacClade 4.0 (Maddison and Maddison, 1992). Second,
partitioned Bremer support values for selected nodes
were calculated for each DNA region using TreeRot v.2
(PBS, e.g., Gatesy et al., 1999). The signiWcance of PBS
values are highly data dependent (Lee, 2000) making
comparisons between data sets problematic (Bremer,
1988; Sanderson and Donoghue, 1996). For this reason
we also followed a method proposed by Lee (2000) to
assess nodal signiWcance. This methodology is similar to
that in conventional Bremer support analysis (Bremer,
1988, 1994) but uses Templetons test (1983) in PAUP*
4.0b10 to statistically compare individual nodes on alternative trees.
2.6. Molecular clock calibration
Divergence times of the tragelaphids were estimated
using the relaxed Bayesian molecular clock method for
multigene data (Thorne and Kishino, 2002). All seven
genes were included and the supermatrix phylogeny was
used as reference tree. The F84 model was applied to
each gene fragment independently. The Markov chain

627

was sampled 10,000 times, every 100th generation. The

burnin was set at 10,000 generations. The following prior
distributions were adapted: 20 Mya (SD D 20 My) for
the expected time between tip and root of the tree if the
nodes are unconstrained (rttm), 0.008 (SD D 0.004) substitutions per site per million years for the mutational
rate at the root of the tree (rtrate). These settings were
determined by approximating the prior distribution of
times and rates. First, arbitrary time settings for rttm
and rttmsd were made followed by the estimation of the
prior distributions of node times. When these approximations resulted in node calibrations that were too
recent (for example in comparison to the fossil dating),
the value of rttm was increased. This process was iterated until a time estimate was obtained approximating
currently accepted values. The values of rtrate and
rtratesd were determined a priori. For example, the tragelaphids are thought to have originated approximately
15 Mya (Hill et al., 1985)we therefore use this value in
combination with the branch length estimates to determine the rate of evolution. In our case, the amount of
evolution from each tip to the root for each gene was
determined, and a median amount of evolution among
genes separating roots and tips was estimated. The
median was calculated at 0.12 substitutions per site and
this value was divided by the 15 million years to equate
to 0.008. The highest possible number of time units
between the tip and the root of the tree was set at 50 My
and the prior on the rate autocorrelation was set at 0
allowing for the rate of each gene fragment to vary. Fossil calibration points were available for two of the deeper
nodes and these were incorporated as upper and lower
constraints. The Wrst of these is the split between Bovinae and Antilopinae (16.423.8 Mya) as used by Hassanin and Douzery (2003 and references therein). Our
second calibration point is based on the emergence of
the Bovinae tribes (11.519.7 Mya, Hill et al., 1985; also
see Hassanin and Douzery, 2003). To explore possible
inconsistencies between the mitochondrial and nuclear
partitions, divergence times obtained using each partition were compared to each other.

3. Results
3.1. Data description
The aligned supermatrix contained 7 DNA regions
and 4682 characters, 3149 bp non-protein coding and
1533 bp protein coding characters. The diVerent DNA
regions were characterized by variation in base composition and modes of evolution (Table 2). As typically
expected the mitochondrial partition contained a higher
proportion of variable characters (28%) when compared
to that of the nuclear partition (12.7%). As expected
given the higher proportion of variable characters in the

628

S. Willows-Munro et al. / Molecular Phylogenetics and Evolution 35 (2005) 624636

Table 2
Patterns of sequence variability among the independent DNA regions, mitochondrial partition, nuclear partition, and supermatrix
DNA
partition

Total
(bp)

% Protein
coding

% Variable
characters

Gamma distribution
(
value)

Nucleotide frequencies
%A

%T

%C

%G

12S rRNA
16S rRNA
Cyt b
MGF
PRKCI
SPTBN
THY
Mitochondrial
Nuclear
Supermatrix

591
348
1140
674
667
765
667
2079
2603
4682

0
0
100
0
11
16
34
55
15
33

15.6
23.9
35.8
14.1
9.60
16.1
9.90
28.0
12.7
19.5

0.736
0.489
2.023
0.499
NA
0.505
0.139
0.799
0.275
0.623

36.1
34.9
31.9
32.3
34.3
23.2
30.1
34.9
29.8
31.6

22.7
24.6
25.9
34.6
38.7
27.2
32.6
23.7
32.3
28.6

24.7
21.3
28.9
18.1
11.2
24.8
19.3
28.0
18.8
22.9

16.5
19.2
13.4
15.0
16.0
24.8
18.1
13.4
19.1
16.9

In the case of nuclear sequence data % protein coding refers to Xanking exon regions.

