Wang Et Al 2004 Canid Ancestry Proof

Seediscussions,stats,andauthorprofilesforthispublicationat:http://www.researchgate.net/publication/267156823

Phylogeny,classification,andevolutionaryecologyoftheCanidaeCHAPTERJANUARY2004

CITATIONS6

DOWNLOADS297

VIEWS54

4AUTHORS,INCLUDING:

XiaomingWangNaturalHistoryMuseumofLosAngeles145PUBLICATIONS1,143CITATIONS

SEEPROFILE

BlaireVanValkenburghUniversityofCalifornia,LosAngeles73PUBLICATIONS1,823CITATIONS

SEEPROFILE

Availablefrom:XiaomingWangRetrievedon:27June2015

Studies and reconstructions of dire wolf (Canis dirus) and Grey wolf (Canis lupus) from late Pleistocene Rancholebrea Tarpits,Los Angeles, California. Illustration by Pat Ortega.

Zubir-02.qxd 20#12#03 7:29 PM Page 36

The evolutionary history of canids (Family Canidae)is a history of successive radiations repeatedly occu-pying a broad spectrum of niches ranging from large,pursuit predators to small omnivores, or even to her-bivory. Three such radiations were first recognized byTedford (1978), each represented by a distinct sub-family (Fig. 2.1). Two archaic subfamilies, Hespero-cyoninae and Borophaginae, thrived in the middleto late Cenozoic from about 40 to 2 million yearsago (Ma) (Wang 1994; Wang et al. 1999). All livingcanids belong to the final radiation, SubfamilyCaninae, which achieved their present diversity onlyin the last few million years (Tedford et al. 1995).

Canids originated more than 40 Ma in the lateEocene of North America from a group of archaic car-nivorans, the Miacidae (Wang and Tedford 1994,1996). They were confined to the North Americancontinent during much of their early history, play-ing a wide range of predatory roles that encompassthose of the living canids, procyonids, hyaenids, andpossibly felids. By the latest Miocene (about 78 Ma),members of the Subfamily Caninae were finally ableto cross the Bering Strait to reach Europe (Crusafont-Pair 1950), commencing an explosive radiation andgiving rise to the modern canids of the Old World. Atthe formation of the Isthmus of Panama, 3 Ma,canids arrived in South America and quickly estab-lished themselves as one of the most diverse groupsof carnivorans on the continent (Berta 1987, 1988).With the aid of humans, Canis lupus dingo was trans-ported to Australia late in the Holocene. Since thattime, canids have become truly worldwide predators,unsurpassed in distribution by any other group ofcarnivorans.

Here, in the context of this volume on the moderncanids, we place more emphasis on the subfamilyCaninae, the latest of the three successive radiationsof the Canidae. We do not attempt to cite all of thereferences in canid palaeontology and systematics,most of which have been summarized in the papersthat we cite at the end of each section. Certain phy-logenetic relationships of the Caninae are controver-sial, as reflected in the different conclusions reachedon the basis of evidence from palaeontological/morphological or molecular research as presentedbelow.

What is a canid?Canids possess a pair of carnassial teeth (the upperfourth premolar and lower first molar) in the form ofa shearing device, and thus belong to the OrderCarnivora. Within the Carnivora, canids fall into the Suborder Caniformia, or dog-like forms. TheCaniformia are divided into two major groups thathave a sister relationship: Superfamily Cynoidea,which includes Canidae, and Superfamily Arctoidea,which include the Ursidae, Ailuridae, Procyonidae,and Mustelidae, as well as the aquatic Pinnipediaand the extinct Amphicyonidae.

As a cohesive group of carnivorans, living canidsare easily distinguished from other carnivoran fami-lies. Morphologically there is little difficulty in rec-ognizing living canids with their relatively uniformand unspecialized dentitions. However, the canids asexemplified by the living forms are narrowly defined.Only a small fraction of a once diverse group has

CHAPTER 2

AncestryEvolutionary history, molecular systematics, andevolutionary ecology of Canidae

Xiaoming Wang, Richard H. Tedford, Blaire Van Valkenburgh,and Robert K. Wayne

Zubir-02.qxd 20#12#03 7:29 PM Page 37

38 Biology and conservation of wild canids

3236

Orel-lan

Chad-ronian

Arikareean

Miocene Pliocene

Barstovian Claren-donian Blancan Irv.RLB

Pleist.+Holo.

HemphillianHeming-fordian

Whit-neyan

OligoceneEocene

28

RhizocyonOtarocyon

Archaeocyon

Hesperocyon

Paraenhydrocyon

Mesocyon

Cynodesmus

Caedocyon

Enhydrocyon

EctopocynusOsbornodon

SunkahetankaPhilotrox

Cynarctoides

Cormocyon

Phlaocyon

DesmocyonParacynarctus

MetatomarctusEuoplocyon

PsalidocyonProtomarctusTephrocyon

TomarctusAelurodonParatomarctus

CarpocyonProtepicyon

Leptocyon

New genus

VulpiniCanini

Vulpes

Eucyon

S. Am. caninesCaninae

BorophaginaeHesperocyoninae

UrocyonOtocyon

Xenocyon

Jackal-like CanisCoyote-like Canis

Wolf-like Canis

Cuon

Lycaon

Epicyon

Cynarctus

Oxetocyon

24 20 16 12 8EEEEEE LLLLLLM

4 Ma

3236 28 24 20 16 12 8 4 Ma

Borophagus

Figure 2.1 Simplified phylogenetic relationships of canids at the generic level. Species ranges are indicated by individual bars enclosed within grey rectangles, detailed relationships among species in a genus is not shown. Relationships for theHesperocyoninae is modified from Wang (1994, fig. 65), that for the Borophaginae from Wang et al. (1999, fig. 141), and that for the Caninae from unpublished data by Tedford et al.

Zubir-02.qxd 20#12#03 7:29 PM Page 38

survived to the present day (Fig. 2.1). Canids in thepast had departed from this conservative pattern suf-ficiently that paleontologists had misjudged somecanids as procyonids. Similarly the extinct bear-dog Family Amphicyonidae, which belongs to theArctoidea, is often placed within the Canidae,because of its unspecialized dentition.

