ArticlePDF Available

Molecular phylogeny of African hinge-back tortoises (Kinixys): Implications for phylogeography and taxonomy (Testudines: Testudinidae)

Authors:

Abstract

We examine the phylogeography, phylogeny and taxonomy of hinge-back tortoises using a comprehensive sampling of all currently recognized Kinixys species and subspecies and sequence data of three mitochondrial DNA fragments (2273 bp: 12S rRNA, ND4 + adjacent DNA coding for tRNAs, cytb) and three nuclear loci (2569 bp: C-mos, ODC, R35). Combined and individual analyses of the two data sets using Bayesian and Maximum Likelihood methods suggest that the savannah species of Kinixys are paraphyletic with respect to the rainforest species K. homeana and K. erosa, and that the rainforest species may be derived from a savannah-living ancestor. The previously recognized savannah species K. belliana was a conglomerate of three deeply divergent clades that we treat here as distinct species. We restrict the name K. belliana (Gray, 1830) to hinge-back tortoises ranging from Angola to Burundi, while five-clawed hinge-back tortoises from the northernmost part of the formerly recognized range of K. belliana, together with four-clawed tortoises from West Africa, are assigned to the species K. nogueyi (Lataste, 1886). These two species are allied to K. spekii, whereas Southeast African and Malagasy hinge-back tortoises formerly lumped together with K. belliana represent the distinct species K. zombensis Hewitt, 1931, which is sister to K. lobatsiana. The latter two species together constitute the sister group of the rainforest species K. homeana and K. erosa. Mitochondrial data suggest that K. natalensis has a basal phylogenetic position in a clade embracing K. belliana sensu stricto, K. nogueyi and K. spekii, while nuclear data and the two data sets combined favour a sister group relationship of K. natalensis to all other hinge-back tortoises. Phylogeographic structure is present in all wide-ranging species and correlates in K. homeana and K. erosa with the Dahomey Gap and former rainforest refugia. The Malagasy population of K. zombensis is weakly differentiated from its South African conspecifics and further sampling is needed to determine whether there is support for the subspecific distinctness of Malagasy tortoises.
1
Museum of Zoology (Museum fu
¨r Tierkunde), Senckenberg Dresden, Dresden, Germany;
2
Department of Herpetology, Port
Elizabeth Museum, Humewood, South Africa;
3
Department of Zoology, Nelson Mandela Metropolitan University, Port Elizabeth,
South Africa;
4
Chelonian Biodiversity and Conservation, Department of Biodiversity and Conservation Biology, University of the
Western Cape, Bellville, South Africa;
5
LÕAssociation du Refuge des Tortues, Bessie
`res, France;
6
Department of Biology and
Wildlife Diseases, Faculty of Veterinary Hygiene and Ecology, University of Veterinary and Pharmaceutical Sciences, Brno, Czech
Republic;
7
Department of Evolutionary Biology, Zoological Institute, Technical University of Braunschweig, Braunschweig,
Germany;
8
Pietermaritzburg, KwaZulu-Natal, South Africa;
9
Kidogo Safaris, Windhoek, Namibia
Molecular phylogeny of African hinge-back tortoises (Kinixys): implications
for phylogeography and taxonomy (Testudines: Testudinidae)
Carolin Kindler
1
,William R. Branch
2,3
,Margaretha D. Hofmeyr
4
,Je
´ro
ˆme Maran
5
,Pavel S
ˇiroky
´
6
,Miguel Vences
7
,
James Harvey
8
,J.Susanne Hauswaldt
7
,Alfred Schleicher
9
,Heiko Stuckas
1
and Uwe Fritz
1
Abstract
We examine the phylogeography, phylogeny and taxonomy of hinge-back tortoises using a comprehensive sampling of all currently recognized
Kinixys species and subspecies and sequence data of three mitochondrial DNA fragments (2273 bp: 12S rRNA, ND4 + adjacent DNA coding
for tRNAs, cytb) and three nuclear loci (2569 bp: C-mos, ODC, R35). Combined and individual analyses of the two data sets using Bayesian and
Maximum Likelihood methods suggest that the savannah species of Kinixys are paraphyletic with respect to the rainforest species K. homeana
and K. erosa, and that the rainforest species may be derived from a savannah-living ancestor. The previously recognized savannah species
K. belliana was a conglomerate of three deeply divergent clades that we treat here as distinct species. We restrict the name K. belliana (Gray, 1830)
to hinge-back tortoises ranging from Angola to Burundi, while five-clawed hinge-back tortoises from the northernmost part of the formerly
recognized range of K. belliana, together with four-clawed tortoises from West Africa, are assigned to the species K. nogueyi (Lataste, 1886).
These two species are allied to K. spekii, whereas Southeast African and Malagasy hinge-back tortoises formerly lumped together with K. belliana
represent the distinct species K. zombensis Hewitt, 1931, which is sister to K. lobatsiana. The latter two species together constitute the sister group
of the rainforest species K. homeana and K. erosa. Mitochondrial data suggest that K. natalensis has a basal phylogenetic position in a clade
embracing K. belliana sensu stricto, K. nogueyi and K. spekii, while nuclear data and the two data sets combined favour a sister group relationship
of K. natalensis to all other hinge-back tortoises. Phylogeographic structure is present in all wide-ranging species and correlates in K. homeana
and K. erosa with the Dahomey Gap and former rainforest refugia. The Malagasy population of K. zombensis is weakly differentiated from its
South African conspecifics and further sampling is needed to determine whether there is support for the subspecific distinctness of Malagasy
tortoises.
Key words: Africa – Madagascar – numt – revision – species delineation – subspecies
Introduction
Hinge-back tortoises (genus Kinixys Bell, 1827) are the only
extant chelonians with a movable hinge in the carapace,
allowing a more or less complete closure of the rear part of
their shell. This hinge is positioned between the fourth and
fifth costal and the seventh and eighth peripheral bones.
Kinixys species are confined to sub-Saharan Africa and
Madagascar and are true land tortoises occurring in open or
forested habitats (Figs S1 and S2). Their carapacial lengths
range from 15 to 32 cm (Loveridge and Williams 1957; Ernst
and Barbour 1989; Iverson 1992; Ernst et al. 2000; Branch
2008). While species and subspecies delineation have fre-
quently changed over the past 50 years (Loveridge and
Williams 1957; Wermuth and Mertens 1961, 1977; Broadley
1981, 1989, 1992, 1993; Ernst and Barbour 1989; Iverson 1992;
Ernst et al. 2000; Fritz and Havas
ˇ2007; Iverson et al. 2007;
Branch 2008; Rhodin et al. 2010), all authors have agreed
explicitly or implicitly that there are two distinct species
groups, one comprising rainforest species and the other
savannah species. These species groups are thought to repre-
sent reciprocally monophyletic clades (Iverson 1992; Iverson
et al. 2007). Most recent studies and check-lists recognized two
rainforest species and four savannah species (but see McCord
et al. 2005 and Rhodin et al. 2008, who treated K. belliana
nogueyi as a full species), even though subspecific structuring
within K. belliana has remained controversial (Table 1). One
of the contentious issues is the validity of the Malagasy
subspecies K. b. domerguei. This taxon is restricted to a small
region in extreme northwestern Madagascar and was described
by Vuillemin (1972) as representing a new genus and new
species (Madakinixys domerguei). However, its conspecifity
with K. belliana was soon recognized (Wermuth and Mertens
1977; Obst 1978). Although it is generally assumed that
K. belliana was introduced to Madagascar (Wermuth and
Mertens 1961, 1977; Bour 1978, 1987, 2006; Raselimanana and
Vences 2003; Fritz and Havas
ˇ2007; Rhodin et al. 2010), some
authors have continued to treat the Malagasy population as a
distinct subspecies (Bour 1987, 2006; Rhodin et al. 2010),
leaving open the possibility that it originated from an older,
natural, colonisation event.
Corresponding author: Uwe Fritz (uwe.fritz@senckenberg.de)
Contributing authors: Carolin Kindler (carolin.kindler@senckenberg.
de), William R. Branch (wrbranch@bayworld.co.za), Margaretha
D. Hofmeyr (mdhofmeyr@uwc.ac.za), Je
´roˆ me Maran (jerome.
maran@wanadoo.fr), Pavel S
ˇiroky´ (sirokyp@vfu.cz), Miguel Vences
(m.vences@tu-bs.de), James Harvey (james_harvey@telkomsa.net),
J. Susanne Hauswaldt (s.hauswaldt@tu-bs.de), Alfred Schleicher
(kidogo@iway.na), Heiko Stuckas (heiko.stuckas@senckenberg.de)
2012 Blackwell Verlag GmbH
Accepted on 28 March 2012
J Zool Syst Evol Res doi: 10.1111/j.1439-0469.2012.00660.x
J Zool Syst Evol Res (2012) 50(3), 192–201
Until now, the taxonomy of Kinixys species was assessed
only by morphological means, although a few publications
have used mitochondrial and nuclear DNA sequences of single
Kinixys species when addressing more general questions.
Le et al. (2006) and Fritz and Bininda-Emonds (2007) used
sequences of three mitochondrial and two nuclear genes of
K. erosa and K. homeana for unravelling the large-scale
phylogeny of land tortoises (Testudinidae), and Sasaki et al.
(2006) applied SINE loci of the same species as outgroups in a
phylogenetic study of freshwater turtles (Geoemydidae). In a
pilot study on barcoding of turtle and tortoise species, Reid
et al. (2011) used COI sequences labelled as K. homeana and
K. natalensis.
Our present paper is the first molecular investigation based
on a comprehensive sampling of all Kinixys species and
subspecies. Here we follow tentatively the taxonomy of
Rhodin et al. (2010) in that we treat the Malagasy and
Southeast African populations of K. belliana as the distinct
subspecies K. b. domerguei and K. b. zombensis, respectively.
We use sequence data of three mitochondrial genes and three
nuclear loci to assess the phylogeny and phylogeography of
hinge-back tortoises. By doing so, we re-examine the validity
and relationships of the studied taxa, and in particular whether
the savannah and rainforest species are reciprocally mono-
phyletic and whether Malagasy hinge-back tortoises are
genetically distinct from their continental African conspecifics.
Materials and Methods
Sampling and gene selection
Eighty-six tissue and blood samples of hinge-back tortoises were
studied, representing all currently recognized taxa including Kinixys
belliana domerguei and K. b. zombensis (Table S1). Three mitochon-
drial genes were sequenced that have previously been shown to unravel
phylogeny and differentiation of terminal chelonian taxa (e.g. Le et al.
