Mitogenomic relationships of placental mammals and

Gene 421 (2008) 37–51

Contents lists available at ScienceDirect

Gene

j ourna l homepage: www.e lsev ie r.com/ locate /gene

Mitogenomic relationships of placental mammals and molecularestimates of their divergences

Ulfur Arnason a,⁎, Joseph A. Adegoke a,b, Anette Gullberg a, Eric H. Harley c, Axel Janke a, Morgan Kullberg a

a Division of Evolutionary Molecular Systematics, University of Lund, S-223 62 Lund, Swedenb Department of Zoology, Obafemi Awolowo University, Ile-Ife, Nigeriac Division of Chemical Pathology, Department of Clinical Laboratory Sciences, University of Cape Town, South Africa

Abbreviations: aa, amino acid(s); mt, mitochondrianucleotide(s); K/T boundary, Cretaceous/Tertiary bound⁎ Corresponding author. Tel.: +46 462229259; fax: +4

E-mail address: [email protected] (U. Arnason

0378-1119/$ – see front matter © 2008 Elsevier B.V. Alldoi:10.1016/j.gene.2008.05.024

A B S T R A C T
A R T I C L E I N F O
Article history:
Molecular analyses of the re Received 12 February 2008Received in revised form 1 April 2008Accepted 26 May 2008Available online 10 June 2008
Keywords:MammaliaPlacentaliaPrimatesMitochondrial EvePhylogenyBiogeographyTemporal estimates of eutherian divergences

lationships of placental mammals have shown a progressive congruence betweenmitogenomic and nuclear phylogenies. Some inconsistencies have nevertheless persisted, notably with respect tobasal divergences. The current study has aimed to extend the representation of groups, whose position in theplacental tree has been difficult to establish in mitogenomic studies. Both ML (maximum likelihood) and Bayesiananalyses identified four basal monophyletic groups, Afroplacentalia (=Afrotheria: Hyracoidea, Proboscidea, Sirenia,Tenrecidea, Tubulidentata, Macroscelidea, Chrysochloridea), Xenarthra, Archontoglires (Primates, Dermoptera,Scandentia, Lagomorpha, Rodentia) and Laurasiaplacentalia (Lipotyphla, Chiroptera, Pholidota, Carnivora,Perissodactyla, Artiodactyla, Cetacea). All analyses joined Archontoglires and Laurasiaplacentalia on a commonbranch (Boreoplacentalia), but the relationship between Afroplacentalia, Xenarthra and Boreoplacentalia was notconclusively resolved. The phylogenomic hypothesis with a sister group relationship between Notoplacentalia(Afroplacentalia/Xenarthra) and Boreoplacentalia served as the basis for estimating the times of placentaldivergences using paleontologicallywell-supportedmammalian calibrationpoints. These estimates placed the basalplacental divergence between Boreoplacentalia and Notoplacentalia at ≈102 MYA (million years ago). The currentestimates of ordinal placental divergences are congruent with recent estimates based on nuclear data, butinconsistent with paleontological notions that have placed the origin of essentially all placental orders within aninterval of 5–10MY in the early Tertiary. Among less deep divergences the estimates placed the split betweenGorillaandPan/Homo at ≈11.5MYAand that betweenPan andHomo at ≈8MYA. As a consequence of these estimates,whichare in accord with recent progress in primate paleontology, the earliest divergences among recent humans becomeplaced ≈270,000 years ago, i.e. ≈100,000 years earlier than the traditional age of “Mitochondrial Eve”. Comparisonbetween the two new mt genomes of Hylomys suillus (short-tailed gymnure) patently demonstrates theinconsistency that may exist between taxonomic designations and molecular difference, as the distance betweenthese two supposedly conspecific genomes exceeds that of the three elephantid genera Elephas, Mammuthus andLoxodonta. In accordance with the progressive use of the term Placentalia for extant orders and extinct taxa fallingwithin this group we forward new proposals for the names of some superordinal clades of placental mammals.

© 2008 Elsevier B.V. All rights reserved.

1. Introduction

Mitogenomics, i.e. phylogenetic analyses based on the concatenatedaa (amino acid) sequences of proteins or on the nt (nucleotide) sequen-ces of protein-coding genes from complete mitochondrial (mt) geno-mes, have provided valuable information on mammalian relationshipsat different levels. The applicability of mt sequences to the examinationof phylogenetic relationships is partly due to the unequivocal orthologyof these sequences and partly to the facility with which they can beacquired. However, since the amount of data that can be obtained from

l; MYA, million years ago; nt,ary.6 46137578.).

rights reserved.

mt genomes is finite, the conclusive solution to tenuous evolutionaryrelationships will by necessity rely upon the accumulation of nucleardata. It is nevertheless likely thatmitogenomic analyseswill still remainvaluable for studying phylogenetic relationships in temporal windowswithin which mt sequences generally provide greater resolution thannuclear data of similar sizes. Thematernal inheritance of themt genomeis another feature that makes these genomes valuable, notably forpopulation studies including comparisons with Y chromosomal data.

More than 200 mammalian species are to date represented by com-plete mt genomes. This collection includes the three extant monotremespecies and representatives of all recent orders of both marsupial andplacental mammals. In the currentmitogenomic studywe address phylo-genetic questions among placental mammals that have proved con-tentious, even in analyses of extensive nuclear data. In order to increasethe comprehensiveness of the study we have made particular efforts to

mailto:[email protected]

http://dx.doi.org/10.1016/j.gene.2008.05.024

http://www.sciencedirect.com/science/journal/03781119

38 U. Arnason et al. / Gene 421 (2008) 37–51

extend the sampling of basal groups. The phylogenetic questions addres-sed include (i) the position of the Erinaceidae (hedgehogs andmoon rats),(ii) the Epitheria hypothesis, i.e. whether there is a basal split betweenXenarthra and other placental orders (=Epitheria, McKenna, 1975),(iii) the support for a basal split between anAfrican clade (Afroplacentalia)and remaining placentals, and (iv) the potential split between Xenarthra/Afroplacentalia (=Notoplacentalia) and other placental orders(Boreoplacentalia).

The sequencing of the mt genome of the hedgehog aimed to deter-mine the position of Lipotyphla in the tree of placental mammals(Krettek et al., 1995). Lipotyphla was at this time recognized as encom-passing the families Erinaceidae, Tenrecidae, Chrysochloridae, Soleno-dontidae, Talpidae and Soricidae (Butler, 1988), an arrangement that isno longer tenable, as two of these families, Tenrecidae and Chryso-chloridae, fall within Afroplacentalia. The study of Krettek et al. (1995)showed a basal split between the hedgehog and other placental mam-mals. This position of Erinaceidae has been maintained in essentially allsubsequent mitogenomic studies that have been performed withoutconstraining Erinaceidae to a particular position. Prior to the currentstudy Erinaceidae was represented by the mt genomes of two species,Erinaceus europaeus and Hemiechinus auratus, both included in theErinaceinae, and one species, Echinosorex gymnurus, belonging to theHylomyinae. Here we add two specimens of Hylomys suillus (short-tailed gymnure) to the mitogenomic data set, thereby splitting the longHylomyinae branch in an effort to clarify the position of Erinaceidae. Theestablishment of this position directly concerns the resolution of basalrelationships amongplacentalmammals becausemitogenomic analyseshave generally tended to promote a basal split between Erinaceidae andremaining taxa.

A variant of the Epitheria hypothesis has postulated a sister grouprelationship between Xenarthra/Pholidota and remaining eutherians(Novacek, 1992). This notion is inconsistent with mitogenomic resultsthat have placed Pholidota as the probable sister group to Carnivora inan extended Cetferungulata (Arnason et al., 2002; Arnason and Janke,2002). The number of xenarthran species included in the currentanalyses exceeds that of previous mitogenomic studies, a circum-stance that should promote the identification of the position ofXenarthra in the placental tree and allow improved molecular esti-mates of the time of basal xenarthran divergences.

The establishment of themonophyly of Rodentia and Glires was fora long time problematic in mitogenomic studies (e.g. D'Erchia et al.,1996; Arnason et al. 2002; Horner et al., 2007) and the currentlyestablished tree differs in some crucial respects (e.g. Glires and thepositions of Erinaceidae and Primates) from previous mitogenomictrees that served as the basis for molecular estimates of the rela-tionships among placental mammals (Arnason et al., 1996, 1998,2000a, 2002; Arnason and Janke, 2002). Also hypotheses with eitherAfroplacentalia or Notoplacentalia as sister to remaining placentalmammals have been difficult to distinguish from the Epitheria hypo-thesis in both mitogenomic and nuclear studies. The current additionof newmt genomes has inter alia aimed to split the hyracoid, sirenian,proboscid and chrysochlorid branches in an effort to establish diffe-rent nodes within the African clade. The inclusion of these taxa mayalso help to stabilize the position of Afroplacentalia in the tree ofplacental mammals. It will at the same time draw attention to an earlymorphologically based suggestion of phylogenetic affinities betweenTubulidentata and Hyracoidea and Proboscidea, suggested by Le GrosClark and Sonntag (1926).

2. Materials and methods

2.1. Data set and analytical approaches

The mt genomes included in the study, 11 of which are new, arelisted in Table 1 together with their accession numbers. The sequenceof the mt genome of the tapir was established using cloned restriction

fragments from a DNA preparation enriched with respect to mt DNAfollowing the procedure described by Arnason et al. (1991). Thenumber of repeats in the control region of that molecule correspondsto the number counted in a particular clone (Xu, 1996). Other newgenomes were PCR amplified using a selection of conserved and spe-cific primers. The sequences of the repeats in control regions of thesemolecules correspond therefore to a majority rule consensus.

The phylogenetic analyses were carried out on the concatenatedsequences of protein-coding genes using the aa and nt sequences (1stplus 2nd codon positions) of the twelve heavy strand encoded protein-coding genes. The light strand encodedNADH6 genewas not included asit deviates markedly in composition from the other protein-codinggenes. The final alignmentwas 3617 aa long after removal of ambiguoussites. Sequences of all placental taxa except for some Hominoidea,Erinaceidae and one sloth passed the x2 test for sequence homogeneity.

The analyseswere performedusing the TREE-PUZZLE (Schmidt et al.,2002), PHYLIP (Felsenstein, 1993), PAUP⁎ (Swofford, 1998), PHYML(GuindonandGascuel, 2003) andMrBayes version3.1 (Huelsenbeck andRonquist, 2001) program packages. Maximum likelihood (ML) analysesof aawere carriedoutusing TREE-PUZZLEapplying themtREV-24model(Yang et al., 1998). The GTR model of nt evolution (Lanave et al., 1984)was used for distance and likelihood analyses as suggested bymodeltest(Posada and Crandall, 1998). All ML analyses were carried out assumingfour classes of gamma distribution (Yang, 1994) and one class ofinvariable sites. First codon positions of leucine (C or T) were recoded asY in the nt analyses. Analyses with MrBayes were carried out using thesamemodels for aa and nt evolution as in theML analyses. The programsampled every 100th generation from 5,000,000 generations of MCMC(Markov chainMonte Carlo), discarding the first 500,000 generations asburninwith each generation assumingone cold and three heated chains(i.e. Metropolis-coupled MCMC or MCMCMC). Two independent runswere performed for convergence.

Divergence times were estimated (aa rates heterogeneity data set)using the “r8s” program package version 1.70 (Sanderson, 2002),applying nonparametric rate smoothing. Ages of different divergenceswere also estimated under a Bayesian framework using the programMULTIDIVTIME as implemented in the T3 package (ftp://abacus.gene.ucl.ac.uk/pub/T3) and applying the same calibration points. 100,000generations of MCMC (Markov chain Monte Carlo) were sampled afterdiscarding the first 100,000 generations as burnin. MULTIDIVTIMEanalyses of aa sequences were performed assuming a gamma distri-bution with eight classes of the mtREV-24 model.

2.2. Placental calibration points

The five placental calibration points used in the current study areA/C-60, E/R-55, C/F-52, O/P-30, and C/P-48. A/C-60 (Arnason et al.,1996) refers to the divergence between ruminant Artiodactyla andCetacea, here set at 60–65 MYA. E/R-55, here set at 53–58 MYA,signifies the perissodactyl divergence between Equidae and Rhinocer-otidae. This reference was originally set at 50 MYA (Arnason et al.,1998), but with the inclusion of the Tapiridae in the tree, and takinginto account the tapiroid fossil record (Prothero and Schoch,1989), theage of this calibration point is now placed at 55 MYA. C/F-52 refers tothe divergence between caniforms and feliforms set at 52 MYA (Flynnand Galiano, 1982), while O/P-30 rests upon the divergence betweenOtarioidea and Phocidae (Arnason et al., 2006) and the age, ≈28MY, ofthe oldest phocid fossil diagnostic for this divergence.