How do we know a canid when we see one? A keyregion of the anatomy used to define canids is themiddle ear region, an area that distinguishes mostfamilies of carnivorans (Hunt 1974), perhaps as a result of a widespread trend of ossifying bullarelements in independent lineages. Canids are char-acterized by an inflated entotympanic bulla that isdivided by a partial septum along the entotympanicand ectotympanic suture (Fig. 2.2). Other featurescharacteristic of canids is the loss of a stapedial arteryand the medial position of the internal carotid artery that is situated between the entotympanic and petrosal for most of its course and contained within the rostral entotympanic anteriorly (Wangand Tedford 1994). These basicranial character-istics have remained more or less stable through-out the history of canids, allowing easy identificationin the fossil record when these structures arepreserved.

Evolutionary historyAmong the living families within the Order Carnivora,the Canidae are the most ancient. The family arose inthe late Eocene, when no other living families ofcarnivorans had yet emerged (two archaic families,Miacidae and Viverravidae, have a much older his-tory but none survive to the present time). Further-more, canids still maintain some features that areprimitive among all carnivorans, to the extent thatdog skulls are often used to illustrate a generalizedmammal in zoological classrooms. Dentally, canidsare closest to the ancestral morphotype of Carnivora.Canids have a relatively unreduced dental formula of3142/3143 [numbers in sequence represent incisors,canines, premolars, and molars in the upper (left halfbefore the oblique) and the lower (right half after theoblique) teeth] and relatively unmodified molarsexcept for the morphology of the carnassials (P4, m1)typical of all carnivorans. In contrast, all other car-nivoran families generally have a more reduceddental formula and highly modified cusp patterns.

From this mesocarnivorous (moderately carniv-orous) conservative plan, canids generally evolvetowards a hypercarnivorous (highly carnivorous) orhypocarnivorous (slightly carnivorous) dental pat-tern. In the hypercarnivorous pattern (Fig. 2.4(b,d) )

Ancestry 39

Bullar septum

Ectotympanic

Caudal entotympanicInternal carotid artery

Anterior

Dorsal

PetrosalLa

tera

l

Figure 2.2 Ventrolateral view ofbasicranial morphology of aprimitive canid, Hesperocyongregarius, showing bullarcomposition and position of theinternal carotid artery (see text forexplanations). The ventral floor ofthe bulla is dissected away(isolated oval piece on top) toreveal the middle ear structuresinside the bulla and the internalseptum. Modified from Wang andTedford (1994, fig. 1).

Zubir-02.qxd 20#12#03 7:29 PM Page 39


Hesperocyon

Leptocyon

Tribe C

aniniTr

ibe

Vulp

ini

DesmocyonCynarctoides

Phlaocyon

Euoplocyoncyon

Cynodesmus

Osbornodon

Paraenhydrocyon

Hesperocyonnaei

Caninae

Borop

hagin

ae

5 5

10 10

15 15

20 20

25 25

30Ma

30Ma

Enhydrocyon

Cynarctus

Paratomar

Aelurodon

Carpocyon

Epicyon

P4 M1 M2

Vulpes

Eucyon

Cerdocyon

Urocyon

CuonCanisS. Am. canines

Canis clade

Borophagus

Vulpes clade

Figure 2.3 Dental evolution of representative canids as shown in upper cheek teeth (P4M2). Generally the most derivedspecies in each genus is chosen to enhance a sense of dental diversity. Species in the Hesperocyoninae are: Hesperocyongregarius; Paraenhydrocyon josephi; Cynodesmus martini; Enhydrocyon crassidens; and Osbornodon fricki. Species in theBorophaginae are: Cynarctoides acridens; Phlaocyon marslandensis; Desmocyon thomsoni; Cynarctus crucidens; Euoplocyonbrachygnathus; Aelurodon stirtoni; Paratomarctus temerarius; Carpocyon webbi; Epicyon haydeni; and Borophagus diversidens.Species in the Caninae are: Leptocyon gregorii; Vulpes stenognathus; Urocyon minicephalus; Cerdocyon thous; Eucyon davisi;Canis dirus; and Cuon alpinus. All teeth are scaled to be proportional to their sizes.

there is a general tendency to increase the size of thecarnassial pair at the expense of the molars behind (see Enhydrocyon, Aelurodon, Borophagus, and Cuon inFig. 2.3). This modification increases the efficiencyof carnassial shear. A hypocarnivorous pattern(Fig. 2.4(a,c) ) is the opposite, with development of thegrinding part of the dentition (molars) at the expenseof carnassial shear (see Cynarctoides, Phlaocyon, andCynarctus in Fig. 2.3). This configuration was only pos-sible in the sister-taxa Borophaginae and Caninae,which share a bicuspid m1 talonid (Fig. 2.4(c) ). One ofthe major trends in canid evolution is the repeated

development of hyper- and hypocarnivorous forms(see below).

HesperocyoninaeThe Subfamily Hesperocyoninae is the first majorclade with a total of 28 species. Its earliest membersare species of the small fox-like form, Hesperocyon,that first appears in the late Eocene (Duchesnean,3740 Ma) (Bryant 1992) and became abundant inthe latest Eocene (Chadronian). By Oligocene time

Zubir-02.qxd 20#12#03 7:29 PM Page 40

(Orellan and Whitneyan, 3430 Ma), early membersof four small clades of the hesperocyonines hademerged: Paraenhydrocyon, Enhydrocyon, Osbornodon,and Ectopocynus. Hesperocyonines experienced theirmaximum diversity of 14 species during the lateOligocene (early Arikareean in 3028 Ma), and

reached their peak predatory adaptations (hypercar-nivory) in the earliest Miocene (late Arikareean) withadvanced species of Enhydrocyon and Paraenhydrocyon.The last species of the subfamily, Osbornodon fricki,became extinct in the early Barstovian (15 Ma), reach-ing the size of a small wolf.

Ancestry 41

ProtoconeProtocone

Parastyle

P4M1

M2

ParastyleParacone Paracone

Metacone

TalonidHypoconidProtoconid

Paraconid

Hypoconid

Talonid

Entoconid

Protoconid

Protostylid

Paraconid

Metaconid

m1 m2m3

ProtoconeProtocone

Parastyle

Metacone

Hypocone

Hypocone

Paracone(a) (b)

(c) (d)

Figure 2.4 Hypercarnivorous (b, Aelurodon and d, Euoplocyon) and hypocarnivorous (a, Phlaocyon and c, Cynarctus) dentitions. Inhypercarnivorous forms, the upper cheek teeth (b) tend to emphasize the shearing part of the dentition with an elongated andnarrow P4, an enlarged parastyle on a transversely elongated M1, and a reduced M2. On the lower teeth (d), hypercarnivory isexemplified by a trenchant talonid due to the increased size and height of the hypoconid at the expense of the entoconid (reducedto a narrow and low ridge), accompanied by the enlargement of the protoconid at the expense of the metaconid (completely lost inEuoplocyon) and the elongation of the trigonid at the expense of the talonid. In hypocarnivorous forms, on the other hand, the upperteeth (a) emphasize the grinding part of the dentition with a shortened and broadened P4 (sometimes with a hypocone along thelingual border), a reduced parastyle on a quadrate M1 that has additional cusps (e.g. a conical hypocone along the internalcingulum) and cuspules, and an enlarged M2. The lower teeth (c) in hypocarnivorous forms possess a basined (bicuspid) talonid onm1 enclosed on either side by the hypoconid and entoconid that are approximately equal in size. Other signs of hypocarnivory on thelower teeth include widened lower molars, enlarged metaconids, and additional cuspules such as a protostylid.

