Taxonomic Issues in Bears 33

Home , Ailuropoda, Ailuropodinae, Amur leopard, Asian black bear, Eurasian brown bear, Spectacled bear

REVIEW: TAXONOMIC ISSUES IN BEARS 33

Int. Zoo Yb. (2010) 44: 33–46 DOI:10.1111/j.1748-1090.2009.00087.x

Taxonomic issues in bears: impacts on conservation in zoos and the wild, and gaps in current knowledge A. C. KITCHENER Department of Natural Sciences, National Museums Scotland, Chambers Street, Edinburgh EH1 1JF, United Kingdom, and Institute of Geography, School of Geosciences, University of Edinburgh, Drummond Street, Edinburgh EH8 9XP, United Kingdom E-mail: [email protected]

Taxonomy is essential for underpinning conservation actually comprises two or more species, some science and action, and the international and national of which are critically endangered, poor taxo- implementation of protective legislation. However, many of the current scientific species and subspecies names for nomic research may inadvertently compro- bears have a poor scientific basis. Poor understanding of mise the survival of those threatened ursid taxonomy could compromise conservation both in populations. In captivity, hybridization may the wild and in captivity; all eight ursid species are listed occur accidentally between unrecognized on the Convention on International Trade of Endangered subspecies and even species, ultimately wast- Species of Wild Fauna and Flora and 75% are Endan- gered or Vulnerable. Although there has been much ing huge amounts of resources as well as molecular research on ursids in recent years, this has affecting the conservation of the species. mainly focused on phylogenetic relationships, including Conversely, if too many species or subspecies resolution of whether the Giant panda Ailuropoda mela- are recognized, captive populations may suf- noleuca is an ursid. Some phylogeographical studies have provided new insights into geographical variation fer from inbreeding from small founder po- of some bear species, but these studies are often only pulations, when they would benefit from the regional, or lack sufficient samples, or use only mtDNA. mixing of needlessly separated captive gene There is an urgent need for integrated molecular and pools. However, we must remember that morphological studies of geographical variation of all taxonomy and systematics are fluid – they bear species in order to establish a robust taxonomy for the Ursidae for enhanced conservation management and represent, to a greater or lesser extent, a action. consensus based on current evidence, and today’s taxonomies will continue to change Key-words: bears; conservation management; conser- and, hopefully, improve as better evidence vation science; taxonomy. accumulates. In this paper, I examine the taxonomic INTRODUCTION issues that currently affect the Ursidae. Most Although taxonomy is often regarded as species are widespread, occurring in a wide rather dull, it provides the essential under- variety of habitats, and some species, such as pinning of all conservation. If species and the Brown bear Ursus arctos and the Amer- subspecies do not have scientific names, it is ican black bear Ursus americanus, show not possible to provide full legal protection enormous phenotypic variation in size and for them under national or international law coloration that has led to the recognition of (e.g. O’Brien & Mayr, 1991). Therefore, it is very many species and subspecies in the past. fundamental for conservation science and For Brown bears, some 232 modern [of action, as well as the implementation of which 83 were described by C. Hart Merriam national and international protective legisla- in North America (Hall, 1981)] and 39 fossil tion, that the taxonomy of species and subspecies and subspecies have been described species is accurate. If a widespread species (see Erdbrink, 1953), which Kurteń&

Int. Zoo Yb. (2010) 44: 33–46. c 2010 The Authors. Journal compilation c 2010 The Zoological Society of London 34 BEARS AND CANIDS

Anderson (1980) considered ‘a waste of and populations. Combining molecular and taxonomic effort, which, as far as we know, geographical data is known as phylogeogra- is unparalleled’. There are many problems phy, and it has developed into a potent tool for that affect our understanding of the taxonomy examining geographical genetic variation and of ursids and many other mammals. Firstly, population subdivision. The analysis of an- the scientific names on which species and cient DNA (aDNA) even allows us to look subspecies are based are poorly founded back in time at the genetics of past populations scientifically. These names date mainly from and even extinct species, providing further the 19th century when none, one or only a insight into how the distribution of genetic handful of specimens were examined to de- variation has arisen today (Leonard et al., termine the diagnostic characteristics and 2000; Loreille et al., 2001; Barnes et al., differences from related species. It should be 2002; Valdiosera et al., 2007; Krause et al., noted that normally there is a requirement for 2008). New morphometric techniques, such as a holotype, which is the specimen that is geometric morphometrics, allow us to examine selected as the voucher on which the scien- subtle differences in the shapes and sizes of tific name is based (Mayr & Ashlock, 1991; skulls, the normal part of a skeleton that is the Groves, 2001), but a wide range of specimens subject of taxonomic research, including re- needs to be examined to determine which moving and/or controlling for the influence of characters are diagnostic (i.e. typical or char- size, which often reflects a phenotypic re- acteristic) for the taxon, and which are the sponse to the available local food resources. result of individual or other kinds of variation In other words, well-fed bears grow bigger! in the population. Therefore, it is not surpris- Sometimes the molecular and morphological ing that these descriptions capture only a approaches are in strong accord, in other cases fraction of the normal range of individual they disagree, often because the old morpho- variation within a species or subspecies. logical taxonomy is poorly based in science. Where variation is clinal (i.e. changes gradu- The best studies involve a combination of ally over a geographical range) it is easy to morphological and molecular techniques, see how the analysis of a small number of without recourse to previous taxonomic ar- samples distant from each other might result rangements, as checks against each approach. in several new species or subspecies being However, whatever the approach, good inadvertently recognized. Over time, this pro- taxonomic research requires statistically sig- cess has led inevitably to a proliferation of nificant samples that cover the entire range of names, for which there is little or no scientific the species. Using current museum collec- basis. tions, this may be very hard to achieve. Normally taxonomic differences are based Inevitably there are collecting biases that on diagnostic morphological characters, such reflect the interest and collecting activities of as coloration, size, shape, presence/absence of individuals, and the accessibility of the geo- characters, and combinations of measure- graphical range of a species. Opportunities to ments, particularly from skulls, but account enhance the number of specimens available must also be taken of sex, age, season, for all kinds of research must be taken in etc. (Pocock, 1941; Erdbrink, 1953). In recent order to provide a better resource base for years, new techniques and technologies have taxonomic and other research. In the mean- been developed which have greatly benefited time other complementary approaches, such taxonomy and systematics. New molecular as dynamic biogeographical modelling, techniques allow us to examine genetic varia- should be considered to provide a framework tion over wide geographical areas regardless of against which we can examine our current sex, age and local phenotypic responses to the knowledge of the geographical variation environment. These molecular studies com- within each ursid species. pare the sequencing of base pairs of mitochon- As is typical for many mammalian fa- drial and nuclear DNA between individuals milies, the bears suffer from numerous

