Carnivores (Carnivora) Eduardo Eizirika,b,* and William J. Murphyc this mammalian order has had its origin in Laurasia. aFaculdade de Biociências, PUCRS, Av. Ipiranga, 6681, Porto Alegre, Only in the Miocene (23–5 Ma) do carnivoran families RS 90619-900, Brazil; bInstituto Pró-Carnívoros, Av. Horácio Neto, appear in the fossil records of Africa and South America, c 1030, Atibaia SP 12945-010, Brazil; Department of Veterinary indicating an initial period of intercontinental dispersal Integrative Biosciences, Texas A&M University, College Station, TX in this group. Here we review the current understand- 77843-4458, USA *To whom correspondence should be addressed (eduardo.eizirik@ ing of the phylogenetic relationships and divergence pucrs.br) times among carnivoran families, focusing exclusively on living lineages, and emphasizing results from recent studies. Abstract Living members of the Order Carnivora are grouped into two monophyletic suborders: Feliformia and Living members of the mammalian Order Carnivora have Caniformia. 7 e former traditionally included Families been traditionally placed in 11 families making up two sub- Viverridae (e.g., civets and genets), Herpestidae (mon- orders: Feliformia and Caniformia. Recent analyses based gooses), Hyaenidae (hyenas), and Felidae (cats), while on morphological and molecular data have identifi ed add- the latter comprised the Families Canidae (dogs, wolves, itional groups of species that warrant family-level recogni- and foxes), Mustelidae (e.g., otters, weasels, and badgers), tion, leading to major changes in the current understanding Ursidae (bears), Procyonidae (e.g., raccoons and coatis), of carnivoran evolutionary history and taxonomy. There Otariidae (sea lions and fur seals), Phocidae (true seals), are presently 16 recognized families, whose relationships and Odobenidae (Walrus). Otariidae, Phocidae, and are now well understood. The carnivoran timetree indicates Odobenidae are highly adapted for marine life and have that Feliformia and Caniformia diverged from each other been historically grouped in a taxon called Pinnipedia. ~55 million years ago (Ma). Within each suborder, suprafa- 7 e monophyly and phylogenetic placement of pin- milial nodes span a broad range of divergence times, from nipeds have been contentious for many years ( 2, 5, 6), 53 to 22 Ma. but this issue seems to be mostly settled now ( 7–10). It is now clear that Pinnipedia is monophyletic and that 7 e Order Carnivora contains a diverse set of mammals, including well-known species such cats, dogs, lions, bears, and seals, as well as enigmatic animals such as the stink badgers (Mydaus spp.), the African Palm Civet (Nandinia binotata), and the Fossa (Cryptoprocta ferox). 7 ere are currently 286 recognized species of living car- nivorans (e.g., Fig. 1), which vary widely in morphology, ecology, physiology, and behavior (1, 2). Size range among carnivoran species is broader than in any other mamma- lian order, with body weight varying 1000 times among its representatives. 7 e Order Carnivora has a relatively rich paleontological record. 7 e earliest fossils, dating from the Paleocene (66–56 Ma), are usually placed in the extinct families Viverravidae and Miacidae (3), both of which likely comprise early branching lineages relative to the living taxa (4). 7 e carnivoran fossil record from Fig. 1 An African Wild Dog (Lycaon pictus), representing the the Paleocene to the Oligocene (34–23 Ma) is conA ned to Family Canidae (Caniformia), from Kruger National Park, South Eurasia and North America (4), supporting the view that Africa. Credit: E. Eizirik. E. Eizirik and W. J. Murphy. Carnivores (Carnivora). Pp. 504–507 in e Timetree of Life, S. B. Hedges and S. Kumar, Eds. (Oxford University Press, 2009). HHedges.indbedges.indb 550404 11/28/2009/28/2009 11:30:22:30:22 PPMM Eukaryota; Metazoa; Vertebrata; Mammalia; Carnivora 505 Mephitidae 9 Mustelidae 12 Procyonidae Ailuridae 7 Otariidae 14 Odobenidae 5 11 Caniformia Phocidae 3 Ursidae Canidae Felidae 8 1 Prionodontidae 4 Hyaenidae 10 Herpestidae 13 6 2 Eupleridae Feliformia Viverridae Nandiniidae Paleogene Neogene CENOZOIC 50 25 0 Million years ago Fig. 2 A timetree of carnivores (Carnivora). Divergence times are shown in Table 1. it is contained in Arctoidea (pinnipeds + Ursidae + carnivoran with unique morphological and behavioral Mustelidae + Procyonidae + Red Panda + skunks (see characteristics oJ en placed in Viverridae, Herpestidae, or later)), which is nested within the carnivoran Suborder in its own monotypic family (1, 2, 26). 7 e last few years Caniformia. have seen a surge in studies on these and other topics of In addition to the argument over the relationships of the carnivoran phylogeny (e.g., 7–10, 18–23) most of which Pinnipedia, several other aspects of the carnivoran tree used concatenations of multiple nuclear and/or mitochon- have been contentious over the last few decades, leading drial genes. 7 is has led to a consistent resolution of most to the production of a large body of literature on the phyl- suprafamilial nodes (Fig. 2), settling many of the disputes ogeny of this mammalian order (e.g., 4–30). Most of the brieP y outlined earlier. 