2006; Fritz and Bininda-Emonds 2007; Vargas-Ramı
´rez et al. 2010a,b;
Wiens et al. 2010; Fritz et al. 2011), viz. the partial 12S ribosomal
RNA (12S rRNA) gene, the partial NADH dehydrogenase subunit 4
(ND4) gene, and the cytochrome b(cytb) gene. The DNA sequence
containing the partial ND4 gene embraced also the flanking DNA
coding for tRNA-His, tRNA-Ser and tRNA-Leu. The DNA sequence
containing the cytbgene included also approximately 20 bp of the
adjacent DNA coding for tRNA-Thr. In addition to these mtDNA
data, up to three nuclear loci were generated for 35 samples being
representative for phylogenetic clades revealed by mitochondrial
sequences. The nuclear genomic loci were the partial genes coding
for the oocyte maturation factor Mos (C-mos) and for ornithine
decarboxylase (ODC), and the intron 1 of the RNA fingerprint protein
35 (R35) gene. These three loci are increasingly applied for phyloge-
netic investigations of turtles and tortoises (e.g. Georges et al. 1998;
Fujita et al. 2004; Le et al. 2006; Vargas-Ramı
´rez et al. 2010b; Wiens
et al. 2010; Fritz et al. 2011). While the complete mitochondrial data
set could be generated for all but one sample, of which only two genes
were available, only two nuclear loci of some specimens could be
sequenced due to bad DNA quality or small sample size. Remaining
samples and DNA are stored at )80C in the tissue collection of the
Museum of Zoology, Dresden.
Laboratory procedures
Total genomic DNA was extracted using either the DTAB method
(Gustincich et al. 1991), the innuPREP DNA Mini Kit, the innuPREP
Blood DNA Mini Kit (both Analytik Jena AG, Jena, Germany) or the
NucleoSpin 8 Tissue Core Kit (Macherey Nagel, Du
¨ren, Germany).
The partial 12S rRNA gene was PCR-amplified using the universal
primers L1091 and H1478; for the DNA fragment comprising the
partial ND4 gene plus flanking DNA coding for tRNAs, the primers
ND4 672 and H-Leu were used. The cytbgene was initially amplified in
two fragments overlapping by approximately 300 bp using the primer
pairs CytbG + mt-E-Rev2 and mt-c2 + mt-f-na. The primers for the
ND4 and cytbgenes are standard primers for turtles and tortoises. Gel
electrophoretic examination of the PCR products of the 12S rRNA
gene showed, besides the target amplicon of approximately 400 bp
length, the presence of a shorter band of approximately 200 bp length,
which is why a preparative gel electrophoresis was applied (see below).
While amplification and sequencing of the ND4 gene worked perfectly,
the cytbprimers yielded in several cases mismatching sequences
suggestive of numts, as known to occur in testudinids (Fritz et al.
2010). Therefore, the original approach was abandoned and the cytb
gene was amplified in one fragment by combining the primers CytbG
and mt-f-na. Since this primer combination produced also regularly the
putative numt sequence, the forward primer CytbG was replaced by
the specifically designed primer Kinixys_cytb_for. This primer did not
work in a few samples of Kinixys erosa; then, the forward primer mt-a-
neu was used instead. The authenticity of obtained mtDNA sequences
was verified as in Fritz et al. (2010). Identification of numts obtained
from the primer combination CytbG and mt-f-na was unambiguous
because these sequences contained frameshift mutations and stop
codons; the numts will be discussed in more detail elsewhere. For
amplification of the nuclear genes, the following primers were used:
Cmos1 and Cmos3 for the C-mos gene, the chicken primers of Friesen
et al. (1999) for ODC, and the primers R35Ex1 and R35Ex2 for the
intron 1 of the R35 gene. For all primer sequences and their original
references, see Table S2.
PCR was carried out in a total volume of 25 ll containing 0.2 llTaq
polymerase (5 u ll; Bioron, Ludwigshafen, Germany), 1·buffer as
recommended by the supplier, 0.4 lM of each primer, and 0.2 mM of
each dNTP (Fermentas, St. Leon-Rot, Germany). Alternatively, for
challenging samples a total volume of 20 ll containing 0.2 ll GoTaq
Flexi DNA Polymerase (5 u ll; Promega, Madison, WI, USA) was
used according to the recommendations by the supplier. PCR
protocols are summarized in Table S3. PCR products were purified
using the ExoSAP-IT enzymatic cleanup (USB Europe GmbH,
Staufen, Germany) or, for the partial 12S rRNA gene, a preparative
gel electrophoresis and the peqGOLD Gel Extraction Kit (peqlab,
Erlangen, Germany). Purified PCR products were then sequenced on
an ABI 3130xl Genetic Analyzer (Applied Biosystems, Foster City,
CA, USA) using the BigDye Terminator v3.1 Cycle Sequencing Kit
(Applied Biosystems). Cycle sequencing reactions were purified by
ethanol sodium acetate precipitation or by using Sephadex (GE
Healthcare, Mu
¨nchen, Germany). For sequencing the 12S rRNA and
ND4 genes, the same primers as for PCR were applied. For the cytb
gene, the internal primers mt-c-For2 and mt-E-Rev2 were used; the
Table 1. Competing schemes for the taxonomy of Kinixys taxa. The
rainforest species group embraces K. erosa and K. homeana, whereas
all other species are placed in the savannah species group
Broadley (1992, 1993),
Ernst et al. (2000),
Fritz and Havas
ˇ(2007),
Branch (2008)
Rhodin et al.
(2010)
Kinixys belliana
belliana (Gray, 1830)
Valid Valid
Kinixys belliana nogueyi
(Lataste, 1886)
Valid Valid
Kinixys belliana domerguei
(Vuillemin, 1972)
Synonym of
K. b. belliana
Valid
Kinixys belliana zombensis
Hewitt, 1931
Synonym of
K. b. belliana
Valid
Kinixys erosa (Schweigger,
1812)
Valid Valid
Kinixys homeana Bell, 1827 Valid Valid
Kinixys lobatsiana Power, 1927 Valid Valid
Kinixys natalensis Hewitt,
1935
Valid Valid
Kinixys spekii Gray, 1863 Valid Valid
Molecular phylogeny of hinge-back tortoises 193
J Zool Syst Evol Res (2012) 50(3), 192–201
2012 Blackwell Verlag GmbH
forward primer Kinixys_cytb_for was applied for completing the 5¢-
end of some samples. For sequencing the nuclear genes, the same
primers as for PCR were used. However, for the ODC and R35 of
challenging samples the newly designed primers Kinixys_ODC_
Seq_F + Kinixys_ODC_Seq_R and Kinixys_R35_Seq_F + Kinixys_
R35_Seq_R were used (Table S2).
Alignment, partitioning and phylogenetic analyses
All sequences were aligned and inspected using BioEdit 7.0.5.2 (Hall
1999) and mega 4.0.2 (Tamura et al. 2007); uncorrected pdistances of
the cytbgene were calculated with the latter software. Homologous
GenBank sequences of Manouria emys,Stigmochelys pardalis and
Testudo graeca were downloaded as outgroups and included in the
alignments. Asian tortoises of the genus Manouria represent together
with the species of the North American genus Gopherus a highly
distinct clade being sister to all other living testudinids, while
S. pardalis and T. graeca are two species representing clades more
closely allied to Kinixys (Le et al. 2006; Fritz and Bininda-Emonds
2007). Since no ODC sequences were available for all three outgroups,
these were generated for the present study as described above.
GenBank accession numbers of newly generated sequences and
GenBank sequences of outgroups are listed in Table S1.
Phylogenetic relationships were inferred for three data sets: (1) the
concatenated mitochondrial genes of 86 samples, corresponding to a
total alignment of 2273 bp; (2) the concatenated nuclear DNA
sequences of 35 samples, corresponding to a total alignment of
2569 bp; and (3) a supermatrix of 4842 bp length, in which the
mitochondrial data of the same samples were added to their nuclear
sequences (for individual gene partitions, see Table S4).
Trees were calculated using the Maximum Likelihood (ML)
approach as implemented in RAxML 7.0.3 (Stamatakis 2006); the
partition scheme of Table S4 and the GTR+G model across all
partitions were applied. Five independent ML searches were
conducted using different starting conditions and the fast bootstrap
algorithm to examine the robustness of the branching patterns by
comparing the best-scored trees. Subsequently, 1000 non-parametric
thorough bootstrap replicates were calculated and plotted against the
tree with the highest likelihood value. In addition, phylogenetic
relationships were inferred by using MrBayes 3.1.2 (Ronquist and
Huelsenbeck 2003) and the implemented Metropolis-coupled Markov
chain Monte Carlo algorithm with two parallel runs, each with one
cold and three heated chains. The chains run for 10
7
generations,
with every 100th generation sampled. The best-fit evolutionary
models were determined for each data set and individual genes using
the Akaike information criterion of modeltest 2.3 (Posada and
Crandall 1998). However, using partitioned alignments the two runs
did not converge onto a stationary distribution and the estimated
evolutionary distances were by far too large. This problem is well
known for Bayesian inference (Brown et al. 2010; Marshall 2010).
To overcome this problem, analyses were rerun using unpartitioned
alignments, with the GTR+I+G model for the mitochondrial data
set and the supermatrix, and the HKY+I+G model for the nuclear
data. Then, convergence was obtained as evinced by average
standard deviations of split frequencies approaching zero. For
generating the final 50% majority rule consensus tree, a burn-in of
4·10
4
was used to sample only the most likely trees. Manouria
emys was used for tree rooting.
To examine the phylogenetic information of individual genes,
exploratory RAxML analyses were also run using each mtDNA and
nDNA fragment and the same conditions as above.
Results
Phylogeny
The trees obtained from the two methods to infer phylogeny
yielded largely congruent results. Differences concerned only
the branching patterns of few weakly supported nodes,
especially in the poorly resolved trees based on the nuclear
data set (Fig. S3).
Based on morphological evidence, previous authors distin-
guished between two species groups. One species group was
thought to comprise the two rainforest species Kinixys
homeana and K. erosa, while the second group should embrace
all the remaining Kinixys species living mainly in savannah
habitats. Accordingly, it has been hypothesized that these two
species groups are reciprocally monophyletic (Iverson 1992;
Iverson et al. 2007). In the trees based on mtDNA sequences
(Fig. 1), the two rainforest species K. homeana and K. erosa
constituted a well-supported clade. Each of the two species
corresponded to a clade with maximum support. Within either
species, geographic structuring was present, with three weakly
to well-supported clades within K. homeana and five clades
within K. erosa. Contrary to expectation, however, the savan-
nah species were not monophyletic and sister to the two
rainforest species. Rather, the savannah species were paraphy-
letic with respect to K. homeana and K. erosa. Moreover,
sequences assigned to K. belliana occurred in three deeply
divergent, well-supported clades.
All Malagasy tortoises (K. b. domerguei) shared the same
haplotype. This clade was with high support sister to South
African K. b. zombensis; their successive sister taxon was with
maximum support K. lobatsiana (Fig. 1). These three savan-
nah taxa together constituted with high support the sister
group of the two rainforest species K. homeana and K. erosa.
The remaining samples of savannah taxa corresponded to a
well-supported clade being sister to the mixed savannah-
rainforest clade.
