The evolution of armadillos, anteaters and sloths

depicted by nuclear and mitochondrial phylogenies:

implications for the status of the enigmatic fossil
Eurotamandua.
Frédéric Delsuc, François Catzeflis, Michael Stanhope, Emmanuel Douzery

To cite this version:

Frédéric Delsuc, François Catzeflis, Michael Stanhope, Emmanuel Douzery. The evolution of armadil-
los, anteaters and sloths depicted by nuclear and mitochondrial phylogenies: implications for the status
of the enigmatic fossil Eurotamandua.. Proceedings of the Royal Society B: Biological Sciences, Royal
Society, The, 2001, 268 (1476), pp.1605-15. �10.1098/rspb.2001.1702�. �halsde-00192975�

HAL Id: halsde-00192975

https://hal.archives-ouvertes.fr/halsde-00192975
Submitted on 30 Nov 2007

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
 Final version accepted for publication in The Proceedings of the Royal Society of London B [20.04.2001]
Main text = 4753 words
______________________________________________________________________

The evolution of armadillos, anteaters, and sloths depicted by nuclear and

mitochondrial phylogenies: implications for the status of the enigmatic fossil

Eurotamandua.

Frédéric DELSUC1* , François M. CATZEFLIS1, Michael J. STANHOPE2 and Emmanuel

J. P. DOUZERY1

1
Laboratoire de Paléontologie, Paléobiologie et Phylogénie, Institut des Sciences de

l’Evolution, Université Montpellier II, Montpellier, France

2
Biology and Biochemistry, Queens University, 97 Lisburn Road, Belfast BT9 7BL, UK.

Present address: Bioinformatics, SmithKline Beecham Pharmaceuticals, UP1345, 1250

South Collegeville Road, Collegeville PA 19426, USA

*
Author for correspondence ([email protected])

Running title: Evolution of Xenarthra

1
ABSTRACT

The mammalian order Xenarthra (armadillos, anteaters and sloths) is one of the four

major clades of placentals, but it remains poorly studied from the molecular phylogenetics

perspective. We here present a study encompassing most of the order’s diversity, to establish

xenarthrans intra-ordinal relationships, discuss the evolution of their morphological

characters, search for their extant sister-group, and specify the timing of their radiation with

special emphasis on the status of the controversial fossil Eurotamandua. Sequences of three

genes (nuclear exon 28 of the von Willebrand Factor [vWF]; mitochondrial 12S and 16S

rRNAs) are compared for eight of the 13 living genera. Phylogenetic analyses confirm the

order’s monophyly and that of its three major lineages: armadillos (Cingulata), anteaters

(Vermilingua), and sloths ("Tardigrada", renamed in "Folivora") and our results strongly

support the grouping of hairy xenarthrans (anteaters and sloths) into Pilosa. Within placentals,

Afrotheria might be the first lineage to branch off, followed by Xenarthra. The morphological

adaptative convergence between New World xenarthrans and Old World pangolins is

confirmed. Molecular datings place the early emergence of armadillos around the

Cretaceous/Tertiary boundary, followed by the divergence between anteaters and sloths in the

Early Eocene. These Tertiary dates contradict the concept of a very ancient origin of modern

xenarthran lineages. They also question the placement of the purported fossil anteater

(Eurotamandua) from the Middle Eocene of Europe, with the Vermilingua, and instead

suggest the independent and convergent evolution of this enigmatic taxon.

Keywords: Molecular phylogeny — Xenarthra — Pilosa — 12S rRNA — 16S rRNA — von

Willebrand Factor (vWF) — local molecular clocks — Eurotamandua.

2
INTRODUCTION

The mammalian order Xenarthra (ex Edentata; see Glass 1985) is composed of three

major lineages corresponding to three distinct living morphotypes: armadillos (Cingulata:

Dasypodidae), anteaters (Vermilingua: Myrmecophagidae) and sloths (Bradypodidae and

Megalonychidae). Xenarthrans include 30 living species classified in 13 genera, almost all

endemic of South and Central Americas, the nine-banded armadillo (Dasypus novemcinctus)

being the only one to occur in North America. Extant species provide only a glimpse of

xenarthran past diversity since 218 fossil genera are recognised (McKenna & Bell 1997). This

order was still greatly diversified no later than 10000 years ago, when the majority of genera

became extinct possibly because of human impact (Patterson & Pascual 1972; Fariña 1996).

The radiation of Xenarthra occurred between Palaeocene and Eocene at an epoch where

South America was isolated from other continental masses (Patterson & Pascual 1972). Their

fossil record is almost restricted to the American continent (Carrol 1988) with three striking

exceptions: a "primitive edentate" (Ernanodon antelios) from the Upper Palaeocene of China

(Ding 1979), a purported anteater (Eurotamandua joresi) discovered in the Middle Eocene of

Europe (Storch 1981), and an unidentified Eocene sloth from Antarctica (Vizcaíno & Scillato-

Yané 1995). These unexpected reports complicated the reconstruction of xenarthran

biogeographical history and raised doubts on their confinement to the New-World. More

specifically, the affinities of Eurotamandua with Xenarthra are still highly debated (Gaudin

1999; Rose 1999).

Despite highly specialised and distinct morphologies of the three lineages, the

monophyly of the order Xenarthra is classically recognised (McKenna & Bell 1997). All

living and fossil xenarthrans exhibit a dental reduction by the loss of enamel, reaching its

paroxysm in anteaters which totally lack teeth (Carrol 1988). The name "Edentata" has been

used to recall this evolutionary trend, but it leads to confusion because it originally referred to

3
the grouping of New-World xenarthrans and Old-World pholidotans (pangolins) which are

also lacking teeth (Glass 1985). One exclusive morphological synapomorphy of Xenarthra is

the presence of "xenarthry" constituted by additional atypical articulations between vertebrae

(Engelmann 1985; Patterson et al. 1992; Rose & Emry 1993; Gaudin 1999). From the

molecular point of view, xenarthran monophyly is strongly supported by a three consecutive

amino acids deletion in the eye lens αA-crystallin protein, that is unique among studied

eutherian mammals (de Jong et al. 1985; van Dijk et al. 1999).

The order Xenarthra is of crucial importance in understanding mammalian

phylogenetics. Following McKenna's (1975) assumption that Xenarthra represents the sister-

group to all other living eutherians, thereby named Epitheria, numerous molecular workers

have used its representatives to root eutherian phylogenies (see e.g. Allard et al. 1996).

Gaudin et al. (1996), however, emphasised the weakness of the morphological

synapomorphies defining epitherians. The comparison of complete mitochondrial genomes

suggested a close relationship between Xenarthra and either Ferungulata (Cetartiodactyla,

Perissodactyla and Carnivora: Arnason et al. 1997) or Afrotheria (Proboscidea, Hyracoidea,

Sirenia, Macroscelidea, Chrysochloridae and Tenrecidae; Waddell et al. 1999a,b). Recently,

nuclear DNA sequences showed that Xenarthra represents one of the four major clades of

placentals (Madsen et al. 2001; Murphy et al. 2001).

Within Xenarthra, the intra-ordinal relationships and their timing remain unclear.

Morphological data suggest the grouping of anteaters and sloths into Pilosa, a clade defined by

the replacement of the carapace by a coat (Engelmann 1985; Patterson et al. 1992; McKenna

& Bell 1997). However, the possibility of an early emergence of anteaters from the ancestral

xenarthran stock is suggested by studies on fossils (Carrol 1988), ear region (Guth 1961), and

cephalic arterial pattern (Bugge 1979). From the molecular perspective, the first

immunological (Sarich 1985) and protein (de Jong et al. 1985) studies, as well as the analysis

4
of partial mitochondrial 12S and 16S rRNA genes (Höss et al. 1996), did not clarify intra-

xenarthran relationships. Data from the mitochondrial ND1 favoured the grouping of

armadillos and sloths to the exclusion of anteaters (Cao et al. 1998), contra the eye lens αA-

crystallin which provided support for a monophyletic Pilosa, but suggested the paraphyly of

armadillos (van Dijk et al. 1999). Most recently, Murphy et al. (2001), in an analysis of a data

set including five xenarthran genera, present a tree with strong support for Pilosa, but with no

statistical analysis of xenarthran issues or estimates of dates for the various splitting events

within the order.

To better understand xenarthran evolution, we here present a study encompassing most

of the order’s diversity, based on three genes already used to explore mammalian

phylogenetics: the exon 28 of the von Willebrand Factor (vWF) — a single copy nuclear

marker —, and the mitochondrial 12S and 16S rRNAs (e.g. Huchon et al. 1999; Montgelard et

al. 1997; Höss et al. 1996). The aims of this study are (1) to investigate the intra-ordinal

relationships of Xenarthra using a larger taxonomic diversity than in previous molecular

studies; (2) to discuss the evolution of morphological characters in light of the phylogeny

obtained; and (3) to specify the chronology of their radiation in relation to South American

biogeography, with special reference to the occurrence of the purported fossil anteater

Eurotamandua in Europe.