mtDNA partition, the consistency (CI) and retention

(RI) index values of the mitochondrial markers (CI:
0.48610.5547, RI: 0.58820.7311) were lower than those
of the nuclear markers (CI: 0.72090.8261, RI: 0.8400
0.9149; Table 3) indicating higher levels of homoplasy in
the mtDNA data.
3.2. Distribution of support among the diVerent data
partitions
In an attempt to compare how well each independent
DNA region recovered the phylogeny, eight nodes of
interest (labelled AH in Figs. 1 and 2) were identiWed.
These nodes elucidate associations among all extant
tragelaphid species and were supported across diVerent
methods in the supermatrix phylogeny (Fig. 2).
Apart from the cytochrome b and to a lesser extend
the SPTBN data which performed reasonably well in
recovering the associations among the tragelaphid species, the separate analysis of individual DNA regions
resulted in several unresolved nodes (Table 3). The 95%
posterior intervals estimated by the Bayesian analysis of
the mitochondrial, nuclear partitions, and the supermatrix resulted in considerably fewer unique topologies relative to those occurring within the same posterior
intervals of the individual DNA regions (Table 3). This
decrease in the number of topologies reXects the increase
in resolution obtained from the combined analysis of
DNA regions (Buckley et al., 2002). Importantly, many
of the independent DNA regions, however, provided
positive character support for nodes when analyzed as
part of the supermatrix. For example, analysis of the
MGF data set was only able to retrieve node A, but PBS
values for this data set (Fig. 3) indicate that the supermatrix topology was not negatively inXuenced by the inclusion of these data. This is also supported by Templetons
test suggesting that the potential phylogenetic noise
from the inclusion of the MGF region to the supermatrix did not reduce support at any of the 8 selected nodes
(Fig. 3). In fact, nodes A and G (which were signiWcantly

supported in the supermatrix analysis, P > 0.05) were not

signiWcantly supported once the MGF data set was
removed, suggesting that this data set probably contains
hidden support for these nodes.
Nodes A, C, E, and G in the supermatrix topology
were statistically supported by Templetons test (Fig. 3).
These nodes were also consistently retrieved in both the
mitochondrial and nuclear topologies (bootstrap values
>75% and posterior probability values >0.95). The distribution of probability values retrieved by Templetons
tests also illustrates how diVerent DNA regions support
diVerent nodes. For example, the removal of only 3 DNA
regions resulted in non-signiWcant P-values at node E
suggesting that the phylogenetic signal in support of this
node was contained within the cytochrome b, 16S rRNA,
and SPTBN data sets (i.e., node E was unaVected by the
removal of 12S rRNA, MGF, PRKCI, and THY).
3.3. The phylogenetic utility of indels
Indels among ingroup taxa (insertions/deletions
involving two or more consecutive bp with clearly
deWned boundaries) were only present in the nuclear
intron data. These were mapped onto the supermatrix
topology (Fig. 2) to determine the phylogenetic information content of these markers. Within the ingroup it was
found that deletions outnumbered insertions 4 to 1 (Fig.
2); the T. imberbis lineage was characterized by the greatest number of unique deletions. This is consistent with
the early divergence of this group from the other tragelaphid taxa (see below). At a deeper phylogenetic level
the Bovinae tribes retrieved a deletioninsertion ratio of
5:1 and in this instance a single indel was homoplasious
involving a 2 bp deletion at position 39903991 in the
MGF data set (present in B. taurus and B. tragocamelus).
3.4. Congruence of diVerent data partitions
The topologies derived from the mtDNA and nuclear
DNA partitions were largely congruent. Pair-wise ILD

Table 3
Summary of the topologies obtained from analysis of the independent DNA regions, mitochondrial partition, nuclear partition, and supermatrix
Data partition

16S rRNA

Cyt b

MGF

PRKCI

SPTBN

THY

Mitochondrial

Nuclear

Supermatrix

Pars.
ML
Bayesian
Pars.
ML
Bayesian
Pars.
ML
Bayesian
Pars.
ML
Bayesian
Pars.
ML
Bayesian
Pars.
ML
Bayesian
Pars.
ML
Bayesian
Pars.
ML
Bayesian
Pars.
ML
Bayesian
Pars.
ML
Bayesian

No. of pars.
informative
characters

No. of equally
parsimonious
trees

Tree length

CI

RI

51

94

142

0.5510

0.5882

61

151

0.5547

0.7311

290

904

0.4861

0.6346

37

116

0.7547

0.8632

17

104

52

0.8261

0.9149

37

142

0.7736

0.8667

31

24

75

0.7209

0.8400

550

1699

0.4753

0.5527

145

491

0.7162

0.8215

700

2203

0.5043

0.5938

No. of topologies
in 95% posterior
interval

14,096

2015

212

7299

13,242

517

10,078

22

18

Nodes
A

51
51
0.70
72
73
0.99
81
100
1.00
100
99
1.00
96
93
1.00
89
80
1.00
100
100
1.00
98
98
1.00
100
100
1.00
100
100
1.00

X
X
X
77
63
1.00
86
88
0.99
X
88
X
X
X
X
X
57
X
X
X
X
95
95
1.00
X
X
X
93
71
0.97

X
X
X
X
X
X
67
79
1.00
X
X
X
X
X
X
92
85
0.97
X
X
X
78
87
1.00
85
94
1.00
95
99
1.00

X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
96
94
1.00
81
75
1.00

X
X
X
X
X
X
95
94
1.00
X
X
X
X
X
X
95
96
1.00
X
X
X
94
94
1.00
65
86
1.00
99
100
1.00

X
X
X
66
70
0.95
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
58
75
1.00
62
76
1.00