This major clade of savannah taxa only embraced four well-
supported minor clades. One of these, corresponding to the
two sequences of K. natalensis, was sister to a more inclusive
clade embracing three well-supported subordinated clades,
each of which showed clear geographic substructuring. The
basal branching pattern of these three subordinated clades
remained poorly resolved, which is why the 50% majority rule
consensus of the Bayesian analysis placed them in a polytomy
(Fig. 1). One of these three clades contained sequences of
K. b. nogueyi and of three K. b. belliana from the Central
African Republic; therein, sequences of K. b. nogueyi from
Ghana and Senegal clustered in a maximally supported
subclade with one of the sequences of K. b. belliana. Two
other K. b. belliana from the Central African Republic were
placed with maximum support in another subclade, together
with two sequences of K. b. nogueyi from Cameroon. The
second of the three clades contained sequences of K. b. belliana
from Angola and Burundi (plus a sequence of a pet tortoise of
unknown provenance), and the third clade was formed by
sequences of K. spekii. Sequences of K. spekii from Namibia,
South Africa, Zambia and Zimbabwe were only weakly
differentiated, whereas a sequence from the Democratic
Republic of the Congo was clearly distinct.
Compared to the mitochondrial trees, the trees based on the
nuclear data set were only poorly resolved and many branch-
ing patterns were only weakly supported (Fig. S3). However,
the three nuclear loci also suggested a paraphyly of the
savannah taxa with respect to the rainforest species. Here, all
savannah taxa except K. natalensis were placed together with
the two rainforest species K. homeana and K. erosa in a well-
supported major clade, which was sister to K. natalensis.
Sequences of the two rainforest species were not clearly distinct
and appeared, with weak support, to be the sister group of all
savannah taxa except K. natalensis. Apart from the fact that
K. homeana and K. erosa and K. b. domerguei and K. b. zombensis,
194 Kindler, Branch, Hofmeyr, Maran, S
ˇiroky
´, Vences, Harvey, Hauswaldt, Schleicher, Stuckas and FRITZ
J Zool Syst Evol Res (2012) 50(3), 192–201
2012 Blackwell Verlag GmbH
7474 homeana Ghana
5498 zombensis South Africa
1317 spekii
K8 domerguei Madagascar
K5 domerguei Madagascar
K10 domerguei Madagascar
4252 erosa Congo-Brazzaville
K14 domerguei Madagascar
5155 erosa Cameroon
6320 erosa Congo-Brazzaville
6335 erosa Congo-Brazzaville
7457 spekii South Africa
6323 erosa Congo-Brazzaville
7454 spekii South Africa
6344 erosa Congo-Brazzaville
5223 spekii Democratic Republic of the Congo
2090 homeana
K13 domerguei Madagascar
6365 homeana Cameroon
203 homeana
6369 erosa Cameroon
K6 domerguei Madagascar
6324 erosa Congo-Brazzaville
5152 erosa Cameroon
4251 erosa Congo-Brazzaville
2092 homeana
7455 natalensis South Africa
5154 homeana Cameroon
7462 spekii South Africa
K4 domerguei Madagascar
K1 domerguei Madagascar
6326 erosa Congo-Brazzaville
7475 lobatsiana South Africa
4246 belliana Central African Republic
7471 lobatsiana South Africa
7470 lobatsiana South Africa
6367 homeana Cameroon
7468 erosa Ghana
K3 domerguei Madagascar
K18 domerguei Madagascar
6327 erosa Congo-Brazzaville
7272 belliana Angola
7034 spekii South Africa
7467 lobatsiana South Africa
7458 spekii South Africa
7476 erosa Ghana
K11 domerguei Madagascar
7472 nogueyi Ghana
6381 nogueyi Cameroon
ZCSH514 domerguei Madagascar
5610 erosa Congo-Brazzaville
7469 nogueyi Ghana
7506 spekii Namibia
6358 erosa Cameroon
6366 homeana Cameroon
7466 homeana Ghana
624 belliana
7507 spekii Namibia
5063 domerguei Madagascar
5184 nogueyi Cameroon
910 belliana Burundi
7473 erosa Ghana
5222 nogueyi Senegal
6947 belliana Angola
2093 spekii
7032 spekii Zimbabwe
485 nogueyi
6364 homeana Cameroon
K7 domerguei Madagascar
K9 domerguei Madagascar
7031 spekii Zambia
K12 domerguei Madagascar
6370 erosa Cameroon
4248 belliana Central African Republic
6368 homeana Cameroon
5677 belliana Angola
204 homeana
861 homeana
7453 spekii South Africa
4244 belliana Central African Republic
6371 erosa Cameroon
6359 erosa Cameroon
5499 zombensis South Africa
7460 nogueyi Senegal
K2 domerguei Madagascar
7461 natalensis South Africa
.66/81
.70/
1.0/96
1.0/93
*
1.0/96
*
1.0/96
.79/93
.94/88
.56/72
.57/70
1.0/98
1.0/97
*
.52/
*
.98/88
1.0/99
‘belliana’
‘belliana’
homeana
erosa
lobatsiana
spekii
natalensis
RAINFOREST
SAVANNAH
0.03
*
*
*
*
*
*
*
*
*
*
*
*
*
1.0/99
**
Fig. 1. Phylogeny of Kinixys species as inferred from Bayesian analysis with MrBayes 3.1.2, based on a 2273-bp-long alignment of mtDNA (12S rRNA, ND4 + tRNAs, cytb+ tRNA-Thr). Outgroups
(Manouria emys,Stigmochelys pardalis,Testudo graeca) removed for clarity; tree was rooted with M. emys. For explanation of sample codes preceding taxon names, see Table S1. Numbers along branches
indicate Bayesian posterior probabilities and thorough bootstrap values as obtained using RAxML 7.0.3. Asterisks indicate maximum support under both methods; dashes, branch not found. Support
values are not shown for some terminal clades with short branches. Samples of K. b. belliana clustering in distinct clades in boldface. On the right, current species delineation and species groups (rainforest
and savannah species groups) indicated
Molecular phylogeny of hinge-back tortoises 195
J Zool Syst Evol Res (2012) 50(3), 192–201
2012 Blackwell Verlag GmbH
respectively, were not reciprocally monophyletic, all mito-
chondrial clades were also recovered from the analyses of the
nuclear DNA sequences, albeit mostly with weak statistical
support.
The trees based on the supermatrix of concatenated mito-
chondrial and nuclear DNA sequences (Fig. 2) largely agreed
with the trees based on mtDNA only, with one exception. In
the mitochondrial trees, K. natalensis was placed in the same
clade as certain other savannah taxa (K. b. belliana from
Angola, Burundi and the Central African Republic; K. b. no-
gueyi;K. spekii), whereas the trees based on nuclear DNA
sequences suggested K. natalensis as the most basal species of
Kinixys. The nuclear signal enforced the latter topology also in
the trees based on the supermatrix, but now with decidedly
weak support for the monophyly of all of the remaining
species. In any case, each data set revealed the savannah taxa
as paraphyletic with respect to the rainforest species.
For gaining deeper insight into the conflicting placement of
K. natalensis, ML trees of individual gene partitions were
compared. With respect to the three trees based on each
mtDNA fragment, the placement of K. natalensis was consis-
tent with the combined analyses of mtDNA, despite a
generally weaker resolution in the trees based on single
partitions. The individual analyses of the three nuclear
partitions suggested that the signal responsible for the
conflicting placement of K. natalensis in the combined anal-
yses of nuclear DNA and nuclear plus mitochondrial DNA
was caused by C-mos and ODC, which delivered, with
moderate to weak bootstrap support, a sister group relation-
ship of K. natalensis and all other Kinixys species. By contrast,
an ML tree based on R35 sequences returned K. natalensis in
a poorly resolved and weakly supported polytomy together
with all other Kinixys, except K. erosa,K. homeana and
K. lobatsiana.
Uncorrected pdistances of the mitochondrial cytbgene
In several recent studies (Vargas-Ramı
´rez et al. 2010b; Prasch-
ag et al. 2011; Stuckas and Fritz 2011; Fritz et al. 2012a,b)
uncorrected pdistances of the mitochondrial cytbgene were
used as a tool for species delineation of turtles and tortoises
(see Discussion). Therefore, uncorrected pdistances of a 1058-
bp-long alignment (1035 bp cytbgene + 23 bp of the adjacent
DNA coding for tRNA-Thr) of the studied Kinixys samples
were computed using mega 4.0.2 (Tamura et al. 2007).
Pairwise divergences were calculated within and between the
major clades of Figs 1 and 2 (Table 2) and between some finer
entities corresponding to subclades (Table S5). Uncorrected p
distances between the major clades ranged on average between
3.76% (K. b. domerguei +K. b. zombensis versus K. lobatsi-
ana) and 11.59% (K. erosa versus K. natalensis); divergences
within the clades, between 0.25% (K. spekii) and 1.93%
(K. b. belliana from Angola and Burundi plus a pet tortoise
of unknown provenance). The haplotypes of the two
K. natalensis were identical (Table 2).
The divergence between K. b. domerguei and K. b. zomben-
sis was with 1.33% clearly lower than between each of these
two taxa and the closely allied K. lobatsiana (3.76% and
3.77%; Table S5). The distinct subclade embracing sequences
of K. b. nogueyi from Ghana and Senegal and a sequence of a
K. b. belliana from the Central African Republic differed from
its sister group (K. b. belliana from the Central African
Republic; K. b. nogueyi from Cameroon) by 1.47%. The
subclade consisting of a sequence of a K. b. belliana from
Burundi and another pet tortoise differed from its sister group
(K. b. belliana from Angola) by 2.57%, and the distinct
sequence of a K. spekii from the Democratic Republic of the
Congo differed from all the other K. spekii by 1.0%
(Table S5). The mean divergences among the corresponding
three more inclusive clades, viz. (1) K. b. belliana (Central
African Republic) + K. b. nogueyi, (2) K. b. belliana from
Angola and Burundi, and (3) K. spekii, were 4.08–6.76%
(Table 2).
Discussion
The results of our combined and individual analyses of
mitochondrial and nuclear DNA were largely consistent.
However, contradictory evidence was obtained for the phylo-
genetic placement of Kinixys natalensis. While this species was
suggested as the sister taxon of all other Kinixys species by
concatenated nDNA (Fig. S3) and mtDNA + nDNA
(Fig. 2), individual and combined analyses of the mitochon-
drial partitions (Fig. 1) revealed K. natalensis as the sister
taxon of a major clade embracing K. spekii and two further
subordinated clades, one corresponding to K. b. belliana from
Angola and Burundi and the other to K. b. belliana from the
Central African Republic plus K. b. nogueyi. Individual anal-
yses of the nuclear partitions suggested that the signal
responsible for the conflicting position of K. natalensis was
caused by C-mos and ODC, but not by R35.
This situation is indicative of incongruent phylogenetic
information in the individual partitions, which was long
considered to be problematic. However, the paradigm that
only DNA partitions with congruent phylogenetic signal
should be combined has been challenged years ago (Gatesy
et al. 1999) and recently been refuted. By combining incon-
gruent individual data sets, even previously undetected phy-
logenetic signal may be uncovered (Struck 2007; Zanol et al.
2010) – signal which is not revealed by single-gene analyses.
Consequently, congruence among data sets is no longer a
prerequisite for combining partitions (Cunningham 1997;
Gatesy et al. 1999; Zanol et al. 2010). In essence, the conflict-
ing branching patterns of Kinixys reflect the expected and well-
known differences between gene trees and species trees (e.g.