MATERIALS AND METHODS

Taxonomy

Throughout the text, we introduced "Folivora" for the clade containing the sloths,

because the usual terms of "Phyllophaga" (e.g. Engelmann 1985) or "Tardigrada" (e.g.

Patterson et al. 1992) both referred to different groups of protostomians. Etymologically,

"Folivora" means "leaf eater" and is the Latin correspondence for "Phyllophaga".

5
 Data acquisition

Xenarthran samples preserved in 95% ethanol were stored in the Collection of

Mammalian Tissues from the Institut des Sciences de l'Evolution de Montpellier (Catzeflis

1991). Total DNAs were extracted for Chaetophractus villosus (larger hairy armadillo),

Cabassous unicinctus (southern naked-tailed armadillo), Dasypus novemcinctus (nine-banded

armadillo), Cyclopes didactylus (pygmy anteater), Tamandua tetradactyla (collared anteater),

Myrmecophaga tridactyla (giant anteater), Bradypus tridactylus (pale-throated three-toed

sloth) and Choloepus didactylus (southern two-toed sloth).

PCR amplifications of the complete 12S rRNA gene were conducted following

Hassanin & Douzery (1999). The 16S rRNA fragment (722 bp) for Myrmecophaga was

amplified with primers R9 = 5'-CGCCTGTTTACCAAAAACATC-3' (direct) and S1 = 5'-

TTATGCAATTACCGAGATCTGCCA-3' (reverse), and cloned using the pCRTM 2.1 plasmid vector

and Escherichia coli strain INVαF' (Original TA cloning kit, Invitrogen). For the vWF exon

28 of sloths and armadillos, two overlapping fragments (V1/W2: 907 bp, and V2/W1: 962 bp)

were amplified according to Huchon et al. (1999). Three overlapping fragments (V1/W11:

398 bp, V8/W10: 504 bp, and V9/W12: 523 bp) were amplified for anteaters, with the

following direct (V) and reverse (W) primers: V8 = 5'-

GGAGCTGCGGCGCRTCGCCAGCCAGGTGAA-3', V9 = 5'-ACGAGATCATCAGCTACCTCTGYGRCCTG-

3', W10 = 5'-TGTTGAAGCCGGCCTCGCCGATCCTGTC-3', W11 = 5'-

CCGAAGACCTGGAACAGCGTGTAC-3', and W12 = 5'-TGAGGGCCCACRCCRATGGG-3'. PCR

products were purified from 1% agarose gels using Amicon Ultrafree-DA columns

(Millipore), and sequenced on both strands using manual (Thermo Sequenase cycle kit,

Amersham) and automatic sequencing (Big Dye Terminator cycle kit) on an ABI 310 (PE

Applied Biosystems). Three distinct clones were sequenced for the partial 16S rRNA of

Myrmecophaga. Sequences have been deposited in EMBL under accession numbers

6
AJ278151-AJ278161, and AJ297939 (Electronic Appendix I). Two already published vWF

sequences have been updated (Chaetophractus villosus: AF076480 and Bradypus tridactylus:

U31603).

Taxonomic sampling

Since species sampling has been shown to have a major impact on phylogenetic

reconstruction (Philippe & Douzery 1994), we chose a data set encompassing a wide range

of ordinal mammalian diversity. To reduce potential long branch attraction artefacts, we

analysed, whenever possible, at least two species per placental order. The sequences of two

marsupials from Australia (Macropus giganteus) and America (Didelphis virginiana) were

used to unambiguously root eutherian trees. Thirty five mammalian genera were considered

for the three genes. All sequences accession numbers are listed in Electronic Appendix I.

Sequence alignment and phylogenetic analyses

Sequences were manually aligned with the ED editor of the MUST package (Philippe

1993). The alignment of vWF exon 28 sequences started at position 38 of the human

sequence, was 1233 bp long, and non sequenced positions were coded as missing data. Indels

and hypervariable zones of the 12S and 16S rRNAs were removed from subsequent analyses.

Alignments are available upon request to FD.

Phylogenetic reconstructions were conducted with PAUP* 4.0b4a (Swofford 1998),

TREE-PUZZLE 4.02 (Strimmer & von Haeseler 1996), and PAML 3.0b (Yang 2000).

Distance analyses used the Minimum Evolution (ME) criterion with TBR branch swapping on

a Neighbor Joining (NJ) tree based on paralinear Log-Det distances with constant site

removal. Maximum Parsimony (MP) analyses consisted of heuristic searches using 100

random addition sequence replicates with TBR branch swapping. Maximum Likelihood (ML)

analyses with PAUP* were made in two steps in order to speed up computation times. First,

ML parameters were estimated using a heuristic search with nearest neighbour interchange

7
(NNI) branch swapping on a NJ starting tree. Second, a new search was conducted with

random addition of sequences and tree bisection-reconnection (TBR) branch swapping, using

the previously estimated parameters. Substitution rates were described by the GTR (under

PAUP* and PAML) and JTT (under TREE-PUZZLE and PAML) models of sequence

evolution for nucleotides and amino acids respectively. Among sites substitution rate

heterogeneity for DNA and vWF protein was described by a gamma distribution with eight

categories (Γ8) (Yang 1996).

Robustness of the different nodes was estimated by Bootstrap Percentages (BP) after

1000 replications for ME and MP with TBR branch swapping, and after 300 replications using

previously estimated parameters for ML, with NJ starting trees and 1000 TBR rearrangements

per resampling. Reliability Percentages (RP) for the protein dataset were obtained after 10000

quartet-puzzling (QP) steps with TREE-PUZZLE. Bremer Support Indices, i.e. the number of

extra steps required to break the corresponding node under the MP criterion, were computed

with PAUP*. Log-likelihoods of alternative topologies were compared by the ML test of

Kishino & Hasegawa (1989) implemented in PAML 3.0b. Three different GTR+Γ8 models

were assumed for the combination of the three genes.

Molecular datings

The presence of local molecular clocks (Yoder & Yang 2000) was tested for different

data sets, and molecular dates were calculated by PAML 3.0b using maximum likelihood

branch lengths in a locally clocklike tree. Taxa were removed one by one until acceptance by

the likelihood ratio test (LRT) of the hypothesis of three local clocks: one for paenungulates,

one for perissodactyls, and one for the remaining species. Three independent paleontological

calibration points were taken into account to check for molecular datings with cross-

calibrations (Huchon et al. 2000): (1) 55 million years (Myr) for the split between

Ceratotherium and Equus (Waddell et al. 1999b) ; (2) 60 Myr for the split between Procavia

8
and Elephas (Gheerbrant et al. 1996); and (3) 63 Myr for the Arctocyonia – Mesonychia

divergence (Gingerich & Uhen 1998), represented here by the Sus (domestic pig) – Physeter

(sperm whale) split. To discuss our molecular dates in the light of geochronology, we have

used the geological timescale compiled for the Tertiary by McKenna & Bell (1997).

RESULTS

Phylogenetic results

Nuclear exon 28 of vWF

Due to significant base composition heterogeneity (evaluated by a χ2 test in TREE-

PUZZLE) for numerous taxa (including Dasypus, Chaetophractus, Choloepus, Tamandua and

Myrmecophaga within Xenarthra), third codon positions were excluded from all subsequent

analyses. All phylogenetic reconstructions provided strong support for most nodes involving

xenarthrans (Table 1): (i) monophyly of Xenarthra; (ii) monophyly of each of its three

lineages: armadillos, anteaters and sloths; (iii) grouping of anteaters and sloths into a clade

called Pilosa; and (iv) close relationship within anteaters between Myrmecophaga and

Tamandua — the subfamily Myrmecophaginae — to the exclusion of Cyclopes. Within

armadillos, the grouping of Cabassous with Chaetophractus to the exclusion of Dasypus

received mixed support (Table 1). Phylogenetic analyses on vWF amino acids yielded high

support for all nodes within Xenarthra, and provided several exclusive amino acid

substitutions supporting them (Table 1 and Electronic Appendix II).

Mitochondrial 12S and 16S rRNAs

All nodes involving xenarthrans were retrieved by phylogenetic analyses using 12S

and 16S rRNAs (Table 1), with high support (monophyly of Cingulata, Vermilingua, Folivora,

and Myrmecophaginae), moderate support (grouping of anteaters and sloths), or low support

(monophyly of the order). One conflict arose from 12S rRNA relationships within armadillos

9
(Table 1): ME favoured the grouping of Cabassous with Chaetophractus, contra MP and ML

which suggested a sister-group relationship between Cabassous and Dasypus.