X
X
X
X
X
X
90
93
1.00
X
X
X
X
X
X
X
X
X
X
X
X
96
100
1.00
X
X
X
100
98
1.00

X
X
X
X
X
X
58
70
0.79
X
X
X
X
X
0.64
97
90
1.00
X
X
X
50
83
0.90
93
87
1.00
100
100
1.00

No. nodes with 770%

bootstrap or 95%
posterior probability
0
0
0
1
1
2
4
6
4
1
2
1
1
1
1
4
4
4
1
1
1
5
6
5
4
6
6
7
7
8

S. Willows-Munro et al. / Molecular Phylogenetics and Evolution 35 (2005) 624636

12S rRNA

Method

The number of parsimony informative characters, number of trees, optimal tree length, consistency index (CI), and retention index (RI) values are given for each parsimony tree, as well as, the
number of topologies found within the 95% posterior intervals of each partition. Nodes AH correspond to those in Figs. 1 and 2, bootstrap values (>50) are given, X indicates where a node
was not retrieved.

629

630

S. Willows-Munro et al. / Molecular Phylogenetics and Evolution 35 (2005) 624636

Fig. 1. Topologies retrieved by (A) the combined mitochondrial DNA data and (B) the combined nuclear DNA data. Values given above the
branches represent the molecular clock estimates of divergence times for nodes labelled AH with standard deviations in parenthesis. Values below
the branches represent Bayesian posterior probabilities, maximum likelihood bootstrap values, and parsimony bootstrap values.

tests showed that there was no signiWcant conXict

between the mitochondrial and nuclear partitions
(P D 0.70). The crossed SH tests indicated that the mitochondrial and nuclear topologies were signiWcantly
diVerent (P D 0.018 and 0.004) and closer inspection of
the topologies revealed that the only striking diVerence
was in the placement of the bushbuck, T. angasi (Fig. 1).
The supermatrix topology was not, however, rejected by
either the mitochondrial or the nuclear partitions, underscoring the utility of the supermatrix for phylogenetic
inference. The crossed SH tests can allow for the identiWcation of areas of topological conXict for example the
nuclear topology was rejected by the supermatrix data
set due to the association between the two basal species,
T. angasi and T. imberbis. The exclusion of these
two basal taxa resulted in congruence between the
supermatrix and nuclear topologies (P D 0.112).
3.5. The tragelaphid phylogeny
Given the overall stability of the supermatrix analysis,
the lower number of trees found in the 95% credibility
interval (Table 3) and the absence of any major conXict
between this topology and the mtDNA and nuclear

DNA partitions (see above), this phylogeny was selected

as best reXecting the evolutionary relationships of the
tragelaphids (Fig. 2). Our molecular data conWrmed Wndings of previous studies by retrieving a single origin for
the subfamily Bovidae (Essop et al., 1997; Gentry, 1992;
Georgiadis et al., 1990; Matthee and Robinson, 1999;
Hassanin and Douzery, 1999a) and no genetic support
for the recognition of three Tragelaphid genera (Matthee and Robinson, 1999). The tribe Tragelaphini was
recovered as a monophyletic entity, with T. imberbis
basal in the phylogeny with the next most basal lineage
represented by T. angasi. The close but old evolutionary
relationship of these taxa is further supported by the
lack of a unique 31 bp deletion in the SPTBN DNA
region which is a synapomorphy uniting the remaining
seven species. Moreover, T. buxtoni, T. euryceros, T. spekei, and T. scriptus (all species adapted to closed forest
living) formed a well-supported monophyletic clade in
all the analyses. Within this closed forest group the sister taxon relationship between T. spekei and T. euryceros
suggested by earlier mtDNA data (Matthee and Robinson, 1999) was similarly retrieved by our extended DNA
data set. Tragelaphus scriptus was placed as a sister
lineage to the T. spekei and T. euryceros clade with

S. Willows-Munro et al. / Molecular Phylogenetics and Evolution 35 (2005) 624636

631

Fig. 2. Topology retrieved by the combined analysis of all molecular data. Values given above the branches represent the molecular clock estimates of
divergence times for nodes labelled AH with standard deviations in parenthesis. These divergence values represent the age of the most recent common ancestor of the taxa contained within that clade. Values below the branches represent Bayesian posterior probabilities, maximum likelihood
bootstrap values, parsimony bootstrap values. Solid bars indicate deletion events and stippled bars represent insertion events. The single homoplasious deletion is highlighted with *.

T. buxtoni (not included in previous analyses) as the

most basal lineage within the closed forest group. An
association between T. derbianus, T. oryx, and T. strepsiceros, which are united by physiological adaptations
to an arid savannah environment (Alden et al., 1995;
Kingdon, 1982) was not consistently retrieved in all
methods of analysis. The morphologically similar T.
derbianus and T. oryx were grouped as sister taxa,
while the supermatrix suggests a more derived position
for T. strepsiceros. It should be noted, however, that
only two data sets contain unambiguous synapomorphies (cytochrome b and MGF) in support of an association between T. strepsiceros and the closed forest
group (node D). Pair-wise sequence divergence values
among individuals of the same species, ranged from
0.02% within the geographically isolated T. buxtoni

lineage, to 0.65% within the geographically widespread

T. scriptus lineage.
3.6. Molecular clock
Prior speciWcation of the mutational rate could only
be made with reference to the sequence data thereby violating the principles underpinning the Bayesian framework (Yang and Yoder, 2003). Although a prior
distribution of 0.008 (SD D 0.004) substitutions per site
per million years was estimated and used for the rate at
the root, we found that large changes (i.e., 0.30.003) to
this prior distribution, and also to that of the root age
prior, had little or no impact on the resulting divergence
times. Additionally, independent analysis of the mitochondrial and nuclear partitions produced divergence