The availability of well-defined calibration points outside theCetferungulata is limited. Within Archontoglires the most solid cali-bration point, C/P-48 (Arnason et al., 2000b), is defined by the mini-mum age, 48 MY, for the split between Catarrhini and Platyrrhini(Godinot and Mahboubi, 1992). As the effect of any calibration point ismost distinctly marked on the branch to which it belongs, thephylogenetic position of this calibration point is of a particular rele-vance to the estimates of primate divergences.

ftp://abacus.gene.ucl.ac.uk/pub/T3

ftp://abacus.gene.ucl.ac.uk/pub/T3

Table 1Taxa included and their accession numbers

PlacentaliaNotoplacentalia

AfroplacentaliaPaenungulata

Hyracoidea, hyraxesDendrohyrax dorsalis (western tree hyrax, AM904727)Procavia capensis (rock hyrax, AJ428945)Procavia capensis (rock hyrax, AB096865)

Proboscidea, elephantsLoxodonta africana (African elephant, AJ224821)Elephas maximus (Indian elephant, AJ428946)Mammuthus primigenius (woolly mammoth, DQ316067)

Sirenia, dugong, manateesDugong dugon (dugong, AJ421723)Trichechus manatus (Caribbean manatee, AM904728)

AfroinsectiphilaTenrecidea, tenrecs, golden molesEchinops telfairi (lesser hedgehog tenrec, AJ400734)

Tubulidentata, aardvarkOrycteropus afer (aardvark, Y18475)

Macroscelidea, elephant shrewsMacroscelides proboscideus (short-eared elephant shrew, AJ421452)Elephantulus sp. (long-eared elephant shrew, AB096867)

Chrysochloridea, golden molesEremitalpa granti (Grant's golden mole, AM904729)Chrysochloris asiatica (Cape golden mole, AJ428944)Chrysochloris asiatica (Cape golden mole, AB096866)

XenarthraCingulata, armadillosDasypus novemcinctus (nine banded armadillo, Y11832)

Pilosa, anteaters, slothsTamandua tetradactyla (lesser anteater, AJ421450)Choloepus didactylus (two-toed sloth, NC 006924)Bradypus tridactylus (pale-throated sloth, AY960979)

BoreoplacentaliaArchontogliresArchonta

Scandentia, tree shrewsTupaia belangeri (tree shrew, AJ421453)

Dermoptera, flying lemursCynocephalus variegatus (flying lemur, AJ428849)

Primates, lemurs, lorises, tarsiers, monkeys, apesNycticebus coucang (slow loris, AJ309867)Daubentonia madagascariensis (Aye-aye, AM905039)Lemur catta (ring-tailed lemur, AJ421451)Eulemur mongoz (mongoose lemur, AM905040)Tarsius bancanus (tarsier, AF348159)Cebus albifrons (pale-fronted capuchin, AJ309866)Macaca sylvanus (Barbary ape, NC 002764)Macaca mulatta (rhesus monkey, AY612638)Papio hamadryas (hamadryas baboon, Y18001)Hylobates lar (gibbon, X99256)Pongo abelii (Sumatran orangutan, X97707)Gorilla gorilla (gorilla, X93347)Homo sapiens (human, X93334)Pan troglodytes (common chimpanzee, X93335)

GliresRodentia, rodentsRattus norvegicus (brown rat, AJ428514)Mus musculus (mouse, J01420)Cricetulus griseus (Chinese hamster, DQ390542)Jaculus jaculus (Egyptian jerboa, AJ416890)Spalax ehrenbergi (Palestine mole rat, AJ416891)Cavia porcellus (guinea pig, AJ222767)Thryonomys swinderianus (cane rat, AJ301644)Glis glis (fat dormouse, Y11137)Sciurus vulgaris (squirrel, AJ238588)

Lagomorpha, pikas, hares, rabbitsOchotona collaris (pika, AF348080)Ochotona princeps (American pica, AJ537415)Oryctolagus cuniculus (rabbit, AJ001588)Lepus europaeus (brown hare, AJ421471)

LaurasiaplacentaliaLipotyphla, hedgehogs, solenodons, shrews, molesErinaceidae, hedgehogs, moon ratsErinaceus europaeus (hedgehog, X88898)

(continued on next page)

Hemiechinus auritus (long-eared hedgehog, AB099481)Echinosorex gymnurus (moon rat, AF348079)Hylomys suillus (short-tailed gymnure, AM905041, T-1773, Java)Hylomys suillus (short-tailed gymnure, AM905042, T-0675, Malaysia)Solenodon paradoxus (Hispaniolan solenodon, AF434830)Talpa europaea (European mole, Y19192)Mogera wogura (Japanese mole, AB099482)Urotrichus talpoides (Japanese shrew mole, AB099483)Galemys pyrenaicus (Pyrenean desman, AY833419)Crocidura russula (white-toothed shrew, NC 006893)Sorex unguiculatus (long-clawed shrew, AB061527)Episoriculus fumidus (Taiwan brown-toothed shrew, NC 003040)

Chiroptera, batsRousettus aegyptiacus (Egyptian rousette, AB205183)Pteropus dasymallus (Ryukyu flying fox, NC 002612)Rhinolophus monoceros (Formosan Lesser Horseshoe Bat, NC 005433)Artibeus jamaicensis (Jamaican fruit bat, AF061340)Chalinolobus tuberculatus (NZ long-tailed bat, AF321051)Pipistrellus abramus (Japanese house bat, NC 005436)

CetferungulataPholidota, pangolinsManis tetradactyla (long-tailed pangolin, AJ421454)

Carnivora, carnivoresFelis catus (cat, U20753)Acinonyx jubatus (cheetah, AY463959)Canis familiaris (dog, U96639)Vulpes vulpes (red fox, AM181037)Ursus maritimus (polar bear, AJ428577)Ursus americanus, black bear (NC003426)Meles meles, Eurasian badger (AM711900)Gulo gulo, wolverine (AM711901)Odobenus rosmarus (walrus, AJ428576)Zalophus californianus, Californian sea lion (AM181017)Monachus schauinslandi (Hawaiian monk seal, AM181022)Phoca vitulina (harbor seal, X63726)

Perissodactyla, horses, tapirs, rhinocerosesEquus caballus (horse, X79547)Equus asinus (donkey, X97337)Tapirus terrestris (Brazilian tapir, AJ428947)Rhinoceros unicornis (Indian rhinoceros, X97336)Ceratotherium simum (white rhinoceros, Y07726)

Artiodactyla, pigs, camels, ruminants, hippopotamusesSus scrofa (pig, AJ002189)Lama pacos (alpaca, Y19184)Bos taurus (cow, J01394)Pantholops hodgsonii (chiru, NC 007441)Ovis aries (sheep, AF010406)Capra hircus (goat, AF533441)Muntiacus muntjak (muntjak, NC 004563)Cervus unicolor swinhoei (Formosan sambar, NC 008414)Hippopotamus amphibius (hippopotamus, AJ010957)

Cetacea, whales, dolphins, porpoisesBalaenoptera physalus (fin whale, X61145)Balaenoptera musculus (blue whale, X72204)Physeter macrocephalus (sperm whale, AJ277029)

MarsupialiaMonodelphis domestica (short-tailed grey opossum, AJ508398)Caenolestes fuliginosus (silky shrew opossum, NC 005828Sminthopsis douglasi (Julia Creek dunnart, AJ639867)Macropus robustus (wallaroo, Y10524)Isoodon macrourus (bandicoot, AF358864)

Non-mammalian taxaStruthio camelus (ostrich, Y12025)Gallus gallus (chicken, X52392)Boa constrictor (boa, AB177354)Iguana iguana (iguana, AJ278511)

Accession numbers of new mt genomes are shown in bold. In the case of Solenodonparadoxus the accession number does not refer to a complete mt molecule. Thehierarchy of the systematic scheme is based on recent analyses of nuclear data and thecurrent mitogenomic results. Ordinal and superordinal nomenclature is discussed inthe text and under the heading Materials and methods.

Table 1 (continued )

39U. Arnason et al. / Gene 421 (2008) 37–51

A/C-60 and E/R-55 have been shown to be mutually congruent(Arnason et al., 1998) andmitogenomic estimates based on them placethe separation between Metatheria and Eutheria at 130–135 MYA, i.e.≈6–10 MY earlier than the age, ≈124 MY, of the oldest metatherian


(Luo et al., 2003) and eutherian (Archibald et al., 2001) fossils hithertodescribed. The age of the calibration point connected with the sepa-ration between Metatheria and Eutheria was placed in the range 124–138 MYA (Benton and Donoghue, 2007).

2.3. Nomenclature

2.3.1. PlacentaliaRougier et al. (1998) in a discussion of metatherian relationships

applied a terminological distinction between Marsupialia and Meta-theria according to which Marsupialia includes all extant marsupialsand their most recent common ancestor plus extinct taxa that fallwithin this group. In analogywith this scheme (e.g. Archibald andRose,2005; Wible et al., 2007) Placentalia includes all extant placentals andtheir most recent common ancestor, while Eutheria encompassesmammals that share a more recent common ancestor than they dowith Metatheria. As a consequence of this arrangement we applyhere a nomenclature that uses the affix -placentalia rather than -theriaor -eutheria for superordinal groups falling within Placentalia.

2.3.2. Notoplacentalia (Greek notos, south, southern): Afroplacentalia(=Afrotheria, Stanhope et al., 1998)+Xenarthra and their common stem

This construct is parallel to Boreoplacentalia (Greek boreo, north-ern). Nototheria and Boreotheria might have been considered as moreadequate constructs. The name Nototheria is occupied by a differentgroup of mammals, however. The names Atlantogenata (cf. Waddellet al., 1999, 2001) and Xenafrotheria (Asher, 2005) have sometimesbeen used for crown group Afroplacentalia/Xenarthra. Waddell et al.(1999) dated the split between Xenarthra and Afroplacentalia to122 MYA, claiming that they had temporally connected the age of asplit between Xenarthra and Afroplacentalia to the opening-up of theSouth Atlantic. This claim is untenable, however, as this geologicalevent occurred ≈100 MYA. The priority of Atlantogenata relative toother and more appropriate names for Xenarthra/Afroplacentalia istherefore questionable.

2.3.3. Boreoplacentalia: the placental sister group to Notoplacentalia, i.e.placentals other than Xenarthra and Afroplacentalia

The names Boreotheria (Waddell et al., 2001) and Boreoeutheria(Murphy et al., 2001) have been used for this group. However, with theuse of Placentalia for all extant placentals and theirmost recent commonancestor the name Boreoeutheria becomes problematic as it wouldsuggest a position of eutherians within crown group Placentalia. For thesake of conformity we therefore apply the names Boreoplacentalia andNotoplacentalia respectively for the twogroups thatmakeupPlacentalia.

2.3.4. Afroinsectiphila: the sister group to Hyracoidea, Sirenia andProboscidea (= Paenungulata, Simpson, 1945) within Afroplacentalia

Afroinsectiphila includes the orders Tubulidentata, Macroscelidea,Tenrecidea and Chrysochloridea. Tenrecoidea has been used for thegroupingofTenrecidaeandChrysochloridae (Butler,1988; cf. Asher, 2005).However, considering the depth of the origin of the two groups and thepossibility that theymay not constitute sister groupswe consider them astwo different orders. The name Afrosoricida (Stanhope et al., 1998) hasbeen used as identical in content as Tenrecoidea (Tenrecidea+Chryso-chloridea). Afrosoricida is a misnomer, however, as Afrosorex is a juniorsynonym for genus Crocidura (musk shrews, white-toothed shrews), ofthe lipotyphlan family Soricidaewhich falls within Laurasiaplacentalia. Inaddition, the construct Afrosoricida implies that there are African Sorexspecies that fall within it. No such taxa exist, however. The namesAfroinsectiphillia (cf. Waddell et al., 2001), and Afroinsectiphilia used bysome authors, are not in accord with Latin grammar.

Archonta (=Archonta s.s.): Primates, Dermoptera, and Scandentia.Archontoglires (=Archontoglires s.s): Archonta+Glires (Rodentia+Lagomorpha).