Zubir-02.qxd 20#12#03 7:29 PM Page 41

With the exception of the Osbornodon clade, whichacquired a bicuspid m1 talonid, hesperocyonines areprimitively hypercarnivorous in dental adaptationswith tendencies towards reduced last molars andtrenchant (single cusped) talonid heels on the lowerfirst molar. Although never reaching the extremesseen in the borophagines, hesperocyonines hadmodest development of bone cracking adaptationsin their strong premolars. At least three lineages, inall species of Enhydrocyon and in terminal species ofOsbornodon and Ectopocynus, have independentlyevolved their own unique array of bone crackingteeth. Hesperocyonines did not experiment withhypocarnivory.

Members of this subfamily were the topic of amonograph by Wang (1994). Two additional speciesof Osbornodon have since been added to the subfam-ily (Hayes 2000; Wang in press). The evolutionarytransition from plantigrade to digitigrade standingposture in early canids was explored by Wang (1993).A hereditary condition, osteochondroma, in thepostcranials of early canids was documented byWang and Rothschild (1992).

BorophaginaeFrom the primitive condition of a trenchant talonidheel on the lower first molar seen in the hesperocyo-nines, borophagines, and canines shared a basined(bicuspid) talonid acquired at the very beginning oftheir common ancestry (Fig. 2.4(c) ). Along with amore quadrate upper first molar with its hypocone,the basined talonid establishes an ancestral statefrom which all subsequent forms were derived. Sucha dental pattern proved to be very versatile andcan readily be adapted towards either a hyper- orhypocarnivorous type of dentition, both of whichwere repeatedly employed by both borophaginesand canines (Fig. 2.3).

The history of the borophagines also begins witha small fox-like form, Archaeocyon, in the late Oligo-cene. Contemporaneous with larger and more pre-datory hesperocyonines, these early borophagines in the late Oligocene and early Miocene (Arikareean)tended to be more omnivorous (hypocarnivorous) intheir dental adaptations, such as Oxetocyon, Otarocyon,and Phlaocyon. One extreme case, Cynarctoides evolved

selenodont-like molars as in modern artiodactyles,a rare occurrence of herbivory among carnivorans.These early borophagines are generally no largerthan a raccoon, which is probably a good ecologicalmodel for some borophagines at a time when procy-onids had yet to diversify.

After some transitional forms in the early Miocene(Hemingfordian), such as Cormocyon and Desmocyon,borophagines achieved their maximum ecologicaland numerical (i.e. species) diversity in the middleMiocene (Barstovian), with highly omnivorousforms, such as Cynarctus, that were almost ursid-like aswell as highly predatory forms, such as Aelurodon, thatwere a larger version of the living African HuntingDog Lycaon. By then, borophagines had acquiredtheir unique characteristics of a broad muzzle, a bonycontact between premaxillary and frontal, multi-cuspid incisors, and an enlarged parastyle on theupper carnassials (modified from an enlargement ofthe anterior cingulum).

By the end of the Miocene, borophagines hadevolved another lineage of omnivory, although onlymodestly in that direction, in the form of Carpocyon.Species of Carpocyon are mostly the size of jackals tosmall wolves. At the same time, the emergence of thegenus Epicyon from a Carpocyon-like ancestor markedanother major clade of hypercarnivorous boro-phagines. The terminal species of Epicyon, E. haydeni,reached the size of a large bear and holds record asthe largest canid ever to have lived (Fig. 2.5). Closelyrelated to Epicyon is Borophagus, the terminal genusof the Borophaginae. Both Epicyon and Borophagusare best known for their massive P4 and p4 in con-trast to the diminutive premolars in front. This pairof enlarged premolars is designed for cracking bones,mirroring similar adaptations by hyaenids in the OldWorld. Advanced species of Borophagus survived thePliocene but became extinct near the beginning ofthe Pleistocene.

The phylogeny and systematics of the Boropha-ginae were recently revised by Wang et al. (1999),which is the basis of above summary. Munthe (1979,1989, 1998) analysed the functional morphology ofborophagine limb bones and found a diverse arrayof postcranial adaptations, in contrast to the morestereotypical view that the hyaenoid dogs were non-cursorial scavengers only. Werdelin (1989) compared


Zubir-02.qxd 20#12#03 7:29 PM Page 42

the bone-cracking adaptations of borophagine canidsand hyaenids in terms of evolutionary constraintswithin their prospective lineages.

CaninaeAs in the hesperocyonines and borophagines, a small fox-sized species of Leptocyon is the earliestrecognized member of the subfamily Caninae. Besidessharing a bicuspid talonid of m1 and a quadrate M1with the borophagines, Leptocyon is also character-ized by a slender rostrum and elongated lower jaw,and correspondingly narrow and slim premolars,features that are inherited in all subsequent canines.It first appeared in the early Oligocene (Orellan) andpersisted through the late Miocene (Clarendonian).Throughout its long existence (no other canid genushad as long a duration), facing intense competitionfrom the larger and diverse hesperocyonines andborophagines, Leptocyon generally remains smalland inconspicuous, never having more than two orthree species at a time.

By the latest Miocene (Hemphillian), fox-sizedniches are widely available in North America, leftopen by extinctions of all small borophagines. Thetrue fox clade, Tribe Vulpini, emerges at this time andundergoes a modest diversification to initiate primit-ive species of both Vulpes and Urocyon (and theirextinct relatives). The North American Pliocenerecord of Vulpes is quite poor. Fragmentary materials

from early Blancan indicate the presence of a SwiftFox-like form in the Great Plains. Vulpes species werewidespread and diverse in Eurasia during thePliocene (see Qiu and Tedford 1990), resulting froman immigration event independent from that of theCanis clade. Red Fox (Vulpes vulpes) and Arctic Fox(Vulpes lagopus) appeared in North America onlyin the late Pleistocene, evidently as a result of animmigration back to the New World.