Combination of the three genes

The ML phylogram obtained with the combined nuclear and mitochondrial data sets

(2107 bp) is presented in Figure 1. Concatenation resulted in a general increase of robustness

indices (Table 1). Xenarthra, Pilosa, Cingulata, Folivora, Vermilingua and Myrmecophaginae

were clearly monophyletic (BPML > 97), and the grouping of the armadillos Cabassous and

Chaetophractus to the exclusion of Dasypus was favoured (BPML = 60).

Tests of alternative hypotheses

The three alternative branching patterns connecting armadillos, anteaters, and

sloths were compared by Kishino & Hasegawa (1989) likelihood tests. Pilosa appeared to

be the most likely hypothesis in all cases. Alternative hypotheses placing either anteaters or

sloths as sister-group to the remaining xenarthrans involved a drop in log-likelihoods

which is significant when the three markers are combined (Table 2). With regard to

Xenarthra sister-group relationships, their emergence second amongst placentals — after

Afrotheria — was the most likely hypothesis in the vWF and combination cases.

Alternative hypotheses, i.e. Xenarthra sister-group of either remaining placentals

(Epitheria), or Afrotheria, or Ferungulata, all appeared less likely when the three markers

were combined as they involved a drop in log-likelihood by ca. 1 standard error (Table 2).

The morphological hypothesis of "Edentata" (Xenarthra + Pholidota) was significantly

rejected in all cases (Table 2).

Molecular timings

Local clocks have been used in the ML datings because they represent a compromise

between setting a single substitution rate among all lineages (global clock constrained) and

independent rates for each branch (no clock constrained). Amino acid substitutions in vWF

10
behaved locally clocklike for a reduced data set of 18 taxa, including Xenarthra,

Paenungulata, Perissodactyla, and Cetartiodactyla (Lama excluded due to a fast rate of

evolution) with the two marsupials as outgroup (LRT between non-clock and locally clocklike

tree: 2∆lnL = 27.2 with 19 degrees of freedom; P = 0.10). Using respectively the

perissodactyl, cetartiodactyl, and paenungulate calibration points, we obtained average

substitution rates of 0.160, 0.194, and 0.205 % amino acid / Myr / lineage in the vWF protein

(Figure 2). Cross-calibration comparisons between these three points showed that the

cetartiodactyl and paenungulate points were compatible, with the latter yielding the youngest

ages, whereas the perissodactyl point provided the oldest dates.

Within placentals, the separation between Afrotheria (here represented by two

paenungulates: Figure 2) and the remaining groups was estimated to have occurred 122 Myr

(range of datings depending on the choice of calibration points: 115-148 Myr), whereas

Xenarthra diverged from Cetartiodactyla + Perissodactyla some 106 Myr (range: 100-128

Myr). Our estimates then placed the xenarthran radiation — corresponding to the split

between Cingulata and Pilosa — at ca. 63 Myr (Figure 2), around the Cretaceous/Tertiary

(K/T) boundary (range: 59-76 Myr). Within Xenarthra, the following splits have been dated:

(i) Vermilingua vs. Folivora: 54 Myr in Early Eocene (range: 51-65 Myr) ; (ii) Dasypus vs.

other armadillos: 33 Myr in Early Oligocene (31-40 Myr), followed by Chaetophractus vs.

Cabassous: 21 Myr in Early Miocene (20-25 Myr) ; (iii) pygmy anteater vs. other anteaters: 38

Myr in Late Eocene (35-45 Myr), followed by giant vs. collared anteaters: 13 Myr in Middle

Miocene (12-15 Myr) ; (iv) Bradypus vs. Choloepus: 18 Myr in Early Miocene (16-21 Myr).

Datings were also conducted with the three concatenated markers at the DNA level,

assuming three independent ML models, and using the same set of 18 species (LRT: 2∆lnL =

28.8; P = 0.07). The paenungulate and the cetartiodactyl points yielded similar dating

estimates between them (data not shown), and also similar relative to vWF amino acids. For

11
example, within Xenarthra, the first (armadillos vs. others) and second (anteaters vs. sloths)

splits are estimated to be respectively 57-69 Myr and 50-60 Myr old at the DNA level, against

59-63 Myr and 51-54 Myr for vWF amino acids (Figure 2). The perissodactyl calibration

point provided the oldest dating estimates, and was not compatible with the two other points:

it estimated the Hyracoidea / Proboscidea split to have occurred 118 Myr, against 60 Myr for

palaeontology.

DISCUSSION

Molecular systematics and character evolution in Xenarthra

Xenarthra monophyly

The Xenarthra monophyly is unambiguously supported in all reconstructions by the

vWF alone or in combination with the 12S and 16S rRNAs (Table 1). The single common

ancestry of the eight xenarthrans here studied is also defined by three exclusive vWF amino

acid replacements (Table1 and Electronic Appendix II), which should be added to the unique

derived deletion in the αA-crystallin known for seven xenarthrans (van Dijk et al. 1999). The

monophyly of Xenarthra is also supported by morphological characters like accessory

articulations between vertebrae ("xenarthry"), ischiosacral fusion — with a secondary reversal

in Cyclopes likely related to its strictly arboreal life style (Gaudin & Branham 1998) — and

dental simplification by the loss of enamel (Engelmann 1985). These characters are generally

thought to reflect adaptation towards fossoriality and myrmecophagy (Carroll 1988; Gaudin

1999).

Xenarthra intra-ordinal relationships

The monophyly of Cingulata, Vermilingua, and Folivora finds strong support from

vWF and mitochondrial rRNAs sequence comparisons. It is also corroborated by numerous

morphological characters like, for example, the modification of dermal ossicles into articulate

12
plates for Cingulata, the total absence of teeth for Vermilingua, and the presence of paired

perforations of the centra in lumbar vertebrae for Folivora (Engelmann 1985).

Our data, including all anteater and sloth genera, provide strong support for the clade

Pilosa (Tables 1, 2), extending the results of Madsen et al. (2001) and Murphy et al. (2001).

This arrangement is in agreement with studies of the ear region (Patterson et al. 1992) and

other morphological characters which provided remarkable synapomorphies for Pilosa like,

e.g. the interruption of the zygomatic arch or the intra-pelvian location of the testes

(Engelmann 1985). These results contrast with those of Guth (1961), Bugge (1979), Carrol

(1988), and Cao et al. (1998) who claimed that anteaters are the sister-group of the remaining

xenarthrans. In the light of our results, it is likely that earlier studies of the ear region (Guth

1961) and cephalic arterial pattern (Bugge 1979) may have been misled by the extreme

specialisation of the skull towards myrmecophagy in anteaters. Reanalysis of ear characters

(Patterson et al. 1992) supported the Pilosa hypothesis, and the character pointed by Bugge

(1979) for the grouping of armadillos and sloths (i.e. the median course of the internal carotid

artery) is likely to be symplesiomorphic. Carrol (1988) suggested an early emergence of

anteaters by emphasising the fact that no fossil anteaters were found to have retained dermal

ossicles, unlike some mylodontid sloths which possess these relicts of the cingulate-like

xenarthrans' ancestor. The fossil record of anteaters is particularly scarce (Gaudin & Branham

1998), and our results showing that anteaters are the sister-group of sloths suggest that fossil

anteaters with armour relicts remain to be found. Thus, if fossils are taken under

consideration, the replacement of the ancestral carapace by a coat could not be considered as a

synapomorphy defining Pilosa. However, this evolutionary trend has led to a discrete pattern

in living xenarthrans where armoured xenarthrans (armadillos) have retained the ancestral

carapace and hairy xenarthrans (anteaters and sloths) share the possession of a true coat. Our

conclusions also contradict those of Cao et al. (1998) based on the mitochondrial NADH

13
dehydrogenase 1 gene (ND1) which moderately supports a close relationship between the

armadillo Cabassous and the sloth Bradypus to the exclusion of the anteater Tamandua. This

result may be due to the reduced taxonomic sampling and/or to the particular behaviour of this

fast evolving molecule which seems to be subject to convergence (Cao et al. 1998).

Our results confirm the classical arrangement within Vermilingua by favouring the

grouping of the strictly terrestrial Myrmecophaga and semi-arboreal Tamandua into the

subfamily Myrmecophaginae, to the exclusion of the strictly arboreal Cyclopes. Such a

relationship is corroborated by numerous myological (Reiss 1997) and morphological (Gaudin

& Branham 1998) characters.