Lipotyphla (=Lipotyphla s.s.): Erinaceidae, Solenodontidae, Sorici-dae, Talpidae.

Cetferungulata (Arnason et al., 2002): Artiodactyla, Cetacea,Carnivora, Pholidota and Perissodactyla.Cetancodonta (Arnason et al., 2000a): Cetacea+Hippopotamidae.

3. Results and discussion

3.1. The mitogenomic tree and comparison with other hypotheses

The initial step of the phylogenetic analysis was to investigate thetopology of theunrooted tree based on all placental taxa listed inTable 1.This tree is shown in a simplified version in Fig. 1. This unconstrainedanalysis identified the four major Placentalia groups, Afroplacentalia,Xenarthra, Archontoglires and Laurasiaplacentalia. Erinaceidae wassister to remaining taxa on the Laurasiaplacentalia branch. In contrastto the majority of earlier mitogenomic studies Rodentia was recoveredas amonophyletic groupwithin Glires. The position of the placental rootwasexaminedbyconnecting theoutgroupto thefivepositions indicatedin Fig. 1. The ML trees with a root at positions [a], [b] and [c] wereappreciably better supported than those with the root at positions [d]and [e], and rooting at [a] wasmarginally favored relative to hypotheses[b] and [c] (Table 2). The limited difference between trees with roots at[a], [b] or [c] was reflected by Bayesian analyses of a data set thatincluded both ingroup and outgroup taxa, which left these relationshipsunresolved. The unstable (“rogue”, Sullivan and Swofford,1997) positionof Erinaceidae in themitogenomic treewas apparent in that a transfer ofErinaceidae to a basal position in any of the five (a–e) rooted trees hadonly minor effect on the ML values. The difficulties associated withanalyses of the erinaceid mt genomes are further underlined by the factthat the mt distances between Erinacinae and Hylominae are greaterthan those between either Erinaceinae or Hylominae and any othergroup of placental mammals.

The ML aa tree with all placental taxa included is shown in Fig. 2together with Bayesian support values. Consistent with the Bayesiananalyses and the limited resolution among trees [a], [b] and [c] inFig. 1 their relationship has been left unresolved in Fig. 2. WithinAfroplacentalia maximum support (ML and Bayesian) was recordedfor both Paenungulata (Hyracoidea, Sirenia and Proboscidea) andAfroinsectiphila (Tenrecidea, Tubulidentata, Macroscelidea andChrysochloridea). Within Paenungulata the grouping of Sirenia andProboscidea (=Tethytheria, McKenna, 1975) to the exclusion of Hyr-acoidea was significantly supported in agreement with the commonmorphological understanding. It is probable that the extended taxonsampling of the Paenungulata in the current analysis has contributed toimproving the resolution within the group as compared to previousstudies.

Basal relationships within Afroinsectiphila have been difficult toresolve. Both the ML and the Bayesian analyses (aa and nt) favored aposition of Tenrecidea as sister to the remaining taxa, among whichTubulidentata was sister to a branch joining Macroscelidea and Chry-sochloridea. A tree exchanging the positions of Tenrecidea andTubulidentata was not statistically refuted by ML analysis. However,the ML tree that joined Tenrecidea with the Chrysochloridae wasabout one S.E. worse than the best tree. Tenrecidea is currently repre-sented only by the genus Echinops in mitogenomic analyses. Theposition of Tenrecidea was investigated further in a separate analysisthat split the Tenrecidea branch by the addition of genus Oryzorictes.This analysis, which was based on complete sequences of the NADH2gene, upheld a basal split between Tenrecidea and the branch con-taining Tubulidentata, Macroscelidea and Chrysochloridea. It is likelythat analyses of the relationships within Afroinsectiphila may alwaysbe hampered by the monotypic nature of Tubulidentata. Most mole-cular studies have identified a sister group relationship betweenTenrecidea and Chrysochlorideawithin the Afroinsectiphila. The sister

Fig. 1. The positions of the five rooting points tested. The ML analysis identified position[a] as marginally better than [b] and [c] and these three positions were appreciablybetter than [d] and [e]. Afroplacentalia: Hyracoidea, Proboscidea, Sirenia, Tenrecidea,Tubulidentata, Macroscelidea, Chrysochloridea. Archontoglires: Primates, Dermoptera,Scandentia, Lagomorpha, Rodentia. Laurasiaplacentalia: Lipotyphla, Chiroptera, Pholi-dota, Carnivora, Perissodactyla, Artiodactyla, Cetacea. Notoplacentalia: Xenarthra+Afroplacentalia. Boreoplacentalia: Archontoglires+Laurasiaplacentalia.

Table 2Support for different rooting positions in the mitogenomic tree

Tree in Fig. 1 mtREV+4Γ+I cdp1leu2 GTR+4Γ+I

ΔlogL pSH pKH ΔlogL pSH pKH

a [−184774.9] 1.00 1.00 [−169022.2] 1.00 1.00b 7.99 0.42 0.16 3.16 0.61 0.31c 7.70 0.47 0.18 3.02 0.73 0.31d 19.06 0.09 0.07 9.84 0.23 0.15e 15.01 0.18 0.13 8.34 0.28 0.19

Trees a–e refer to the rooting points shown in Fig. 1. The logL values of the best trees areshown in square brackets and the support for alternative trees in ΔlogL pSH and pKHvalues.


group relationship between Macroscelidea and Chrysochlorideafavored in the current analyses is consistent with a recent studythat addressed various placental relationships based on LINE-1 data(Waters et al., 2007).

Within Xenarthra there was a well-defined basal split betweenthe armadillo and a branch that contained the lesser anteater andthe two sloths. The result conforms to current systematic schemes,some of which (cf. McKenna and Bell, 1997) even split Xenarthra intotwo orders, Cingulata (represented by the armadillo) and Pilosa (thetamandua and the two sloths).

The third of the basal groups, Boreoplacentalia, splits into Archon-toglires (i.e. Archonta s.s.: Scandentia, Dermoptera and Primates; andGlires: Rodentia and Lagomorpha), and a branch, Laurasiaplacentalia,containing Lipotyphla, Chiroptera and Cetferungulata. The position ofScandentia (represented by only one taxon) was unstable. Thus the ntanalyses placed it as sister to remaining Archontoglires, while the aadata set tended to join Scandentia and Lagomorpha, thereby disrup-ting the monophyly of Glires. As a result we have maintained amonophyletic Glires in Fig. 2 and shown the relationship betweenGlires, Scandentia and Dermoptera/Primates as unresolved.

As in other mitogenomic analyses Dermoptera grouped togetherwith Simiiformes, making Primates paraphyletic (e.g. Arnason et al.,2002; Arnason and Janke, 2002; Schmitz et al., 2002; Kjer and Honey-cutt, 2007). A close relationship between Dermoptera and Primates iscommon to essentially all molecular studies, albeit not to the extentof placing Dermoptera within Primates itself. Since Scandentia andDermoptera are each represented by the mt genome of only onespecies, this may limit confidence in their correct placement.

Basal relationshipswithin Primates have been contentious, notablywith respect to the position of Tarsius. Earlier systematic schemescommonly split recent primates into two groups, Simiiformes andProsimii, with Tarsius included in the latter. However, current hypo-theses, based on extended morphological and molecular data, favor asplit between Strepsirrhini and Haplorrhini with the latter includingSimiiformes and Tarsius (Martin,1993, 2000; Schmitz et al., 2005). Thebest mitogenomic tree had Tarsius sister to remaining primates. How-ever, neither the tree with Tarsius sister to Simiiformes to the exclu-sion of Strepsirrhini (Haplorrhini hypothesis), nor the least supportedalternative with Tarsius sister to Strepsirrhini (Prosimii hypothesis),was statistically refuted. Of particular interest was to estimate thetime of the divergence between Daubentonia and other MalagasyStrepsirrhini. As apparent in Fig. 2 the common Strepsirrhini branch iswell defined as is also the common Malagasy branch separating theMalagasy Strepsirrhini from the (non-Malagasy) slow loris.

Laurasiaplacentalia includes Lipotyphla (Erinaceidae, Solenodon-tidae, Soricidae and Talpidae), Chiroptera and Cetferungulata (sensuArnason et al., 2002), i.e. Perissodactyla, Pholidota, Carnivora,Artiodactyla and Cetacea. The analysis placed Lipotyphla as sister toChiroptera/Cetferungulata. Within Lipotyphla Erinaceidae was sisterto Talpidae and Soricidae, which were joined on a common branch.

The position of Chiroptera as sister to Cetferungulata was wellsupported. A position of Chiroptera as sister to Perissodactyla, assupported by the retroposon analyses of Nishihara et al. (2006), wassignificantly inconsistent with the mitogenomic results. Chiropterahave traditionally been divided into Megachiroptera and Microchir-optera. The mitogenomic results showing microchiropteran paraphylyby nesting Megachiroptera within Microchiroptera are in agreementwith previous molecular findings (e.g. Jones and Teeling, 2006).

Consistent with previous mitogenomic results (Arnason et al.,2002; Arnason and Janke, 2002) the analyses placed the pangolin(Pholidota) as sister to Carnivora. This positionwas unstable, however,and placing it as sister to Carnivora/Perissodactyla was only margin-ally worse than the best tree. The Bayesian analysis left the pangolin inan unresolved tripartite position together with Lipotyphla andChiroptera/Cetferungulata. The close molecular relationship betweenCarnivora and Perissodactyla agrees with the initial mitogenomicestablishment of this relationship (Xu et al., 1996).

Within Perissodactyla the new mt genome of the tapir served toestablish the depth of the basal Ceratomorpha split between Rhino-cerotidae and Tapiridae. The basal perissodactyl split between Cera-tomorpha and Hippomorpha is of a particular interest to moleculardatings within the eutherian tree because this divergence was one ofthe first explicitly defined placental calibration points applied to esti-mate the time of other eutherian divergences with a scantier fossilrecord (Arnason et al., 1998). The current cetartiodactyl relationshipsare consistent with earlier mitogenomic findings including the sistergroup relationships between Cetacea and Hippopotamidae (Ursingand Arnason, 1998), originally demonstrated in the cytb analyses ofIrwin and Arnason (1994).

Above we have referred to a limited selection of studies that areconnected to certain parts of the phylogenetic tree of placental mam-mals. Instead of extending the discussion of some particular mito-genomic relationships we will briefly summarize the similarities anddissimilarities between the current results and those of some recentstudies based on extensive nuclear data that have primarily addresseddeep placental relationships.

The relationship between Rodentia (represented by the mouse),Primates (Homo) and Artiodactyla (pig) was examined by Jorgensenet al. (2005) using nuclear sequences of full-length cDNA alignments(N700,000 nt) with a piscine (fugu) outgroup. The analyses supporteda sister group relationship between human and pig to the exclusion ofthemouse. Essentially the same relationshipwas addressed in awholegenome study of Cannarozzi et al. (2007), which included the dog andthe cow instead of the pig and used avian and marsupial outgroups.The study consisted of aa and nt analysis of all identifiable homo-logous nuclear protein-coding genes plus examination of the syntenyof these genes. The tree (outgroup, (Rodentia, (Primates, Carnivora/Artiodactyla))) was strongly supported in all analytical approachesapplied. It is noteworthy that the topology obtained in these twocomprehensive nuclear studies corresponds to that of the great majo-rity of earlier mitogenomic studies of placental mammals. Thetree supported by these analyses was not favored in the current



mitogenomic analysis, however. The genomic results of Cannarozziet al. (2007) are continuously updated by the inclusion of additionaltaxa and extended sequences. A recent update from this undertakingshows essentially an unresolved tripartite split between Xenarthra,Afroplacentalia and Boreoplacentalia. Within Boreoplacentalia thetree favors the monophyly of Glires and Archontoglires. Thus, in so faras this tree and the mitogenomic taxon sampling overlap, the resultsare mutually consistent.

Various relationships among placental mammals, including deepdivergences, were recently examined by applying retroposon analyses(Nishihara et al., 2006; Kriegs et al., 2006). The study of Kriegs et al.(2006) reported two retroposon inserts favoring the Epitheriahypothesis. The current results conform to that study with respectto other relationships. The other study (Nishihara et al., 2006) iden-tified an unresolved tripartite basal placental split between Afropla-centalia, Xenarthra and Boreoplacentalia. The mitogenomic findingsare essentially consistent with those of the retroposon analyses (Ni-shihara et al., 2006), except for the position of Chiroptera, which theretroposon analyses placed as sister to Carnivora/Perissodactyla. Pho-lidota was not included in that study.