Preferring more wooded areas, the grey fox Urocyonhas remained in southern North America and MiddleAmerica. Records of the grey fox clade have a more or less continuous presence in North Americathroughout its existence, with intermediate formsleading to the living species Urocyon cinereoargenteus.Morphologically, the living African Bat-eared FoxOtocyon is closest to the Urocyon clade, althoughmolecular evidence suggests that the Bat-eared Foxlies at the base of the fox clade or even lower (Geffenet al. 1992d; Wayne et al. 1997). If the morphologicalevidence has been correctly interpreted, then theBat-eared fox must represent a Pliocene immigrationevent to the Old World independent of other foxes.A transitional form, Protocyon, occurs in southernAsia and Africa in the early Pleistocene.

Advanced members of the Caninae, Tribe Canini,first occur in the medial Miocene (Clarendonian,912 Ma) in the form of a transitional taxon Eucyon.As a jackal-sized canid, Eucyon is mostly distin-guished from the Vulpini in an expanded paroccipi-tal process and enlarged mastoid process, and in the

Ancestry 43

25 cm

Figure 2.5 Reconstruction ofEpicyon saevus (small individual,based on AMNH 8305) andEpicyon haydeni (large individual,composite figure, based onspecimens from Jack SwayzeQuarry). These two species co-occur extensively during the lateClarendonian and earlyHemphillian of Western NorthAmerica. Illustration by MauricioAntn. (From Wang et al. 1999.)

Zubir-02.qxd 20#12#03 7:30 PM Page 43

consistent presence of a frontal sinus. The lattercharacter initiates a series of transformations in theTribe Canini culminating in the elaborate develop-ment of the sinuses and a domed skull in C. lupus. Bylatest Miocene time, species of Eucyon have appearedin Europe (Rook 1992) and by the early Pliocene inAsia (Tedford and Qiu 1996). The North Americanrecords all predate the European ones, suggestinga westward dispersal of this form.

Arising from about the same phylogenetic level asEucyon is the South American clade. Morphologicaland molecular evidence generally agrees that livingSouth American canids, the most diverse group ofcanids on a single continent, belong to a naturalgroup of their own. The South American canids areunited by morphological characters such as a longpalate, a large angular process of the jaw with awidened scar for attachment of the inferior branch ofthe medial pterygoid muscle, and a relatively longbase of the coronoid process (Tedford et al. 1995).By the close of the Miocene, certain fragmentarymaterials from southern United States and Mexicoindicate that taxa assignable to Cerdocyon (Torres andFerrusqua-Villafranca 1981) and Chrysocyon occurin North America. The presence of these derived taxain the North American late Miocene predicts thatancestral stocks of many of the South Americancanids may have been present in southern NorthAmerica or Middle America. They appear in the SouthAmerican fossil record shortly after the formation ofthe Isthmus of Panama in the Pliocene, around 3 Ma(Berta 1987). The earliest records are Pseudalopex andits close relative Protocyon, an extinct large hyper-carnivore, from the Pliocene (Uquian, around2.51.5 Ma) of Argentina. By the latest Pleistocene(Lujanian, 300,00010,000 years ago), most livingspecies or their close relatives have emerged, alongwith the extinct North American Dire Wolf, Canisdirus. By the end of the Pleistocene, all large, hyper-carnivorous canids of South America (Protocyon,Theriodictis) as well as C. dirus had become extinct.

The Canis clade within the Tribe Canini, the mostderived group in terms of large size and hypercar-nivory, arises near the MiocenePliocene boundarybetween 5 and 6 Ma in North America. A series ofjackal-sized ancestral species of Canis thrived in theearly Pliocene (early Blancan), such as Canis ferox,Canis lepophagus, and other undescribed species. At

about the same time, first records of canids begin toappear in the European late Neogene: Canis cipio inthe late Miocene of Spain (Crusafont-Pair 1950),Eucyon monticinensis in the latest Miocene of Italy(Rook 1992), the earliest raccoon-dog Nyctereutesdonnezani, and the jackal-sized Canis adoxus in theearly Pliocene of France (Martin 1973; Ginsburg1999). The enigmatic Canis cipio, only representedby parts of the upper and lower dentition, may per-tain to a form at the Eucyon level of differentiationrather than truly a species of Canis.

The next phase of Canis evolution is difficult totrack. The newly arrived Canis in Eurasia underwentan extensive radiation and range expansion in thelate Pliocene and Pleistocene, resulting in multiple,closely related species in Europe, Africa, and Asia. Tocompound this problem, the highly cursorial wolf-like Canis species apparently belong to a circum-arcticfauna that undergoes expansions and contractionswith the fluctuating climate. Hypercarnivorousadaptations are common in the crown-group ofspecies especially in the Eurasian middle latitudesand Africa. For the first time in canid history, phylo-genetic studies cannot be satisfactorily performed onforms from any single continent because of theirHolarctic distribution and faunal interminglingbetween the new and old worlds. Nevertheless someclades were localized in different parts of Holarctica.The vulpines major centre of radiation was in theOld World. For the canines, North America remaineda centre through the Pliocene producing the Coyoteas an endemic form. A larger radiation yielding thewolves, dhole, African hunting dog, and fossil relat-ives took place on the Eurasian and African conti-nents. During the Pleistocene elements of the largercanid fauna invaded mid-latitude North Americathe last invasion of which was the appearance of theGrey wolf south of the glacial ice sheets in the latestPleistocene (about 100 Ka).

A comprehensive systematic revision of NorthAmerican fossil canines by Tedford et al. (in pre-paration) forms the basis of much of the foregoingsummary. As part of that revision, the phylogeneticframework as derived from living genera waspublished by Tedford et al. (1995). Nowak (1979)published a monograph on the Quaternary Canis ofNorth America. Berta (1981, 1987, 1988) undertookthe most recent phylogenetic analysis of the South


Zubir-02.qxd 20#12#03 7:30 PM Page 44

American canids. Rook (1992, 1994) and Rook andTorre (1996a,b) partially summarized the Eurasiancanids. The African canid records are relatively poorly understood but recent discoveries promise toadvance our knowledge in that continent (Werdelin,personal communication). See also citations belowfor recent molecular systematic studies.