Int. Zoo Yb. (2010) 44: 33–46. c 2010 The Authors. Journal compilation c 2010 The Zoological Society of London REVIEW: TAXONOMIC ISSUES IN BEARS 35 taxonomic uncertainties. While the number of size to hold bamboo stems while feeding recognized species has remained stable over (Pocock, 1941; Salesa et al., 2006b). Spec- the last 140 years, our knowledge of the tacled bears also have a small pseudo-thumb geographical variation in bears, given their for grasping, which might suggest a close phenotypic variation, is still very poor, relationship, but basal arctoids, such as Simo- although some recent studies have improved cyon batalleri (an ancestor of the Red panda our understanding for some species or popu- Ailurus fulgens) from near Madrid, Spain, lations. In this article, I review the current also had a well-developed radial sesamoid, taxonomic status of bear genera, species and which was apparently originally adapted for subspecies, highlighting uncertainties and fu- climbing (Salesa et al., 2006a,b). Therefore, ture areas for research. The impact that these the pseudo-thumb is an ancestral feature or uncertainties have on conservation in the wild plesiomorphy, which cannot be used to infer and in captivity will also be examined. that the Giant panda is an ursid. Thenius (1989) argued that while molecular data are good at establishing phylogenies, particularly GENERA where morphological specializations may The number of genera of bears has varied confuse them, they should not dictate taxon- from one for each species to three in total omy. Thenius (1989) considered that the (Table 1). There have been numerous studies Giant panda should be placed in its own of the phylogenetic relationships between family, the Ailuropodidae, because it has a modern bears based mainly on mitochondrial series of morphological and behavioural au- DNA (mtDNA) (Zhang & Ryder, 1994; tapomorphies (derived features) that clearly Talbot & Shields, 1996a,b; Waits et al., distinguish it from the Ursidae. 1998, 1999; Yu et al., 2007; Krause et al., Resolution of the phylogenetic relation- 2008) but until recently there had been none ships between the ursine bears is difﬁcult, that used nuclear DNA or a combination of because they appear to have radiated rapidly both (Yu et al., 2004; Page`s et al., 2008). All about 5 million years ago (McLellan & the phylogenetic trees show the Giant panda Reinder, 1994; Yu et al., 2007; Krause et al., Ailuropoda melanoleuca as the most basal 2008; Page`s et al., 2008). Most phylogenies with a coalescence time varying between 12 show a close relationship between the Brown and 22 million years ago (Ma), depending on bear and Polar bear Ursus maritimus, while the molecular method used. Then comes the the Asian black bear Ursus thibetanus and Spectacled bear Tremarctos ornatus which American black bears are frequently shown diverged 6–15 Á 6 Ma. The Giant panda has a as sister species, although the American black distinctive adaptation, the pseudo-thumb or bear is sometimes shown to be closer to the radial sesamoid, which has evolved a large Brown/Polar bears (Yu et al., 2007; Krause

PAGE` S ET AL. (2008), WOZENCRAFT (2005); EISENBERG (1981) KRAUSE ET AL. (2008) THIS PAPER Brown bear Ursus arctos Ursus arctos Ursus arctos Polar bear Thalarctos maritimus Ursus maritimus Ursus maritimus American black bear Euarctos americanus Ursus americanus Ursus americanus Asian black bear Selenarctos thibetanus Ursus thibetanus Ursus thibetanus Sun bear Helarctos malayanus Ursus malayanus Helarctos malayanus Sloth bear Melursus ursinus Ursus ursinus Melursus ursinus Spectacled bear Tremarctos ornatus Tremarctos ornatus Tremarctos ornatus Giant panda Ailuropoda melanoleuca Ailuropoda melanoleuca Ailuropoda melanoleuca

Table 1. The changing generic classiﬁcation of bears.

Int. Zoo Yb. (2010) 44: 33–46. c 2010 The Authors. Journal compilation c 2010 The Zoological Society of London 36 BEARS AND CANIDS et al., 2008; Page`s et al., 2008). The greatest also uncertain, being alternately recognized uncertainty remains with the phylogenetic as being in its own genus, Helarctos, or the placement and generic assignment of the most basal Ursus. Its skull and forelimb Sloth bear Melursus ursinus and Sun bear morphology seem to be different from those Helarctos malayanus, each of which are of other ursine bears and I would continue to variously shown as basal members of the support its recognition as H. malayanus until Ursinae. The Sloth bear shows a number of more evidence is available. In a molecular distinct autapomorphies related to its adapta- phylogeny, Page`s et al. (2008) distinguish tions for feeding on termites and ants, includ- Helarctos and Melursus because of the ability ing only two pairs of upper incisors, a of Ursus spp to hibernate. Interestingly, there concave palate, well-developed lips and is a dramatic morphological convergence ﬂanges on the nose, highly reduced molar between the skulls of the Sun and Spectacled and premolar dentition, and a low metabolic bears (Plate 1), which presumably relates to rate (Pocock, 1941; McNab, 1992). Some of similarities in diet and foraging techniques. the most recent phylogenies support the Sloth But why is this generic discussion impor- bear as basal to the rest of the ursines and its tant? It has been suggested that the use of recognition as belonging to a monotypic scarce conservation resources ought to be genus, Melursus (Yu et al., 2007; Krause dictated partly or completely by the taxo- et al., 2008), while others are more uncertain nomic and phylogenetic uniqueness of a (Page`s et al., 2008). The Sun bear’s status is species (e.g. Isaac et al., 2007). Therefore, if

b d

Plate 1. Dorsal and lateral views of skulls of the Sun bear Helarctos malayanus (right: c,d) and Spectacled bear Tremarctos ornatus (left: a,b) have a strikingly similar morphology, which may reﬂect convergent evolution owing to similarities in diet and foraging. r Neil McLean, National Museums Scotland.