In addition to the whole genome study of Cannarozzi et al. (2007)there are other recent and extensive studies that have addressedphylogenetic relationships among placental mammals. One of thesestudies (Nikolaev et al., 2007), which was partly based on codingsequences collected from 218 nuclear genes and partly on conservednon-coding sequences, examined the relationship among 10 placentalorders: Proboscidea, Tenrecidae, Xenarthra, Rodentia, Lagomorpha,Primates, Lipotyphla, Chiroptera, Carnivora and Artiodactyla. The exonbased analysis (68,262 codons) yielded a tree that is identical to theMLmitogenomic tree when the latter is reduced to the same orders asincluded in the nuclear study, and like the current study these exten-sive nuclear analyses could not reject the alternative relationshipsamong the three basal branches of placental mammals.

The next study used data from N30 ongoing genome projects toexamine indel differences in coding nuclear sequences (Murphy et al.,2007). This study, which primarily focused on the relationshipbetween Afroplacentalia, Xenarthra and remaining placental orders,supported a basal split between Notoplacentalia and Boreoplacentalia.The current mitogenomic results are largely similar to these nuclearresults (e.g. Fig. 6 in Murphy et al., 2007) with essentially all diffe-rences limited to one or two steps in the tree.

The hypothesis with a basal split between Boreoplacentalia andNotoplacentalia was more recently corroborated in two comprehen-sive nuclear studies that represented the same eight orders of placen-tal mammals (Hallström et al., 2007; Wildman et al., 2007). The moreextensive of these studies (Hallström et al. 2007) was based onN2×106 nt from 2840 homologous protein-encoding loci, the other(Wildman et al., 2007) on N1.4×106 nt from 1698 loci. Both studies,which are more comprehensive than any other sequence based phy-logenetic studies presented so far, yielded strong support to theBoreoplacentalia/Notoplacentalia tree.

The tenuous nature of basal placental relationships has been appa-rent in the great majority of the studies addressing this phylogeneticissue. The circumstance that data sets of limited alignment sizes andrestricted taxon representation have in some cases yielded strongsupport to a particular phylogeny underlines the stochastic nature ofsuch data sets.

The Boreoplacentalia/Notoplacentalia hypothesis has beenstrongly supported in recent extensive nuclear studies (Hallströmet al., 2007; Murphy et al., 2007; Wildman et al., 2007). Thus theaccumulating molecular evidence in support of the Boreoplacentalia/

Fig. 2. The mitogenomic tree of Placentalia. The unresolved relationship between Boreoplaclimited difference between trees [a], [b] and [c] in Fig. 1. Numbers on nodes indicate Bayesianposition of Solenodon draws attention to the circumstance that this positionwas determinedper site.

Notoplacentalia relationship speaks strongly in favor of accepting thisrelationship as being the overall most probable hypothesis for thebasal placental divergence. The mitogenomic study of Kjer andHoneycutt (2007), which had a somewhat more limited taxonomicrepresentation than the current study, favored a basal split Boreopla-centalia and Notoplacentalia, underlining the point raised by Arnasonet al. (2002) that different representation of taxa may have dramaticeffects on the basal topology of the placental tree. The versatility ofphylogenetic analyses is apparent among the mitogenomic resultspresented in Issue 1 of Systematic Biology, vol. 48 (1999). In this issuethe disparity of the mitogenomic trees is evident even among papersthat include the same authors. Except forWaddell et al. (1999) none ofthe trees is in accord with the Boreoplacentalia/Notoplacentalia hypo-thesis. However, to recover a phylogeny that resembled the Boreo-placentalia/Notoplacentalia tree the authors (Waddell et al., 1999)resorted to the exclusion of orders such as Lipotyphla, Rodentia andPrimates, the inclusion or exclusion of which was known to havepronounced effect on the topology of the placental tree.

3.2. Temporal estimates of basal placental divergences

The results of the mitogenomic estimates of the times of differentplacental divergences have been condensed into Fig. 3. Standarddeviations for the earliest estimates were in a span of 3–5 MYon eachside of the estimates. For estimates younger than 15MYA the spanwasabout 1 MY. In accord with recent and strongly supported nuclearconclusions the estimates are based on the Boreoplacentalia/Noto-placentalia hypothesis. The estimates placed the Metatheria/Eutheriadivergence at ≈135 MYBP and the basal placental divergence betweenBoreoplacentalia and Notoplacentalia at 102 MYA. In this context it isof interest that the 135 MYA estimate was obtained already by Fitchand Langley (1976) in analyses of seven nuclear protein sequencesrepresenting a number of placental orders.

Among nuclear studies there is a growing consensus that basaldivergences among placental mammals took place ≈100–105 MYA(e.g. Springer et al., 2003; Hallström et al., 2007; Murphy et al., 2007;Springer and Murphy, 2007). The current mitogenomic results concurwith this timing. The similarity between the studies of Springer et al.(2003), Murphy et al. (2007) and Springer and Murphy (2007) is notunexpected as the two latter studies used the estimates of Springeret al. (2003) for placing the age of this phylogenetically and biogeo-graphically important event. The 105 MYA estimate obtained bySpringer et al. (2003) is somewhat arbitrary, however, as it was basedon the mean of two figures, 65 MYA and 146 MYA, which were takento be related to this particular event. Thus, 65MYA is the age of the K/Tboundary, proclaimed by some paleontologists to mark the ordinaldiversification of placental mammals, while 146 MYA referred to anestimate by Penny et al. (1999) for the divergence between the hed-gehog (Erinaceidae) and other placentals and hence a tree that wasmarkedly different from that favored by Springer et al. (2003). Asapparent, the selection of some other equally probable figures tocalculate the time of the basal split would have resulted in a differentestimate. Despite the unconventional nature of the approach appliedby Springer et al. (2003) the 105 MYA estimate of the basal placentaldivergence happens to be in accord with estimates based on the use ofpaleontologically well-defined placental calibration points.

The Metatheria/Eutheria split was placed at 170–193 MYA in thestudy of Springer and Murphy (2007). Thus the eutherian branchleading to the basal placental split is ≈45 MY longer in that study thanin the current study and in the more extensive nuclear study ofHallström et al. (2007). The 170–193MYA estimates of theMetatheria/

entalia, Xenarthra and Afroplacentalia is in accord with the Bayesian analyses and theposterior probability. ⁎ indicates maximum (1.0) support. The dashed line showing the

only on the basis of analysis of the NADH2 gene. Scale bar shows number of substitutions



Eutheria divergence and the basal placental split are not in accordwith some recent paleontological conclusions (e.g. Benton and Donog-hue, 2007). It should be borne in mind, however, that the earlierdemarcations of evolutionary divergences are difficult to define bypaleontological approaches. Springer and Murphy (2007) refer toKumar and Hedges (1998) for the ≈180 MYA dating of the Metatheria/Eutheria divergence. This is problematic, however, as that estimate didnot rest upon phylogenetic analyses but on distance values that wereput into a temporal context. Springer andMurphy (2007) note that theeutherian estimates of Kumar and Hedges (1998) were allegedly basedon 658 nuclear genes. However, the estimate of the most basal pla-cental divergence obtained by Kumar and Hedges (1998), in this casethat between Xenarthra and remaining placentals (=the Epitheriahypothesis, placed at 129 MYA), was based on three sequences onlyand that between Metatheria and Eutheria (placed at 173 MYA) on 10sequences. Similarly, the paleontologically untenable dating of8.2 MYA for the divergence between Pongo and Homo was based onsix sequences. In comparison the split between the mouse and the rat,placed at 41 MYA, was based on 343 sequences, i.e. more than 50% ofall sequences used in the study.

Within Notoplacentalia the current estimates placed the splitbetween Afroplacentalia and Xenarthra at 99 MYA. The Paenungulatasplit between Hyracoidea and Sirenia/Proboscidea was placed at78 MYA and that between Sirenia and Proboscidea at 70 MYA. Thedepth of the intra-sirenian split between Trichechus and Dugong,≈48 MYA, is notable as is also the recent estimate, 5–6 MY, of the splitbetween the three elephantid genera Elephas, Loxodonta and Mam-muthus. Similarly, the great depths of basal divergences within Afroin-sectiphila are apparent. Basal divergences within Xenarthra were alsodeep. The split between Dasypodidae and Myrmecophagidae/Brady-podidae was dated to 79 MYA and that between Myrmecophagidaeand Bradypodidae to 68 MYA.

The basal Boreoplacentalia split between Laurasiaplacentaliaand Archontoglires was dated to 99 MYA, the same age as thatof the basal Notoplacentalia split. The Laurasiaplacentalia splitbetween Lipotyphla and remaining orders was dated to 88 MYA.Within the Erinaceidae the mitogenomic distance between the twospecimens ofH. suilluswas striking, yielding an estimate of N10MY forthe divergence between these two supposedly conspecific genomes.In the popular Bar coding segment of the COI gene the two Hylomysspecimens differed at 108 positions. In the same region Pan andHomo differed at 79 positions and the Asiatic and African elephantsat 44.

The Laurasiaplacentalia includes the only recent order, Cetacea,that is definitely of Tertiary age. However, considering the position ofCetacea within Artiodactyla as the sister group of Hippopotamidae(Irwin and Arnason, 1994; Ursing and Arnason, 1998) forming theCetancodonta (Arnason et al., 2000a), the justification for recognizingCetacea as a separate order might even be disputable. The Cetanco-donta split between Hippopotamidae and Cetacea was dated to54 MYA. In the Perissodactyla the divergence between Tapiridae andRhinocerotidae was estimated at 48 MYA. This estimate, which isconsistent with the fossil record of Ceratomorpha (Prothero andSchoch, 1989), corroborates the current placement of the E/R calibra-tion point set at 55 MY.

The Glires split between Rodentia and Lagomorpha was dated to83 MYA and that between South American rodents (Cavia) and itsclosest African relative (Thryonomys) to 53 MYA.

Fig. 3. Temporal estimates within Placentalia based on mitogenomic data and the BorAfroplacentalia; brown: Xenarthra; green: Archontoglires; blue: Laurasiaplacentalia. The estion nt sequences yielded estimates that differed by less than one standard deviation from thedivergences have particularly problematic among taxa where paleontologically solid calibralargely limited to the Laurasiaplacentalia. Outside Laurasiaplacentalia some recently discoverelevance of the current molecular estimates. The primate estimates are in accord with these

3.3. Primate, notably hominid, issues

Within Primates the divergence between Strepsirrhini and Simii-formes was placed at 80 MYA. In the context of the current mito-genomic estimates it is of interest that this particular divergence hasbeen dated to 81.5 MYA in non-molecular calculations based onspecies preservation and the primate fossil record (Tavaré et al., 2002).

The split between Malagasy and non-Malagasy Strepsirrhini wasdated to 70 MYA. The intra-Malagasy split between Daubentonia andthe lemurs was dated to 64 MYA. The split between Platyrrhini (NewWorld monkeys) and Catarrhini (Old World monkeys and great apes)was placed at ≈55 MYA and the catarrhine split between Cercopithe-coidea and Hominoidea at 39 MYA.

For self-evident reasons the relationship between Homo and thegreat apes has attracted the attention of scholars and the interestedlayman. For a long time it was common wisdom that man was onlydistantly related to the great apes, although the early serologicalstudies of George H. F. Nuttall were inconsistent with this under-standing. Actually the study of Nuttall and Strangeways (1904, p. 319)had shown a closer relationship between Homo and Pan than betweenHomo and Gorilla, and with Pongo still more distantly related. Amongother primate results obtained by G. H. F. Nuttall was the unequivocaldemonstration of a closer relationship between Hominoidea andCercopithecoidea than between Hominoidea and Cebidae (Nuttall,1901a,b, 1904, p. 163). In retrospect the very title of Nuttall's (1901a)paper, “The new biological test for blood in relation to zoologicalclassification”, ought to have attracted greater attention within thefield of zoological systematics and classification than became the case.