The phylogeny of Cingulata is still poorly understood from a morphological point of

view (Patterson et al. 1989). Our results suggest the grouping of Cabassous (Priodontini) and

Chaetophractus (Euphractini) to the exclusion of Dasypus (Dasypodini), contra Engelmann

(1985) who claims a close relationship between Dasypus and Cabassous. Our results are in

agreement with a recent study of sperm morphology and morphometry (Cetica et al. 1998)

showing that Cabassous and Chaetophractus have rather similar spermatozoa. However, the

discrepancies existing between vWF, 12S, and 16S rRNA suggest that additional genes for a

greater taxonomic diversity are needed to further resolve relationships within armadillos.

Searching for the Xenarthra sister-group

The vWF and mitochondrial rRNA genes do not provide a clear response to the

difficult question of Xenarthra affinities within Eutheria. However, our analyses including

eight armadillos, anteaters and sloths genera suggest that Xenarthra may constitute the

second offshoot of the placental tree, Afrotheria being the first one to branch off (Figure 1

and Table 2). This is in agreement with the results obtained by Madsen et al. (2001) and

Murphy et al. (2001) on respectively four and five genera. Alternative affinities of

Xenarthra with either epitherians, afrotherians, or ferungulates appear less likely (Table 2),

14
but cannot be significantly rejected. Sequencing of additional molecular markers to provide

an increased number of phylogenetically informative positions concomitant with adequate

taxonomic diversity of Xenarthra is needed to further resolve their position within the

eutherian tree. Interesting to note, the classical hypothesis of "Edentata" (Novacek 1992) is

significantly rejected in all cases (Table 2). Together with other molecular results (de Jong

1998; Madsen et al. 2001; Murphy et al. 2001), this confirms that morphological

similarities between xenarthrans and pangolins are a spectacular example of convergence

in relation to myrmecophagy (Bugge 1979; Rose & Emry 1993).

Molecular timing of the xenarthran radiation

With vWF amino acids, the three calibration points — cetartiodactyls, perissodactyls,

paenungulates — are reciprocally compatible in the local molecular clocks analysis (Figure 2).

One should note that setting 63 Myr for the split between Bos (rather than Sus) and Physeter

would involve 85 Myr for the split between Procavia and Elephas, i.e. a divergence time 25

Myr older than the paleontological one. This is the reason why our molecular datings tend to

yield younger dates than previously published for the Hippopotamus / Physeter split: 41-53

Myr (Figure 2) vs. 53-54 Myr (Montgelard et al. 1997; Ursing & Arnason 1998). However,

our datings of deeper splits become more compatible with other proposed dates because of

increasing intervals between the dates obtained from the three different calibration points. For

example, our estimates of the xenarthran separation from other placentals (Figure 2: 100-128

Myr) are in the range of those of Springer (1997) and Waddell et al. (1999b).

The xenarthran radiation, corresponding to the emergence of armadillos, is here

estimated to have occurred close to the K/T boundary (59-76 Myr) and the separation between

anteaters and sloths during the Palaeocene (51-65 Myr). These results are younger than other

molecular estimations yielding about 80 Myr for the radiation of xenarthran families (review

in Bromham et al. 1999). Indeed, immunological comparisons suggest ca. 80 Myr (Sarich

15
1985), partial 12S and 16S rRNA gave the same result (Höss et al. 1996), and Cao et al.

(1998) even suggest an interval of 65 to 130 Myr based on the deep divergences they observed

in ND1. These different molecular dates suggest that breaking long isolated branches through

increased taxonomic sampling within Xenarthra might have modified the divergence time

estimates.

Our estimates obtained after extensive taxonomic sampling within Xenarthra are

compatible with paleontological data. The first armadillo scutes are from the late Palaeocene

(ca. 58 Myr) of Brazil (Scillato-Yané 1976), the first sloth remains are from the Middle

Eocene of Antarctica (Vizcaíno & Scillatto-Yané 1995), and the first apparition of undoubted

myrmecophagid anteaters is from the Early Miocene of Patagonia (Carlini et al. 1992). Long

gaps in the fossil record are clearly not unusual, and our dating of the origin of anteaters and

sloths (around the Palaeocene/Eocene limit) suggests a gap of about 30 million years in the

South American fossil record of anteaters, and emphasises its incompleteness. Regarding

Myrmecophagidae, the very ancient origin of the pygmy anteater (Cyclopes) explains its very

divergent morphology shaped by ca. 30 Myr of arboreal life-style, since arboreality is

considered ancestral for anteaters (Gaudin & Branham 1998). The separation between the two

modern sloths, unknown as fossils (Patterson & Pascual 1972), is dated by our data at ca. 18

Myr (end of Early Miocene). Two-toed (Choloepus) and three-toed sloths (Bradypus) have

been placed into two distinct families (respectively Megalonychidae and Bradypodidae) on the

basis of their numerous morphological differences and a presumably diphyletic origin (Webb

1985). Our estimation is slightly younger than the 25 Myr of Sarich (1985) but confirms their

considerable divergence. It suggests that their apparent external similarities are the result of

paralellism and that arboreality may have evolved at least twice within Folivora (Höss et al.

1996; Greenwood et al. 2001). The quite ancient dates obtained for the armadillos radiation

strengthen the fact that Cingulata contains strikingly divergent taxa, a result coherent with

16
marked differences observed in the structure of their spermatozoa (Cetica et al. 1998). Finally,

it is noteworthy that numerous cladogenic events within Xenarthra seem to occur close to

transitions between Epochs of the chronostratigraphic scale (Figure 2). Since those transitions

are defined by dramatic environmental and climatic changes (Pascual & Ortiz Jaureguizar

1990), such a synchrony suggests a major role of paleobiogeographic changes in the

diversification of xenarthrans.

Eurotamandua: helping to solve the enigma?

Our molecular estimates of divergence dates for armadillos, anteaters and sloths raise

the question of the occurrence of purported fossil xenarthrans outside of South America.

Indeed, if the presence of a sloth in the Eocene of Seymour Island is compatible with land

connections between Antarctica and the Patagonian province until the Oligocene (Vizcaíno &

Scillatto-Yané 1995), the occurrence of Ernanodon in the Palaeocene of China (Ding 1979)

and Eurotamandua in the Middle Eocene (45 Myr) of Europe (Storch 1981) are more difficult

to explain since South America was an island at these epochs. The true belonging of

Ernanodon and Eurotamandua to Xenarthra have been widely debated (Rose & Emry 1993;

Gaudin 1999). Actually, Ernanodon antelios seems likely to be assigned to a group of endemic

Chinese mammals whose morphological similarities with Xenarthra would reflect convergent

adaptations to fossorial habits (Rose & Emry 1993; Gaudin 1999). However, there is no

consensus about the status of Eurotamandua joresi (Storch 1981). Vermilinguan affinities of

Eurotamandua have been suggested (Storch 1981; Storch & Habersetzer 1991), but several of

the most crucial features for assessing its phylogenetic position remain equivocal (Gaudin &

Branham 1998; Gaudin 1999; Rose 1999). If Eurotamandua belongs to Vermilingua, our

results dating the origin of Vermilingua well nested in the Tertiary (between 54 and 38 Myr:

Figure 2) raise again the question of how an anteater reached Europe at a time where South

America was isolated from all other continental masses? The only possible response is

17
dispersal which seems to be very unlikely in the present state of our knowledge of

paleobiogeography based on plate tectonic models. Thus, our results, contra Höss et al.

(1996), cast doubt on the true belonging of Eurotamandua to Vermilingua. They suggest that

the striking morphological resemblances between this taxon and Myrmecophagidae might be

the result of adaptative convergence towards fossoriality and ant-feeding. This example

underlines, once again, how morphological adaptation to similar ecological niches could be

positively misleading in terms of phylogenetic signal.

The xenarthran status of Eurotamandua was also widely questioned and this fossil was

interpreted as either an endemic Old World xenarthran sister-group of New World xenarthrans

(Szalay & Schrenk 1994), or a Pholidota (Shoshani et al. 1997), or a sister-group to Pilosa

(Gaudin & Branham 1998). However, the evidence for the presence of true xenarthry in

Eurotamandua is far from convincing (Rose & Emry 1993; Gaudin & Branham 1998; Rose

1999). Based on an exhaustive and detailed study of the evolution of xenarthrous vertebrae,

Gaudin (1999) concluded that there is at present little evidence of a close phylogenetic

relationship between true xenarthrans and Eurotamandua. Moreover, on the study of forelimbs

from a second specimen, Rose (1999) argued for a close relationship of Eurotamandua and

Palaeanodonta (a group of extinct fossorial mammals with reduced dentitions). Our results

placing the origin of Pilosa in the Early Tertiary contradict those of Gaudin & Branham

(1998), but our dating of the xenarthrans origin into Cretaceous (between 106 and 63 Myr:

Figure 2) leave opened the possibility that Eurotamandua could be a basal member of

Xenarthra. Only the discovering of additional specimens will provide us with a better

understanding on the evolutionary affinities of this enigmatic fossil.