Int. Zoo Yb. (2010) 44: 33–46. c 2010 The Authors. Journal compilation c 2010 The Zoological Society of London REVIEW: TAXONOMIC ISSUES IN BEARS 37 the Sun and Sloth bears belong to their own selection for the Polar bear phenotype would genera, they would be regarded as being of have occurred, perhaps coupled with genetic greater conservation value than if they were drift, and the populations would have di- in the genus Ursus. Therefore, better phylo- verged. On 16 April 2006, a Polar bear of genetic and taxonomic studies at the generic unusual appearance was shot on Banks Is- level are important not only for determining land, Northwest Territories, Canada, which relationships between species but also they was found to be a Polar bear U. mariti- may be key to determining which species mus Â Grizzly bear U. arctos horribilis hy- benefit from conservation action in the brid (Doupe´ et al., 2007) and there are future. increasing reports of Grizzly bears in this part of the Canadian Arctic, perhaps indicating a SPECIES permanent range extension. It is possible that hybridization could occur increasingly fre- The number of species of bear has remained quently during the current global warming remarkably stable for 140 years, even though episode. However, it is unknown if clade IIb there has been considerable debate as to mtDNA is fixed in Polar bears, which could which genera they should belong and the support this speculation of ancient hybridiza- interrelationships between species. Several tion and temporary introgression, so that different mtDNA clades have been identified further sampling is required over a wide among Brown bears. Currently, mtDNA data geographical area (A. Cooper, pers. comm.). suggest that Brown bears are paraphyletic, Interestingly, using nuclear DNA, and in because the Polar bear is derived from the contrast to their mtDNA data, Page`s et al. same mtDNA clade (clade II) as the ABC (i.e. (2008) found that Brown and Polar bears are Admiralty, Baranof and Chichagof Islands) reciprocally monophyletic, suggesting an Brown bears (Taberlet & Bouvet, 1994; Tal- ancient hybridization event (I. Barnes, pers. bot & Shields, 1996a,b; Waits et al., 1999), comm.). and following conventional taxonomy, two or more Brown bear species ought to be recog- SUBSPECIES nized. However, Brown bears from two apparently distinct mtDNA clades, the western Polar bear U. maritimus (clade I) and eastern (clade III) clades, do The Polar bear is regarded as a monotypic interbreed in contact zones in the wild and species. Although subspecies have been pro- there are apparently no consistent morpholo- posed in the past, none are recognized today. gical differences between the bears from Wilson (1976) found that there was clinal these clades (Taberlet et al., 1998). It is variation in skull morphometrics across the possible that the presence of Brown bear Polar bear’s North American range with the mtDNA in Polar bears may reflect an ancient possible exception of a distinct population in hybridization event that may even have al- South Alaska, which was proposed as a lowed the Polar bear to survive interglacials. subspecies. A recent analysis of population As the climate warmed, Brown bears were genetics was unable to support the recogni- able to move further north (see Doupe´ et al., tion of subspecies but it did recognize some 2007) and Polar bears may have become geographical structuring of the global popu- stranded on land owing to lack of pack ice, lation (Paetkau et al., 1999). so that opportunities for hybridization in- creased, resulting in more or less complete Brown bear U. arctos introgression. We know from stable isotope studies that some Polar bears had a terrestrial With its widespread distribution in the north- rather than a marine diet (Kitchener & Bon- ern hemisphere across a wide range of habi- sall, 1997), supporting their land-locked ex- tats, it is not surprising that the Brown bear istence. As the climate cooled, rapid natural shows a great deal of phenotypic variation.

Int. Zoo Yb. (2010) 44: 33–46. c 2010 The Authors. Journal compilation c 2010 The Zoological Society of London 38 BEARS AND CANIDS

This has led to the description of a plethora of southern refugia of the western clade subspecies based on morphological differ- throughout the LGM, so that today’s pattern ences. Among those traditionally recognized of geographical variation is the result of a are: Eurasian brown bear Ursus arctos arctos, high degree of , philopatry and a drastic Syrian bear Ursus arctos syriacus, the now- reduction in geographical distribution during extinct Atlas bear Ursus arctos crowtheri, the late Holocene, probably owing to human- Isabelline or Red bear Ursus arctos isabellinus, mediated habitat loss and hunting. This study Blue bear Ursusarctospruinosus,Kamchatka emphasizes that today’s phylogeographical bear Ursus arctos beringianus, Hairy-eared patterns may not reflect those in deeper time bear Ursus arctos lasiotus, Kodiak bear Ur- and hence caution should be exercised in sus arctos middendorffi, Grizzly bears Ursus developing conservation management plans arctos horribilis and Ursus arctos dalli, etc. for bears based only on phylogeographies of (Pocock, 1932a,b; Hall, 1981). mtDNA. Morphological investigations have not In Scandinavia, both western and eastern been forthcoming in recent years [e.g. Erd- clades were found to be separated by more brink (1953) is one of the most recent], than 130 km with little evidence for contact possibly owing to the difficulty of obtaining between them (Taberlet & Bouvet, 1994; sufficiently large samples of a highly variable Taberlet et al., 1998). The Scandinavian species, which now occupies a highly frag- populations were originally subdivided into mented distribution in many parts of its four management units based on mtDNA range. However, molecular studies abound, haplotypes but this was subsequently reduced although many of these are restricted to to three by Manel et al. (2004). However, it mtDNA analyses in particular range countries should be noted that <-mediated gene flow or regions (e.g. Saarma et al., 2007), so that a occurs between these sub-populations, so that clear picture of what genetic variation means the pattern of mtDNA haplotypes we see taxonomically is not always possible. The today reflects a high degree of , philopatry lack of information from nuclear DNA has as a result of historical colonization patterns also led to misinterpretations of the meaning following the last glaciation and into the early of mtDNA phylogenies. For example, Taber- Holocene (Taberlet et al., 1995). Similar <- let & Bouvet (1994) identified two main mediated gene flow is seen between North mtDNA lineages in Eurasian Brown bears; a American Brown bear populations, despite , western one, comprising two subclades, and philopatry reinforcing an apparent geographi- an eastern one stretching from eastern Europe cal separation between mtDNA clades (Paet- through northern Eurasia [see also Loreille kau et al., 1998; Barnes et al., 2002). et al. (2001) and Hofreiter et al. (2002) for Overall five mtDNA clades were originally relationships with the extinct Cave bear identified in the Brown bear (Waits et al., Ursus spelaeus]. Contact zones were found 1999) (Table 2) but a recent analysis of aDNA in Scandinavia and central Europe (Taberlet has revealed a sixth basal clade, which was et al., 1998). These subclades were assumed found in subfossil bones from two caves in to have arisen in refugia in the Iberian, Italian Algeria and Morocco (Calvignac et al., 2008). and Balkan peninsulas when populations This study supports the view that the Altas were thought to have been isolated during bear U. a. crowtheri was a distinct subspecies, the Last Glacial Maximum (LGM), which although comparison with Middle Eastern was the peak of the last Ice Age about bears U. a. syriacus would be required to 20 000 years ago. However, Valdiosera et al. confirm this. Other North African finds were (2007) examined ancient mitochondrial DNA from the European western clade (V), which from Brown bear fossils in Europe, from suggests these Brown bears were imported to Spain, France, Germany, Italy and Romania. North Africa, possibly by the Romans. Combined with existing aDNA data, they A small relict population of very small claimed that gene flow occurred between the Brown bears from the inhospitable Gobi desert