In the early 1960sMorris Goodman (cf. Goodman,1996) conductedextensive molecular studies that conclusively demonstrated thegrouping of Gorilla, Pan and Homo to the exclusion of Pongo. A fewyears later Sarich and Wilson (1967) presented a study in which theyadvocated a split between Gorilla, Pan and Homo ≈5 MYA. Thisrelationship was later addressed by Sibley and Ahlquist (1987)applying DNA/DNA hybridization in a study that showed a sistergroup relationship between Pan and Homo to the exclusion of Gorilla.In addition the authors dated the divergence between Pan and Homoto 5 MYA. The study included no reference to Nuttall and Strangeways(1904) and their pioneer findings. Despite the discomfort of a fewheretic primatologists the view of the recent (5 MYA) split betweenPan andHomo gradually became a dogma that remained unchallengedfor some 30 years until reexamined in studies that applied stringentlydefined paleontological calibration points (Arnason et al., 1996, 1998).

The current estimates of hominid divergences placed the splitbetween Pongo and Hominini (Gorilla, Pan, Homo) at 20 MYA, thatbetween Gorilla and Pan/Homo at ≥11.5 MYA and that between Panand Homo at ≥8 MYA. As the result of an altered tree topology thecurrent estimates are all more recent than those of previousmitogenomic studies (e.g. Arnason et al., 1996,1998, 2000a,b; Arnasonand Janke 2002) that have addressed these relationships andchallenged the traditional dogma of a Pan/Homo divergence 5 MYA.

The estimate of the divergence between Gorilla and Pan/Homo≥11.5 MYA is corroborated by the paleontological age, 12.5 MY, of afossil that may precede the divergence between Gorilla and Pan/Homo(Pickford and Senut, 2005), and the age, 10–10.5 MY, of the newlydescribed fossil Chororapithecus abyssinicus with a position on thebranch leading to Gorilla (Suwa et al., 2007). Similarly, the molecularlyrefutable Pan/Homo split 5 MYA has also been paleontologically

eoplacentalia/Notoplacentalia hypothesis. Colorings are in accord with Fig. 2; red:mates were based onMULTIDIVTIME and aa sequences. Estimates performed with r8s oraa results. As discussed by Graur and Martin (2004) molecular estimates of evolutionarytion points are missing. Among placental mammals well-defined calibration points arered primate fossils (see text and Fig. 4) have particular interest for judging the generalfinds.


refuted by the description of fossils older than 5 MY residing onthe branch leading to Pan/Homo (Samburupithecus kiptalami, Ishidaand Pickford, 1997) or Homo: Orrorin tugenensis (Senut et al., 2001),Ardipithecus ramidus (Haile-Selassie, 2001) and Sahelanthropus tscha-densis (Brunet et al., 2002). In addition to the fossil age, ≈6 MY, ofO. tugenensis, its upright gait was essential to the rebuttal of the 5MYAdogma. Thus, there is now paleontological evidence directly con-nected to both the Pan/Homo split and the divergence between Gorillaand Pan/Homo that corroborates the molecular challenge of thetraditional dogma. The positions of the various fossils discussed havebeen included in the hominid tree shown in Fig. 4. The figure includesalso the catarrhine fossil Algeripithecus (Godinot and Mahboubi 1992)that provides the minimum age, 48 MY, for the divergence betweenCatarrhini and Platyrrhini.

The Pan/Homo dogmawas defended in a recent study (Kumar et al.2005) that did not consider the importance of the bipedalism of O.tugenensis for the definition of the Pan/Homo split. Instead the authorsused the geological age of the Oligocene/Miocene boundary, 23.8 MY,as the most probable age of the calibration point applied, viz. thedivergence between Cercopithecoidea and Hominoidea. By includingthis age the authors obtained an estimate of 4.98 MYA as the mostprobable time of the divergence between Pan and Homo. In additionto this dating the authors provided sequential estimates of the Pan/Homo divergence based on age increments of 2 MY for the Cerco-pithecoidea/Hominoidea split, ending at 35MYA. The 35MYage of thecalibration point yielded an upper estimate of 7.02 MY for the Pan/Homo split, thereby providing the reader with a range of 4.98–7.02 MYA for this particular divergence.

Themolecular approach behind the estimate of Kumar et al. (2005)of the Pan/Homo split is in accord with conventional approaches. Thissuggests that the estimates as such were not related to the moleculardata as such but rather to the temporal range within which theauthors chose to place the calibration point used. As compared to thestudy of Kumar et al. (2005) the current estimates placed the diver-gence between Cercopithecoidea and Hominoidea at ≈38 MYA.

The time of the divergence between Pan and Homo is of particularinterest due to the circumstance that the deepest divergence amongrecent humans is taken to be ≈1/30 of that between Pan and Homo. Onthese grounds the mt root of human evolution (Mitochondrial Eve)has traditionally been placed at ≈150,000–170,000 YA (e.g. Vigilantet al., 1991). With the split between Pan and Homo placed at ≈8 MYAand applying the same approach the age of mt Eve becomes placed at≈270,000 YA.

The traditional age of mt Eve has been used to estimate the time ofdivergence between the Neanderthal and the ancestors of recenthumans (Krings et al., 1997). According to Krings et al. (1997) the

Fig. 4. Hominid relationships and the positions of fossils related to hominid divergences.Numbers above open arrowheads show the estimated times of different divergences.Numbers 1–6 show the positions of individual fossils.1:Algeripithecus; 2: Chororapithecus;3: Samburupithecus; 4: Sahelanthropus; 5: Orrorin; 6: Ardipithecus.

separation between the Neanderthals and the ancestors of recenthumans took place ≈600,000 YA, thereby precluding the possibility ofgenetic exchange between the two lineages. The 600,000 YA datinghas subsequently become a standard reference in textbooks and evenin the scientific literature.

The conclusions of Krings et al. (1997) were based on a comparisonbetween one mt Neanderthal sequence (≈370 nt) and a sample of thecorresponding sequences of recent humans. The mean differencewithin the recent sample was 8, while the mean difference betweenthis sample and the Neanderthal sequence was ≥3× larger. Based onthe belief that the mean of 8 differences within the sample of recenthumans was representative for the age of mt Eve the authors conclu-ded that the split between the Neanderthal and the ancestors of recenthumans had taken place ≈600,000 YA. The approach applied by Kringset al. (1997) is inherently wrong, however, because the age of mt Everefers to the deepest divergences among recent humans and not to themean within any arbitrary collection of samples, which, in this parti-cular case, originated to a considerable extent from recent Europeans.If Krings et al. (1997) had been cognizant of the definition of mt Evethey might have recognized that the greatest differences within theirsample of recent humans were at the same level as the mean diffe-rence between the Neanderthal and the recent samples. Thus, incontrast to the statement of Krings et al. (1997), their approach shouldhave led them to conclude that the separation between the Nean-derthal and the ancestors of recent humans only marginally precededthe earliest divergences among recent humans. That interpretation isconsistent with the mt study of Ovchinnikov et al. (2000) that did notprovide statistically conclusive evidence for the position of the Nean-derthal outside recent humans. Provided that the current dating of mtEve is reasonably correct, this might suggest that the divergencebetween the Neanderthals and the ancestors of recent humans tookplace some 25,000–50,000 years earlier than the earliest divergencesamong the latter, i.e. ≈300,000–325,000 YA. That dating is in accordwithWall and Kim's (2007) estimate of 325,000 YA for the split time ofthe Neanderthal–European population based on nuclear data.

The estimate of Wall and Kim (2007) was based on the Nean-derthal data set established by Noonan et al. (2006) and not thatpresented by Green et al. (2006), which Wall and Kim (2007) showedto have contamination rates of 72–85% of recent DNA, therebyinvalidating the scientific value of this data set other than demon-strating the need for general technical thoroughness when workingwith ancient DNA. The conclusions of Wall and Kim (2007) wereupheld in the study of Hodgson and Disotell (2008). The difficultiesassociated with the contamination in the data of Green et al. (2006)draw attention to another problematic study of old DNA, which, in thiscase, was based on a 2400 year old mummy (Pääbo, 1985). Theauthenticity of the mummy sequence described (allegedly establishedby direct cloning of a single fragment that was 3.4 kb long) wasclaimed on the grounds that the sequences flanking its Alu repeat hadnot been described before. Although this is a non sequitur argument itis of interest that, as demonstrated by Del Posso and Guardiola (1989),these sequences actually corresponded to those of a clone that hadalready been described at the time of the mummy study. It thusappears that the “mummy” sequence was that of a clone that wascommonly used as a labeled probe in different laboratories, includingthat inwhich the mummy study was carried out. As politely remarkedby Del Posso and Guardiola (1989) the similarity between themummyDNA and the previously described clone “strikingly contrasts with thefrequency of polymorphism observed in modern population studies”.The author of the mummy study (Pääbo, 1985) did not reply to thecomments of Del Posso and Guardiola (1989). In retrospect it isprobable that the retraction of the “mummy” paper would havecounteracted the publication of a series of uncritical ancient DNAstudies that were later shown to be erroneous.

The current estimates placed the divergence between Gorilla andPan/Homo ≈3.5 MY earlier than that between Pan and Homo. A


temporal interval of 2.1 MY for this divergence was recently used byEbersberger et al. (2007) in a study that examined the sorting ofancestral polymorphisms among the hominid lineages and theproportions of sequence alignments that supported different relation-ships. The difference between the two intervals is immaterial,however, for the results of Ebersberger et al. (2007), which werebased on 23,210 individual DNA sequence alignments and a total ofmore than 12×106 autosomal nt positions. Depending on the data andapproaches used the proportions of the favored relationships coulddiffer slightly, with the most conservative approach showing that≈23% of the data did not identify Pan and Homo as sister groups. Thedata that did not support the Pan/Homo grouping supported inessentially the same proportions a Pan/Gorilla or a Gorilla/Homogrouping. As concluded by Ebersberger et al. (2007) the results, whichare in accord with common population-genetic procedures and allelicsorting, appear to provide a more plausible explanation to thegenomic mosaicism of Gorilla, Pan and Homo than that involvinghybridization at different times among these three taxa as suggestedby Patterson et al. (2006).

Retroposon inserts have become a recent tool to resolve placentalrelationships that have been difficult to establish by using sequencedata. Some of these studies have been referred to above. It would beexpected that retroposon inserts follow the same sorting rules as dogenetic polymorphisms. However, the cleanliness of some retroposonresults seems to suggest that retroposon sorting does not follow thesame pattern as that of alleles as demonstrated for example in thehominid study of Ebersberger et al. (2007). There are neverthelesssome early exceptions from retroposon cleanliness as apparent in thecetartiodactyl study of Shimamura et al. (1997).

3.4. Biogeographical considerations

Any discussion of biogeography is directly connected with anunderlying phylogeny, estimates of divergences related to this phylo-geny and paleontology. Basal relationships in the tree shown in Fig. 3are consistent with recent sequence analyses of extensive nuclear data(Hallström et al., 2007; Wildman et al., 2007). The mitogenomicestimates of these relationships are also in reasonable agreement withthe datings based on this data, while the more limited taxon repre-sentation of the nuclear data sets precludes detailed comparisonbetween the mitogenomic and the nuclear estimates in less basalparts of the placental tree.

The precision of molecular estimates of evolutionary divergenceshas been subjected to extensive discussion both among paleontolo-gists and molecular phylogeneticists. Graur and Martin (2004) chal-lenged several claims coming from the molecular camp to which theybelong themselves. In particular they criticized the application ofsingle calibration points to the dating of divergences that were at agreat distance from the calibration point used. Among the issuesdiscussed by Graur and Martin (2004) was the use of the Diapsida/Synapsida split set at 310 MYA as a single calibration point to date aseries of divergences among placental mammals (e.g. Hedges et al.,1996). The different placental calibration points used in the currentstudy are discussed under Section 2.2. At the establishment of thesecalibration points particular attention was paid to the strength of thepaleontology behind each particular point. However, due to thedifferent solidity of the paleontology of different placental ordersthere is a marked bias in the positions of the calibration points in theplacental tree.