18
ACKNOWLEDGMENTS

This study would not have been feasible without the invaluable contribution of several

people who kindly provided us with xenarthran samples: Danny Devillier, Eric Hansen, Jesus

Mavarez, Jean-Christophe Vié and his team "Faune Sauvage" at Petit-Saut (French Guiana).

We wish to thank Claudine Montgelard and Mark S. Springer for giving access to unpublished

sequences, Ziheng Yang for PAML advices, and Jeff Mauffrey for help with figures. Sergio F.

Vizcaíno and three anonymous referees are thanked for helpful comments. This work has been

supported by the "Mammalian phylogeny" TMR Network (FMRX–CT98–022) of the

European Community, and by the Génopôle Languedoc-Roussillon. Research on French

Guiana mammals was funded by EDF-CNEH (contract GP-0850) and Ministère de

l’Environnement (contract SOFT-Guyane). F. D. is supported by a M.E.N.R.T. grant (99075)

associated with the D.E.A. "Biologie de l’Evolution et Ecologie" (Montpellier, France). This

is the contribution ISEM 2001-040 of the Institut des Sciences de l’Evolution de Montpellier

(UMR 5554 - CNRS).

REFERENCES

Allard M. W., McNiff B. E. & Miyamoto M. M. 1996 Support for interordinal eutherian

relationships with an emphasis on primates and their archontan relatives. Mol.

Phylogenet. Evol. 5, 78-88.

Arnason U., Gullberg A. & Janke A. 1997 Phylogenetic analyses of mitochondrial DNA

suggest a sister group relationship between Xenarthra (Edentata) and Ferungulates. Mol.

Biol. Evol. 14, 762-768.

Bromham L., Phillips M. J. & Penny D. 1999 Growing up with dinosaurs: molecular dates

and the mammalian radiation. Trends Ecol. Evol. 14, 113-118.

19
Bugge J. 1979 Cephalic arterial pattern in New World edentates and Old World pangolins

with special reference to their phylogenetic relationships and taxonomy. Acta Anat. 105,

37-46.

Cao Y., Janke A., Waddell P. J., Westerman M., Takenaka O., Murata S., Okada N., Pääbo S.

& Hasegawa M. 1998 Conflict among individual mitochondrial proteins in resolving the

phylogeny of eutherian orders. J. Mol. Evol. 47, 307-322.

Carlini A. A., Scillatto-Yané G. J., Vizcaíno S. F. & Dozo M. T. 1992 Un singular

Myrmecophagidae (Xenarthra, Vermilingua) de Edad Colhuehuapense (Oligoceno

Tardio-Miocene Temprano) de Patagonia, Argentina. Ameghiniana 29, 176.

Carroll R. L. 1988 Vertebrate paleontology and evolution. New-York: Freeman.

Catzeflis F. 1991 Animal tissue collections for molecular genetics and systematics. Trends

Ecol. Evol. 6, 168.

Cetica P. D., Solari A. J., Merani M. S., De Rosas J. C. & Burgos M. H. 1998 Evolutionary

sperm morphology and morphometry in armadillos. J. Submicrosc. Cytol. Pathol. 30,

309-314.

de Jong W. W., Zweers A., Joysey K. A., Gleaves J. T. & Boulter D. 1985 Protein sequence

analysis applied to xenarthran and pholidote phylogeny. In The Evolution and Ecology

of Armadillos, Sloths and Vermilinguas (ed. G. G. Montgomery), pp. 65-76.

Washington: Smithsonian Institution.

de Jong W. W. 1998 Molecules remodels the mammalian tree. Trends Ecol. Evol. 13, 270-

275.

Ding S. Y. 1979 A new Edentate from the Paleocene of Guangdong. Vert. Palasiatica 17, 62-

64.

20
Engelmann G. F. 1985 The phylogeny of the Xenarthra. In The Evolution and Ecology of

Armadillos, Sloths and Vermilinguas (ed. G. G. Montgomery), pp. 51-63. Washington:

Smithsonian Institution.

Fariña R. A. 1996 Trophic relationships among Lujanian mammals. Evolutionary Theory 11,

125-134.

Gaudin T. J., Wible J. R., Hopson J. A. & Turnbull W. D. 1996 Reexamination of the

morphological evidence for the Cohort Epitheria (Mammalia, Eutheria). J. Mammal.

Evol. 3, 31-78.

Gaudin T. J. & Branham D. G. 1998 The phylogeny of the Myrmecophagidae (Mammalia,

Xenarthra, Vermilingua) and the relationship of Eurotamandua to the Vermilingua. J.

Mammal. Evol. 5, 237-265.

Gaudin T. J. 1999 The morphology of xenarthrous vertebrae (Mammalia: Xenarthra).

Fieldiana: Geology, n. s. 41, 1-38.

Gheerbrant E., Sudre J. & Cappetta H. 1996 A Palaeocene proboscidean from Morocco.

Nature 383, 68-70.

Gingerich P. D. & Uhen M. D. 1998 Likelihood estimation of the time of origin of Cetacea

and the time of divergence of Cetacea and Artiodactyla. Paleontol. Electron. 2, 1-47.

Glass B. P. 1985 History of classification and nomenclature in Xenarthra (Edentata). In The

Evolution and Ecology of Armadillos, Sloths and Vermilinguas (ed. G. G.

Montgomery), pp. 1-3. Washington: Smithsonian Institution.

Greenwood A. D., Castresana J., Feldmaier-Fuchs G. & Pääbo S. 2001 A molecular

phylogeny of two extinct sloths. Mol. Phylogenet. Evol. 18, 94-103.

Guth C. 1961 La région temporale des Edentés. [Ph. D. dissertation] University of Paris,

Paris, France.

21
Hassanin A. & Douzery E. J. P. 1999 The tribal radiation of the family Bovidae (Artiodactyla)

and the evolution of the mitochondrial cytochrome b gene. Mol. Phylogenet. Evol. 13,

227-243.

Höss M., Dilling A., Currant A. & Pääbo S. 1996 Molecular phylogeny of the extinct ground

sloth Mylodon darwinii. Proc. Natl. Acad. Sci. USA 93, 181-185.

Huchon D., Catzeflis F. M. & Douzery E. J. P. 1999 Molecular evolution of the nuclear von

Willebrand Factor gene in mammals and the phylogeny of rodents. Mol. Biol. Evol. 16,

577-589.

Huchon D., Catzeflis F. M. & Douzery E. J. P. 2000 Variance of molecular datings, evolution

of rodents and the phylogenetic affinities between Ctenodactylidae and Hystricognathi.

Proc. R. Soc. Lond. B. 267, 393-402.

Kishino H. & Hasegawa M. 1989 Evaluation of the maximum likelihood estimate of the

evolutionary tree topologies from DNA sequence data, and the branching order in

Hominoidea. J. Mol. Evol. 29, 170-179.

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 adaptative radiations in

two major clades of placental mammals. Nature 409, 610-614.

McKenna M. C. 1975 Toward a phylogenetic classification of the Mammalia. In Phylogeny of

the primates (ed. W. P. Luckett & F. S. Szalay), pp. 21-46. New York: Plenum Press.

McKenna M. C. & Bell S. K. 1997 Classification of Mammals above the species level. New

York: Columbia University Press.

Montgelard C., Catzeflis F. M. & Douzery E. 1997 Phylogenetic relationships of artiodactyls

and ceteacans as deduced from the comparison of cytochrome b and 12S rRNA

mitochondrial sequences. Mol. Biol. Evol. 14, 550-559.

Murphy W. J., Eizirik E., Jonhson W. E., Zhang Y. P., Ryder O. A. & O’Brien S. J. 2001

22
 Molecular phylogenetics and the origins of placentals mammals. Nature 409, 614-618.

Novacek M. J. 1992 Mammalian phylogeny: shaking the tree. Nature 356, 121-125.

Pascual R. & Ortiz Jaureguizar E. 1990 Evolving climates and mammal faunas in Cenozoic

South America. J. Hum. Evol. 19, 23-60.

Patterson B. & Pascual R. 1972 The fossil mammal fauna of South-America. In Evolution,

mammals and southern continents (ed. A. Keast, F. C. Erk & B. P. Glass), pp. 247-309.

Albany: State University of New-York Press.

Patterson B., Segall W. & Turnbull W. D. 1989 The ear region in xenarthrans ( = Edentata,

Mammalia). Part I. Cingulates. Fieldiana: Geology, n. s. 18, 1-46.