632

S. Willows-Munro et al. / Molecular Phylogenetics and Evolution 35 (2005) 624636

Fig. 3. The number of unambiguous synapomorphies, partitioned Bremer support and nodal signiWcance values for each DNA region at nodes AH.
Positive PBS values indicate that a particular gene region contributes phylogenetic signal at that node, while a negative value (shown in black) indicates that a data set conXicts with the supermatrix at a particular node.

times in the same range as those obtained when using all

data (Figs. 1 and 2).
The divergence of the Tragelaphini from the other
bovid tribes was estimated at approximately 14.08 Mya
(SD D 2.60, Fig. 2) while that between the basal bushland
species, T. angasi and T. imberbis from the other tragelaphid antelope was estimated to have occurred approximately 10.89 Mya (SD D 2.87). This was followed by a
period of rapid ecological specialization which gave rise
to those species conWned to moist forest environments
(T. buxtoni, T. euryceros, T. spekei, and T. scriptus), and
those adapted to a more arid savannah environment (T.
derbianus, T.oryx, and T. strepsiceros).

4. Discussion
4.1. Nuclear intron markers
The nuclear introns selected in the present investigation have been applied successfully to explain evolutionary relationships among the Cetartiodactyla (Matthee

et al., 2001), Leporidae (Matthee et al., 2004), and also the

Bovidae (Matthee and Davis, 2001) permitting a rigorous
assessment of intron markers in mammalian systematics.
Several broad conclusions are evident from these studies.
First, it is clear that phylogenetic resolution is positively correlated with the number of nuclear DNA characters employed. All four studies revealed an increase in
phylogenetic resolution when genes are combined. This
is contrary to most of the results obtained from previous
mammalian mtDNA endeavors where increasing the
number of mtDNA genes (characters) had little eVect on
robust resolution (for examples see Irwin et al., 1991 and
Corneli, 2003). Even recent investigations using Bayesian methods on complete mtDNA genomes (Reyes et al.,
2004) reveal weak support for most of the well-supported mammalian inter-ordinal associations as suggested by nuclear DNA studies (Madsen et al., 2001;
Murphy et al., 2001).
Second, the low levels of homoplasy present in the
nuclear DNA introns make them particularly well
suited for retrieving information on rapid radiation
events such as those typically observed in the Pecora

S. Willows-Munro et al. / Molecular Phylogenetics and Evolution 35 (2005) 624636

(Matthee et al., 2001), Bovidae (Matthee and Davis,

2001), and Leporidae (Matthee et al., 2004). This is in
marked contrast to several mtDNA studies on the same
groups (Gatesy et al., 1997; Halanych et al., 1999; Matthee and Robinson, 1999).
Third, given the relatively conservative nature of the
nuclear DNA intron markers, the number of characters
needed to obtain phylogenetic resolution seems to be
directly linked with the age of the speciation events (as
with most other loci). Our tragelaphid data, however,
also suggests a link between the phylogenetic resolution and the rapidity at which these evolutionary radiations occurred. For older radiations (among the
Cetartiodactyla) all individual nuclear DNA fragments
(with low numbers of characters) resulted in wellresolved phylogenies (Matthee et al., 2001), whereas the
results are often equivocal in species characterized by
more recent radiations (such as the Tragelaphini). For
example, in the present study only the combined
nuclear DNA tree produced robust results for the Tragelaphinia group that are thought to have evolved
approximately 14 Mya. In fact, the best resolution on
tragelaphid relationships from any single gene was that
obtained from the mtDNA cytochrome b data (Table
3). This is contrary to the published literature where
cytochrome b failed to provide phylogenetic resolution
in the Bovidae and Leporidae (Halanych et al., 1999;
Matthee and Robinson, 1999). This seems likely to
reXect diVerences in the modes of speciation events
both the Leporidae and Bovidae are thought to have
experienced rapid radiations between 1410 (Leporidae) and 2014 million years ago (Bovidae) (Matthee
and Davis, 2001 and Matthee et al., 2004). In these
instances, reverse mutations could have randomized
the limited phylogenetic signal that accumulated during the rapid radiation. In the Tragelaphini, the diversiWcation probably also started approximately 14 Mya
but speciation was more gradual occurringup till 5
million years ago, which provided more time for the
accumulation of synapomorphies, and less time for the
mtDNA signal to be erased. We are of the opinion that
this accounts for the cytochrome b genes usefulness in
resolving phylogenetic relationships within the spiralhorned antelope.
4.2. Tragelaphid systematics: molecules versus
morphology
The interspeciWc and intraspeciWc (Haltenorth,
1963) variability in many of the morphological features
previously used to delimit the group makes the tracking of species turn-over from the fossil record diYcult.
For example, Gentry (1978) suggested that T. strepsiceros and T. imberbis diverged approximately 3 Mya
which is clearly in conXict with the molecular phylogeny presented herein (Fig. 2). Given that our investiga-