Int. Zoo Yb. (2010) 44: 33–46. c 2010 The Authors. Journal compilation c 2010 The Zoological Society of London REVIEW: TAXONOMIC ISSUES IN BEARS 39

via the Bering land bridge during the Pleisto- GEOGRAPHICAL NOMINAL cene (Talbot & Shields, 1996a,b). However, CLADE DISTRIBUTION SUBSPECIES/SPECIES an aDNA analysis of fossil Brown bear bones I Western Europe, Balkans, arctos, marsicanus from Alaska has revealed that all three southern Scandinavia extant haplotypes occurred in Alaska and the II ABC Islands/Arctic horribilis/U. maritimus Yukon 35 000–45 000 radiocarbon years ago III Eastern Alaska, northern horribilis (Leonard et al., 2000). Therefore, today’s Canada, eastern phylogeographical pattern could only have Siberia, Honshu, occurred relatively recently. A more in-depth Japan study with far larger numbers of samples IV Southern Canada, horribilis northern USA confirmed that three haplotypes co-existed in V Eastern Europe, northern arctos, beringianus, Alaska before 35 000 BP, and two were found Asia, Caucasus, ?syriacus, between 10 000 and 21 000 BP, before the ?Middle East, Tibet, pruinosus, present-day phylogeographical pattern was China, Japan lasiotus, middendorffi, established (Barnes et al., 2002). Brown horribilis bears were absent from Alaska during VI Atlas Mountains, North crowtheri,?syriacus 21 000–35 000 BP and were only able to reco- Africa, ?Middle East lonize North America once the huge carni- VII W Himalayas, Gobi isabellinus vorous Short-faced bear Arctodus simus was forced out by the spreading ice sheets of the Table 2. Mitochondrial DNA clades of the Brown Last Ice Age. This study concluded that the bear Ursus arctos (Waits et al., 1999; Barnes et al., 2002; Galbreath et al., 2007; Calvignac et al., 2008). present phylogeographical pattern was estab- It is unclear whether syriacus from the Middle East lished naturally after the end of the Last Ice should be in clade V or VI because data are not Age or by the possible impact of human available. activities through megafaunal extinctions or later habitat loss and persecution. was identified by Schaller (1998) as probably Kurteń (1973) recognized three North belonging to the Isabelline bear U. a. isabelli- American subspecies; horribilis in most of nus. A recent mtDNA study confirmed this mainland North America, dalli representing identification but also showed that the Isabel- the large coastal bears of British Columbia line bear formed a distinct ancient (seventh) and Alaska and the huge middendorffi from clade close to the Polar bears/ABC Brown Kodiak Island. Using microsatellites (nuclear bears of North America (Galbreath et al., DNA), Paetkau et al. (1998) found that 2007). Moreover, this limited analysis con- despite their distinctive mtDNA, the ABC firmed that the blue bear U. a. pruinosus of bears were not genetically isolated from Tibet, formed a subclade within the wider east- mainland horribilis and the coastal grizzlies ern Eurasian/North American clade (clade V). (dalli) were part of the contiguous continental In North America numerous species and Grizzly bear population. Only the Kodiak subspecies names have been given to local bear showed evidence of little or no genetic populations of Brown bears, and mtDNA exchange with mainland populations. There- studies have identified three haplotypes that fore, it is inappropriate to use mtDNA alone, are geographically separated. These include especially where ,, are highly philopatric, in the ABC Islands (which also share their order to infer colonization patterns of Brown mtDNA haplotype with Polar bears; clade bears and hence to inform management of II), the United States haplotype (clade IV) present populations. and the Alaskan/Canadian haplotype (clade III) (Talbot & Shields, 1996a,b). American black bear U. americanus It was originally assumed that the pattern of three allopatric haplotypes in North Amer- Up to 16 subspecies of American black bear ica reflected three different colonizations have been recognized (Hall, 1981) but this

Int. Zoo Yb. (2010) 44: 33–46. c 2010 The Authors. Journal compilation c 2010 The Zoological Society of London 40 BEARS AND CANIDS would seem to be unlikely, given its former habitat since the beginning of the Neolithic contiguous distribution. Recent phylogeogra- era c. 9000 years ago and the exploitation of phical studies have identified two mtDNA Asian black bears for traditional medicines by lineages; one in the west, mainly coastal in the early civilizations in northern China distribution, while the other seems to occupy since at least the Shang dynasty c. 3500 years the rest of the distribution (Wooding & Ward, ago (Barnes, 1999; Ren, 2000), almost cer- 1997; Stone & Cook, 2000), but these tainly led to the development of this disjunct lineages overlap in south-east Alaska. There distribution. Therefore, by default of their is also evidence of recent contact and hybri- complete isolation Ussuri black bears should dization between them elsewhere (Wooding probably be treated as a distinct subspecies. & Ward, 1997). Other studies are more loca- Other subspecies are described from Pakistan lized. For example, a study of skull morpho- in the west to south-east Asia but it is unclear metrics in the eastern United States and whether these represent clinal variation or Canada showed clinal variation in body size distinct geographical variation. from large animals in the south (Louisiana and eastern Texas) to small animals in the Sloth bear M. ursinus north (Quebec), which appears to reflect The Sloth bear is usually regarded as having differences in abundance of food (Kennedy two subspecies; Melursus ursinus ursinus et al., 2002). Another study examined the from the Indian peninsula and Melursus ursi- subspecific status of black bears in the White nus inornatus from Sri Lanka. They are River National Wildlife Refuge, Arkansas apparently distinguishable on size [mean (Warrilow et al., 2001). However, sampling greatest lengths of skull: India (322 mm, has not been comprehensive enough from n 5 6), Sri Lanka (280 mm, n 5 4): Pocock, throughout the geographical distribution in 1941; Corbet & Hill, 1992]. However, I am any of the published studies carried out so far not aware of any recent molecular studies and to answer the question as to how many the sample sizes for the skull differences are subspecies there may be. It is possible that very small. It is possible that there is clinal the two mtDNA lineages may represent sub- variation in skull size from larger animals in species, although without examining nuclear the north to smaller animals in the south but DNA and morphological data, this is mere this has not been examined so far. speculation. It seems more likely that most variation in American black bears is clinal Sun bear H. malayanus with no subspecies recognized. The Sun bear is also usually regarded as having two subspecies; Helarctos malayanus Asian black bear U. thibetanus malayanus on the Asian mainland and Suma- There are no recent taxonomic or phylogeo- tra, and Helarctos malayanus euryspilus from graphical studies on the Asian black bear (e.g. Borneo (Chasen, 1940; Ellerman & Morri- Pocock, 1932a,b, 1941). Several subspecies son-Scott, 1951; Corbet & Hill, 1992). A have been described, including some on recent analysis of skull morphometrics by islands (Japan ssp Ursus thibetanus japoni- Meijaard (2004) has confirmed the distinc- cus; Taiwan ssp Ursus thibetanus formosa- tiveness of the Bornean population, which nus), which may be distinct. The Ussuri black has a smaller skull size [mean condylobasal bear Ursus thibetanus ussuricus, like the lengths of skull: Borneo (206 Á 3 mm), Suma- Amur tiger Panthera tigris altaica, is sepa- tra (227 Á 3 mm), Asia (235 Á 3 mm)] but a rated from the rest of its conspecifics by more relatively longer upper toothrow. However, than 1500 km (Servheen et al., 1998). there was no distinction between Sumatran Although the Asian black bear’s distribution and mainland samples. Meijaard (2004) re- was once contiguous from the Russian Far ferred also to a limited unpublished molecular East to southern Asia, the loss of woodland study on 300 base pairs of mtDNA from