Patterson B., Segall W., Turnbull W. D. & Gaudin T. J. 1992 The ear region in xenarthrans (=

Edentata, Mammalia). Part II. Sloths, Anteaters, Palaeanodonts, and a Miscellany.

Fieldiana: Geology, n. s. 24, 1-79.

Philippe H. 1993 MUST: a computer package of management utilities for sequences and trees.

Nucl. Acids Res. 21, 5264-5272.

Philippe H. & Douzery E. 1994 The pitfalls of molecular phylogeny based on four species, as

illustrated by the Cetacea/Artiodactyla relationships. J. Mammal. Evol. 2, 133-151.

Reiss K. Z. 1997 Myology of the feeding apparatus of myrmecophagid anteaters (Xenarthra:

Myrmecophagidae). J. Mammal. Evol. 4, 87-117.

Rose K. D. & Emry R. J. 1993 Relationships of Xenarthra, Pholidota, and fossil "Edentates":

the morphological evidence. In Mammal Phylogeny (Placentals) (ed. F. S. Sazlay, M. J.

Novacek & M. C. McKenna), pp. 81-102. New York: Springer-Verlag.

Rose K. D. 1999 Eurotamandua and Palaeanodonta: convergent or related? Paläontol. Z. 73,

384-401.

23
Sarich V. M. 1985 Xenarthran systematics: albumin immunological evidence. In The

Evolution and Ecology of Armadillos, Sloths and Vermilinguas (ed. G. G.

Montgomery), pp. 77-81. Washington: Smithsonian Institution.

Scillato-Yané G. J. 1976 Sobre un Dasypodidae (Mammalia, Xenarthra) de edad

Riochiquense (Paleoceno superior) de Itaboraí, Brasil. Annais Academia Brasileira de

Cièncias 48, 527-530.

Shoshani J., McKenna M. C., Rose K. D. & Emry R. J. 1997 Eurotamandua is a pholidotan

not a xenarthran. J. Vertebr. Paleontol. 17, 76A.

Springer M. S. 1997 Molecular clocks and the timing of the placental and marsupial radiations

in relation to the Cretaceous-Tertiary boundary. J. Mammal. Evol. 4, 285-302.

Storch G. 1981 Eurotamandua joresi, ein Myrmecophagidae aus dem Eozän der "Grubbe

Messel" bei Darmstadt (Mammalia, Xenarthra). Senckenb. leth. 61, 247-269.

Storch G. & Habersetzer J. 1991 Rückverlagerte Choanen und akzessorische Bulla tympanica

bei rezenten Vermilingua und Eurotamandua aus dem Eozän von Messel (Mammalia:

Xenarthra). Z. Säugetierkunde 56, 257-271.

Strimmer K. & von Haeseler A. 1996 Quartet puzzling: a quartet maximum-likelihood

method for reconstructing tree topologies. Mol. Biol. Evol. 13, 964-969.

Swofford D. L. 1998 PAUP*. Phylogenetic Analysis Using Parsimony (* and Other Methods).

Version 4. Sunderland, Massachusetts: Sinauer Associates.

Szalay F. S. & Schrenk F. 1994 Middle Eocene Eurotamandua and the early differentiation of

the Edentata. J. Vertebr. Paleontol. 14, 48A.

Ursing B. M. & Arnason U. 1998 Analyses of mitochondrial genomes strongly support a

hippopotamus-whale clade. Proc. R. Soc. Lond. B 265, 2251-2255.

24
van Dijk M. A. M., Paradis E., Catzeflis F. & de Jong W. W. 1999 The virtues of gaps:

Xenarthran (Edentate) monophyly supported by a unique deletion in αA-crystallin. Syst.

Biol. 48, 94-106.

Vizcaíno S. F. & Scillato-Yané G. J. 1995 A Eocene Tardigrade (Mammalia, Xenarthra) from

Seymour Island, West Antartica. Antarc. Sci. 7, 407-408.

Waddell P. J., Cao Y., Hauf J. & Hasegawa M. 1999a Using novels phylogenetic methods to

evaluate mammalian mtDNA, including amino acid-invariant sites-logDet plus site

stripping, to detect internal conflicts in the data, with special reference to the positions

of hedgehog, armadillo, and elephant. Syst. Biol. 48, 31-53.

Waddell P. J., Cao Y., Hasegawa M. & Mindell D. P. 1999b Assessing the cretaceous

superordinal divergence times within birds and placental mammals by using whole

mitochondrial protein sequences and an extended statistical framework. Syst. Biol. 48,

119-137.

Webb S. D. 1985 The interrelationships of tree sloths and ground sloths. In The Evolution and

Ecology of Armadillos, Sloths and Vermilinguas (ed. G. G. Montgomery), pp. 105-112.

Washington: Smithsonian Institution.

Yang Z. 1996 Among-site rate variation and its impact on phylogenetic analyses. Trends in

Ecol. Evol. 11, 367-372.

Yang Z. 2000 Phylogenetic Analysis by Maximum Likelihood (PAML), Version 3.0b.

London: University College London.

Yoder A. D. & Yang Z. 2000 Estimation of primate speciation dates using local molecular

clocks. Mol. Biol. Evol. 17, 1081-1190.

25
FIGURE CAPTIONS

Figure 1: Maximum Likelihood phylogram obtained using the combination of complete

mitochondrial 12S rRNA, partial 16S rRNA and positions 1 and 2 of the nuclear vWF exon

28 (2107 sites whose 1150 variable). The GTR model was used (rate matrix: A-C: 5.28, A-G:

11.33, A-T: 4.86, C-G: 3.05, C-T: 23.03 and G-T: 1.00) and among site variation following a

gamma distribution (eight categories) of shape α = 0.44 with θ = 16% of invariable sites.

Branch lengths are proportional to the estimated number of substitution per site, and that

leading to the outgroup has been reduced by one-half. Xenarthrans are connected by bold

lines. Node labels are bootstrap percentages obtained after 300 ML replications. Note that the

combination representing Elephantidae is a chimera between Elephas maximus (asian

elephant) for 12S rRNA and vWF, and Loxondonta africana (african elephant) for 16SrRNA.

Parameters for 12S rRNA: 838 sites with 485 variable, rate matrix: A-C: 3.80, A-G: 10.32, A-

T: 4.67, C-G: 0.41, C-T: 28.00, G-T: 1.00 ; θ = 5% ; α = 0.30. Parameters for 16S rRNA: 447

sites with 194 variable, rate matrix: 31.84, 33.33, 20.59, 0.64, 123.70 ; θ = 34% ; α = 0.41.

Parameters for vWF (codon positions 1 + 2): 822 sites with 494 variable, rate matrix: 2.27,

6.66, 1.48, 2.64, 5.90, 1.00 ; θ = 12% ; α = 0.77.

Figure 2: Molecular estimates of divergence dates for the xenarthran radiation based on a

reduced dataset of vWF amino acids (415 sites of which 319 were variable). The JTT model

of protein evolution and a gamma distribution (eight categories ; α = 0.94) were used as

likelihood assumptions. The maximum likelihood phylogram constrained to be clocklike was

obtained with three local clocks: one for the branch leading to perissodactyls (Ceratotherium

and Equus ; rate: 0.09), one for the branch leading to paenungulates (Procavia and Elephas ;

rate: 0.45), and one for the remaining branches (default rate: 1.00). Divergence dates (in

26
million years; Myr) are indicated, and were deduced from three calibration points: CPER (55

Myr: split between Ceratotherium and Equus), CPAE (60 Myr: split between Procavia and

Elephas), and CCET (63 Myr: split between Sus and Bos + Hippopotamus + Physeter).

Rectangles at nodes represent the range of dates estimated using the three calibration points.

Youngest and oldest estimates are respectively provided by the paenungulate and

perissodactyl points. Intermediate values are given by the cetartiodactyl point and are

indicated near the nodes. Cross-calibration results: (i) the perissodactyl split is estimated to

have occurred either 43 [calibration by CPAE] or 46 [CCET] Myr, (ii) paenungulate split: 64

[CCET] or 77 [CPER] Myr, and (iii) cetartiodactyl split: 59 [CPAE] or 76 [CPER] Myr. The split

between Macropus and Didelphis has not been dated, because arbitrary dates can be obtained

by assuming different relative rates on the two sides of the root (Yoder & Yang 2000). The

time scale is given above the tree. The main geological epochs are indicated and delimited by

dashed lines: U C = Upper Cretaceous ; Tertiary: Pa = Palaeocene, E = Eocene, O =

Oligocene, M = Miocene, and Pl = Plio-Pleistocene. The Cretaceous/Tertiary boundary (K/T)

is represented by the vertical striped line. Fossil records of Xenarthra (Eurotamandua

excluded) are figured by continuous lines below the tree. Broken lines indicate gaps in the

fossil record suggested by our molecular estimates.