633

tion is based on full taxonomic representation,

corroborating nuclear and mtDNA evidence, and is
unaVected by method of phylogenetic construction
used, we consider our supermatrix phylogeny to be
particularly stable. This is especially true for the monophyletic status of the tribe, the monophyly of the four
closed forest species (T. buxtoni, T. euryceros, T. spekei,
and T. scriptus), and the basal position of T. imberbis
and T. angasi. These associations, however, are in
marked disagreement with the morphological data.
Although we cannot exclude the possibility that independent lineage sorting, gene duplication/deletion and/
or hybrid speciation could have contributed to the conXicting placement of T. angasi and T. strepsiceros (Fig.
1), the remaining evolutionary relationships are well
supported in both data sets (nuclear and mtDNA).
Based on the molecular supermatrix topology it seems
reasonable to suggest that many of the morphological
similarities among taxai.e., the cranial similarity
between T. spekei and T. angasi (Kingdon, 1982; Roberts, 1952), and T. imberbis and T. strepsiceros (Alden
et al., 1995; Kingdon, 1982), the presence of horns in
both sexes of T. euryceros, T. oryx, and T. derbianus
(Gentry, 1990), and the similarity in general body form
and pelt coloration uniting T. strepsiceros and T.
imberbis (Walker, 1964)are due to convergence and
not recent common ancestry.
4.3. Speciation within Tragelaphus
Although the timing of divergence among tragelaphid species may be estimated using a molecular clock,
the factors driving speciation can only be examined by
comparing the tragelaphid phylogeny with paleoclimatic changes coupled to vicariance in Africa. Our
results suggest that the global cooling trend which
peaked 14 Mya (Cerling et al., 1997; Zachos et al., 2001)
coincided with the diversiWcation among species
belonging to this bovid tribe. During the mid-Miocene,
declining global temperatures saw an increase in seasonality leading to the expansion of savannah grasslands and the contraction of humid forests into
geographically isolated forest islands surrounded by
drier bushland and savannah (Lindsay, 1998). The most
primitive members of the group, T. imberbis and T. angasi both occur in dense woodland, with the former generally found at higher altitudes (Nowak, 1999). It seems
probable that the fragmentation of dense forest habitats
resulted in the isolation of T. imberbis and T. angasi.
The expansion of savannah grassland prompted the
development of open savannah specialists such as T.
strepsiceros and the two eland species (T. derbianus and
T. oryx) between 14.08 and 10.89 Mya (Fig. 2). The subsequent diversiWcation of species adapted to a more
tropical/wet environment (T. buxtoni, T. euryceros, T.
spekei, and T. scriptus) at approximately 7.31 Mya is,

634

S. Willows-Munro et al. / Molecular Phylogenetics and Evolution 35 (2005) 624636

however, less clear. The latter period coincides with a

vegetational shift leading to the global expansion of C4
ecosystems (Cerling et al., 1997). It seems likely therefore that the diversiWcation within the closed forest
group was driven by the continued fragmentation of
forest habitats, as this speciation was not accompanied
by a shift in diet. In summary, we propose that the rapid
diversiWcation of the tragelaphids was likely to have
been due to a combination of factors including paleoclimatic shifts, vicariance, and the sudden availability of
open niches on the African continent.

Acknowledgments
We acknowledge E. Harley (University of Cape
Town), M. Harman (Powell Cotton Museum), J. Sakwa
(University of Pretoria), and the Pretoria Zoological
Gardens for providing samples. John Gatesy and three
anonymous reviewers are also thanked for their useful
comments on the manuscript. Financial assistance was
provided by the National Research Foundation of
South Africa (GUN 2053662).

References
Alden, P.C., Estes, R.D., Schlitter, D., MacBride, B., 1995. National
Audubon Society Field Guide to African Wildlife. Knopf, New
York.
Allard, M., Miyamoto, M., Jarecki, L., Kraus, F., Tennant, M., 1992.
DNA systematics and evolution of the artiodactyls family Bovidae.
Proc. Natl. Acad. Sci. USA 89, 39723976.
Akaike, H., 1974. A new look at the statistical model identiWcation.
IEEE Trans. Autom. Contr. 19, 716723.
Anderson, S., de Bruijn, M.H., Coulson, A.R., Eperon, I.C., Sanger, F.,
Young, I.G., 1982. Complete sequence of bovine mitochondrial
DNA. Conserved features of the mammalian genome. J. Mol. Biol.
156, 683717.
Ansell, W.F.H., 1971. Order Artiodactyla. In: Meester, J., Setzer, H.W.
(Eds.), The Mammals of Africa: An IdentiWcation Manual. Smithsonian Institute Press, Washington.
Bremer, K., 1988. The limits of amino acid sequence data in angiosperm phylogenetic reconstruction. Evolution 42, 795803.
Bremer, K., 1994. Branch support and tree stability. Cladistics 10, 295
304.
Bronner, G.N., HoVmann, M., Taylor, P.J., Chimimba, C.T., Best, P.B.,
Matthee, C.A., Robinson, T.J., 2003. A revised systematic checklist
of the extant mammals of the southern African subregion. Durban
Mus. Novit. 28, 56106.
Buckley, T.R., Arensburger, P., Simon, C., Chambers, G.K., 2002. Combining data, bayesian phylogenetics, and the origin of the New Zealand Circada genera. Syst. Biol. 51, 418.
Cerling, T.E., Harris, J.M., MacFadden, B.J., Leakey, M.G., Quade,
J., Eisenmann, V., Ehleringer, J.R., 1997. Global vegetation
change through the Miocene/Pliocene boundary. Nature 389,
153158.
Corneli, P.S., 2003. Complete mitochondrial genomes and eutherian
evolution. J. Mammal. Evol. 9, 281305.
Cracraft, J., Helm-Bychowski, K., 1991. Parsimony and phylogenetic inference using DNA sequences: some methodological