Molecular systematicsThe ancient divergence of dogs from other carnivores is reaffirmed by molecular data. DNADNA hybridization of single copy DNA clearly shows them as the first divergence in the suborderCaniformia that includes seals, bears, weasel, andraccoon-like carnivores (Fig. 2.6). This basal place-ment is further supported by mitochondrial DNAsequence studies (Vrana et al. 1994; Slattery andBrien 1995; Flynn and Nedbal 1998), and recentlystudies of DNA sequences from nuclear genes(Murphy et al. 2001). Based on molecular clockcalculations, the divergence time was estimated as 50 million years before present (Wayne et al. 1989).This value is consistent with the first appearance ofthe family in the Eocene, although it is somewhatmore ancient than the date of 40 million years sug-gested by the fossil record (see above). Consideringthat first appearance dates generally post-date actualdivergence dates because of the incompleteness of therecord (e.g. Marshall 1977), the agreement betweenfossil and molecular dates is surprisingly good.

Evolutionary relationships within the familyCanidae have been reconstructed using comparativekaryology, allozyme electrophoresis, and mitochon-drial DNA protein coding sequence data (Wayne andBrien 1987; Wayne et al. 1997, 1987a,b). Further,relationships at the genus level have been studiedwith mtDNA control region sequencing (a non-coding, hypervariable segment of about 1200 bp inthe mitochondrial genome) and microsatellite loci(hypervariable single copy nuclear repeat loci)(Geffen et al. 1992; Bruford and Wayne 1993; Girmanet al. 1993; Gottelli et al. 1994; Vil et al. 1997, 1999).The protein-coding gene phylogeny, which is largelyconsistent with trees based on other genetic approaches, shows that the wolf genus Canis is amonophyletic group that also includes the Dhole orAsian Wild Dog (Cuon alpinus). The Grey wolf, coyote

Ancestry 45

Dog

Jackal

Arctic fox

Spotted shunk

Striped skunk

Mink

Weasel

Ferret

Otter

Raccoon

Red panda

Brown bear

Malayan sun bear

Spectacled bear

Giant panda

Harbor seal

Sea lion

Walrus

Spotted genet

Civet

Palm civet

Mongoose

Spotted hyena

Domestic cat

Jungle cat

Ocelot

Geoffroey cat

Lion

Leopard

Cheetah

Striped hyena

1020304050

MYBP

Felidae

Hyaenidae

Herpestidae

Viverridae

Odobenidae

Feliformia

Caniformia

OlariidaePhocidae

Ursidae

Procyonidae

Mustelidae

Mephitidae

Canidae

Figure 2.6 Relationship of carnivores based on DNAhybridization data (Wayne et al. 1989). Family and suborder groupings are indicated. Time scale in millions ofyear before present (MYBP) is based on comparisons of DNAsequence divergence to first appearance times in the fossilrecord.

Zubir-02.qxd 20#12#03 7:30 PM Page 45

(Canis latrans) and Ethiopian wolf (Canis simensis)form a monophyletic group, with the Golden Jackal(C. aureus) as the most likely sister taxon (Fig. 2.7). TheBlack-backed and Side-striped jackals are sister taxa,but they do not form a monophyletic group with theGolden jackal and Ethiopian wolf. Basal to Canis and

Cuon are the African hunting dog (Lycaon pictus) and aclade consisting of two South American canids, thebush dog (Speothos venaticus) and the maned wolf(Chrysocyon brachyurus). Consequently, although theAfrican hunting dog preys on large game as does theGrey wolf and dhole, it is not closely related to either


Harbor sealIsland grey fox

Grey foxRaccoon dog

Bat-eared fox

Fennec fox

Red foxKit fox

Arctic fox

Grey wolf

Coyote

Ethiopian wolf

Golden jackal

Black-backed jackal

Dhole

Side-striped jackal

Crab-eatingfox

Argentine grey foxDarwins fox

Culpeo fox

Hoary fox

Pampas fox

Sechuran fox

Small-eared dog

MYBP

Bush dog

Maned wolf

African hunting dog

Red fox-likecanids

Wolf-likecanids

SouthAmericanfoxes

10 6 2

Figure 2.7 Consensus tree of 26 canid species based on analysis of 2001 bp of DNA sequence from mitochondrial proteincoding genes (Wayne et al. 1997). See Geffen et al. (1992) for a more detailed analysis of the Red-fox like canids. Time scale in millions of year before present (MYBP) is based on comparisons of DNA sequence divergence to first appearance times in the fossil record.

Zubir-02.qxd 20#12#03 7:30 PM Page 46

Ancestry 47

species but is sister to the clade containing thesespecies. This phylogeny implies that the trenchantheeled carnassial now found only in Speothos, Cuon,and Lycaon, evolved at least twice or was primitive andlost in other wolf-like canids and the maned wolf.

The South American canids do not form a mono-phyletic group. Speothos and Chrysocyon are sister taxathat group with the wolf-like canids rather than theSouth American foxes. The large sequence divergencebetween the bush dog and maned wolf and betweenthese taxa and the South American foxes suggeststhat they diverged from each other 67 Ma, wellbefore the Panamanian land bridge formed about23 Ma. Thus, three canid invasions of SouthAmerica are required to explain the phylogeneticdistribution of the extant species. These invasions are today survived by (1) the bush dog, (2) the manedwolf, and (3) the South American foxes. Further,within the South American foxes, divergence valuesbetween crab-eating fox (Cerdocyon thous), the Short-eared fox (Atelocynus microtis), and other SouthAmerican foxes, suggest they may have divergedbefore the opening of the Panamanian land bridge aswell (Wayne et al. 1997). The fossil record supportsthe hypothesis that the crab-eating fox had its origin outside of South America, as the genus hasbeen described from late Miocene deposits of NorthAmerica (36 Ma) (Berta 1984, 1987, see above).Consequently, only the foxes of the genus Pseudalopex,Lycalopex, and perhaps Atelocynus, might have aSouth American origin. Further, the generic distinc-tion given to Pseudalopex and Lycalopex does not reflectmuch genetic differentiation, and in the absence ofappreciable morphologic differences, the geneticdata suggest these species should be assigned to asingle genus.

A fourth grouping in the tree consists of other fox-like taxa, including Vulpes, Alopex, and Fennecus(Fig. 2.7) (Geffen et al. 1992; Mercure et al. 1993;Wayne et al. 1997). The Arctic Fox, Alopex, is a closesister to the Kit Fox, Vulpes macrotis and both share thesame unique karyotype (Wayne et al. 1987a). Basal toVulpes is Fennecus, suggesting an early divergence ofthat lineage. Finally, Otocyon, Nyctereutes, and Urocyonappear basal to other canids in all molecular and kary-ological trees (Wayne et al. 1987a). The first two taxaare monospecific whereas the third includes theIsland Fox, Urocyon littoralis and the grey fox,

U. cinereoargenteus. The three genera diverged early inthe history of the family, approximately 812 Ma assuggested by molecular clock extrapolations.