Int. Zoo Yb. (2010) 44: 33–46. c 2010 The Authors. Journal compilation c 2010 The Zoological Society of London REVIEW: TAXONOMIC ISSUES IN BEARS 41

Sumatran and Bornean Sun bears by L. Waits, phological data, as well as providing a frame- who found five clades, which did not fully work for focused future research. This accord with geography, suggesting that some approach has been taken for the Tiger gene flow may have occurred between the Panthera tigris, which is traditionally re- islands during the LGM. Further molecular garded as having eight (or even nine) sub- and morphological research could help to species, apparently supported by molecular confirm these tentative conclusions. data (Kitchener, 1999; Luo et al., 2004). However, morphological data suggest per- Spectacled bear T. ornatus haps two lineages that derive from the Asian mainland and the Sunda islands (Kitchener, The spectacled bear is treated as a monotypic 1999; Maza´k, in press). A biogeographical species (e.g. Mondolfi, 1989), although there model by Kitchener & Dugmore (2000), is considerable individual variation in facial which took into account changes in habitat markings. The Spectacled bear is the only since the LGM 20 000 years ago, supported survivor of the radiation of New World Short- both morphological and molecular analyses faced bears, subfamily Tremarctinae, which (Luo et al., 2004; Maza´k, in press), even included the formidable Short-faced bear though their conclusions as to subspecies ´ A. simus, of North America (Kurten & An- were not in accord. This approach would be derson, 1980; Krause et al., 2008). of great benefit to ursids, particularly the very Five other specific or subspecific names widespread Brown, American black and have been given to Spectacled bears based on Asian black bears. Indeed, the results from differences in body proportions, claw length, Kitchener & Dugmore (2000) are broadly facial markings, and skull sizes and propor- applicable to the Asian black bear, which tions. For example, Hornaday (1911) gave would suggest that perhaps only two subspe- the name Ursus ornatus thomasi to a Spec- cies would be recognized; the mainland Ur- tacled bear from the Andes of southern Co- sus thibetanus thibetanus and the Japanese lombia, which was living at the New York black bear U. t. japonicus. It is likely that the Zoological Gardens, USA, on the basis of a Taiwanese black bear would have been con- lack of white facial markings. However, this tiguous with the mainland population when diagnostic character as with the others listed sea levels were lower during the LGM, and above are examples of individual variation. the Ussuri black bear would once have been in direct contact with Chinese populations Giant panda A. melanoleuca and was derived from them following the Until recently the Giant panda was regarded end of the last Ice Age. The now complete as a monotypic species but a subspecies, isolation of the Ussuri black bear probably Ailuropoda melanoleuca qinlingensis, from warrants its recognition as a distinct subspe- the Qinling Mountains, Shaanxi province, cies, U. t. ussuricus, even though the cause of has been described based on genetic differ- its separation is likely to have been human ences from Sichuan Giant pandas Ailuropoda activity. Molecular and morphological stu- melanoleuca melanoleuca (Wang et al., dies could test whether mainland variation is 2003; Wan et al., 2005). The proposed new largely clinal and whether the Japanese black subspecies differs morphologically, having a bear is distinctive. browner pelage and smaller skull with larger molars (Wan et al., 2005). CONSEQUENCES OF LACK OF KNOWLEDGE DYNAMIC BIOGEOGRAPHY Lack of a clear taxonomy for bears may have Given the limitations of current samples, important consequences for conservation at biogeographical modelling may offer ways the levels of species, subspecies and local of interpreting existing molecular and mor- populations, both in captivity and in the wild.

Int. Zoo Yb. (2010) 44: 33–46. c 2010 The Authors. Journal compilation c 2010 The Zoological Society of London 42 BEARS AND CANIDS

For example, the recognition of fewer mono- may prevent mixing of depleted gene pools typic genera (e.g. if the Sun bear were con- owing to local population bottlenecks. The sidered congeneric with other Ursus spp) recent confirmation of the Gobi bear as being could affect whether some species are con- effectively a malnourished Isabelline bear U. sidered for conservation action, especially if a. isabellinus (Galbreath et al., 2007), may resources are limited. The current uniqueness lead us to question whether conservation of the Sun bear H. malayanus could disappear resources should be directed to a tiny popula- instantly if reclassified as Ursus malayanus, tion at the margin of the main subspecies downgrading it in the conservation stakes range. In contrast the recent proposal of a new (see differing points of view: Krause et al., Giant panda subspecies (Wang et al., 2003; 2008; Page`s et al., 2008). Clearly, we must Wan et al., 2005) could have important con- take great care in whether to assign species to sequences for the conservation of this species unique genera or not in the light of new data, (if inadvertent hybridization has not already and it is important to take all evidence into occurred), which would in theory require account rather than only short fragments of double the enclosure space to cater for two mtDNA. subspecies. Great care must be taken to be Poor knowledge of subspecies often results certain that new taxa really do exist before in inadvertent hybridization in captivity so adding to the pile of scientific names that that conservation effort is wasted, and source burden the bears. stock may no longer be available from the Reintroductions may also suffer from lack wild or the remaining pure captive population of knowledge about the intraspecific taxon- may be too small to prevent high levels of omy of a species. For example, following the inbreeding. A classic example was the dis- discovery of two mtDNA subclades of Brown covery that the Association of Zoos and bears in Europe (Taberlet & Bouvet, 1994), Aquariums (AZA) Asian lion Panthera leo the decision was made to reinforce the persica Species Survival Plans Program was seriously threatened Pyrenean brown bear totally compromised by unrecorded hybridi- population (the western subclade Ia) by zation with African lions (O’Brien et al., introductions of animals from Croatia (sub- 1987); and the Amur leopard Panthera par- clade Ib), which was widely criticized at the dus orientalis captive-breeding programme time because of the genetic mixing of two will always be contaminated by the influence different subclades. However, later studies of of a founder that was not an Amur leopard fossil Brown bears showed that the current (Uphyrkina et al., 2002). The decision by the geographical distribution of these subclades AZA to allow hybridization of Sloth bear could not be supported, so that they are an subspecies in North American zoos is surely artefact of post-glacial colonization and/or insupportable if we want to maintain com- human impact (Hofreiter et al., 2004; Valdio- mon standards and the integrity of ex situ sera et al., 2007). Similarly, the isolation of conservation programmes. A huge proportion the mtDNA haplotype II on the ABC Islands of enclosure space in European zoos is taken does not constitute the recognition of these up by Brown bears (and Asian black bears) of bears as a unique subspecies, because unknown or mixed origin, and while many of <-mediated gene flow occurs at a relatively these zoos will want to maintain ‘local’ high level with mainland Grizzly bears (Paet- Brown bears, much enclosure space could be kau et al., 1998). We must take care that freed up for bears of known threatened spe- genetic evidence for colonization history is cies and subspecies that desperately require not confused with the unnecessary recogni- extra holders (e.g. Sloth bears) for the estab- tion of subspecies. lishment of viable captive populations. Even though Polar bears are monotypic, Lack of recognition of subspecies in the there is some genetic substructuring (Paetkau wild may result in loss of unique populations, et al., 1999) and some big differences in size or the recognition of too many subspecies and shape of skulls (Wilson, 1976) within the