strategies. In: Miyamoto, M.M., Cracraft, J. (Eds.), Phylogenetic

Analysis of DNA Sequences. Oxford University Press, New
York.
Cummings, M.P., Handley, S.A., Myers, D.S., Reed, D.L., Rokas, A.,
Winka, K., 2003. Comparing bootstrap and posterior
probability values in the four-taxon case. Syst. Biol. 52, 477487.
De Donto, M., Gallagher Jr., D.S., Davis, S.K., Stelly, D.M., Taylor,
J.F., 2001. The assignment of PRKCI to bovine chromosome
1q34 ! q36 by FISH suggests a new assignment to human chromosome 3. Cytogenet. Cell. Genet. 95, 7981.
Essop, M.F., Harley, E.H., Baumgarten, I., 1997. A molecular phylogeny of some Bovidae based on restriction-site mapping of mitochondrial DNA. J. Mammal. 78, 377386.
Flores-Villela, O., Kjer, K.M., Benabib, M., Sites Jr., J.W., 2000. Multiple data sets, congruence, and hypothesis testing for the phylogeny
of basal groups of the lizard genus Sceloporus (Squamata, Phrynosomatidae). Syst. Biol. 49, 713739.
Gatesy, J., Hayashi, C., Cronin, M.A., Arctander, P., 1996. Evidence
from milk casein genes that cetaceans are close relatives of hippopotamid artiodactyls. Mol. Biol. Evol. 13, 954963.
Gatesy, J., Amato, G., Vrba, E.S., Schaller, G., DeSalle, R., 1997. A cladistic analysis of mitochondrial ribosomal DNA from the Bovidae.
Mol. Phylogenet. Evol. 7, 303319.
Gatesy, J., Milinkovitch, M., Waddell, V., Stanhope, M., 1999. Stability
of cladistic relationships between Cetacea and higher level artiodactyl taxa. Syst. Biol. 48, 620.
Gatesy, J., Arctander, P., 2000. Hidden morphological support for
the phylogenetic placement of Pseudoryx nghetinhensis with
bovine bovids: a combined analysis of gross anatomical evidence and DNA sequences from Wve genes. Syst. Biol. 3, 515
538.
Gatesy, J., Matthee, C., DeSalle, R., Hayashi, C., 2002. Resolution of a
supertree/supermatrix paradox. Syst. Biol. 51, 652664.
Gatesy, J., Amato, G., Norell, M., Desalle, R., Hayashi, C., 2003. Combined support for wholesale taxic atavism in Gavialine Crocodylians. Syst. Biol. 52, 403422.
Gentry, A.W., 1978. Evolution of African Mammals. Harvard University Press, Cambridge.
Gentry, A.W., 1990. Evolution and the dispersal of African Bovidae.
In: Bubenik, G.A., Bubenik, A.B. (Eds.), Horns, Proghorns and
Antlers: Evolution, Morphology, Physiology and Social SigniWcance. Springer-Verlag, New York.
Gentry, A.W., 1992. The subfamilies and tribes of the family Bovidae.
Mammal Rev. 22, 132.
Georgiadis, N.J., Kat, P.W., Oketch, H., Patton, J., 1990. Allozyme
divergence within the Bovidae. Evolution 44, 21352149.
Graur, D., Higgins, D.G., 1994. Molecular evidence for the inclusion of
cetaceans within the order Artiodactyla. Mol. Biol. Evol. 11, 357
364.
Halanych, K.M., Dembaski, J.K., Jansen van Vuuren, B., Klein,
D.R., Cook, J.A., 1999. Cytochrome b phylogeny of North
American hares and jackrabbits (Lepus, Logomorpha) and the
eVects of saturation in ingroup taxa. Mol. Phylogenet. Evol. 11,
213221.
Haltenorth, T., 1963. KlassiWkation der sugetiere: Artiodactyla I. In:
Handbuch der Zoologie 8, pp. 1167.
Hassanin, A., Douzery, E.J.P., 1999a. Evolutionary aYnities of the
enigmatic saola (Pseudoryx nghetinhensis) in the context of the
molecular phylogeny of Bovida. Proc. R. Soc. Lond. B 266, 893
900.
Hassanin, A., Douzery, E.J.P., 1999b. The tribal radiation of the
family Bovidae (Artiodactyla) and the evolution of the mitochondrial cytochrome b gene. Mol. Phylogenet. Evol. 13, 227
243.
Hassanin, A., Douzery, E.J.P., 2003. Molecular and morphological
phylogenies of ruminantia and the alternative position of Moschidae. Syst. Biol. 52, 206228.