The estimates shown in Fig. 3 place the origin of all placentalorders except Cetacea in the Cretaceous. The time of the origin ofplacental orders was addressed by Foote et al. (1999) on the basis ofcalculations of rates of fossil preservation. Based on these calculationsFoote et al. (1999) found it unlikely that placental orders in generalarose much earlier than their oldest fossil record, an understandingthat led to the conclusion that the great majority of extant placental

orders had arisen in Tertiary times. According to the authors thisconclusionwas maintained even after the most generous treatment ofthe opposing hypothesis, viz. a Cretaceous origin of placental orders.Considering the fact that the minimum paleontological age of theCetacea is ≥53MY (e.g. Gingerich, 2005) and that it is conceivable thatthe divergence between Hippopotamidae and Cetacea, and ruminantArtiodactyla and Cetancodonta took place some million years earlier,the conclusions of Foote et al. (1999) imply that the divergencesbetween essentially all extant orders of placental mammals werecompressed into a temporal window of less than 10 MY. If so, it isunlikely that even the most extensive molecular data available willhave the power of resolving the ordinal and superordinal tree ofplacental mammals. In the context of the calculations of preservationcarried out by Foote et al. (1999) it is of interest that Tavaré et al.(2002) have also performed calculations of divergence times based onpreservation. In this case the calculations of primate divergences ledto conclusions that agree with molecular estimates and recent pro-gress in primate paleontology

Although calculations of the kind performed by Foote et al. (1999)may have relevance for orders with exceptionally extensive fossilrecord the generality of the conclusions is in disharmony with thepresence of four placental orders, Primates, Rodentia, Carnivora andTenrecidea, on Madagascar, despite their non-existent fossil record.Another case that underlines the possible extension of a missing fossilrecord is the early Miocene South American fossil Necrolestespatagonensis, which was recently subjected to a detailed morpholo-gical analysis by Asher et al. (2007). In spite of the completeness of thefossil its phylogenetic position was not established. A particularproblem with Necrolestes is that it shares a series of characters withplacentals (e.g. Chrysochloridea) as well as with Metatheria and to alesser extent with Prototheria. Irrespective of the precise position ofNecrolestes in the mammalian tree it is probable that some 80 MY ormore of its evolutionary history remain to be detected. The Necrolestescase may suggest that that the S American fossil record is still largelyunrevealed.

The presence of placental mammals onMadagascar in spite of theirlacking fossil record, and the hitherto untraced fossil record of Ne-crolestes in S America, constitute two cases that might deserve con-sideration in the context of any no-fossil/no-life hypothesis. Similarly,the circumstance that no recognized orders of extant placentalmammals have arisen during the last 50 MY makes it unlikely thatordinal differences among placental mammals should have originatedde novo in any particular 5–10 MY epoch, including that following theK/T boundary.

Recently Kitazoe et al. (2007) presented time estimations ofvarious divergences among placental mammals. In Fig. 1 of that paperthe authors presented estimates of placental divergences based ontwo different approaches. One approach placed the basal placentaldivergence at 122 MYA, the other at 84 MYA. Not unreasonably, basedon the 84 MYA estimate, the authors placed the origin of mostplacental orders close to the K/T boundary when applying theirfavored MVS-FIR approach. As an example of the refinement of theanalyses the authors depict in their Fig. 2 the changes in evolutionaryrates from the root of the placental tree to recent Laurasiaplacentalia,as represented by the fin whale. The resolution provided by the MVS-FIR approach allowed the demonstration of an extraordinary accel-eration of the evolution of Laurasiaplacental mt genomes during aperiod of three MY, 73–70 MYA. Actually the total of evolutionarychanges accumulating during this interval was at a similar level as thataccumulating during the 15 MY period 25–10 MYA. While thesophistication of the MVS-FIR approach should not be disregarded,Kitazoe et al. (2007) might have considered that the Laurasiaplacentalcalibration points that were included affect in particular the estimatesof the Laurasiaplacental splits, while the estimate of the split betweenLaurasiatheria and Archontoglires is affected by the contribution ofboth groups. Therefore, a too-recent position of the rabbit/pika


Archontoglires calibration point used by Kitazoe et al. (2007) wouldcompress the length of the common Laurasitherian branch, therebycreating the impression of an acceleration in the 73–70 MYA interval.

Kitazoe et al. (2007) placed the age of the rabbit/pika calibrationpoint in the interval 36–55 MYA. The lower bound (55 MYA) may betoo recent, however, as indicated by the fact that all of their appro-aches, including MVS-FIR, placed this divergence at the constrainedage of 55 MYA. The current estimates placed this divergence at62 MYA, an estimate that also Kitazoe et al. (2007) might haveobtained provided they had not constrained the maximum age of it to55 MY. With respect to Kitazoe et al.'s (2007) placement of thiscalibration point at 55 MYA it is of interest that Douzery et al. (2003),who reported divergence times of placental orders close to the K/Tboundary, placed its age at 37 MYA.

Kitazoe et al. (2007) refer to some papers that have reported theroot of living placental mammals to be within the range of 100–140 MYA. Waddell et al. (1999) is one of these references. However, inthat paper the basal placental divergence is estimated at 169 MYA,even in the absence of the hedgehog, which usually occupied a stillmore basal position in the mitogenomic studies performed at thattime. Thus the studies of Waddell et al. (1999) and Kitazoe et al. (2007,which includes Waddell as coauthor) cover a range of 85 MY or morefor the basal mitogenomic divergence of placental mammals. Waddellet al. (1999) estimated the split between the armadillo and the Africanelephant to 122 MYA. This estimate was then arbitrarily placed at100 MYA to fit the geological time of the opening-up of the SouthAtlantic, thereby creating a new calibration point, “Atlantogenata”.The “Atlantogenata” calibration point set at 100 MYA was in turnused to recalibrate other divergences in the mitogenomic tree of 12species. With this approach Waddell et al. (1999) inter alia moved thedivergence between the rabbit and remaining species from 151.5 to124 MYA. After referring to the estimates of Springer (1997) andKumar and Hedges (1998) for general consistency, Waddell et al.(1999) declared that the dating of the split between the armadillo andthe elephant was one of the first demonstrated coincidences betweenan estimated splitting time and the age of a major biogeographicevent. As just mentioned the correctness of this statement is open toquestion, and consequently also the use of the name “Atlantogenata”,because the estimate ofWaddell et al. (1999) actually placed the age ofthe divergence between Xenarthra (armadillo) and Afrotheria (Africanelephant) at a date that was temporally inconsistent with the ope-ning-up of the South Atlantic. As also evident the “Atlantogenata”hypothesis is unrelated to the Boreoplacentalia/Notoplacentaliahypothesis, as “Atlantogenata” does not refer to the basal split amongplacental mammals but to a split within Notoplacentalia itself.

Considering the problems associated with the conclusions of Footeet al. (1999) and the adaptability of the study of Kitazoe et al. (2007)we will base our biogeographical discussion on the assumption thatthere is a reasonable consistency between the mitogenomic estimatesof the different divergences in the placental tree and the evolutionaryhistory behind these divergences. It should be borne in mind, how-ever, that the mitogenomic results as such are representative only forthis particular data set and the current taxon sampling.

Apart from recent hypotheses that have involved Gondwana in thediscussion of the dispersal of placental mammals (Arnason et al.,2000b; Murphy et al., 2001) it has been commonly believed that basalplacental diversification took place in Laurasia. This notion has partlybeen based on the location of the earliest stem eutherians on thissupercontinent (Ji et al., 2002) and partly on the notion of phylo-genetic links between the 85–90 MY old Asian located “Zhelestidae”and Laurasiaplacentalia and Glires (Lagomorpha) (e.g. Archibald,2003). More specifically there are suggestions for a relationship bet-ween zhelestids and lipotyphlans/cetferungulates and betweenzalambdalestids and Glires (Archibald et al., 2001). These conclusionshave been challenged, however, in other morphological studies thathave placed the “Zhelestidae” as stem eutherians, i.e. outside

placentals (Asher et al., 2005; Wible et al., 2007). The latter of thesestudies included a tree that showed the relationship among someplacental superordinal groups and orders and the times of theirdivergences, which were placed close to the K/T boundary or later(Wible et al., 2007). Phylogenetic conclusions related to stem euthe-rians can neither be corroborated nor challenged by molecularapproaches. The pronounced and persisting inconsistency betweenthe molecular trees of placental mammals and the correspondingmorphological phylogenetic tree is troublesome, however, for theunification of themorphological andmolecular views on the evolutionof placental mammals.

The Notoplacentalia/Boreoplacentalia relationship allows for twoprimary hypotheses related to the early dispersal of placentalmammals.One of these hypotheses assumes placental origin in Laurasia, the otherbasal placental splits on southern continents (Gondwana).

In accord with the understanding of early eutherian evolution inLaurasia combinedwith theNotoplacentalia/Boreoplacentaliahypothesisit would be logical to assume a Laurasian split between Notoplacentaliaand Boreoplacentalia coincident with dispersal of ancestral Notoplacen-talia to Africa followed by a split between Xenarthra and Afroplacentaliaconnected to the geological separation of S America and Africa.

During the last few years there has been a remarkable and continuousprogress in thedelineationof thegeological history following thebreak-upof Pangaea (Scotese, 2002; ODSN Plate Tectonic Reconstruction Service;Gondwana Animation, International Institute for Aerospace Survey andEarth Science). This information underlines the presence of pathways ineither direction between Laurasia (Europe) and Africa 90–110 MYA andlater, even though these pathwaysmay not have been continuous at anyparticular time. It is thus apparent that thewaters separating EuropeandAfrica at this time were not an impassable hindrance for dispersal ofplacental mammals between Laurasia and Africa.

The Boreoplacentalia/Notoplacentalia scenario under discussionmay satisfy the recent notion (cf. Wible et al., 2007) that recentplacental orders fall outside stem eutherians. The temporal frame-work of the scenario is not consistent with the conclusions of Wible etal. (2007), however. The difficulties in reaching agreement betweenmolecular and morphologically based phylogenies become still moreapparent when the placental relationships presented by Wible et al.(2007) are compared with those of recent molecular studies, some ofwhich are based on millions of nucleotides representing each indivi-dual taxon. If we assume that the current molecular picture of pla-cental relationships is reasonably correct the inconsistency betweenthis picture and the morphological tree of recent taxa may raise thequestion if morphological approaches are more suitable for obtainingcorrect phylogenies among extinct taxa than among recent ones wherecomplete specimens are available for morphological comparisons.

The Notoplacentalia/Boreoplacentalia relationship and the molecu-lar estimates resulting from this tree have not previously been placed inthe context of a Gondwanan origin of placental mammals, as both thestudy of Arnason et al. (2000b) and that of Murphy et al. (2001) basedtheir discussion on other phylogenies. The focus of the studies ofArnasonet al. (2000b) andMurphyet al. (2001) differed in that the studyof Arnason et al. (2000b) primarily addressed the complex biogeographyof thePrimates,whileMurphyet al. (2001) concentrated theirdiscussionon a basal divergence betweenAfroplacentalia and remainingplacentalsand its implications for placental biogeography.

It has been traditionally understood that early primate evolutiontook place in Asia. That notion is consistent with the present occur-rence of the two other Archonta orders, Dermoptera and Scandentia,in SE Asia. This might also suggest a Laurasian (SE Asian) Haplorrhinisplit between Tarsiiformes and Simiiformes with a subsequentdispersal of Simiiformes to Africa and later to S America. Similarlythe Laurasian origin of the Strepsirrhini assumes an early dispersal toAfrica followed by a subsequent colonization of Madagascar givingrise to the Lemuriformes. The problem with this scenario is not anearly primate dispersal from Laurasia to Africa, however, but the time


of primate colonization of Madagascar, which the mitogenomicestimates place in the interval of 70–63 MYA.

The current extension of the primate hypothesis proposed byArnason et al. (2000b) rests upon the same primate relationships asunderlying that paper but the estimates of the different divergencesdiffer as the result of the changed topology of the placental tree. Theupdated hypothesis posits that stem Archonta (Primates/Dermoptera/Scandentia) dispersed to S America ≈90MYA and that their subsequentevolution and dispersalwas closely related to the break-up of S America,Antarctica, the Indian subcontinent, and Madagascar. According to thisscheme the ancestors of both Dermoptera and Scandentia becameisolated on the Indian subcontinent at the break-up of Antarctica andMadagascar/Indian subcontinent and reachedAsiavia thedockingof thesub-Indian continent with Asia 60–65 MYA. The outlined scenariorelates also the Strepsirrhini/Haplorrhini split to the break-up of sou-thern continents.With respect to Malagasy Strepsirrhini their ancestorsare postulated tohavebecome isolatedonMadagascar at thebreak-upofMadagascar and the Indian subcontinent, with the Lorisiformes takingthe Indian route to Asia and later to Africa. Consistent with a basalHaplorrhini split between Tarsius and Simiiformes the proposed schemeposits that also this divergence is related to the break-up of southerncontinents with Tarsius taking the Indian route to Asia. This would leavestem-Simiiformes in S America and connect the Platyrrhini/Catarrhinisplit with Simiiformdispersal from SAmerica to Africa, giving rise to theCatarrhini. The estimate of this split, ≈55 MYA, is not refuted by the ageof the oldest African simiiform fossils hitherto detected (Godinot andMahboubi, 1992).