Int. Zoo Yb. (2010) 44: 33–46. c 2010 The Authors. Journal compilation c 2010 The Zoological Society of London REVIEW: TAXONOMIC ISSUES IN BEARS 43

SPECIES SPECIES GENERIC STATUS STATUS SUBSPECIES STATUS REFERENCES Brown/grizzly bear 11 11 ?? – very uncertain; distinct Paetkau et al. (1998), Ursus arctos geographical variation, but Galbreath et al. lacking comprehensive global (2007), Calvignac molecular and morphological et al. (2008) study Polar bear Ursus 11; formerly 11 11 – subspecies not usually Wilson (1976), maritimus Thalarctos recognized; geographical Paetkau et al. variation fairly well (1999) understood American black bear 11, formerly 11 ?? – very uncertain; Wooding & Ward Ursus americanus Euarctos comprehensive morphological (1997), Stone & and molecular studies would Cook (2000), be beneficial Kennedy et al. (2002) Asian black bear 11, formerly 11 ? – uncertain, especially in China Pocock (1941) Ursus thibetanus Selenarctos and South East Asia; requires comprehensive morphological and molecular studies Sloth bear Melursus 1, sometimes 11 ? – uncertain; two subspecies Corbet & Hill (1992) ursinus placed in recognized, but requires Ursus confirmation from molecular and further morphological research Sun bear Helarctos 1, sometimes 11 ? – uncertain; two subspecies Meijaard (2004) malayanus placed in recognized, but requires Ursus confirmation from further morphological and molecular research Spectacled bear 11; subfamily 11 1 – usually regarded as Mondolfi (1989) Temarctos ornatus Tremarctinae monotypic, but much individual pelage variation; molecular and morphological research recommended Giant panda 11; subfamily 11 ? – recent molecular research Thenius (1989), Wang Ailuropoda Ailuropodinae; suggests two subspecies, but et al. (2003), Wan melanoleuca sometimes further morphological and et al. (2005) placed in molecular research monotypic recommended family, Ailuropodidae

Table 3. Current knowledge of the status of family, genera, species and subspecies of all ursid species: 11, high degree of stability and certainty; 1, moderate certainty, but subject to change in the light of new evidence; ??, high degree of uncertainty; ?, moderate degree of uncertainty. global population. Is this variation in mor- ogy and genetics point to adaptation to the phology the result of phenotypic responses to local environment. different local food sources, or does it have a Clearly we cannot wait until all taxonomic genetic component? Given that captive con- research is completed to a common level of servation of Polar bears is likely to become a scientiﬁc rigour, so informed decisions will vital component of their future survival dur- have to be made in the short to medium term ing the current global warming episode, we by bringing together specialists in genetics must take care not to mix Polar bears from and morphology, as well as ﬁeld workers and different local populations if their morphol- zoo managers in order to pool current

Int. Zoo Yb. (2010) 44: 33–46. c 2010 The Authors. Journal compilation c 2010 The Zoological Society of London 44 BEARS AND CANIDS knowledge effectively. However, much taxo- through basic photography, they can record nomic research still remains to be done. seasonal and interannual changes in coat colour as fur is moulted and is subject to WHAT NEXT? fading and wear through the year. These kinds of contributions from zoos are espe- There is clear stability over the number of cially important for rare subspecies and/or ursid species but phylogenies continue to be bears with known geographical origin. All reﬁned and phylogeographical studies are these data separately make tiny contributions increasing, although they remain focused to our overall knowledge, but without them largely on mtDNA. Table 3 reviews the we will never make progress and succeed in current knowledge of genera, species and our mission to have effective conservation subspecies of all ursids and recommends action in zoos and in the wild. further areas for research. It is clear that all species require urgently comprehensive studies of genetic (both mitochondrial and nu- REFERENCES clear DNA) and morphological variation in BARNES, G. L. (1999): The rise of civilization in East Asia: the archaeology of China, Korea and Japan. order to ascertain which of the many de- London: Thames & Hudson. scribed subspecies have taxonomic validity BARNES, I., MATHEUS, P., SHAPIRO, B., JENSEN,D.& for ensuring the effectiveness of conservation COOPER, A. (2002): Dynamics of Pleistocene population action and implementation of national and extinctions in Beringian brown bears. Science 295: international legislation. Every opportunity 2267–2270. CALVIGNAC, S., HUGHES, S., TOUGARD, C., MICHAUX, J., should be taken for the ethical and legal THEVENOT, M., PHILIPPE, M., HAMDINE,W.& HA¨ NNI,C. collection of samples for these studies to (2008): Ancient DNA evidence for the loss of a highly enhance those existing in museums and la- divergent brown bear clade during historical times. boratories worldwide. Molecular Ecology 17: 1962–1970. CHASEN, F. N. (1940): A handlist of Malaysian mammals. In conclusion taxonomic issues are just one Bulletin of the Rafﬂes Museum No. 15: 1–209. important factor that is potentially compro- CORBET,G.B.& HILL, J. E. (1992): The mammals of the mising conservation of bears in captivity, Indomalayan region: a systematic review. London: Nat- along with issues relating to enhanced long- ural History Museum Publications, and Oxford: Oxford evity, lack of opportunities for breeding and University Press. DOUPE´, J. P., ENGLAND, J. H., FURZE,M.& PAETKAU,D. poor enclosure design and other welfare (2007): Most northerly observation of a grizzly bear issues. In this paper I have tried to explain (Ursus arctos) in Canada: photographic and DNA evi- why it is important for both captive and wild dence from Melville Island, Northwest Territories. Arctic conservation of bears that our knowledge of 60(3): 271–276. EISENBERG, J. F. (1981): The mammalian radiations. bear taxonomy is as good as it can be, so that Chicago, IL: Chicago University Press. we can ensure that scarce conservation re- ELLERMAN,J.R.& MORRISON-SCOTT, T. C. S. (1951): sources are appropriately targeted without Checklist of Palaearctic and Indian mammals 1758– wasting effort through accidental hybridiza- 1946. London: Trustees of the British Museum. ERDBRINK, D. P. (1953): A review of fossil and recent tion or unnecessary splitting of populations. bears of the Old World with remarks on their phylogeny Molecular methods can be very powerful in based on their dentition 2. Deventer, the Netherlands: looking at relationships between populations, Jan de Lange. but they are not a panacea and should be GALBREATH, G. J., GROVES,C.P.& WAITS, L. P. (2007): viewed critically in relation to other kinds of Genetic resolution of composition and phylogenetic placement of the isabelline bear. Ursus 18(1): 129–131. evidence. Zoo managers can assist advances GROVES, C. P. (2001): Primate taxonomy. Washington, in taxonomy by supplying samples for mole- DC: Smithsonian Institution Press. cular research and ensuring that dead bears HALL, E. R. (1981): The mammals of North America II are preserved, whenever possible, in museum (2nd edn). New York, NY: John Wiley. HOFREITER, M., CAPELLI, C., KRINGS, M., WAITS, L., collections for morphological research. Zoos CONARD, N., MU¨ NZEL, S., RABEDER, G., NAGEL, D., can themselves contribute basic information PAUNOVIC, M., JAMBRSIC´, G., MEYER, S., WEISS,G.& such as body measurements and weights, and PA¨ A¨ BO, S. (2002): Ancient DNA analyses reveal high