S. Willows-Munro et al. / Molecular Phylogenetics and Evolution 35 (2005) 624636

Hill, A., Drake, R., Tauxe, L., Monaghan, M., Barry, J.C., Behrensma,
A.K., Curtis, G., Jacobs, B.F., Jacobs, L., Johnson, N., 1985. Neogene paleontology and geochronology of Baringo Basin, Kenya. J.
Hum. Evol. 14, 759773.
Hillis, D.M., 1999. SINEs of the perfect character. Proc. Natl. Acad.
Sci. USA 96, 99799981.
Hipp, A.L., Hall, J.C., Sytsma, K.J., 2004. Congruence versus phylogenetic accuracy: revisiting the incongruence length diVerence test.
Syst. Biol. 53, 8189.
Huelsenbeck, J.P., Ronquest, F., 2001. MrBayes: Bayesian inference of
phylogeny. Bioinformatics 17, 754755.
Irwin, D.M., Arnason, U., 1994. Cytochrome b gene of marine mammals: phylogeny and evolution. J. Mammal. Evol. 2, 3755.
Irwin, D.M., Kocher, T.D., Wilson, A.C., 1991. Evolution of the cytochrome b gene of mammals. J. Mol. Evol. 32, 128144.
Kingdon, J., 1982. East African Mammals: An Atlas of Evolution in
Africa, Vol. IIIC. Academic Press, London.
Kingdon, J., 1997. The Kingdon Field Guide to African Mammals.
Academic Press, London.
Kluge, A.G., 1989. A concern for evidence and a phylogenetic hypothesis of relationships among Epicrates (Boidae, Serpentes). Syst. Zool.
38, 725.
Leach, A.D., Reeder, T.W., 2002. Molecular systematics of the
eastern fence lizard (Sceloporus undulates): A comparison of
parsimony, likelihood, and bayesian approaches. Syst. Biol. 51,
4468.
Lee, M.S.Y., 2000. Tree robustness and clade signiWcance. Syst. Biol. 49,
829836.
Lindsay, J.A., 1998. Past climates of southern Africa. In: Bridgeman,
H.A., Hobbs, J.E., Lindsay, J.A. (Eds.), Climates of the Southern
Continents: Present, Past and Future. Wiley, England.
Luckett, W.P., Hong, H., 1998. Phylogenetic relationships between the
orders Artiodactyla and Cetacea: a combined assessment of morphological and molecular evidence. J. Mammal. Evol. 5, 127182.
Maddison, W.P., Maddison, D.R., 1992. MacClade: Analysis of phylogeny and character evolution, version 4.0. Sinauer, Sunderland, MA.
Madsen, O., Scally, M., Douady, C.J., Kao, D.J., DeBry, R.W., Adkins,
R., Amrine, H.M., Stanhope, M.J., de Jong, W.W., Springer, M.S.,
2001. Parallel adaptive radiations in two major clades of placental
mammals. Nature 409, 610614.
Matthee, C.A., Robinson, T.J., 1999. Cytochrome b phylogeny of family Bovidae: resolution within the Alcephini, Antilopini, Neotragini
and Tragelaphini. Mol. Phylogenet. Evol. 12, 3146.
Matthee, C.A., BurzlaV, J.D., Taylor, J.F., Davis, S.K., 2001. Mining the
mammalian genome for artiodactyl systematics. Syst. Biol. 50, 367
390.
Matthee, C.A., Davis, S.K., 2001. Molecular insights into the evolution
of the family Bovidae: a nuclear DNA perspective. Mol. Biol. Evol.
18, 12201230.
Matthee, C.A., Jansen van Vuuren, B., Bell, D., Robinson, T.J., 2004. A
molecular supermatrix of the rabbits and hares (Leporidae) allows
for the identiWcations of Wve intercontinental exchanges during the
Miocene. Syst. Biol. 53, 433447.
McKenna, M.C., 1975. Toward a phylogenetic classiWcation of the
Mammalia. In: Luckett, W.P., Szalay, F.S. (Eds.), Phylogeny of the
Primates. Plenum, New York.
Miyamoto, M.M., Cracraft, J., 1991. Phylogenetic inference, DNA
sequence analysis and the future of molecular systematics. In:
Miyamoto, M.M., Cracraft, J. (Eds.), Phylogenetic Analysis of
DNA Sequences. Oxford University Press, New York.
Miyamoto, M.M., Fitch, W.M., 1995. Testing species phylogeneties and
phylogenetic methods with congruence. Syst. Biol. 44, 6476.
Montgelard, C., CatzeXis, F.M., Douzery, E., 1997. Phylogenetic relationships of artiodactyla and cetaceans as deduced from the comparison of cytochrome b and 12S rRNA mitochondrial sequences.
Mol. Biol. Evol. 14, 550559.