27
TABLES

Table 1. Indices of robustness for the nodes of the Xenarthra subtree, obtained for 12S rRNA, partial 16S rRNA, vWF exon 28 (codon positions

1 + 2), and combination of the three genes, with three methods of phylogenetic reconstruction (ME: distance, MP: maximum parsimony, ML:

maximum likelihood). Reliability percentages (RP) after Quartet Puzzling and number of Exclusive (non homoplasic) Synapomorphies (ExS)

observed for the vWF exon 28 amino acid data set are also indicated for each node. Position and nature of these exclusive synapomorphies are

indicated in Electronic Appendix II.

12S rRNA Partial 16S rRNA vWF (1 + 2) Combination vWF a.a.

BSI ME MP ML BSI ME MP ML BSI ME MP ML BSI ME MP ML RP ExS

Cingulata (armadillos) + 25 100 100 100 +4 100 94 96 + 10 100 99 99 + 36 100 100 100 99 2

Cabassous + Chaetophractus 0 75 - - +1 85 61 48 +2 92 67 50 +1 83 61 60 90 2

Vermilingua (anteaters) +3 88 73 96 +4 99 80 84 + 11 100 99 100 + 16 97 100 100 98 3

Tamandua + Myrmecophaga + 22 100 100 100 +3 100 92 87 + 32 100 100 100 + 44 100 100 100 99 5

Folivora (sloths) + 15 100 99 100 +5 100 95 95 + 34 100 100 100 + 33 100 100 100 99 6

Pilosa (anteaters + sloths) +3 78 53 75 +2 73 43 59 +4 97 89 93 +6 87 88 97 93 2

Xenarthra +3 43 29 48 0 50 24 62 + 17 100 100 100 + 27 100 100 100 93 3

Notes: BSI = Bremer Support Indices. ME = bootstrap percentages obtained after 1000 replications of minimum evolution on paralinear Log-Det

distances. MP = bootstrap percentages obtained after 1000 replications of standard parsimony. ML = bootstrap percentages obtained after 100
replications of maximum likelihood. RP = reliability percentages obtained with TREE-PUZZLE after 10000 QP steps. A dash indicates that the

node does not appear in the corresponding majority-rule consensus bootstrap tree.
 Table 2. Results of Kishino & Hasegawa (1989) likelihood tests of alternative topologies computed for the nuclear vWF (codon positions 1 + 2),

the mitochondrial 12S and 16S rRNAs, and their combination. δ = ∆lnL / S. E. refers to the ratio between the difference in log-likelihood (∆lnL)

of two phylogenetic hypotheses and its standard error (S. E.). P refers to the one-tailed significance level of the Kishino & Hasegawa test as

performed in PAML ( *: significantly worse at the 5% level).

12S rRNA Partial 16S rRNA vWF (1+2) Combination

Phylogenetic a priori hypotheses - lnL δ P - lnL δ P - lnL δ P - lnL δ P

INTRA XENARTHRA
RELATIONSHIPS

Vermilingua + Folivora (= Pilosa) 11216.44 Best — 4602.55 Best — 8569.98 Best — 24587.14 Best —
Vermilingua + Cingulata 11223.96 1.26 0.10 4605.04 0.93 0.18 8576.64 1.35 0.09 24603.76 2.06 0.02 *
Folivora + Cingulata 11223.93 1.25 0.11 4604.51 0.58 0.28 8576.42 1.26 0.10 24603.14 1.92 0.03 *

XENARTHRA SISTER-GROUP
RELATIONSHIPS

Xenarthra + Pholidota ("Edentata") 11225.16 1.84 0.03 * 4619.23 2.54 < 0.01 * 8585.99 1.86 0.03 * 24620.34 3.04 < 0.01 *
Xenarthra + Epitheria 11216.94 0.44 0.33 4602.55 Best — 8574.29 1.09 0.14 24591.23 1.07 0.14
Xenarthra 2nd offshoot 11216.46 0.08 0.47 4604.30 0.75 0.23 8569.98 Best — 24587.14 Best —
Xenarthra + Afrotheria 11216.94 0.44 0.33 4604.13 0.62 0.27 8574.29 1.09 0.14 24591.31 1.11 0.13
Xenarthra + Ferungulata 11216.44 Best — 4609.68 0.99 0.16 8572.31 0.41 0.34 24594.88 0.95 0.17
 Electronic Appendix I: Taxonomic sampling used for this study. Accession numbers and corresponding references are given for each sequence
used. Tissue numbers refer to the collection of mammalian tissues of the "Institut des Sciences de l'Evolution" de Montpellier (Catzeflis 1991).

Species Common name 12S rRNA Partial 16S rRNA vWF exon 28

MARSUPIALIA
DIDELPHIMORPHIA
Didelphis virginiana Virginia opossum Z29573, Janke et al. (1994) Z29573, Janke et al. (1994) AF226848, Madsen et al. (2001)
DIPROTODONTIA
Macropus giganteus Eastern grey kangaroo X86941, Douzery & Catzeflis (1995) — AJ224670, Huchon et al.(1999)
Macropus robustus Wallaroo — Y10524, Janke et al. (1997) —

PLACENTALIA
CHRYSOCHLORIDAE
Amblysomus hottentotus Hottentot golden mole M95108, Allard & Miyamoto (1992) U97336, Springer et al. (1997) U97534, Springer et al. (1997)
TENRECIDAE
Echinops telfairi Lesser hedgehog-tenrec AF069540, Stanhope et al. (1998) AF069540, Stanhope et al. (1998) AF076478, Stanhope et al. (1998)
EULIPOTYPHLA
Erinaceus europaeus West European hedgehog X88898, Krettek et al. (1995) X88898, Krettek et al. (1995) U97536, Springer et al. (1997)
Scalopus aquaticus Eastern mole AF069539, Stanhope et al. (1998) AF069539, Stanhope et al. (1998) AF076479, Stanhope et al. (1998)
MACROSCELIDEA
Elephantulus rufescens Rufous elephant shrew U97339, Springer et al. (1997) U97339, Springer et al. (1997) U31612, Porter et al. (1996)
SIRENIA
Dugong dugon Dugong U60185, Lavergne et al. (1996) AF179291, Springer et al. (in press) U31608, Porter et al. (1996)
PROBOSCIDEA
Elephas maximus Asian elephant X93602, Lavergne et al. (1996) — U31611, Porter et al. (1996)
Loxodonta africana African elephant — AJ224821, Hauf et al. (2000) —
HYRACOIDEA
Procavia capensis Common rock hyrax U60184, Lavergne et al. (1996) U97335, Springer et al. (1997) U31619, Porter et al. (1996)
TUBULIDENTATA
Orycteropus afer Aardvark U97338, Springer et al. (1997) U97338, Springer et al. (1997) U31617, Porter et al. (1996)
CARNIVORA
Canis familiaris Domestic dog Y08507, Ledje & Arnason (1996) U96639, Kim et al. (1998) L16903, Mancuso et al. (unp.)
Felis catus Domestic cat U20753, Lopez et al. (1996) U20753, Lopez et al. (1996) U31613, Porter et al. (1996)
PRIMATES
 Homo sapiens Human J01415, Anderson et al. (1981) J01415, Anderson et al. (1981) X06828, Bonthron & Orkin (1988)
Ateles sp. † Spider monkey AF069978, Horovitz et al. (1998) U39011, Horovitz & Meyer (1995) —
Ateles belzebuth White-bellied spider monkey — — AF061059, Stanhope et al. (1998)
CETARTIODACTYLA
Bos Taurus Domestic cow J01394, Anderson et al. (1982) J01394, Anderson et al. (1982) X63820, Bakhshi et al. (1992)
Sus scrofa Domestic pig AJ002189, Ursing & Arnason (1998a) AJ002189, Ursing & Arnason (1998a) S78431, Bahnak et al. (1992)
Physeter catodon Sperm whale Arnason et al. (1993)* AJ277029, Arnason et al. (2000) AF108834, Gatesy et al. (1999)
Hippopotamus amphibius Common hippopotamus AJ010957, Ursing & Arnason (1998b) AJ010957, Ursing & Arnason (1998b) AF108832, Gatesy et al. (1999)
Lama guanicoe Guanaco Y08809, Montgelard et al. (1997) AJ010815, Douzery et al. (unp.) —
Lama glama Lama — — AF108835, Gatesy et al. (1999)
PHOLIDOTA
Manis sp. † Pangolin U61079, Springer & Douzery (1996) U97340, Springer et al. (1997) U97535, Springer et al. (1997)
RODENTIA
Mus musculus Domestic mouse X84382, Hänni et al. (1995) V00711, Bibb et al. (1981) AJ238390, Huchon et al. (1999)
Cavia porcellus Domestic Guinea pig AJ222767, D'Erchia et al. (1996) AJ222767, D'Erchia et al. (1996) AJ224663, Huchon et al. (1999)
PERISSODACTYLA
Equus asinus Donkey X97337, Xu et al. (1996) X97337, Xu et al. (1996) U31610, Porter et al. (1996)
Ceratotherium simum White rhinoceros X86942, Douzery & Catzeflis (1995) Y07726, Xu & Arnason (1997) U31604, Porter et al. (1996)
CHIROPTERA
Megaderma lyra False vampire bat AF069538, Stanhope et al. (1998) AF069538, Stanhope et al. (1998) U31616, Porter et al. (1996)
Dobsonia moluccensis Bar-backed fruit bat U93065, Hollar & Springer (1997) AF179290, Teeling et al. (2000) U31609, Porter et al. (1996)
XENARTHRA
Dasypus novemcinctus Nine-banded armadillo Y11832, Arnason et al. (1997) Y11832, Arnason et al. (1997) AJ278158, tissue T-1830, this study
Chaetophractus villosus Larger hairy armadillo U61080, Springer & Douzery (1996) AF069534, Stanhope et al. (1998) AF076480, Stanhope et al. (1998)
Cabassous unicinctus Southern naked-tailed AJ278151, tissue T-1641, this study Z48940, Höss et al. (1996) AJ278159, tissue T-1641, this study
armadillo
Myrmecophaga tridactyla Giant anteater AJ278153, tissue T-2080, this study AJ297939, tissue T-2080, this study AJ278157, tissue T-2080, this study
Tamandua tetradactyla Collared anteater AJ278154, tissue T-1640, this study Z48946, Höss et al. (1996) AJ278161, tissue T-2038, this study
Cyclopes didactylus Pygmy anteater AJ278155, tissue T-1724, this study Mark Springer (unpublished) AJ278156, tissue T-1724, this study
Bradypus tridactylus Pale-throated three-toed sloth AF038022, Springer et al. (unp.) AF069535, Stanhope et al. (1998) U31603, Porter et al. (1996)
Choloepus didactylus Southern two-toed sloth AJ278152, tissue T-1722, this study Z48942, Höss et al. (1996) AJ278160, tissue T-1722, this study