In sum, the living Canidae is divided into fivedistinct groupings. These include the wolf-like canids,which consists of the Coyote, Grey wolf, Jackals,dhole, and African hunting dog. This clade is associ-ated with a group containing bush dog and manedwolf in some trees and further, this larger grouping isassociated with the South American foxes (Wayne et al. 1997). The Red Fox group is a fourth independentclade containing Alopex, Vulpes, and Fennecus.Finally, three lineages have long distinct evolutionaryhistories and are survived today by the Raccoon Dog,Bat-eared Fox, and grey fox. Assuming an approxi-mate molecular clock, the origin of the modernCanidae begins about 1012 Ma and is followed bythe divergence of wolf and fox-like canids about6 Ma. The South American canids are not a mono-phyletic group and likely owe their origin to threeseparate invasions. This group included the manedwolf, bush dog, crab-eating fox, and the other SouthAmerican canids, which diverged from each otherabout 36 Ma.

Morphological and molecularphylogeniesTedford et al. (1995) performed a cladistic analysis ofliving canids on morphological grounds. The result is a nearly fully resolved relationship based on an 18 taxa by 57 characters matrix at the generic level.This relationship recognizes three monophyleticclades in the canines: the fox group (tribe Vulpini),the South American canine group, and the wolfgroup containing hypercarnivorous forms (the lattertwo form the tribe Canini). Recent molecular studies(presented above), on the other hand, contradictsome of these arrangements while maintaining otherparts in the morphological tree (Fig. 2.8(a) ).

Trees derived from 2001 bp of mitochondrial DNA(Wayne 1997, p. 239 and Fig. 2.7 of this chapter) tendto places the foxes near the basal part, the SouthAmerican canines in the middle, and the wolves andhunting dogs towards the terminal branches, a pat-tern that is consistent with the morphological tree.

Zubir-02.qxd 20#12#03 7:30 PM Page 47

The detailed arrangements, however, differ in a num-ber of ways. The foxes are generally in a paraphyleticarrangement in contrast to a monophyletic clade inthe morphological tree. The grey fox and bat-earedfox are placed at the base despite their highly deriveddental morphology compared with other foxes.

Similarly, South American canines are no longermonophyletic under molecular analysis but form atleast two paraphyletic branches. A glaring discr-epancy is the Asiatic raccoon dog being allied to thefoxes in the molecular analysis despite its numerousmorphological characters shared with some South


Vulp

es

Uroc

yon

Otoc

yon

Pse

udal

opex

Pseu

dalo

pex

Dusi

cyon

Lyca

lope

x

Chry

socy

on

Cerd

ocyo

n

Nyct

ereu

tes

Atel

ocyn

us

Speo

thos

Cani

s

Cuon

Lyca

on

Uroc

yon

Vulp

es

Otoc

yon

Nyct

ereu

tes

Speo

thos

Chry

socy

on

Cerd

ocyo

n

Atel

ocyn

us

Pseu

dalo

pex

Lyca

lope

x

Sech

uran

fox

Lyca

on

Cani

s

Cuon

Molecular tree(Wayne et al. 1997)

Morphological tree(Tedford et al. 1995)

Combined analysis(Wayne et al. 1997)

(a)

(b)

Figure 2.8 (a) Contrasting caninerelationships from recent morphological(Tedford et al. 1995) and molecularstudies (Wayne et al. 1997 and Fig. 2.7);(b) combined analysis of 2001 bp ofcanid mtDNA and 57 morphologicalcharacters (Wayne et al. 1997, fig. 7).

Zubir-02.qxd 20#12#03 7:30 PM Page 48

Ancestry 49

American forms. Finally, molecular data suggestindependent origins for the Asiatic and Africanhunting dogs in contrast to a sister relationship inthe morphological tree supported by a large numberof characters related to hypercarnivory.

Not surprisingly, there are increased agreementsbetween the molecular and morphological resultswhen the two data sets are combined in a total evi-dence analysis (Fig. 2.8(b) ). Under such conditions,the South American canines (except Nyctereutes)become monophyletic, as does the clade includingthe wolf, dhole, and African hunting dog.

Evolutionary ecologyIterative evolution of hypercarnivoryOne of the most remarkable features of canid historyis their repeated tendency to evolve both hypocar-nivorous and hypercarnivorous forms. As notedabove, hypercarnivorous species evolved within eachsubfamily, and hypocarnivorous species evolvedwithin two of the three (all but the Hesperocyo-ninae). Hypocarnivory was most fully expressed inthe Borophaginae, where at least 15 species showed atendency towards a dentition similar to that of livingraccoons (Wang et al. 1999). Among the Caninae, thetendency has not been quite as strong, with only asingle lineage, Nyctereutes, developing a markedlyhypocarnivorous dentition. However, all three sub-families include multiple species of apparent hyper-carnivores with enhanced cutting blades on theircarnassials, reduced grinding molars, and enlargedcanines and lateral incisors. When and why didhypercarnivory evolve within each subfamily?

In two of the three subfamilies, Hesperocyoninaeand Caninae, the evolution of hypercarnivoryappears to have occurred at least partly in response toa reduced diversity of other hypercarnivorous taxa.The Hesperocyoninae evolved hypercarnivory earlyin their history (Figs 2.1 and 2.7) and the mostadvanced forms appear in the early Miocene (about2420 Ma) at a time when the two previously domi-nant carnivorous families had vanished. These twofamilies were the Nimravidae, an extinct group ofsaber-tooth cat-like forms, and the Hyaenodontidae,a group of somewhat dog-like predators included in the extinct order Creodonta. The nimravids

and hyaenodontids dominated the North Americanguild of large, predatory mammals in the late Eoceneto mid-Oligocene (3729 Ma), but faded rapidly in thelate Oligocene, and were extinct in North America byabout 25 Ma (Van Valkenburgh 1991, 1994). Duringmost of their reign, hesperocyonines existed at lowdiversity and small (fox-size) body size, but as thehyaenodontids and nimravids declined in the lateOligocene, the early canids seem to have radiated toreplace them. Most of these hypercarnivorous canidswere jackal-size (less than 10 kg), with only the lastsurviving species, Osbornodon fricki, reaching the sizeof a small wolf (Wang 1994). In the early Miocene,large hypercarnivores immigrated from the OldWorld in the form of hemicyonine bears (Ursidae) and temnocyonine bear-dogs (Amphicyonidae). Thesubsequent decline to extinction of the hesperocy-onines might have been a result of competition withthese new predators (Van Valkenburgh 1991, 2001).