Maddison 1997; Rosenberg 2002; Degnan and Rosenberg
2009), which are exacerbated by the different inheritance
modes of mitochondrial and nuclear DNA (Ballard and
Whitlock 2004).
Based on morphological criteria, two rainforest species
(K. erosa,K. homeana) and four savannah species of hinge-
back tortoises (K. belliana,K. lobatsiana,K. natalensis,K. spe-
kii) are currently recognized. Each of these ecological groups is
thought to represent a reciprocally monophyletic group
(Iverson 1992; Iverson et al. 2007). Only within K. belliana
are two to four subspecies distinguished (Broadley 1993; Fritz
and Havas
ˇ2007; Branch 2008; Rhodin et al. 2010; Table 1).
However, our molecular data contradict this classification, and
according to our results the two ecological groups are also not
reciprocally monophyletic. Rather, the savannah species are
paraphyletic with respect to the two rainforest species K. erosa
and K. homeana, suggesting that the rainforest species may
derive from a savannah-dwelling ancestor. Furthermore,
K. belliana is not monophyletic, but consists of three highly
distinct major clades that were independently revealed by the
mitochondrial and nuclear DNA data sets (Figs 1 and S3).
196 Kindler, Branch, Hofmeyr, Maran, S
ˇiroky
´, Vences, Harvey, Hauswaldt, Schleicher, Stuckas and FRITZ
J Zool Syst Evol Res (2012) 50(3), 192–201
2012 Blackwell Verlag GmbH
624 belliana
7272 belliana Angola
4248 ‘belliana’ Central African Republic
Manouria
6947 belliana Angola
7460 nogueyi Senegal
7454 spekii South Africa
7455 natalensis South Africa
7506 spekii Namibia
6326 erosa Congo-Brazzaville
7461 natalensis South Africa
5677 belliana Angola
4252 erosa Congo-Brazzaville
Stigmochelys
5498 zombensis South Africa
2092 homeana
5499 zombensis South Africa
5184 nogueyi Cameroon
4246 ‘belliana’ Central African Republic
5222 nogueyi Senegal
7457 spekii South Africa
7031 spekii Zambia
6364 homeana Cameroon
7475 lobatsiana South Africa
K7 domerguei Madagascar
7507 spekii Namibia
485 nogueyi
910 belliana Burundi
1317 spekii
5223 spekii Democratic Republic of the Congo
Testudo
6324 erosa Congo-Brazzaville
7032 spekii Zimbabwe
6368 homeana Cameroon
4251 erosa Congo-Brazzaville
7467 lobatsiana South Africa
2093 spekii
K11 domerguei Madagascar
.95
94
*
*
*
1.0/99
1.0/91
*
*
**
*
1.0/99
*
*
*
.53/59
.63/77
*
*1.0/98
1.0/94
.69/50
*
1.0/90
*
*
1.0/88
0.02
*
homeana
erosa
zombensis
lobatsiana
spekii
natalensis
nogueyi
belliana
Fig. 2. Phylogeny of Kinixys species as inferred from Bayesian analysis, based on the supermatrix (2273 bp of mtDNA: 12S rRNA, ND4 + tRNAs, cytb+ tRNA-Thr plus 2569 bp of nDNA: C-mos,
ODC, R35). On the right, new species delineation. For further explanations, see Fig. 1 and Discussion
Molecular phylogeny of hinge-back tortoises 197
J Zool Syst Evol Res (2012) 50(3), 192–201
2012 Blackwell Verlag GmbH
One of the three clades within what is currently identified with
K. belliana is, together with K. lobatsiana, the sister group of
the two rainforest species; the other two clades are most closely
allied to K. spekii. This situation argues for the revision of the
current species delineation of K. belliana.
Moreover, our analyses (Figs 1, 2 and S3) confirm the
evolutionary distinctiveness of K. lobatsiana,K. natalensis and
K. spekii; three species formerly lumped together with K. bel-
liana (Loveridge and Williams 1957), but whose taxonomic
disparity had been revealed by Broadley (1981, 1989, 1992,
1993). Kinixys belliana is the only species currently thought to
be polytypic. Accordingly, several parapatric subspecies are
regarded as valid (Table 1; Fig. S1), and our findings neces-
sitate examining whether some of these subspecies, or unrec-
ognized taxa within K. belliana, may represent distinct species.
The phylogenetic placement of K. belliana sequences alone
suggests that two or three distinct species are involved, one
being sister to K. lobatsiana and allied to the rainforest species,
and one or two further species being related to K. spekii. One
of the clades allied to K. spekii comprises sequences of
K. b. belliana from Angola and Burundi, and the other
sequences from the northern part of the range of K. b. belliana
(in our tree represented by samples from the Central African
Republic) and of K. b. nogueyi. The question is whether these
two clades could be conspecific with K. spekii. This possibility
can be excluded with respect to the clade comprising sequences
of K. b. belliana from Angola and Burundi, because these
tortoises occur sympatrically with K. spekii (Fig. S1). By
contrast, K. b. nogueyi is allopatric from all other savannah
taxa, and the same is true for tortoises from the northern part
of the range of K. b. belliana.
In absence of direct evidence, such as sympatric occurrence,
several recent studies have used uncorrected pdistances of the
mitochondrial cytbgene as a proxy for assessing the taxonomic
status of turtles and tortoises. However, congeneric chelonian
species differ by uncorrected pdistances of 1.5–18.3% (Vargas-
Ramı
´rez et al. 2010b; Praschag et al. 2011; Stuckas and Fritz
2011; Fritz et al. 2012a,b), and this wide range implies that a
rigid threshold is not helpful for species delineation. Rather,
the threshold value needs to be adjusted by comparison with
closely related species. Among congeneric tortoise species
(family Testudinidae), uncorrected pdistances range from
3.7% to 12.7% (Fritz et al. 2012a), and these values resemble
the ones of Kinixys (Table 2). Particularly helpful is to use the
closest related sympatric species as a yardstick for allopatric or
parapatric taxa. Among the taxa that occur sympatrically with
K. spekii (Fig. S1) is the clade corresponding to sequences of
K. b. belliana from Angola and Burundi. Consequently, there
is direct evidence that K. spekii and the hinge-back tortoise
from Angola and Burundi represent distinct species under the
strict Biological Species Concept (Mayr 1942; Coyne and Orr
2004). When the uncorrected pdistances among K. spekii and
the two allied K. belliana clades are compared (Table 2), it is
obvious that the mean divergences between K. spekii and the
clade consisting of K. b. belliana from the Central African
Republic plus K. b. nogueyi (6.76%) and between the latter
clade and the K. b. belliana from Angola and Burundi (4.86%)
exceed the value between the sympatric K. spekii and K. b. bel-
liana from Angola and Burundi (4.08%). This suggests that
each of the three clades of K. belliana represents a distinct
species. This is corroborated by the mean divergence of 3.76%
between K. b. domerguei +K. b. zombensis and their sister
species K. lobatsiana, resembling the value of 4.08% between
K. spekii and K. belliana from Angola and Burundi. By
contrast, the divergences within each of the three clades
hitherto lumped together under K. belliana are decidedly lower
(Table S5). Between K. b. domerguei and K. b. zombensis a
value of 1.33% occurs; between the two subclades of
K. b. belliana from the Central African Republic plus
K. b. nogueyi, a value of 1.47%; and between the two
subclades of K. b. belliana from Angola and Burundi, 2.57%,
so that we suggest that this level of divergence represents
phylogeographic variation within one and the same species, as
does the value of 1.0% between a K. spekii from the
Democratic Republic of the Congo and the samples of
K. spekii from Namibia, South Africa, Zambia, and Zimbabwe.
These findings do have taxonomic implications (Fig. 2).
Firstly, the name Kinixys belliana (Gray, 1830) has to be
restricted to only one species-level clade. This name is
currently identified with five-clawed hinge-back tortoises from
the western part of the speciesÕrange (Loveridge and Williams
1957; Fritz and Havas
ˇ2007; Branch 2008; Rhodin et al. 2010).
Therefore, we restrict this name to the taxon represented in our
sampling by tortoises from Angola and Burundi. Secondly,
fore-foot claw number seems to be of only limited taxonomic
use in the north-central part of the range of the K. belliana
complex. Until now, four-clawed West African hinge-back
tortoises were identified with the allopatric subspecies
K. b. nogueyi (Lataste, 1886). Our data imply that this taxon
is conspecific with five-clawed tortoises from the Central
Table 2. Mean uncorrected pdistances (percentages ± standard errors) within and between major mitochondrial clades of Kinixys using a 1058-
bp-long alignment of mtDNA (1035 bp cytbgene + 23 bp of the adjacent DNA coding for tRNA-Thr). Below the diagonal, divergences between
clades are given; on the diagonal, divergences within clades in boldface. Kinixys b. belliana (1) refers to tortoises from the Central African
Republic; K. b. belliana (2), Angola and Burundi (plus a pet tortoise of unknown provenance)
n homeana erosa
domerguei +
zombensis lobatsiana
belliana (1) +
nogueyi belliana (2) spekii natalensis
K. homeana 13 0.55 ± 0.14
K. erosa 19 7.09 ± 0.75 0.62 ± 0.13
K. b. domerguei +
K. b. zombensis 19 8.52 ± 0.84 8.70 ± 0.78 0.26 ± 0.06
K. lobatsiana 4 8.55 ± 0.78 8.78 ± 0.77 3.76 ± 0.56 0.30 ± 0.11
K. b. belliana (1) +
K. b. nogueyi 10 9.75 ± 0.82 11.35 ± 0.95 9.43 ± 0.83 10.23 ± 0.91 1.05 ± 0.20
K. b. belliana (2) 5 9.99 ± 0.81 11.20 ± 0.89 9.58 ± 0.85 10.37 ± 0.92 4.86 ± 0.55 1.93 ± 0.31
K. spekii 13 8.88 ± 0.84 10.54 ± 0.91 8.72 ± 0.87 9.48 ± 0.91 6.76 ± 0.51 4.08 ± 0.58 0.25 ± 0.06
K. natalensis 2 10.54 ± 0.90 11.59 ± 0.94 9.97 ± 0.95 10.36 ± 0.89 8.42 ± 0.81 8.36 ± 0.83 7.15 ± 0.78 0
198 Kindler, Branch, Hofmeyr, Maran, S
ˇiroky
´, Vences, Harvey, Hauswaldt, Schleicher, Stuckas and FRITZ
J Zool Syst Evol Res (2012) 50(3), 192–201
2012 Blackwell Verlag GmbH
African Republic, which are combined here with the four-
clawed West African tortoises as the distinct species K. nogueyi
(Lataste, 1886). Thirdly, the formerly recognized subspecies
K. b. domerguei (Vuillemin, 1972) and K. b. zombensis Hewitt,
1931 together represent another distinct species. Unlike
K. belliana and K. nogueyi, this species is not allied to
K. spekii, but rather to K. lobatsiana and the rainforest species
K. erosa and K. homeana. The valid name for this Malagasy
and Southeast African savannah species is K. zombensis
Hewitt, 1931 because Bour (1978) gave precedence to this
name over the simultaneously published name K. zuluensis
Hewitt, 1931 (ICZN 1999: Article 24, First Reviser Principle).