The age of the split between Lagomorpha and Rodentia was datedto ≈93 MYA and the deepest Rodentia divergence to ≈88 MYA. Thesplit between African rodents and their closest relative, the (African)cane rat was dated to 53 MYA. This suggests that Rodentia had a longtime of evolution before their colonization of S America. It is of interestthat the estimate of the split between African and S American rodentsis close to that between Platyrrhini and Catarrhini. This might suggestthat rodents and primates exploited a similar route for their S Atlanticdispersal, albeit in opposite directions.

As mentioned above Murphy et al. (2001) based their biogeographi-cal discussion on a basal placental split between Afroplacentalia andXenarthra/Boreoplacentalia, which they associated with the geologicalseparation between Africa and S America ≥100 MYA. The phylogenyupon which Murphy et al. (2001) based their discussion has now beenreplaced by the Boreoplacentalia/Notoplacentalia hypothesis and theirdiscussion is therefore not relevant in all details, among which theauthors mentioned the potential phylogenetic connection betweensome Boreoplacentalia and the 85–90 MYold fossils of zalambdalestidsand zhelestids. Having placed the ancestors of Boreoplacentalia in SAmerica the authors outlined two possible routes along which Boreo-placentalia might have entered Laurasia, one along the Proto-AntillesIsland Arc, the other via the Indian subcontinent. A dispersal ofBoreoplacentalia to Laurasia at this time via the Indian subcontinent isproblematic, however, in the light of the fact that that the Indiansubcontinent did not dock with Asia until ≈60–65 MYA.

The Notoplacentalia/Boreoplacentalia hypothesis as such doesneither support nor refute a postulate suggesting a Gondwanan originof placental mammals and a later placental dispersal to Laurasia alongthe Proto-Antilles Island Arc, from Africa to Europe, or via the Indiansubcontinent to Asia. However, considering the extension of the basaleutherian fossil record in Laurasia it is probable that the placentalsarose on this continent although their definite sister group stillremains to be identified.

4. Conclusions

(1) The current mitogenomic analyses identified four basal cladesof Placentalia: Afroplacentalia, Xenarthra, Archontoglires andLaurasiaplacentalia. Archontoglires and Laurasiaplacentalia

joined on a common branch, Boreoplacentalia. The relationshipbetween Afroplacentalia, Xenarthra and Boreoplacentalia wasnot resolved as compared to recent phylogenomic analyses thatpreferably join Afroplacentalia and Xenarthra on a commonbranch (Notoplacentalia).

(2) The position of Erinaceidae (hedgehogs, moon rats) has beendifficult to establish in the mitogenomic placental tree. Thecurrent analyses with an extended erinaceid sampling placedthe Erinaceidae together with other lipotyphlans on a commonbranch that was sister to remaining Laurasiaplacentalia. Thecost of moving the Erinaceidae to other positions in the treewas limited, a circumstance that is presumably related to theextreme mitogenomic distances within this family.

(3) Mitogenomic estimates of divergence times placed the basaldivergence among placental mammals at ≈102 MYA. The sameestimates suggest that Cetacea is probably the only placentalorder that has originated in Tertiary times, i.e. after the K/T(Cretaceous/Tertiary) boundary. The estimates challenge somepaleontological notions that claim that the fossil record yields alargely correct picture of the origin and evolution of placentalorders.

(4) The molecular estimates of hominid divergences placed thesplit between Pan and Homo at ≥8 MYA and that betweenGorilla and Pan/Homo at ≥11.5 MYA. Both these datings are inaccord with recent progress in primate paleontology.

(5) Based on the previous understanding of a Pan/Homo split 5MYAthe age of “Mitochondrial Eve” has been placed at ≈170,000 YA.With the age of the Pan/Homo split placed at ≥8 MYA the age ofmt Eve becomes moved to ≈270,000 years.

(6) Earlier studies haveplaced the split between theNeanderthal andthe lineage leading to recent humans about four times the age ofmt Eve. Examination of the assumptions upon which theseconclusions were based showed that these assumptions wereinherently wrong, however. The conceptual problems connectedwith the earliest Neanderthal study have also drawn attention tothe problematic nature of some other studies of ancient DNA.

Acknowledgements

We thank professors Birger Bergh (Latin) and Kent Larsson (Geology)for valuable discussions and comments. The studywas supportedby TheSwedishResearchCouncil, TheNilsson-Ehle Foundation, TheErik Philip-Sörensen Foundation and by The TMR program of the EuropeanCommission, Grant ERB-FMRX-CT98-0221.

References

Archibald, J.D., 2003. Timing and biogeography of the eutherian radiation: fossils andmolecules compared. Mol. Phylogenet. Evol. 28, 350–359.

Archibald, J.D., Rose, K.D., 2005. Wombwith a view: the rise of placentals. In: Rose, K.D.,Archibald, J.D. (Eds.), The Rise of Placental Mammals. The Johns Hopkins UniversityPress, Baltimore, pp. 1–8.

Archibald, J.D., Averianov, A.O., Ekdale, E.G., 2001. Late Cretaceous relatives of rabbits,rodents, and other extant eutherian mammals. Nature 414, 62–65.

Arnason, U., Janke, A., 2002. Mitogenomic analyses of eutherian relationships.Cytogenet. Genome Res. 96, 20–32.

Arnason, U., Gullberg, A., Widegren, B., 1991. The complete nucleotide sequence of themitochondrial DNA of the fin whale. J. Mol. Evol. 33, 556–568.

Arnason, U., Gullberg, A., Janke, A., Xu, X., 1996. Pattern and timing of evolutionarydivergences among hominoids based on analyses of complete mtDNAs. J. Mol. Evol.43, 650–661.

Arnason, U., Gullberg, A., Janke, A., 1998. Molecular timing of primate divergences asestimated by two nonprimate calibration points. J. Mol. Evol. 47, 718–727.

Arnason, U., Gullberg, A., Gretarsdottir, S., Ursing, B., Janke, A., 2000a. Themitochondrialgenome of the sperm whale and a new molecular reference for estimatingeutherian divergence dates. J. Mol. Evol. 50, 569–578.

Arnason, U., Gullberg, A., Schweizer Burguete, A., Janke, A., 2000b. Molecular estimatesof primate divergences and new hypotheses for primate dispersal and the origin ofmodern humans. Hereditas 133, 217–228.

Arnason, U., et al., 2002. Mammalian mitogenomic relationships and the root of theeutherian tree. Proc. Natl. Acad. Sci. U. S. A. 99, 8151–8156.


Arnason, U., et al., 2006. Pinniped phylogeny and a new hypothesis for their origin anddispersal. Mol. Phylogenet. Evol. 41, 345–354.

Asher, R.J., 2005. Insectivoran-grade placental mammals: character evolution and fossilhistory. In: Rose, K.D., Archibald, J.D. (Eds.), The Rise of Placental Mammals. TheJohns Hopkins University Press, Baltimore, pp. 50–70.

Asher, R.J., et al., 2005. Stem Lagomorpha and the antiquity of Glires. Science 307,1091–1094.

Asher, R.J., Horovitz, I., Martin, T., Sánchez-Villagra, M.R., 2007. Neither a rodent nor aplatypus: a reexamination of Necrolestes patagonensis Ameghino. Am. Mus. Novit.3546, 1–40.

Benton, M.J., Donoghue, C.J., 2007. Paleontological evidence to date the tree of life. Mol.Biol. Evol. 24, 26–53.

Brunet, G.M., et al., 2002. A new hominid from the Upper Miocene of Chad, CentralAfrica. Nature 418, (37 authors)145–151.

Butler, P.M., 1988. Phylogeny of the insectivores. In: Benton, M.J. (Ed.), The Phylogenyand Classification of the Tetrapods. Mammals, vol. 2. Clarendon Press, Oxford,pp. 117–141.

Cannarozzi, G., Schneider, A., Gonnet, G., 2007. A phylogenomic study of human, dog,and mouse. PLoS Comput. Biol. 3 (1), e2.

Del Posso, G., Guardiola, J.,1989.MummyDNA fragment identified. Nature339, 431–432.D'Erchia, A.M., Pesole, G., Saccone, C., Arnason, U., 1996. The complete mitochondrial

genome of the guinea pig: rodent polyphyly and phylogeny of mammals. Nature381, 597–600.

Douzery, E.J.P., Delsuc, F., Stanhope, M.J., Huchon, D., 2003. Local molecular clocks inthree nuclear genes: divergence times for rodents and other mammals andincompatibility among fossil calibrations. J. Mol. Evol. 57, S201–S213.

Ebersberger, I., Galgoczy, P., Taudien, S., Taenzer, S., Platzer, M., von Haeseler, A., 2007.Mapping human genetic ancestry. Mol. Biol. Evol. 24, 2266–2276.

Felsenstein, J., 1993. Phylogenetic Inference Programs (PHYLIP), University of Washing-ton and University Herbarium, University of California, Seattle WA/Berkeley, CA.

Fitch, W.M., Langley, C.H., 1976. Protein evolution and the molecular clock. Fed. Proc. 35,2092–2097.

Flynn, J.J., Galiano, H.,1982. Phylogeny of Early Tertiary Carnivora, with a description of anew species of Protictis from the Middle Eocene of northwestern Wyoming. Am.Mus. Novit. 2725, 1–64.

Foote, M., Hunter, J.P., Janis, C.M., Sepkoski Jr., J.J., 1999. Evolutionary and preservationalconstraints on origins of biologic groups: divergence times of eutherian mammals.Science 283, 1310–1314.

Gingerich, P.D., 2005. Cetacea. In: Rose, K.D., Archibald, J.D. (Eds.), The Rise of PlacentalMammals. The Johns Hopkins University Press, Baltimore, pp. 234–252.

Godinot, M., Mahboubi, M., 1992. Earliest known simian primate found in Algeria.Nature 357, 324–326.

Gondwana Animation, International Institute for Aerospace Survey and Earth Science,Enschede, http://kartoweb.itc.nl/gondwana.

Goodman, M., 1996. Epilogue: a personal account of the origins of a new paradigm. Mol.Phylogenet. Evol. 5, 269–285.

Graur, D., Martin, W., 2004. Reading the entrails of chickens: molecular timescales ofevolution and the illusion of precision. Trends Genet. 20, 80–86.

Green, R.E., et al., 2006. Analysis of one million base pairs of Neanderthal DNA. Nature444, 330–336.

Guindon, S., Gascuel, S., 2003. A simple, fast and accurate algorithm to estimate largephylogenies by maximum likelihood. Syst. Biol. 52, 696–704.

Haile-Selassie, Y., 2001. Late Miocene hominids from the Middle Awash, Ethiopia.Nature 412, 178–181.

Hallström, B., Kullberg, M., Nilsson, M., Janke, A., 2007. Phylogenomic data analysesprovide evidence that Xenarthra and Afrotheria are sistergroups. Mol. Biol. Evol. 24,2059–2068.

Hedges, S.B., Parker, P.H., Sibley, C.G., Kumar, S., 1996. Continental breakup and theordinal diversification of birds and mammals. Nature 381, 226–229.

Hodgson, J.A., Disotell, T.R., 2008. No evidence of a Neanderthal contribution to modernhuman diversity. Genome Biol. 9, 206.

Horner, D.S., Lefkimmiatis, K., Reyes, A., Gissi, C., Saccone, C., Pesole, G., 2007.Phylogenetic analyses of complete mitochondrial genome sequencessuggest a basal divergence of the enigmatic rodent Anomalurus. BMC Evol.Biol. 8 (7), 16.

Huelsenbeck, J.P., Ronquist, F., 2001. MRBAYES: Bayesian inference of phylogenetic trees.Bioinformatics 17, 754–755.

Irwin, D.M., Arnason, U., 1994. Cytochrome b gene of marine mammals: phylogeny andevolution. J. Mamm. Evol. 2, 37–55.

Ishida, H., Pickford, M., 1997. A new late Miocene hominoid from Kenya: Samburu-pithecus kiptalami gen. et sp. nov. C. R. l'Acad. Sci. Paris, Ser. 2A 325, 823–829.

Ji, Q., Luo, Z., Yuan, C., Wible, J.R., Zhang, J., Georgi, J.A., 2002. The earliest knowneutherian mammal. Nature 416, 816–822.

Jones, G., Teeling, E.C., 2006. The evolution of echolocation in bats. Trends Ecol. Evol. 21,149–156.