Hypercarnivory appears late in the history of theCaninae and represents at least several independentradiations in South America, North America, and theOld World (Figs 2.1 and 2.7). As was true of the hes-perocyonine example, the South American radiationof large hypercarnivorous canids occurred at a time(2.50.01 Ma) when cat-like predators were rare orabsent. It followed the elevation of the Panamanianland bridge around 23 Ma that allowed immigra-tion between the previously separated continents.The canids that first entered South America founda depauperate predator community, consisting of onebear-like procyonid carnivoran, three species of car-nivorous didelphid marsupials, one of which was thesize of a coyote, and a gigantic, predaceous groundbird (Marshall 1977). With the possible exception ofthe rare ground bird, none of these species was aspecialized hypercarnivore. Between 2.5 Ma and10,000 years ago, 16 new species of canids appearedin South America, at least seven of which had tren-chant heeled carnassials and clearly were adapted forhypercarnivory (Berta 1988; Van Valkenburgh 1991).They represent three different endemic genera,Theriodictis, Protocyon, and Speothos. In addition,there were three large wolf-like species of Canis inSouth America, Canis gezi, Canis nehringi, and Canisdirus, all of which were probably hypercarnivorousbut retained a bicuspid heel on their carnassials. Ofthese only the Dire Wolf, C. dirus, evolved in NorthAmerica. All but one of these ten hypercarnivorous

Zubir-02.qxd 20#12#03 7:30 PM Page 49

canids of South America went extinct at the end ofthe Pleistocene (Van Valkenburgh 1991). The solesurvivor, the bush dog (Speothos) is rarely sighted.

In the Old World, the evolution of hypercarnivo-rous canines occurred within the last 4 million yearsand did not coincide with an absence of cats. Largecats, both sabertooth and conical tooth forms, arepresent throughout the Plio-Pleistocene when thehighly carnivorous species of Canis, Cuon, Lycaon,and Xenocyon appear (Turner and Antn 1996).However, their evolution might be a response to thedecline of another group of hypercarnivores, wolf-like hyaenids. Hyaenids were the dominant dog-likepredators of the Old World Miocene, reaching adiversity of 22 species between 9 and 5 Ma, but thendeclining dramatically to just five species by about4 Ma (Werdelin and Turner 1996). Their decline mayhave opened up ecospace for the large canids andfavored the evolution of hypercarnivory.

The remaining episode of hypercarnivory incanids occurred in the Borophaginae between 15 and4 Ma (Van Valkenburgh et al. in press). As was true ofthe Caninae, the hypercarnivorous species do notevolve early in the subfamilys history. Instead, theyappear in the latter half of the subfamilys lifespanand only become prevalent in the last third (midlate Miocene; Figs 2.1 and 2.7). In the late Miocene,

borophagine canids were the dominant dog-likepredators of North America, having replaced theamphicyonids and hemicyonine bears that hadthemselves replaced the hesperocyonines some tenmillion years earlier (Van Valkenburgh 1999). In thecase of the Borophaginae, the evolution of hyper-carnivory appears more gradual than in the othertwo subfamilies, and is not easily ascribed to oppor-tunistic and rapid evolution into empty ecospace.

In all three subfamilies, there is a pattern of greaterhypercarnivory and increasing body size with time(Fig. 2.9). Even in the Hesperocyoninae, where hyper-carnivory evolves very early, large species with themost specialized meat-eating dentitions appear later(Wang 1994). This directional trend towards the evo-lution of large, hypercarnivorous forms is apparentin other groups of dog-like carnivores, such asthe amphicyonids (Viranta 1996) and hyaenids(Werdelin and Solounias 1991; Werdelin and Turner1996), and may be a fundamental feature of carni-vore evolution. The likely cause is the prevalence ofinterspecific competition among large, sympatricpredators. Interspecific competition tends to bemore intense among large carnivores because preyare often difficult to capture and can represent asizable quantity of food that is worthy of stealing anddefending. Competition appears to be a motive for


0

10

20

30

40

0

5

10

15

0

10

20

30

40

50

N

Millions of years before present

Hesperocyoninae Borophaginae Caninae

3730 3022 2215 3424 2414 142 3423 2312 125 50

Figure 2.9 Iterative evolution of large hypercarnivores. Number (N) of hypocarnivorous (white), mesocarnivorous (grey), and large (20 kg) hypercarnivorous (black) species over time in each of the three subfamilies. The few hesperocyonine species withtrenchant-heeled carnassials estimated to have been less than 20 kg in mass were assigned to the mesocarnivorous categorybecause they are assumed not have taken prey as large or larger than themselves. For the Hesperocyoninae and Borophaginae,their stratigraphic ranges were broken into thirds; for the Caninae, four time divisions were used because of the large number ofspecies appearing in the past 5 million years. Species were assigned to dietary categories and body mass was estimated on thebasis of dental morphology as described in Van Valkenburgh (1991) and Wang et al. (1999).

Zubir-02.qxd 20#12#03 7:30 PM Page 50

The last one million yearsAll of the canids that are extant today evolved wellprior to the late Pleistocene extinction event approx-imately 11,000 years ago. The same could be said ofmost, if not all, extant carnivores. In the New World,the end-Pleistocene event removed numerous largemammals, including both herbivores (e.g. camels,horses, proboscideans) and carnivores (e.g. SabertoothCat, Dire Wolf, Short-faced Bear). In the Old World,many of the ecological equivalents of these speciesdisappeared earlier, around 500,000 years ago(Turner and Antn 1996). Consequently, all extantcarnivore species evolved under very differentecological circumstances than exist at present. Forexample, the Grey wolf today is considered the toppredator in much of Holarctica, but it has only heldthis position for the last ten to eleven thousand years.For hundreds of thousands of years prior to that time,the wolf coexisted with 11 species of predator as large or larger than itself (Fig. 2.10). Now there arebut three, the Puma, Black Bear, and Grizzly Bear, andwolves are usually dominant over the first two speciesat least (Van Valkenburgh 2001). Thus, for most of itsexistence, the Grey wolf was a mesopredator ratherthan a top predator, and so its morphology andbehaviour should be viewed from that perspective.Given the greater diversity and probable greaterabundance of predators in the past, interspecificcompetition was likely more intense than at pre-sent. Higher tooth fracture frequencies in latePleistocene North American predators provide indi-rect evidence of heavy carcass utilization and strongfood competition at that time (Van Valkenburghand Hertel 1993). Intense food competition wouldfavour group defence of kills and higher levels ofinterspecific aggression. Perhaps the sociality of thewolf and the tendency of some carnivores to kill butnot eat smaller predators are remnant behavioursfrom a more turbulent past.