Further investigations are warranted to delineate the exact
distribution ranges of K. zombensis and in particular of
K. belliana sensu stricto and K. nogueyi.
We studied from the putative continental African range of
K. zombensis only two samples from the southernmost tip of
its distribution (Fig. S1). These samples differed in the cytb
gene by an uncorrected pdistance of 1.33% from 17 Malagasy
samples, a value that resembles the phylogeographic variation
within other Kinixys species (Tables 2 and S5). Malagasy
hinge-back tortoises are thought to represent a distinct taxon
by some authors (Bour 1987, 2006; Rhodin et al. 2010), while
others believe that Malagasy tortoises are introduced and not
taxonomically distinct (e.g. Ernst et al. 2000; Raselimanana
and Vences 2003; Fritz and Havas
ˇ2007; Branch 2008). The
low molecular divergence between Malagasy and South
African samples of K. zombensis renders the validity of the
Malagasy taxon problematic. It is well known that the fauna
of Madagascar has been severely altered since the first human
settlers arrived approximately 2300 years ago (Goodman and
Benstead 2003; Burney et al. 2004), and Raselimanana and
Vences (2003) assume that K. zombensis was introduced 1000–
1500 years ago. We believe that the mitochondrial uniformity
of our 17 Malagasy samples, together with the tiny distribu-
tion range on Madagascar, supports the hypothesis that the
species was introduced. In all other Kinixys taxa with
comparable sample sizes greater variation is evident (Fig. 1;
Tables 2 and S5), suggestive of a founder effect in the
Malagasy population, as expected for an introduced popula-
tion. We predict that the same genetic lineage will be
discovered in the more northern continental distribution range
of K. zombensis. The populations of several other chelonians
from Madagascar (Pelomedusa subrufa and Pelusios castano-
ides), the Seychelles (Pelusios castanoides and P. subniger), and
even Guadeloupe (Lesser Antilles, P. castaneus) have also
been shown to represent recent, probably anthropogenic,
colonisations (Silva et al. 2010; Vargas-Ramı
´rez et al. 2010b;
Fritz et al. 2011).
Phylogeographic structure was evident not only in each
savannah species having a wide distribution range, but also in
the two rainforest species K. homeana and K. erosa (Figs 1 and
2). In K. homeana, two weakly to well-supported mitochon-
drial clades from Ghana were discovered and another one
from Cameroon; in K. erosa, two clades from Cameroon, a
clade from Ghana, and two further clades from Congo-
Brazzaville. Pronounced phylogeographic structure was also
reported in a South American tortoise species occupying
savannah habitats (Chelonoidis carbonaria), whereas the South
American rainforest species C. denticulata is phylogeograph-
ically weakly structured, despite an extensive range (Vargas-
Ramı
´rez et al. 2010a). This difference could be related to
distinct histories of the rainforests in South America and
Africa. For South America, the formerly widely accepted
Pleistocene forest refugia hypothesis has been challenged in
recent years. By contrast, it is still widely accepted that African
rainforests repeatedly contracted during the Pleistocene (Ham-
ilton and Taylor 1991; Maley 1996; Kastner and Gon
˜i 2003;
Pennington et al. 2004; Primack and Corlett 2005; Penner
et al. 2011). The distinct clades of K. homeana and K. erosa
correlate with the Dahomey Gap, a 300-km-wide aisle sepa-
rating until today the West African and Central African
rainforests (Primack and Corlett 2005), and with the location
of forest refugia during the last major arid phase approxi-
mately 18 000 years ago (compare our Fig. S2 with Fig. 5 of
Maley 1996). This suggests that, as with Gabon vipers (Bitis
gabonica and B. rhinoceros; Lenk et al. 2001) and frogs
(Penner et al. 2011), vicariance has shaped the observed extant
structure.
Acknowledgements
Thanks for help with laboratory work go to Anja Rauh. Mario
Vargas-Ramı
´rez and Thomas Datzmann assisted with some phyloge-
netic calculations. Raffael Ernst suggested literature about African
rainforests. Annegret Kindler produced the maps. Graham Alexander
and Olda Mudra provided some photos. Jens Junggebauer and Olda
Mudra assisted with sampling. William R. Branch and Margaretha D.
Hofmeyr thank the National Research Foundation (South Africa) for
support through the Incentive Funding Programme.
References
Ballard JWO, Whitlock MC (2004) The incomplete natural history of
mitochondria. Mol Ecol 13:729–744.
Bour R (1978) Les tortues actuelles de Madagascar (Re
´publique
malgache): liste syste
´matique et description de deux sous-espe
`ces
nouvelles. Bull Soc Et sci Anjou Nouv Se
´r10:141–154.
Bour R (1987) Les tortues terrestres et dÕeau douce de Madagascar et
des ıˆles voisines. II. Cle de de
´termination des tortues malgaches
actuelles. Bull Assoc Avanc Rech Scient Madagascar Re
´g malg
20:1–11.
Bour R (2006) Kinixys belliana domerguei (Vuillemin, 1972), la tortue a
`
dos articule
´de Madagascar: donne
´es historiques et e
´cologiques.
Che
´loniens 3:42–55.
Branch B (2008) Tortoises, Terrapins and Turtles of Africa. Struik,
Cape Town, South Africa.
Broadley DG (1981) A review of the populations of Kinixys (Testu-
dinidae) occurring in south-eastern Africa. Ann Cape Prov Mus
(Nat Hist) 13:195–216.
Broadley DG (1989) Kinixys belliana, BellÕs Hinged Tortoise;
‘‘Kinixys spekii’’, SpekeÕs Hinged Tortoise; ‘‘Kinixys lobatsiana’’,
Lobatse Hinged Tortoise; Kinixys erosa, Serrated Hinged
Tortoise; Kinixys homeana, HomeÕs Hinged Tortoise; Kinixys
natalensis, Natal Hinged Tortoise. In: Swingland IR, Klemens
MW (eds), The Conservation Biology of Tortoises. IUCN, Gland,
Switzerland, pp 49–61.
Broadley DG (1992) The savannah species of Kinixys (Testudinidae). J
Herpetol Assoc Afr 40:12–13.
Broadley DG (1993) A review of the Southern African species of
Kinixys Bell (Reptilia: Testudinidae). Ann Transvaal Mus 36:
41–52.
Brown JM, Hedtke SM, Lemmon AR, Lemmon EM (2010) When
trees grow too long: investigating the causes of highly inaccurate
Bayesian branch-length estimates. Syst Biol 59:145–161.
Burney DA, Burney LP, Godfrey LR, Jungers WL, Goodman SM,
Wright HT, Jull AJT (2004) A chronology for late prehistoric
Madagascar. J Hum Evol 47:25–63.
Coyne JA, Orr HA (2004) Speciation. Sinauer Associates, Sunderland,
MA.
Cunningham CW (1997) Can three incongruence tests predict when
data should be combined? Mol Biol Evol 14:733–740.
Molecular phylogeny of hinge-back tortoises 199
J Zool Syst Evol Res (2012) 50(3), 192–201
2012 Blackwell Verlag GmbH
Degnan JH, Rosenberg NA (2009) Gene tree discordance, phyloge-
netic inference and the multispecies coalescent. Trends Ecol Evol
24:332–340.
Ernst CH, Barbour RW (1989) Turtles of the World. Smithsonian
Institution Press, Washington, DC.
Ernst CH, Altenburg RGM, Barbour RW (2000) Turtles of the World.
World Biodiversity Database, CD-ROM Series, Windows, Version
1.2. Biodiversity Center of ETI, Amsterdam, The Netherlands.
Friesen VL, Congdon BC, Kidd MG, Birt TP (1999) Polymerase chain
reaction (PCR) primers for the amplification of five nuclear introns
in vertebrates. Mol Ecol 8:2147–2149.
Fritz U, Bininda-Emonds ORP (2007) When genes meet nomenclature:
tortoise phylogeny and the shifting generic concepts of Testudo and
Geochelone. Zoology 110:298–307.
Fritz U, Havas
ˇP (2007) Checklist of chelonians of the world. Vertebr
Zool 57:149–368.
Fritz U, Daniels SR, Hofmeyr MD, Gonza
´lez J, Barrio-Amoro
´s CL,
S
ˇiroky´ P, Hundsdo
¨rfer AK, Stuckas H (2010) Mitochondrial
phylogeography and subspecies of the wide-ranging sub-Saharan
leopard tortoise Stigmochelys pardalis (Testudines: Testudinidae) – a
case study for the pitfalls of pseudogenes and GenBank sequences.
J Zool Syst Evol Res 48:348–359.
Fritz U, Branch WR, Hofmeyr MD, Maran J, Prokop H, Schleicher
A, S
ˇiroky´ P, Stuckas H, Vargas-Ramı
´rez M, Vences M, Hundsdo
¨r-
fer AK (2011) Molecular phylogeny of African hinged and helmeted
terrapins (Testudines: Pelomedusidae: Pelusios and Pelomedusa).
Zool Scr 40:115–125.
Fritz U, Alcalde L, Vargas-Ramı
´rez M, Goode EV, Fabius-Turoblin
DU, Praschag P (2012a) Northern genetic richness and southern
purity, but just one species in the Chelonoidis chilensis complex. Zool
Scr 41:220–232.
Fritz U, Stuckas H, Vargas-Ramı
´rez M, Hundsdo
¨rfer AK, Maran J,
Pa
¨ckert M (2012b) Molecular phylogeny of Central and South
American slider turtles: implications for biogeography and system-
atics (Testudines: Emydidae: Trachemys). J Zool Syst Evol Res
50:125–136.
Fujita MK, Engstrom TN, Starkey DE, Shaffer HB (2004) Turtle
phylogeny: insights from a novel nuclear intron. Mol Phylogenet
Evol 31:1031–1040.
Gatesy J, OÕGrady P, Baker RH (1999) Corroboration among data
sets in simultaneous analysis: hidden support for phylogenetic
relationships among higher level artiodactyl taxa. Cladistics 15:271–
313.
Georges A, Birrel J, Saint KM, McCord WP, Donnellan SC (1998) A
phylogeny for side-necked turtles (Chelonia: Pleurodira) based on
mitochondrial and nuclear gene sequence variation. Biol J Linn Soc
67:213–246.
Goodman SM, Benstead JP (eds) (2003) The Natural History of
Madagascar. University of Chicago Press, Chicago, IL.
Gustincich S, Manfioletti G, del Sal G, Schneider C, Carninci C (1991)
A fast method for high-quality genomic DNA extraction from whole
human blood. BioTechn 11:298–302.
Hall TA (1999) bioedit: a user-friendly biological sequence alignment
editor and analysis program for Windows 95 98 NT. Nucl Acids
Symp Ser 41:95–98.
Hamilton AC, Taylor D (1991) History of climate and forests in tropical
Africa during the last 8 million years. Climatic Change 19:65–78.
ICZN [International Commission on Zoological Nomenclature] (1999)
International Code of Zoological Nomenclature, 4th edn. Interna-
tional Trust for Zoological Nomenclature, London.