unp. : unpublished.
* : This sequence was manually copy from the reference cited because it has not been yet deposited in EMBL.
† : The species name of the specimens are unknown.
Abridged references: Allard & Miyamoto (1992) Mol. Biol. Evol. 9: 778-786; Anderson et al. (1981) Nature 290: 457-465; Anderson et al. (1982) J.
Mol. Evol. 156: 683-717; Arnason et al. (1993) Biochem. Syst. Ecol. 21: 115-122; Arnason et al. (1997) Mol. Biol. Evol. 14: 762-768; Arnason et al.
(2000) J. Mol. Evol. 50: 569-578; Bahnak et al. (1992) Biochem. Biophys. Res. Commun. 182: 561-568; Bakhshi et al. (1992) Biochim. Biophys. Acta
1132: 325-328; Bibb et al. (1981) Cell 26: 167-180; Bonthron & Orkin (1988) Eur. J. Biochem. 171: 51-57; D'Erchia et al. (1996) Nature 381: 597-600;
Douzery (1993) C. R. Acad. Sci. Paris, Sciences de la vie 316: 1511-1518; Douzery & Catzeflis (1995) J. Mol. Evol. 41: 622-636; Höss et al. (1996) Proc.
Natl. Acad. Sci. U.S.A. 93: 181-185; Gatesy et al. (1999) Syst. Biol. 48: 16-20; Hänni et al. (1995) Isr. J. Zool. 41: 131-146; Hauf et al. (2000) Zoology
102: 184-195; Hollar & Springer (1997) Proc. Natl. Acad. Sci. U.S.A. 94: 5716-5721; Horovitz & Meyer (1995) Mol. Phylogenet. Evol. 4: 448-456;
Horovitz et al. (1998) Am. J. Phys. Anthropol. 106: 261-281; Huchon et al. (1999) Mol. Biol. Evol.16: 577-589; Huchon et al. (2000) Proc. R. Soc. Lond. B
267: 393-402; Janke et al. (1994) Genetics 137: 243-256; Janke et al. (1997) Proc. Natl. Acad. Sci. U.S.A. 94: 1276-1281; Kim et al. (1998) Mol.
Phylogenet. Evol. 10: 210-220; Krettek et al. (1995) J. Mol. Evol. 41: 952-957; Lavergne et al. (1996) Mol. Phylogenet. Evol. 6: 245-258; Ledje &
Arnason (1996). J. Mol. Evol. 43: 641-649; Lopez et al. (1996) Genomics 33: 229-246; Madsen et al. (2001) Nature 409: 610-614; Mancuso et al. (1989) J.
Biol. Chem. 264: 19514-19527; Montgelard et al. (1997) Mol. Biol. Evol. 14: 550-559; Porter et al. (1996) Mol. Phylogenet. Evol. 5: 89-101; Springer &
Douzery (1996) J. Mol. Evol. 43: 357-373; Springer et al. (1997) Nature 388: 61-64; Stanhope et al. (1998) Proc. Natl. Acad. Sci. U.S.A. 95: 9967-9972;
Teeling et al. (2000) Nature 403: 188-192; Ursing & Arnason (1998a) J. Mol. Evol. 47: 302-306; Ursing & Arnason (1998b) Proc. R. Soc. Lond. B. 265:
2251-2255; Xu et al. (1996) Mol. Biol. Evol. 13: 1167-1173; Xu & Arnason (1997) Mol. Phylogenet. Evol. 7: 189-194.
Electronic Appendix II. Position and nature of the exclusive Amino-Acids synapomorphies
observed in the vWF exon 28 protein sequences alignment supporting nodes of the Xenarthra subtree.
Clades Positions Substitutions
Cingulata (armadillos) 25 Asp, Glu => Gly
162 Phe, Val, Leu, Ala, Gly => Pro
Cabassous + Chaetophractus 254 Gly, Ala, Glu, Leu, Ile, Thr, Arg, Asn => Ser
294 Asp, Ala => His
Vermilingua (anteaters) 35 Phe, His, Tyr, Leu => Glu
360 Gly, Asp => His
375 Asn, Ser => Gly
Tamandua + Myrmecophaga 2 Ala, Glu, Lys, Arg, Gly => Ser
212 Gln, Arg, His, Pro => Ala
223 Asp => Gly
303 Thr, Ala, Ser, Met => Glu
320 Ala, Thr, Pro, Val => Arg
Folivora (sloths) 168 Leu => Met
233 Gln, Arg, Pro, Val, Met, Ile, Leu => Ala
234 Pro, Gln, Arg, His, Phe => Ala
235 Pro, Ser, Thr, His => Ala
307 Tyr => Phe
375 Asn, Ser => Thr
Pilosa (anteaters + sloths) 121 Thr, Ile => Ala
154 Pro => Ser
Xenarthra 173 Val, Ala => Ile
331 Arg, Gln, Glu, His, Asp => Gly
371 Met => Leu
 Didelphis
Macropus MARSUPIALIA

Elephantulus

98 Orycteropus
10 Amblysomus
56
12 Echinops AFROTHERIA
Elephantidae
99
Dugong
84
Procavia
100 Dasypus
100
Cabassous X
Cingulata E
60 Chaetophractus
100 N
Bradypus A
10
Folivora P R
Choloepus T
97 I
Cyclopes L H
100 R
Myrmecophaga Vermilingua O
100 A
28 S
Tamandua A
Mus RODENTIA
Erinaceus
63 EULIPOTYPHLA
Scalopus
19 6
Cavia
RODENTIA
47 Ateles
100 PRIMATES
Homo
11 Canis
100 CARNIVORA
Felis
Ceratotherium
100 PERISSODACTYLA
41 Equus
25
Dobsonia
77 CHIROPTERA
Megaderma
26
Manis PHOLIDOTA
Lama
32 43
Sus
93
Hippopotamus CETARTIODACTYLA
80 Bos
67
0.1 Physeter
 K- T
98 65 55 34 24 5 Myr

Macropus

Didelphis

0.05 Dasypus
33
substitution/sit
Chaetophractus
21
63 Cabassous

Bradypus
18
Choloepus
54
Cyclopes

Tamandua
106 38
13 Myrmecophaga

Sus
C
CET
Bos
63
48
Hippopotamus
122
44
90
Physeter

C Ceratotherium
PER
55
Equus

C Procavia
PAE
60
Elephas

UC Pa E O M Pl
Cingulata
Folivora
FOSSIL RECORDS
Vermilingua
† Eurotamandua