Int. Zoo Yb. (2010) 44: 33–46. c 2010 The Authors. Journal compilation c 2010 The Zoological Society of London REVIEW: TAXONOMIC ISSUES IN BEARS 45 mitochondrial DNA sequence diversity and parallel tigris). PLoS Biology 2: 2275–2293. Available at http:// morphological evolution of late Pleistocene cave bears. biology.plosjournals.org/perlserv/?request=get-doc Molecular Biology and Evolution 19: 1244–1250. umentdoi=10.1371%2Fjournal.pbio.0020442. HOFREITER, M., SERRE, D., ROHLAND, N., RABEDER, G., MCLELLAN,B.& REINDER, D. C. (1994): A review of bear NAGEL, D., CONARD, N., MU¨ NZEL,S.& PA¨ A¨ BO, S. (2004): evolution. International Conference on Bear Research Lack of phylogeography in European mammals before and Management 9(1): 85–96. the last glaciation. Proceedings of the National Academy MCNAB, B. K. (1992): Rate of metabolism in the termite- of Sciences USA 101: 12963–12968. eating sloth bear (Ursus ursinus). Journal of Mammalogy HORNADAY, W. T. (1911): The spectacled bear. Bulletin of 73: 168–172. the New York Zoological Society No. 45: 747–748. MANEL, S., BELLEMAIN, E., SWENSON,J.E.& FRANÇIS,O. ISAAC,N.B.,TURVEY,S.T.,COLLEN,B.,WATERMAN,C.& (2004): Assumed and inferred spatial structure of popula- BAILLIE, J. E. M. (2007): Mammals on the EDGE: tions: the Scandinavian brown bears revisited. Molecular conservation priorities based on threat and phylogeny. Ecology 13: 1327–1331. PLoS ONE 2(3): e296. DOI: 10.1371/journal.pone. MAYR,E.& ASHLOCK, P. D. (1991): Principles of 0000296. Available at http://www.plosone.org/arti systematic biology. New York, NY: McGraw-Hill. cle/info%3Adoi%2F10.1371%2Fjournal.pone.0000 MAZA´ K, J. (In press): Craniometric variation in the tiger 296. (Panthera tigris): implications for patterns of diversity, KENNEDY, M. L., KENNEDY, P. K., BOGAN,M.A.& WAITS, taxonomy and conservation. Mammalian Biology. DOI: J. L. (2002): Geographic variation in the black bear 10.1016/j.mambio.2008.06.003. (Ursus americanus) in the eastern United States and MEIJAARD, E. (2004): Craniometric differences among Canada. The Southwestern Naturalist 47: 257–266. Malayan sun bears (Ursus malayanus); evolutionary and KITCHENER, A. C. (1999): Tiger distribution, phenotypic taxonomic implications. The Raffles Bulletin of Zoology variation and conservation issues. In Riding the tiger. 52(2): 665–672. Tiger conservation in human-dominated landscapes: 19– MONDOLFI, E. (1989): Notes on the distribution, habitat, 39. Seidensticker, J., Christie, S. & Jackson, P. (Eds). food habits, status and conservation of the spectacled Cambridge: Cambridge University Press. bear (Tremarctos ornatus Cuvier) in Venezuela. Mamma- KITCHENER,A.C.& BONSALL, C. (1997): AMS radio- lia 53: 525–544. carbon dates for some extinct Scottish mammals. Quar- O’BRIEN,S.J.& MAYR, E. (1991): Bureaucratic mischief: ternary Newsletter. No. 83: 1–11. recognizing endangered species and subspecies. Science KITCHENER,A.C.& DUGMORE, A. J. (2000): Biogeogra- 251: 1187–1188. phical change in the tiger, Panthera tigris. Animal O’BRIEN, S. J., JOSLIN, P., SMITH III, G. L., WOLFE, R., Conservation 3: 113–124. SCHAFFER, N., HEATH, E., OTT-JOSLIN, J., RAWAL,P.P., KRAUSE, J., UNGER, T., NOÇON, A., MALASPINAS, A.-S., BHATTACHARJEE,K.K.& MARTENSON, J. S. (1987): Evi- KOLOKTRONIS, S.-O., STILLER, M., SOIBELZON, L., SPRIGGS, dence for African origins of founders of the Asiatic lion H., DEAR, P. H., BRIGGS, A. W., BRAY, S. C. E., O’BRIEN, species survival plan. Zoo Biology 6: 99–116. S. J., RABEDER, G., MATHEUS, P., COOPER, A., SLATKIN, M., PAETKAU, D., SHIELDS,G.F.& STROBECK, C. (1998): Gene PA¨ A¨ BO,S.& HOFREITER, M. (2008): Mitochondrial gen- flow between insular, coastal and interior populations of omes reveal an explosive radiation of extinct and extant brown bears in Alaska. Molecular Ecology 7: 1283– bears near the Miocene–Pliocene boundary. BMC Evolu- 1292. tionary Biology 8: 220–232. PAETKAU, D., AMSTRUP, S. C., BORN, E. W., CALVERT,W., KURTEŃ, B. (1973): Transberingian relationships of Ursus DEROCHER, A. E., GARNER, G. W., MESSIER, F., STIRLING, arctos Linne´ (brown and grizzly bears). Commentationes I., TAYLOR, M. K., WIIG,Ø.& STROBECK, C. (1999): Biologicae 65: 1–10. Genetic structure of the world’s polar bear populations. KURTEŃ,B.& ANDERSON, E. (1980): Pleistocene mammals Molecular Ecology 8: 1571–1584. of North America. New York, NY: Columbia University PAGE` S, M., CALVIGNAC, S., KLEIN, C., PARIS, M., HUGHES, Press. S. & HA¨ NNI, C. (2008): Combined analysis of fourteen LEONARD,J.A.,WAYNE,R.K.& COOPER, A. (2000): Popula- nuclear genes refines the Ursidae phylogeny. Molecular tion genetics of Ice Age brown bears. Proceedings of the Phylogenetics and Evolution 47: 73–83. National Academy of Sciences USA 97: 1651–1654. POCOCK, R. I. (1932a): The black and brown bears of LOREILLE, O., ORLANDO, L., PATOU-MATHIS, M., PHILIPPE, Europe and Asia. Journal of the Bombay Natural History M., TABERLET,P.& HA¨ NNI, C. (2001): Ancient DNA Society 35: 771–823. analysis reveals divergence of cave bear, Ursus spelaeus, POCOCK, R. I. (1932b): The black and brown bears of and brown bear, Ursus arctos, lineages. Current Biology Europe and Asia. Journal of the Bombay Natural History 11: 200–203. Society 36: 101–138. LUO, S. J., KIM, J. H., JOHNSON,W.E.,VAN DER WALT, J., POCOCK, R. I. (1941): The fauna of British India including MARTENSON, J., YUKHI, N., MIQUELLE, D. G., UPHYRINKA, Ceylon and Burma. Mammalia vol. II Carnivora (con- O., GOODRICH, J. M., QUIGLEY, H. B., TILSON, R., BRADY, tinued from vol. 1). Suborders Aeluroidea (part) and G., MARTELLI, P., SUBRAMANIAM, V., MCDOUGAL, C., Arctoidea. London: Taylor and Francis. HEAN, S., HUANG, S. Q., PAN, W., KARANTH, U. K., REN, G. (2000): Decline of the mid- to late Holocene SUNQUIST, M., SMITH,J.L.& O’BRIEN, S. J. (2004): forests in China: climatic change or human impact? Phylogeography and genetic ancestry of tigers (Panthera Journal of Quaternary Science 15: 273–281.