Iverson JB (1992) A Revised Checklist with Distribution Maps of the
Turtles of the World. Privately Printed, Richmond, IN.
Iverson JB, Brown RM, Akre TS, Near TJ, Le M, Thomson RC,
Starkey DE (2007) In search of the tree of life for turtles. Chelon Res
Monogr 4:85–106.
Kastner TP, Gon
˜i MA (2003) Constancy in the vegetation of the
Amazon Basin during the late Pleistocene: evidence from the organic
matter composition of Amazon deep sea fan sediments. Geology
31:291–294.
Le M, Raxworthy CJ, McCord WP, Mertz L (2006) A molecular
phylogeny of tortoises (Testudines: Testudinidae) based on mito-
chondrial and nuclear genes. Mol Phylogenet Evol 40:517–531.
Lenk P, Kalyabina S, Wink M, Joger U (2001) Evolutionary relationships
among the true vipers (Reptilia: Viperidae) inferred from mitochon-
drial DNA sequences. Mol Phylogenet Evol 19:94–104.
Loveridge A, Williams EE (1957) Revision of the African tortoises and
turtles of the suborder Cryptodira. Bull Mus Comp Zool 115:163–
557.
Maddison WP (1997) Gene trees in species trees. Syst Biol 46:523–
536.
Maley J (1996) The African rain forest – main characteristics of
changes in vegetation and climate from the Upper Cretaceous to the
Quaternary. Proc R Soc Edinb Biol 104B:31–73.
Marshall DC (2010) Cryptic failure of partitioned Bayesian phylo-
genetic analyses: lost in the land of long trees. Syst Biol 59:108–
117.
Mayr E (1942) Systematics and the Origin of Species from the
Viewpoint of a Zoologist. Columbia University Press, New York,
NY.
McCord WP, Joseph-Ouni M, Tabaka C (2005) Chelonian illustra-
tions #18: African hinge-back tortoises. Reptilia (GB) 38:71–74.
Obst FJ (1978) Beitra
¨ge zur Kenntnis der Testudiniden Madagaskars.
Zool Abhandl Staatl Mus Tierkd Dresden 35:31–54.
Penner J, Wegmann M, Hillers A, Schmidt M, Ro
¨del M-O (2011) A
hotspot revisited – a biogeographical analysis of West African
amphibians. Divers Distrib 17:1077–1088.
Pennington RT, Lavin M, Prado DE, Pendry CA, Pell SK, Butter-
worth CA (2004) Historical climate change and speciation: neo-
tropical seasonally dry forest plants show patterns of both Tertiary
and Quaternary diversification. Philos Trans R Soc Lond B Biol Sci
359:515–538.
Posada D, Crandall KA (1998) modeltest: testing the model of DNA
substitution. Bioinfomatics 14:817–818.
Praschag P, Stuckas H, Pa
¨ckert M, Maran J, Fritz U (2011)
Mitochondrial DNA sequences suggest a revised taxonomy of
Asian flapshell turtles (Lissemys Smith, 1931) and the validity of
previously unrecognized taxa (Testudines: Trionychidae). Vertebr
Zool 61:147–160.
Primack RB, Corlett R (2005) Tropical Rain Forests: An Ecological
and Biogeographical Comparison. John Wiley & Sons, Hoboken,
NJ.
Raselimanana AP, Vences M (2003) Introduced reptiles and amphibians.
In: Goodman SM, Benstead JP (eds), The Natural History of
Madagascar. The University of Chicago Press, Chicago, IL, pp 949–
951.
Reid BN, Le M, McCord WP, Iverson JB, Georges A, Bergmann T,
Amato G, Desalle R, Naro-Maciel E (2011) Comparing and
combining distance-based and character-based approaches for
barcoding turtles. Mol Ecol Res 11:956–967.
Rhodin AGJ, van Dijk PP, Parham JF (2008) Turtles of the world:
annotated checklist of taxonomy and synonymy. Chelon Res
Monogr 5:000.1–000.38.
Rhodin AGJ, van Dijk PP, Iverson JB, Shaffer HB (2010) Turtles of
the world, 2010 update: annotated checklist of taxonomy, synon-
ymy, distribution, and conservation status. Chelon Res Monogr
5:000.85–000.164.
Ronquist F, Huelsenbeck JP (2003) MrBayes 3: Bayesian phyloge-
netic inference under mixed models. Bioinformatics 19:1572–1574.
Rosenberg NA (2002) The probablity of topological concordance of
gene trees and species trees. Theor Popul Biol 61:225–247.
Sasaki T, Yasukawa Y, Takahashi K, Miura S, Shedlock AM, Okada
N (2006) Extensive morphological convergence and rapid radiation
in the evolutionary history of the family Geoemydidae (Old World
pond turtles) revealed by SINE insertion analysis. Syst Biol 55:912–
927.
Silva A, Rocha S, Gerlach J, Rocamora G, Dufrenne A, Harris DJ
(2010) Assessment of mtDNA genetic diversity within the terrapins
Pelusios subniger and Pelusios castanoides across the Seychelles
islands. Amphibia-Reptilia 32:583–588.
Stamatakis A (2006) RAxML-VI-HPC: maximum likelihood-based
phylogenetic analyses with thousands of taxa and mixed models.
Bioinformatics 22:2688–2690.
Struck TH (2007) Data congruence, paedomorphosis and salaman-
ders. Front Zool 4:22.
200 Kindler, Branch, Hofmeyr, Maran, S
ˇiroky
´, Vences, Harvey, Hauswaldt, Schleicher, Stuckas and FRITZ
J Zool Syst Evol Res (2012) 50(3), 192–201
2012 Blackwell Verlag GmbH
Stuckas H, Fritz U (2011) Identity of Pelodiscus sinensis revealed by
DNA sequences of an approximately 180-year-old type specimen
and a taxonomic reappraisal of Pelodiscus species (Testudines:
Trionychidae). J Zool Syst Evol Res 49:335–339.
Tamura K, Dudley J, Nei M, Kumar S (2007) mega 4: molecular
evolutionary genetics analysis (mega) software version 4.0. Mol Biol
Evol 24:1596–1599.
Vargas-Ramı
´rez M, Maran J, Fritz U (2010a) Red- and yellow-footed
tortoises (Chelonoidis carbonaria,C. denticulata) in South American
savannahs and forests: do their phylogeographies reflect distinct
habitats? Org Divers Evol 10:161–172.
Vargas-Ramı
´rez M, Vences M, Branch WR, Daniels SR, Glaw F,
Hofmeyr MD, Kuchling G, Maran J, Papenfuss TJ, S
ˇiroky´ P,
Vieites DR, Fritz U (2010b) Deep genealogical lineages in the widely
distributed African helmeted terrapin: evidence from mitochondrial
and nuclear DNA (Testudines: Pelomedusidae: Pelomedusa subrufa).
Mol Phylogenet Evol 56:428–440.
Vuillemin S (1972) Note sur Madakinixys domerguei n. gen., n.
sp. (Testudinidae). Ann Univ Madagascar Sci Nat Math 9:170–
181.
Wermuth H, Mertens R (1961) Schildkro
¨ten, Krokodile, Bru
¨ckenech-
sen. VEB Gustav Fischer, German Democratic Republic.
Wermuth H, Mertens R (1977) Testudines, Crocodylia, Rhyncho-
cephalia. Das Tierreich 100:I–XXVII + 1–174.
Wiens JJ, Kuczynski CA, Stephens PR (2010) Discordant mitochon-
drial and nuclear gene phylogenies in emydid turtles: implications
for speciation and conservation. Biol J Linn Soc 99:445–461.
Zanol J, Halanych KM, Struck TH, Fauchald K (2010) Phylogeny of
the bristle worm family Eunicidae (Eunicida, Annelida) and the
phylogenetic utility of noncongruent 16S, COI and 18S in combined
analyses. Mol Phylogenet Evol 55:660–676.
Supporting Information
Additional Supporting Information may be found in the online
version of this article:
Figure S1. Distribution ranges of the savannah taxa of
Kinixys and sampling sites.
Figure S2. Distribution ranges of the rainforest species of
Kinixys and sampling sites.
Figure S3. Phylogeny of Kinixys species as inferred from
Bayesian analysis, based on a 2569-bp-long alignment of
nDNA (C-mos, ODC, R35).
Table S1. Kinixys samples and outgroups used in the present
study.
Table S2. Primers used for PCR and sequencing.
Table S3. PCR protocols for mitochondrial and nuclear
genes.
Table S4. Partitions in the mitochondrial and nuclear
alignments used for phylogenetic inference.
Table S5. Mean uncorrected pdistances (percent-
ages ± standard errors) within and between major mitochon-
drial clades and subclades using a 1058-bp-long alignment
(1035 bp cytbgene + 23 bp of the adjacent DNA coding for
tRNA-Thr) of Kinixys sequences.
Please note: Wiley-Blackwell are not responsible for the
content or functionality of any supporting materials supplied
by the authors. Any queries (other than missing material)
should be directed to the corresponding author for the article.
Molecular phylogeny of hinge-back tortoises 201
J Zool Syst Evol Res (2012) 50(3), 192–201
2012 Blackwell Verlag GmbH
... It is home to 14 species of tortoises in six different genera, with strong indications that the current taxonomy, based on morphology, underestimates diversity (Hofmeyr, Vamberger, Branch, Schleicher, & Daniels, 2017). Using mitochondrial and nuclear DNA (mtDNA and nDNA), Kindler et al. (2012) distinguished eight sub-Saharan Kinixys species in contrast to six previously recognized by morphology. The importance of molecular studies to reveal chelonian diversity is exemplified by the freshwater turtle Pelomedusa, formerly considered monotypic (P. ...
... We used MEGA v.7 to calculate the uncorrected p-distances for the Cytb gene among clades, within each clade, as well as the p-distances between outgroups. We used Cytb as the reference gene to calculate p-distances, as it has been widely used in testudine studies (Cunningham, 2002;Fritz et al., 2012;Kindler et ...
... Phylogenetic studies of chelonians (tortoises and turtles) often rely on mtDNA loci because nDNA markers frequently provide incongruent and poorly resolved trees with poor support (e.g., Kindler et al., 2012;Petzold et al., 2014). Caccone et al. (2004) found that nDNA divergence was nearly 30 times slower than that of mtDNA for Galápagos tortoises. ...