7 e Giant Panda was established eB ort has been traditionally focused on the Caniformia, as the most basal extant ursid, and the Red Panda is now and particularly on the positions of the Giant Panda placed in its own monotypic family (Ailuridae), nested (Ailuropoda melanoleuca), Red Panda (Ailurus fulgens), in Arctoidea. Skunks and stink badgers (Mydaus spp.) and skunks (originally placed in the Mephitinae within are closely related, and together constitute the Family Mustelidae, but now recognized as comprising a separate Mephitidae, which is not immediately connected to the family, Mephitidae). On the feliform side, the monophyly Mustelidae (14, 15, 28). Another recent A nding is that of the Family Viverridae has been challenged multiple Mustelidae and Procyonidae are each other’s closest times, starting with the proposition that the African relatives, an observation which is supported by several Palm Civet (N. binotata) was actually not a member of studies (e.g., 8, 15, 24). 7 ere is a core group in Arctoidea this family, but rather the only living representative of the containing Mephitidae, Ailuridae, and Mustelidae + most basal extant lineage of the Feliformia (e.g., 4, 24, 25). Procyonidae, whose internal structure has still not been Another feliform whose phylogenetic a1 nities have his- conA dently resolved (Fig. 2). Pinnipedia is now seen as torically been enigmatic is the Fossa (C. ferox), a Malagasy the closest relative of this core clade, with Ursidae being HHedges.indbedges.indb 550505 11/28/2009/28/2009 11:30:24:30:24 PPMM 506 THE TIMETREE OF LIFE Table 1. Divergence times (Ma) and their confi dence/credibility intervals (CI) among carnivores (Carnivora). Timetree Estimates Node Time Ref. (19) Ref. (22) Ref. (23) Ref. (30) Ref. (31) Ref. (32) Ref. (33) Time CI Time CI Time CI Time CI Time CI Time CI Time CI 152.9– – ––––– – 55.060–5157.562–5246.257–35 249.0– – 54.6–––43.354–33– – – – – – 344.0––––44––––––––– 441.8– – 47.0–––36.547–29– – – – – – 540.0––––40––––––––– 639.7– – 44.2–––35.245–28– – – – – – 738.0––––38––––––––– 837.833.335–3242.3––––––––––– 935.0––––35––––––––– 1033.8––38.3–––29.238–23–––––– 1130.0––––30––––––––– 1228.0––––28––––––––– 1328.0––31.6–––24.432–18–––––– 14 22.0 – – – – 22 – – – – – – – – – Note: Node times in the timetree represent the mean of time estimates from different studies. the most basal lineage in Arctoidea. Canidae is indeed studies (Fig. 2). Felidae and Prionodontidae are each other’s the most basal family in the Caniformia, supporting the closest relatives, as are Eupleridae and Herpestidae. traditional view that it is the only extant lineage of the Hyaenidae is the closest relative of the Eupleridae + Superfamily Cynoidea. Herpestidae clade. 7 e relative position of Viverridae Among feliforms, recent studies have led to major (now restricted to a monophyletic core group) has not changes in the prevailing views on phylogenetic structure been conA dently established with high support, though and evolutionary history. Two separate studies published most studies indicate that it is more closely related to the in 2003 have shown further evidence of viverrid para- Hyaenidae + Eupleridae + Herpestidae clade (Fig. 2). phyly, and identiA ed novel lineages that are now recog- Although many studies have addressed carnivoran nized as valid families (19, 20). Asian linsangs (Prionodon relationships, few have assessed the age of the inferred spp.), traditionally part of the Viverridae, have been clades using molecular data. 7 e results reviewed here shown to be the closest relative of the Felidae (19), and are drawn mostly from four recent studies, which have are now placed in their own family, Prionodontidae (21). separately addressed each of the two carnivoran sub- Another remarkable A nding was that all Malagasy car- orders (19, 22, 23, 30). 7 e basal divergence between nivores (including the Fossa), traditionally placed in the Feliformia and Caniformia seems to have occurred Viverridae or Herpestidae, comprise a separate, endemic between the Paleocene and the middle Eocene (49–40 monophyletic lineage (20), which is now recognized as Ma), with the dates used here being derived from large Family Eupleridae. In addition, all recent studies that studies involving all placental mammal clades (31, 32) or included Nandinia conA rmed that this taxon is indeed multiple vertebrate groups (33). Within Feliformia, the the most basal feliform, and now constitutes its own divergence between Nandiniidae and the other lineages family, Nandiniidae (1). As a whole, these recent stud- was estimated by one study to be ~43 Ma ( 30), and by ies have challenged not only the monophyly of tradition- another to be ~55 Ma (22); the latter may be an over- ally recognized Viverridae, but also the monophyly of estimate given the branch length observed in multiple Herpestidae, restructuring the feliform phylogeny to a studies between the feliform–caniform split and this large degree.