635

Moore, W.S., 1995. Inferring phylogenies from mtDNA variation:

mitochondrial-gene trees versus nuclear-gene trees. Evolution 49,
718726.
Murphy, W.J., Eizirik, E., Johnson, W.E., Zhang, Y.P., Ryder, O.A.,
OBrien, S.J., 2001. Molecular phylogenetics and the origins of placental mammals. Nature 409, 614618.
Novacek, M.J., 1982. Information for molecular studies from anatomical and fossil evidence on higher eutherian phylogeny. In: Goodman, M. (Ed.), Macromolecular Sequences in Systematic and
Evolutionary Biology. Plenum, New York.
Nowak, R.M., 1999. Walkers Mammals of the World, XI ed. The
Johns Hopkins University Press, Baltimore and London.
Nylander, J.A., Ronquist, F., Huelsenbeck, J.P., Nieves-Aldrey, J.L.,
2004. Bayesian phylogenetic analysis of combined data. Syst. Biol.
53, 4767.
OLeary, M.A., Geisler, J.H., 1999. The position of Cetecae within
Mammalia: phylogenetic analysis of morphological data from
extinct and extant taxa. Syst. Biol. 48, 455490.
Posada, D., Crandall, K.A., 1998. MODELTEST version 3.06. Brigham
Young University.
Reyes, A., Gissi, C., CatzeXis, F., Nevo, E., Pesole, G., Saccone, C., 2004.
Congruent mammalian trees from mitochondrial and nuclear genes
using Bayesian methods. Mol. Biol. Evol. 21, 397403.
Roberts, A., 1952. The Mammals of South Africa. Trustees of The
Mammals of South Africa book fund, Johannesburg.
Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning: A
Laboratory Manual, second ed. Cold Spring Harbour Laboratory
Press, New York.
Sanderson, M.J., Donoghue, M.J., 1996. The relationship between
homoplasy and conWdence in a phylogenetic tree. In: Sanderson,
M.J., HuVord, L. (Eds.), Homoplasy: The Recurrence of Similarity
in Evolution. Academic Press, San Diego.
Shimodaira, H., Hasegawa, M., 1999. Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Mol. Biol.
Evol. 16, 11141116.
Simpson, G.G., 1945. The principles of classiWcation and a classiWcation of mammals. Bull. Am. Mus. Nat. Hist. 85, 1350.
Sorenson, M.D., 1999. TreeRot. Version 2. Boston University: Boston,
MA, USA.
Springer, M.S., DeBry, R.W., Douady, C., Amrine, H.M., Madsen, O.,
de Jong, W.W., Stanhope, M.J., 2001. Mitochondrial versus nuclear
gene sequences in deep-level mammalian phylogeny reconstruction.
Mol. Biol. Evol. 18, 132143.
Suzuki, Y., Glazko, G.V., Nei, M., 2002. Overcredibility of molecular
phylogenies obtained by bayesian phylogenetics. Proc. Natl. Acad.
Sci. USA 99, 1613816143.
SwoVord, D.L., 1991. When are phylogenies estimates from molecular
and morphological data congruent. In: Miyamoto, M.M., Cracraft,
J. (Eds.), Phylogenetic Analysis of DNA Sequences. Oxford University Press, New York.
SwoVord, D.L., 2003. PAUP*: Phylogenetic Analysis Using Parsimony* (and other methods). Version 4.0. Sunderland, MA: Sinauer.
Templeton, A.R., 1983. Phylogenetic inference from restriction site
endonuclease cleavage site maps with particular reference to the
humans and apes. Evolution 37, 221244.
Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins,
D.G., 1997. The Clustal X windows interface: Xexible strategies for
multiples sequence alignment aided by quality analysis tools.
Nucleic Acids Res. 25, 48764882.
Thorne, J.L., Kishino, H., 2002. Divergence time and evolutionary rate
estimation with multilocus data. Syst. Biol. 51, 689702.
Ursing, B.M., Arnason, U., 1998. Analyses of mitochondrial genomes
strongly support a hippopotamus-whale clade. Proc. R. Soc. Lond.
B 265, 22512255.
Van Valen, L., 1968. Monophyly or diphyly in the origin of whales.
Evolution 23, 3741.

636

S. Willows-Munro et al. / Molecular Phylogenetics and Evolution 35 (2005) 624636

Vrba, E.S., 1985. African Bovidae: evolutionary events since the Miocene. S. Afr. J. Sci. 81, 263266.
Waddell, P.J., Kishino, H., Ota, R., 2001. A phylogenetic foundation
for comparative mammalian genomics. Genome Informat. 12, 141
154.
Walker, E.P., 1964. Mammals of the World. Johns Hopkins Press, Baltimore.
Wilcox, T.P., Zwickl, D.J., Heath, T.A., Hillis, D.M., 2002. Phylogenetic
relationships of the dwarf boas and a comparison of Bayesian and

bootstrap measures of phylogenetic support. Mol. Phylogenet.

Evol. 25, 361371.
Yang, Z., Yoder, A.D., 2003. Comparison of likelihood and Bayesian
methods for estimating divergence times using multiple gene loci
and calibration points, with application to a radiation of cute-looking mouse lemur species. Syst. Biol. 52, 705716.
Zachos, J., Pagani, M., Sloan, L., Thomas, E., Billups, K., 2001. Trends,
rhythms, and aberrations in global climate 65 Ma to present. Science 292, 686693.