Article
Full-text available
The high level of phenotypic diversity in southern African tent tortoises (Psammobates tentorius complex) has for decades prevented systematists from developing a stable taxonomy for the group. Here, we used a comprehensive DNA sequence dataset (mtDNA: Cytb, ND4, ND4 adjacent tRNA-His, and tRNA-Ser, 12S, 16S; and nDNA: PRLR gene) of 455 specimens, and the latest phylogenetic and species delimitation analytical procedures, to unravel the long-standing P. tentorius complex systematic puzzle. Our results for mtDNA and nDNA were incongruent, with the poorly supported nDNA phylogeny differentiating the three recognized subspecies, and showing potential hybridization in some regions. In contrast, the concatenated mtDNA phylogeny identified seven operational taxonomic units, with strong support. Clades 1, 4, 5, and 7 corresponded to tortoises identified as P. t. tentorius, clade 3 to P. t. tri-meni, and clades 2 and 6 to P. t. verroxii. Our analyses showed conflicting topolo-gies for the placement of C6 (P. t. verroxii north of the Orange River), with stronger support for it being sister to C2 + C3 than to the other clades. Clades 1, 2, and 6 had significantly higher genetic diversity than clades 3, 4, 5, and 7, perhaps because these clades inhabit substantially larger areas. The potential for future cladogenic radiations seems high in C1 and C6, particularly in C6 for which the within-clade diversification level was highest. Further research involving microsatellite DNA, phy-logeographic evaluations, and morphological variation among clades is crucial for understanding the adaptive radiation of the P. tentorius complex and for modifying their taxonomy.
... Interspecific mean shape differences in adults (>100 mm SCL; Table 1) were tested with NP-MANOVA, as described above, and shape was graphically compared with thin-plate splines. Adult shape variation was initially explored via a phylogenetic PCA on a pruned topology from Kindler et al. (2012) with branch lengths adjusted according to the age of the oldest known Kinixys fossil (ca., 19-20 Ma;Lapparent de Broin, 2000;Meylan & Auffenberg, 1986). Using the physignal function, the phylogenetic signal was determined to be weak (Blomberg's K = 0.059; p = .849). ...
Article
Full-text available
Changes in the structural association of skeletal traits are crucial to the evolution of novel forms and functions. In vertebrates, such rearrangements often occur gradually and may precede or coincide with the functional activation of skeletal traits. To illustrate this process, we examined the ontogeny of African hinge-back tortoises (Kinixys spp.). Kinixys species feature a moveable "hinge" on the dorsal shell (carapace) that enables shell closure (kinesis) when the hind limbs are withdrawn. This hinge, however, is absent in juveniles. Herein, we describe how this unusual phenotype arises via alterations in the tissue configuration and shape of the carapace. The ontogenetic repatterning of osseous and keratinous tissue coincided with shifts in morphological integration and the establishment of anterior (static) and posterior (kinetic) carapacial modules. Based on ex vivo skeletal movement and raw anatomy, we propose that Kinixys employs a "sliding hinge" shell-closing system that overcomes thoracic rigidity and enhances the protective capacity of the carapace. Universal properties of the vertebrate skeleton, such as plasticity, modularity, and secondary maturation processes, contributed to adaptive evolutionary change in Kinixys. We discuss a hypothetical model to explain the delayed emergence of skeletal traits and its relevance to the origins of novel form-to-function relationships.
... In the cyt b gene, a frequently used marker for estimating interspecific divergence in turtles (e.g., Kindler et al. 2012;Iverson et al. 2013), P. shipian sp. nov. ...
Article
Full-text available
We describe a new species from the Pelodiscus axenaria complex from Hunan and Jiangxi Provinces, China. Also, the application of the name P. axenaria (Zhou, Zhang & Fang, 1991) is clarified by designating a neotype for this species. Besides its genetic divergence, the new species differs from all other Pelodiscus species, including the two other taxa constituting the complex (P. axenaria, P. huangshanensis), in the following combination of morphological traits: (1) small adult size, <15 cm carapace length; (2) carapace distinctly keeled, more or less strongly tuberculated, usually olive clay-coloured and adorned with greenish black marbling; (3) plastron yellowish white, typically immaculate except for a blurred-edged blotch behind each axilla that does not extend to the entoplastron and some slight black suffusion along its anterior border; (4) underside of the leathery margin of the carapace with varying amounts of dark pigmentation; (5) head olive clay-coloured with numerous black splotches; (6) chin grey brown with pale stipples, throat dark grey, finely spotted with black; (7) neck with a wide yellow lateral band stretching from the tympanum posteriorly, which tends to fade with age; (8) entoplastron boomerang-shaped, the amount of bending of the transverse bar between the two posteriolaterally directed rami >90°.
... A good example among tortoises would be the genus Kinixys (Bell, 1827). Six morpho-species previous recognized were shown to contain cryptic species and that the actual number of Kinixys species was eight (Kindler et al., 2012). Chersobius signatus (Gmelin, 1789) was initially believed to comprise two subspecies, Chersobius signatus signatus (Iverson, 1992) and Chersobius Signatus cafer (Bour, 1980), based on morphology, but mtDNA and nDNA sequence data showed that the two "subspecies" were phylogenetically invalid (Daniels et al., 2010). ...
Article
Full-text available
Morphology paper We examined morphometric differentiation in the highly polymorphic and taxonomi- cally confusing Psammobates tentorius (tent tortoise) species complex in southern Africa, using morphological data together with previously published DNA sequence and microsatellite DNA data. We used the approximate Bayesian computation-based simulation analysis to test alternative phylogenetic hypotheses supported by differ- ent markers, using a total evidence dataset for the tent tortoise species complex. The results supported a scenario, which is better and more inclusive in explaining its ge- nealogical history. Our total evidence-based Bayesian BPP species delimitation analy- sis (mtDNA + nDNA + morphology) suggested a seven species scheme should be applied to the P. tentorius species complex. Nonetheless, based on comparative results obtained from sequence, microsatellite, ecological niche, and morphological data, a four-species scheme (among the seven mtDNA clades) appears to be the best taxo- nomic solution for the systematic puzzle of the P. tentorius species complex, namely, “C1 + C4 + C5 + C7,” “C3,” “C2,” and “C6.” We detected several possible hybridiza- tions between C1 and C2 in their intergradation zone that confirmed the possible cross-over between them. Lastly, we provide a summarized flow chart of the analyses carried out for determining the optimal taxonomic solution for this phenotypically dif- ficult group, as it may also be useful for resolving complex taxonomic issues in other species groups. KEYWORDS alternative phylogenetic hypothesis, approximate Bayesian computation analysis, hidden diversity, morphology, species delimitation
... When the uncorrected p distances of the three subclades within the Orinoco lineage (ORN-1, ORN-2, ORN-3) are compared to the three subclades of the Amazon lineage (EB, AUM, MJ), values ranging from 7.6 to 9.8% are obtained (Table S5). These distances clearly exceed the levels of genetic divergence between many congeneric chelonian species (e.g., Fritz et al., 2012;Kindler et al., 2012Kindler et al., , 2016Petzold et al., 2014; see also the reviews in Vargas-Ramírez et al., 2010, Michels and Vargas-Ramírez, 2018, and Torstrom et al., 2014, confirming the species status of each lineage. Yet, the conflicting mitochondrial and nuclear results for the matamatas from the Branco and Essequibo Rivers suggest past gene flow and old mitochondrial introgression from the Amazon lineage into the Orinoco lineage. ...
Article
• Illegal wildlife trade is a major global threat to biodiversity. Environmental authorities around the world face many challenges enforcing and implementing laws intended to protect wild species. Wildlife forensic protocols can help to solve many of these implementation and enforcement issues. • Matamata turtles comprise two different species that look identical (Chelus fimbriata, Chelus orinocensis), which are geographically separated into the Amazon and Orinoco river basins, respectively, in South America. The trade in matamata turtles (Chelus sp.) is illegal in Colombia and law enforcement officers need to identify confiscated individuals to the species level in order to return them to the correct river basin. This is a time-sensitive problem because increasing time in captivity can lead to mortality and imposes logistical and financial burdens on resource managers. • This article presents the development of a reliable, potentially field-based, fast (~2 h), and cost-effective (~USD$1 per sample identified) real-time polymerase chain reaction assay to identify live matamata turtles to the species level, which facilitates accurate and fast translocation back to the correct river-of-origin. • This assay was applied in a real law enforcement scenario where confiscated matamata turtles were quickly identified as C. orinocensis, showing that they should be returned to the Orinoco basin. This research highlights the practicality and usefulness of these kinds of protocols to solve wildlife trade issues faced by law enforcement officers, and provides a blueprint for the development and implementation of field-based molecular protocols to enhance the enforcement and implementation of laws intended to protect wildlife subject to illegal trade. • Given the widespread detection of illegal trade in wildlife and the threat that it poses to biodiversity, it is critical that these types of protocols and collaborative efforts among all stakeholders continue to be developed and implemented.
... Geochelona is a recently named clade of testudinids (Vlachos and Rabi 2018) that corresponds to the Geochelone complex of a previous phylogeny (Parham et al. 2006). In Afro-Arabia, geochelonans represent the most diverse clade of testudinids in their fossil record as well as extant chelonian fauna, and are, in part, comprised of forms with peculiar morphologies, such as members of the genus Kinixys, which are the only known extant turtles with a movable hinge on their carapace that enables an almost full closure of the posterior portion of the shell (Kindler et al. 2012). ...
Article
Turtles of the clade Testudinoidea have a rather scarce fossil record in Afro-Arabia, ranging from the late Eocene up to the Quaternary. The vast majority of testudinoid fossils from Afro-Arabia are ascribed to Testudinidae, which has had a continuous presence in the area since the late Eocene. Geoemydidae is poorly documented by fragments found throughout the Neogene across northern Africa and the Middle East. Emydidae is absent from the fossil record of this area. All valid named taxa pertain to testudinids. Within Testudinidae, the majority of known fossil species are members of the clade Geochelona, while a few others belong to the clade Testudona. Four fossil taxa are members of now-extinct genera, five are members of extant genera, and seven cannot be assigned to a known genus with certainty. The fossil record also documents that several extant genera had a much broader distribution during the Neogene and Quaternary. Endemic insular lineages were formerly present on the Canary Islands, Cape Verde islands, and on several islands in the Western Indian Ocean. The highest known diversity of testudinoids seems to have existed during the Neogene; however, definitive conclusions are hampered by the extremely poor Paleogene record and large, unsampled areas of Afro-Arabia. A taxonomic review of the 22 named Afro-Arabian taxa finds 16 nomina valida, 1 nomen invalidum, and 5 nomina dubia.
... With respect to the cyt b gene, the average uncorrected p distance between the two clades (2.6%) falls into the range observed within other tortoise species (0.9%-3.4%), while between distinct congeneric tortoise species average values of 3.7%-12.7% are observed (see the reviews in Fritz et al., 2012 andKindler et al., 2012). The two clades of C. angulata occupy different, largely mutually exclusive climatic niches ( Figure 5). ...
Article
Full-text available
Using range-wide sampling and 1,143 bp of mtDNA (cytochrome b gene) and 14 microsatellite loci, we examined genetic differentiation in the widely distributed Southern African angulate tortoise (Chersina angulata). We found evidence for two genealogical lineages that differ in both genetic marker systems and their preferred habitat conditions. According to a fossil-calibrated molecular clock for all African tortoise lineages using 1,870 bp mitochondrial and 1,416 bp nuclear DNA, the two lineages of C. angulata diverged in the Pliocene (approx. 3.8 million years ago). Species distribution models reveal that the ranges of the two lineages shifted little since the Last Glacial Maximum, which is in agreement with the demographic population descriptors suggestive of stationary populations that did not experience expansion. One lineage occurs in the west, and the other in the south of the extant distribution range. In the geographic contact zone, the two lineages hybridize extensively, providing evidence for their conspecificity under the biological species concept. Each lineage could be recognized as a distinct subspecies, but the ill-defined geographic origins of the type material of the available names prevent their identification with any taxon. With respect to the nuclear genomic markers, the western lineage shows further north-south substructuring. A few genetically mismatched tortoises are interpreted as human translocations. Our study underlines that confiscated or captive angulate tortoises of unknown geographic provenance should not be released without prior genetic screening to avoid genetic pollution of wild populations.