The only canid to go extinct in the North Americanend Pleistocene was the Dire Wolf, C. dirus. The Greywolf, Coyote, and several foxes survived. In additionto the Dire Wolf, two bears and three cats wentextinct, all of which were very large (Fig. 2.10). Canwe learn something about the causes of currentpredator declines by examining the winners andlosers in the late Pleistocene? Examination of the

Ancestry 51

much intraguild predation because the victim oftenis not eaten ( Johnson et al. 1996; Palomares and Caro1999; Van Valkenburgh 2001). Larger carnivores tendto dominate smaller ones and so selection shouldfavour the evolution of large body size. Large bodysize in turn selects for a highly carnivorous dietbecause of energetic considerations. As shown byCarbone et al. (1999), almost all extant carnivoresthat weigh more than 21 kg take prey as large orlarger than themselves. Using an energetic model,they demonstrated that large body size brings with itconstraints on foraging time and energetic return.Large carnivores cannot sustain themselves on rela-tively small prey because they would expend moreenergy in hunting than they would acquire. By tak-ing prey as large or larger than themselves, theyachieve a greater return for a given foraging bout.Killing and consuming large prey is best done with ahypercarnivorous dentition and so the evolutionof large body size and hypercarnivory are linked. Ofcourse, this does not preclude the evolution ofhypercarnivory at sizes less than 21 kg, but it seemsrelatively rare. It has occurred in the Canidae as evi-denced by the hesperocyonines and the extant ArcticFox, V. lagopus, and Kit Fox, V. macrotis. However, thetwo extant foxes do not have trenchant-heeledcarnassials despite their tendency towards a highlycarnivorous diet, and this may reflect regular, oppor-tunistic consumption of fruits and invertebrates(Van Valkenburgh and Koepfli 1993).

Returning to the questions of when and whyhypercarnivory evolves among canids, it seems thatwhen and why are intertwined. That is, because ofintraguild competition and predation, selectionfavours the evolution of larger size in canids and asa consequence, hypercarnivory. However, when thisoccurs it is largely a function of other members of thepredator guild. In the case of the Hesperocyoninae,it occurred relatively early in their history becausepreviously dominant large hypercarnivores were indecline or already extinct. In the case of the Boroph-aginae and Caninae, it did not occur until much laterbecause other clades held the large hypercarnivorousroles for much of the Miocene. In all these examples,it appears as though the rise of large hypercarnivorouscanids reflects opportunistic replacement rather thancompetitive displacement of formerly dominant taxa(Van Valkenburgh 1999).

Zubir-02.qxd 20#12#03 7:30 PM Page 51

loser species reveals that they tended to be the morespecialized members of their clades; they were larger(Fig. 2.10) and tended to be more dentally special-ized for hypercarnivory (Van Valkenburgh andHertel 1998). Remarkably, two of the species thatwent extinct, the Dire Wolf and Sabertooth Cat(Smilodon fatalis), are five times more common in theRancho La Brea tar pit deposits than the next mostcommon carnivore, the Coyote (C. latrans). This sug-gests that the Dire Wolf and Sabertooth Cat weredominant predators at this time, comparable to thenumerically dominant African Lion and Spottedhyena of extant African ecosystems. The extinctionof the apparently successful Dire Wolf and Sabertooth

Cat implies there was a major perturbation to theecosystem in the late Pleistocene. Their demise andthat of the other large hypercarnivores suggest thatlarge prey biomass dropped to extremely low levels.Supporting this are the parallel extinctions of 10 of the 27 species of raptors and vultures (VanValkenburgh and Hertel 1998).

In the late Pleistocene, the largest meat-eaters,both avian and mammalian, were the most vulnera-ble. Is this the case today for canids? Of the threelarge hypercarnivorous canids, the dhole, Grey wolf,and African hunting dog, only the hunting dog ishighly endangered. Among living canids in general,species that appear to be most at risk tend to be insu-lar (Darwins Fox, Channel Islands Fox) or restrictedto limited habitats (Ethiopian Wolf), or just verypoorly known species (e.g. Short-eared Zorro, bushdog). Indeed, it is a bit difficult to answer the questionof which of the living species are most endangeredbecause we have so little information on many of thesmaller taxa. Nevertheless, it does seem that the endPleistocene extinction is not a good analogue for whatis happening at present, at least in terms of which ismost vulnerable. Then, it was the largest, most abun-dant, and most carnivorous. Now it seems more oftento be smaller mesocarnivores that are at risk due tosmall population size exacerbated by habitat loss. Inboth the end-Pleistocene and at present, the hand ofhumanity looms large as a cause of predator declines.Initially, the damage was largely due to overhuntingof both prey and predator, and to this we have addedsignificant habitat loss. Survivors of the current crisisare likely to be both dietary and habitat generalists,such as the Coyote.

AcknowledgementsWe thank David Macdonald and Claudio Sillero forthe invitation to write this contribution and DavidMacdonald for his critical review. This research isfunded in part by grants from the National ScienceFoundation (DEB 9420004; 9707555).


0

100

200

300

400

Vulpes vulpesProcyon lotor

Lynx canadensisLynx rufus

Felis pardalisTaxidea taxusCanis latransCuon alpinus

Gulo guloMiracinonyx trumani

Canis lupusCanis dirus

Puma concolorPanthera onca

Arctodus pristinusTremarctos floridanus

Ursus americanusHomotherium serum

Smilodon fatalisArctodus simus

Ursus arctosPanthera atrox

Body mass

Figure 2.10 North American Pleistocene carnivoransarranged by body mass. Black bars represent extant species,and white bars represent extinct species. Arrow indicates the Grey wolf (Canis lupus). Data from Van Valkenburgh andHertel (1998).

Zubir-02.qxd 20#12#03 7:30 PM Page 52

Wang Et Al 2004 Canid Ancestry Proof

Documents

Transcript of Wang Et Al 2004 Canid Ancestry Proof