Jorgensen, F.G., Hobolth, A., Hornshoj, H., Bendixen, C., Fredholm, M., Schierup, M.H.,2005. Comparative analysis of protein coding sequences from human, mouse andthe domesticated pig. BMC Biol. 28 (3), 2.

Kitazoe, Y., Kishino, H., Waddell, P.J., Nakajima, N., Okabayashi, T., Watabe, T., Okuhara,Y., 2007. Robust time estimation reconciles views of the antiquity of placentalmammals. PloS ONE 2, e384.

Kjer, K.M., Honeycutt, R.L., 2007. Site specific rates of mitochondrial genomes and thephylogeny of Eutheria. BMC Evol. Biol. 7, 8.

Krettek, A., Gullberg, A., Arnason, U., 1995. Sequence analysis of the completemitochondrial DNA molecule of the hedgehog, Erinaceus europaeus, and thephylogenetic position of the Lipotyphla. J. Mol. Evol. 41, 952–957.

Kriegs, J.O., Churakov, G., Kiefmann, M., Jordan, U., Brosius, J., Schmitz, J., 2006.Retroposed elements as archives for the evolutionary history of placentalmammals. PLoS Biol. 4 (4), e91.

Krings, M., Stone, A., Schmitz, R.W., Krainitzki, H., Stoneking, M., Paabo, S., 1997.Neandertal DNA sequences and the origin of modern humans. Cell 90, 19–30.

Kumar, S., Hedges, S.B., 1998. A molecular timescale for vertebrate evolution. Nature392, 917–920.

Kumar, S., Filipski, A., Swarna, V., Walker, A., Hedges, S.B., 2005. Placing confidencelimits on the molecular age of the human-chimpanzee divergence. Proc. Natl. Acad.Sci. U. S. A. 102, 18842–18847.

Lanave, C., Preparata, G., Saccone, C., Serio, G., 1984. A new method for calculatingevolutionary substitution rates. J. Mol. Evol. 20, 86–93.

Le Gros Clark, W.E., Sonntag, C.F., 1926. A monograph of Orycteropus afer. Proc. Zool. Soc.Lond. 30, 445–485.

Luo, Z.–X., Qiang, A., Wible, J.R., Yuan, C., 2003. An Early Cretaceous tribosphenicmammal and metatherian evolution. Science 302, 1934–1940.

Martin, R.D., 1993. Primate origins: plugging the gaps. Nature 363, 223–234.Martin, R.D., 2000. Origins, diversity and relationships of lemurs. Int. J. Primatol. 21,

1021–1049.McKenna,M.C.,1975. Toward a phylogenetic classification of theMammalia. In: Luckett,W.P.,

Szalay, F.S. (Eds.), Phylogeny of the Primates. Plenum Press, New York, pp. 21–46.McKenna, M.C., Bell, S.K., 1997. Classification of Mammals above the Species Level.

Columbia University Press, New York.Murphy, W.J., et al., 2001. Resolution of the early placental mammal radiation using

Bayesian phylogenetics. Science 294, 2348–2351.Murphy, W.J., Pringle, T.H., Crider, T.A., Springer, M.S., Miller, W., 2007. Using genomic

data to unravel the root of the placental mammal phylogeny. Genome Res. 17,413–421.

Nikolaev, S., Montoya-Burgos, J.I., Margulies, E.H., 2007. Early history of mammals iselucidated with the ENCODE multiple species sequencing data. NISC ComparativeSequencing Program, Rougemont, J., Nyffeler, B., Antonarakis, S.E., PLoS Genet. 5, 3(1).

Nishihara, H., Hasegawa, M., Okada, N., 2006. Pegasoferae, an unexpected mammalianclade revealed by tracking ancient retroposon insertions. Proc. Natl. Acad. Sci. U. S. A.103, 9929—9934.

Noonan, J.P., et al., 2006. Sequencing and analysis of Neanderthal genomic DNA. Science314, 1113—1118.

Novacek, M.J., 1992. Mammalian phylogeny: shaking the tree. Nature 356, 121—125.Nuttall, G.H.F., 1901a. The new biological test for blood in relation to zoological

classification. Proc. Royal. Soc. Lond. LXIX, 150—153.Nuttall, G.H.F., 1901b. The new biological test for blood—its value in legal medicine and

in relation to zoological classification. J. Trop. Med. IV, 405–408.Nuttall, G.H.F., 1904. Tests with precipitins. In: Nuttall, G.H.F. (Ed.), Blood Immunity and

Blood Relationships, Section VI. Cambridge University Press, Cambridge, pp. 152–311.Nuttall, G.H.F., Strangeways, T.S.P., 1904. On the results of 500 quantitative tests with

precipitating antisera upon the bloods of Primates, Insectivora, Carnivora, Ungulata,Cetacea, Marsupialia, and Aves. In: Nuttall, G.H.F. (Ed.), Blood Immunity and BloodRelationships, Section VII. Cambridge University Press, Cambridge, pp. 312–335.

ODSN Plate Tectonic Reconstruction Service, GEOMAR, Kiel, http://www.odsn.de/odsn/services/paleomap/paleomap.html.

Ovchinnikov, et al., 2000.Pääbo, S., 1985. Molecular cloning of ancient Egyptian mummy DNA. Nature 314,

644–645.Patterson, N., Richter, D.J., Gnerre, S., Lander, E.S., Reich, D., 2006. Genetic evidence for

complex speciation of humans and chimpanzees. Nature 441, 1103–1108.Penny, D., Hasegawa, M., Waddell, P.J., Hendy, M.D., 1999. Mammalian evolution: timing

and implications from using the logdeterminant transform for proteins of differingamino acid composition. Syst. Biol. 48, 76–93.

Pickford, M., Senut, B., 2005. Hominoid teeth with chimpanzee- and gorilla-like featuresfrom the Miocene of Kenya: implications for the chronology of ape-humandivergence and biogeography of Miocene hominoids. Anthropol. Sci. 113, 95–102.

Posada, D., Crandall, K.A., 1998. Modeltest: testing the model of DNA substitution.Bioinformatics 14, 817–818.

Prothero, D.R., Schoch, R.M., 1989. Origin and evolution of Perissodactyla. In: Prothero,D.R., Schoch, R.M. (Eds.), The Evolution of Perissodactyls. Oxford University Press,New York, pp. 504–529.

Rougier, G.W., Wible, J.R., Novacek, M.J., 1998. Implications of Deltatheridium specimensfor early marsupial history. Nature 396, 459–463.

Sanderson, M.J., 2002. Estimating absolute rates of molecular evolution and divergencetimes: a penalized likelihood approach. Mol. Biol. Evol. 19, 101–109.

Sarich, V.M., Wilson, A.C., 1967. Immunological time scale for human evolution. Science158, 1200–1203.

Schmidt, H.A., Strimmer, K., Vingron, M., von Haeseler, A., 2002. TREE-PUZZLE:maximum likelihood phylogenetic analysis using quartets and parallel computing.Bioinformatics 18, 502–504.

Schmitz, J., Ohme, M., Suryobroto, B., Zischler, H., 2002. The colugo (Cynocephalusvariegatus, Dermoptera): the primates' gliding sister?Mol. Biol. Evol.19, 2308–2312.

Schmitz, J., Roos, C., Zischler, H., 2005. Primate phylogeny: molecular evidence fromretroposons. Cytogenet. Genome Res. 108, 26–37.

Scotese, C.R., 2002. PALEOMAP website, http://www.scotese.com.Senut, B., Pickford, M., Gommery, D., Mein, P., Cheboi, K., Coppens, Y., 2001. First

hominid from the (Lukeino Formation, Kenya. C. R. l'Acad. Sci. Paris, Ser. 2A 332,137–144.

Shimamura, M., et al., 1997. Molecular evidence from retroposons that whales form aclade within even-toed ungulates. Nature 388, 622–624.

Sibley, C.G., Ahlquist, J.E., 1987. DNA hybridization evidence of hominoid phylogeny:results from an expanded data set. J. Mol. Evol. 26, 99–121.

http://kartoweb.itc.nl/gondwana

http://www.odsn.de/odsn/services/paleomap/paleomap.html

http://www.odsn.de/odsn/services/paleomap/paleomap.html

http://www.scotese.com


Simpson, G.G., 1945. The principles of classification and a classification of mammals.Bull. Am. Mus. Nat. Hist. 85, 1–350.

Springer, M.S., 1997. Molecular clocks and the timing of the placental and marsupialradiations in relation to the Cretaceous–Tertiary boundary. J.Mamm. Evol. 4, 285–302.

Springer, M.S., Murphy, W.J., 2007. Mammalian evolution and biomedicine: new viewsfrom phylogeny. Biol. Rev. Camb. Philos. Soc. 82, 375–392.

Springer, M.S., Murphy, W.J., Eizirik, E., O'Brien, S., 2003. Placental mammaldiversification and the Cretaceous–Tertiary boundary. Proc. Natl. Acad. Sci. U. S. A.100, 1056–1061.

Stanhope, M.J., et al., 1998. Molecular evidence formultiple origins of Insectivora and fora new order of endemic African insectivore mammals. Proc. Natl. Acad. Sci. U. S. A.95, 9967–9972.

Sullivan, J., Swofford, D.L., 1997. Are guinea pigs rodents? The importance of adequatemodels in molecular phylogenetics. J. Mamm. Evol. 4, 77–86.

Suwa, G., Kono, R.T., Katoh, S., Asfaw, B., Beyene, Y., 2007. A new species of great apefrom the late Miocene epoch in Ethiopia. Nature 448, 921–924.

Swofford, D.L., 1998. Phylogenetic analysis using parsimony (⁎and other methods).Version 4. Sinauer Associates, Sunderland, MA.

Tavaré, S., Marshall, C.R., Will, O., Soligo, C., Martin, R.D., 2002. Using the fossil record toestimate the age of the last commonancestor of extant primates. Nature 416, 726–729.

Ursing, B.M., Arnason, U., 1998. Analyses of mitochondrial genomes strongly support ahippopotamus–whale clade. Proc. R. Soc. Lond., B. 265, 2251–2255.

Vigilant, L., Stoneking, M., Harpending, H., Hawkes, K., Wilson, A.C., 1991. Africanpopulations and the evolution of human mitochondrial DNA. Science 253,1503–1507.

Waddell, P.J., Cao, Y., Hasegawa, M., Mindell, D.P., 1999. Assessing the CretaceousSuperordinal divergence timeswithin birds and placental mammals by usingwholemitochondrial protein sequences and an extended statistical framework. Syst. Biol.48, 119–137.

Waddell, P.J., Kishino, H., Ota, R., 2001. A phylogenetic foundation for comparativemammalian genomics. Genome Inf. 12, 141–154.

Wall, J.D., Kim, S.K., 2007. Inconsistencies in Neanderthal genomic DNA sequences. PloSGenet. 3, e175. doi:10.1371/journal.pgen.0030175.

Waters, P.D., Dobigny, G., Waddell, P.J., Robinson, T.J., 2007. Evolutionary history of LINE-1 in the major clades of placental mammals. PloS ONE 17 (2), e158.

Wible, R.J., Rougier, G.W., Novacek, M.J., Asher, R.J., 2007. Cretaceous eutherians andLaurasian origin for placental mammals near the K/T boundary. Nature 447,1003–1006.

Wildman, D.E., et al., 2007. Genomics, biogeography, and the diversification of placentalmammals. Proc. Natl. Acad. Sci. U. S. A. 104, 14395–14400.

Yang, Z., 1994. Maximum likelihood phylogenetic estimation from DNA sequences withvariable rates over sites: approximate methods. J. Mol. Evol. 39, 306–314.

Yang, Z., Nielsen, R., Hasegawa, M., 1998. Models of amino acid substitution andapplications to mitochondrial protein evolution. Mol. Evol. Biol. 15, 1600–1611.

Xu, X., 1996. Studies of mammalian mitochondrial genomes with special emphasis onthe Perissodactyla. PhD Thesis, Univ of Lund, Sweden.

Xu, X., Janke, A., Arnason, U., 1996. The complete mitochondrial DNA sequence of thegreater Indian rhinoceros, Rhinoceros unicornis, and the phylogenetic relationshipamong Carnivora, Perissodactyla and Artiodactyla (+Cetacea). Mol. Biol. Evol. 13,1167–1173.

http://dx.doi.org/10.1371/journal.pgen.0030175

Documents

Mitogenomic relationships of placental mammals and