... When the uncorrected p distances of the three subclades within the Orinoco lineage (ORN-1, ORN-2, ORN-3) are compared to the three subclades of the Amazon lineage (EB, AUM, MJ), values ranging from 7.6 to 9.8% are obtained (Table S5). These distances clearly exceed the levels of genetic divergence between many congeneric chelonian species (e.g., Fritz et al., 2012;Kindler et al., 2012Kindler et al., , 2016Petzold et al., 2014; see also the reviews in Vargas-Ramírez et al., 2010, Michels and Vargas-Ramírez, 2018, and Torstrom et al., 2014, confirming the species status of each lineage. Yet, the conflicting mitochondrial and nuclear results for the matamatas from the Branco and Essequibo Rivers suggest past gene flow and old mitochondrial introgression from the Amazon lineage into the Orinoco lineage. ...
Article
The matamata is one of the most charismatic turtles on earth, widely distributed in northern South America. Debates have occurred over whether or not there should be two subspecies or species recognized due to its geographic variation in morphology. Even though the matamata is universally known, its natural history, conservation status and biogeography are largely unexplored. In this study we examined the phylogeographic differentiation of the matamata based on three mitochondrial DNA fragments (2168 bp of the control region, cytochrome oxidase subunit I, and the cytochrome b gene), one nuclear genomic DNA fragment (1068 bp of the R35 intron) and 1661 Single Nucleotide Polymorphisms (SNPs). Our molecular and morphological analyses revealed the existence of two distinct, genetically deeply divergent evolutionary lineages of matamatas that separated in the late Miocene (approximately 12.7 million years ago), corresponding well to the time when the Orinoco Basin was established. As a result of our analyses, we describe the genetically and morphologically highly distinct matamata from the Orinoco and Río Negro Basins and the Essequibo drainage as a species new to science (Chelus orinocensis sp. nov.). Chelus fimbriata sensu stricto is distributed in the Amazon Basin and the Mahury drainage. Additionally, the analyses revealed that each species displays phylogeographic differentiation. For C. orinocensis, there is moderate mitochondrial differentiation between the Orinoco and the Río Negro. For C. fimbriata, there is more pronounced differentiation matching different river systems. One mitochondrial clade was identified from the Amazon, Ucayali, and Mahury Rivers, and another one from the Madeira and Jaci Paraná Rivers. The C. orinocensis in the Essequibo and Branco Rivers have haplotypes that constitute a third clade clustering with C. fimbriata. Phylogenetic analyses of the R35 intron and SNP data link the matamatas from the Essequibo and Branco with the new species, suggesting past gene flow and old mitochondrial introgression. Chelus orinocensis is collected for the pet trade in Colombia and Venezuela. However, neither the extent of the harvest nor its impact are known. Hence, it is crucial to gather more information and to assess its exploitation throughout its distribution range to obtain a better understanding of its conservation status and to design appropriate conservation and management procedures. RESUMEN La matamata es una de las tortugas más carismáticas del mundo, ampliamente distribuida en el norte de Sudamérica. Debido a su variación morfológica geográfica, se debate sobre el reconocimiento de dos subespecies o especies. A pesar de que la matamata es universalmente conocida, su historia natural, estado de conservación y biogeografía han sido muy poco estudiados. En este estudio examinamos la diferenciación filogeográfica de las matamatas en base a tres fragmentos de ADN mitocondrial (2168 pb de la región de control, la subunidad I del citocromo oxidasa y el gen del citocromo b), un fragmento de ADN genómico nuclear (1068 pb del intrón R35) y 1661 polimorfismos de nucleótido único (SNP). Nuestros análisis moleculares y morfológicos revelaron la existencia de dos linajes evolutivos distintos de matamatas, genéticamente divergentes que se separaron en el Mioceno tardio (hace aproximadamente 12.7 millones de años), correspondiendo al tiempo en que se estableció la cuenca del Orinoco. Como resultado de nuestros análisis, describimos las genéticamente y morfológicamente distintas matamatas de las cuencas del Orinoco, Río Negro y Essequibo como una especie nueva para la ciencia (Chelus orinocensis sp. nov.). Chelus fimbriata sensu stricto se distribuye en la cuenca del Amazonas y en el drenaje del Mahury. Adicionalmente, los análisis revelaron que cada especie muestra diferenciación filogeográfica. Para C. orinocensis, hay una moderada diferenciación mitocondrial entre el Orinoco y el Río Negro. Para C. fimbriata, hay una diferenciación más pronunciada, concordando con los diferentes sistemas fluviales. Se identificó un clado de los ríos Amazonas, Ucayali y Mahury y otro de los ríos Madeira y Jaci Paraná. Las C. orinocensis de los ríos Essequibo y Branco tienen haplotipos que constituyen un tercer clado que se agrupa con C. fimbriata. Los análisis filogenéticos del intrón R35 y los datos de SNP asocian las matamatas de Essequibo y Branco con la nueva especie, sugiriendo flujo de genes pasado e introgresión mitocondrial antigua. Chelus orinocensis se colecta para el comercio de mascotas en Colombia y Venezuela. Sin embargo, ni se conoce el alcance de las colectas ni su impacto. Por lo tanto, es crucial recopilar más información y evaluar su explotación en todo su rango de distribución, comprender mejor su estado de conservación y para diseñar acciones apropiadas de conservación y manejo.
... Scale counts could not be made on the specimen from Handa Farm, although the photographed snake had a very conspicuous V-shaped marking on the head, 25 well-separated dark conspicuous rhombic vertebral markings occasionally with a white contour against a paler background, and lacked dorsolateral stripes (Fig. 36). It was therefore assigned to C. rhombeatus, which is widespread in Angola (Rasmussen 2005;Marques et al. 2018 (Kindler et al. 2012). However, only northern Angolan material was included and it is very likely that K. spekii may enter southern Angola (see Marques et al. 2018), thus the identification as K. belliana is tentative. ...
Article
Full-text available
Bicuar National Park (BNP) is a protected area in southwestern Angola where biodiversity has been poorly studied. BNP is located on the Angolan plateau on Kalahari sands, in a transition zone between the Angolan Miombo Woodland and the Zambezian Baikiaea Woodland ecoregions. Herpetological surveys were conducted in BNP and surrounding areas, through visual encounter surveys, trapping, and opportunistic collecting of specimens from 2015 to 2018. The regional herpetofauna is described here based on these surveys, literature records, and additional unpublished records. In total, 16 amphibian, 15 lizard, 18 snake, two testudine, and one crocodilian species were observed from the recent surveys, and in combination with historical records the species counts are 21, 36, 32, four, and one species for these herpetofauna groups, respectively. Important observations include the first record of Xenocalamus bicolor bicolor (Günther, 1868), the second records of Sclerophrys poweri (Hewitt, 1935) and of Amblyodipsas ventrimaculata (Roux, 1907), and the fourth record of Monopeltis infuscata (Broadley, 1997) for Angola. Additionally, the type locality of Hyperolius benguellensis (Bocage, 1893) is discussed. A part of the material could not be confidently identified to species level, reflecting the taxonomic uncertainty associated with the Angolan herpetofauna. Fossorial herpetofauna was well represented, reflecting adaptation to sandy soils, the dominant substrate in the area. The likely presence of endemic and poorly known species in BNP reinforces the importance of the park for the conservation of Angolan biodiversity. Further surveys are necessary for a more comprehensive understanding of the park's fauna and biogeographic affinities, and to improve conservation planning. Resumo.-O Parque Nacional do Bicuar (BNP) é uma área protegida no sudoeste de Angola cuja biodiversidade se encontra pouco estudada. Localiza-se no planalto de Angola em areias do Calaári, numa zona de transição entre as ecorregiões de Mata de Miombo Angolana e Mata de Baikiaea Zambeziana. Neste trabalho foram realizados levantamentos de herpetofauna no BNP e arredores, através de levantamentos de encontro visual, armadilhagem e recolha oportunística de espécimes entre 2015 e 2018. Aqui é apresentada uma descrição da herpetofauna da região baseada nestes levantamentos, em registos bibliográficos, e outros registos não publicados. Os dados recentes resultaram num total de 16 espécies de anfíbios, 15 espécies de lagartos, 18 espécies de cobras, duas espécies de quelónios, e uma espécie de crocodilo. A combinação destes dados com registos históricos resulta num total de 21, 36, 32, quatro, e uma espécies destes grupos herpetológicos, respectivamente. Entre os resultados mais importantes estão o primeiro registo de Xenocalamus bicolor bicolor Günther, 1868, o segundo registo de Sclerophrys poweri (Hewitt, 1935) e de Amblyodipsas ventrimaculata (Roux, 1907), e o quarto registo de Monopeltis infuscata Broadley, 1997 para Angola. A localidade-tipo de Hyperolius benguellensis (Bocage, 1893) é também discutida. Uma parte do material não pôde ser identificado com certeza ao nível da espécie, uma consequência da incerteza taxonómica associada à herpetologia angolana. A herpetofauna fossorial está bem representada, reflectindo uma adaptação a solos arenosos, o substrato dominante na área. A presença provável de espécies endémicas e pouco conhecidas no BNP reforça a importância do parque para a conservação da biodiversidade de Angola. Mais levantamentos contribuirão para um melhor conhecimento da fauna do parque e das suas afinidades biogeográficas e para um melhor planeamento de estratégias de conservação. Palavras-chave. Anfíbios, répteis, fossorial, levantamentos de biodiversidade, áreas protegidas, areias do Calaári, província da Huíla Official journal website: amphibian-reptile-conservation.org Correspondence. *
Article
Full-text available
This paper gives a historical overview of the African rain forest from its origins, towards, the end of the Cretaceous period. The areas around the Gulf of Guinea, in particular from Ivory Coast to Nigeria and especially Cameroon, Gabon and Congo, appear to have been already occupied at this time by wet tropical forest formations mainly composed of Angiosperms. In the course of the Tertiary period the combined effect of the equator being situated further north than now and the development of the Antarctic ice cap favoured the development of wet tropical conditions over a large part of North Africa. Towards the end of the Tertiary, the equator reached its present position and the northern hemisphere ice caps appeared, and these phenomena resulted in the disappearance of the forest formations spread across the north of Africa, and the concentration of these formations near the equatorial zone around the Gulf of Guinea and in the Congo-Zaire basin. From 800 000 years ago onwards the marked glacial variations at middle and high latitudes in both hemispheres, lowered temperatures in equatorial areas and brought arid climates at times of maximum glacial extension. -Author
Article
Half-Title PageTitle PageCopyright PageTable of ContentsPreface to the first editionPreface to the second editionAcknowledgments