Int. Zoo Yb. (2010) 44: 33–46. c 2010 The Authors. Journal compilation c 2010 The Zoological Society of London 46 BEARS AND CANIDS

SAARMA, U., HO, S. Y. W., PYBUS, O. G., KALJUSTE, M., VALDIOSERA, C. E., GARCIÁ, N., ANDERUNG, C., DALEŃ, L., TUMANOV, I. L., KOJOLA, I., VOROBIEV, A. A., MARKOV,N. CRE´GUT-BONNOURE, E., KAHLKE, R.-D., STILLER, M., I., SAVELJEV, A. P., VALDMANN, H., LYAPUNOVA, E. A., BRANDSTRO¨ M, M., THOMAS, M. G., ARSUAGA, J. L., ABRAMOV, A. V., MA¨ NNIL, P., PUTCHKOVSKIY,S.V.& GO¨ THERSTRO¨ M,A.& BARNES, I. (2007): Staying out in the RO˜ KOV, A. M. (2007): Mitogenetic structure of brown cold: glacial refugia and mitochondrial DNA phylogeo- bears (Ursus arctos L.) in northeastern Europe and a new graphy in ancient European brown bears. Molecular time frame for the formation of European brown bear Ecology 16: 5140–5148. lineages. Molecular Ecology 16: 401–413. WAITS, L., PAETKAU,D.& STROBECK, C. (1998): Genetics SALESA, M. J., ANTO´ N, M., PEIGNE´,S.& MORALES,J. of the bears of the world. In Bears. Status survey and (2006a): Evidence of a false thumb in a fossil carnivore conservation action plan: 25–32. Servheen, C., Herrero, clarifies the evolution of pandas. Proceedings of the S. & Peyton, B. (Eds). Gland, Switzerland: IUCN. National Academy of Sciences USA 103: 379–382. WAITS, L. P., SULLIVAN, J., O’BRIEN,S.J.& WARD,R.H. SALESA, M. J., SILICEO, G., ANTO´ N, M., ABELLA, J., (1999): Rapid radiation events in the family Ursidae MONTOYA,P.& MORALES, J. (2006b): ‘‘Anatomy of the indicated by likelihood phylogenetic estimation from ‘‘false thumb’’ of Temarctos ornatus (Carnivora, Ursidae, multiple fragments of mtDNA. Molecular Phylogenetics Temarctinae): phylogenetic and functional implications. and Evolution 13: 82–92. Estudios Geolo´gicos 62: 389–394. WAN, Q.-H., WU,H.& FANG, S.-G. (2005): A new SCHALLER, G. B. (1998): Wildlife of the Tibetan steppe. subspecies of giant panda (Ailuropoda melanoleuca) Chicago, IL: Chicago University Press. from Shaanxi, China. Journal of Mammalogy 86: 397– SERVHEEN, C., HERRERO,S.& PEYTON, B. (Eds) (1998): 402. Bears. Status survey and conservation action plan. WANG, Q.-H., FANG, S.-G., WU,H.& FUJIHARA, T. (2003): Gland, Switzerland: IUCN. Genetic differentiation and subspecies development of SHIELDS, G. F., ADAMS, D., GARNER, G., LABELLE, M., the giant panda as revealed by DNA fingerprinting. PIETSCH, J., RAMSAY, M., SCHWARTZ, C., TITUS,K.& Electrophoresis 24: 1353–1359. WILLIAMSON, S. (2000): Phylogeography of mitochon- WARRILOW, J., CULVER, M., HALLERMAN,E.& VAUGHAN, drial DNA variation in brown bears and polar bears. M. (2001): Subspecificaffinity of black bears in the Molecular Phylogenetics and Evolution 15: 319–326. White River National Wildlife Refuge. The American STONE,K.D.& COOK, J. A. (2000): Phylogeography of Genetics Association 92: 226–233. black bears (Ursus americanus) of the Pacific Northwest. WILSON, D. E. (1976): Cranial variation in polar bears. In Canadian Journal of Zoology 78: 1218–1223. Bears: their biology and management. A selection of TABERLET,P.& BOUVET, J. (1994): Mitochondrial DNA papers from the third international conference on bears: polymorphism, phylogeography, and conservation genet- 447–453. Pelton, M. R., Lentfer, J. W. & Folk, G. E. ics of the brown bear Ursus arctos in Europe. Proceed- (Eds). Morges, Switzerland: IUCN. ings of the Royal Society London B 255: 195–200. WOODING,S.& WARD, R. (1997): Phylogeography and TABERLET, P., SWENSON, J. E., SANDEGREN,F.& BJA¨ RVALL, Pleistocene evolution in the North American black bear. A. (1995): Localization of a contact zone between two Molecular Biology and Evolution 14: 1096–1105. highly divergent mitochondrial DNA lineages of the WOZENCRAFT, W. C. (2005): Order Carnivora. In Mammal brown bear Ursus arctos in Scandinavia. Conservation species of the world. A taxonomic and geographic Biology 9: 1255–1261. reference (3rd edn): 532–628. Wilson, D. E. & Reeder, TABERLET, P., FUMAGALLI, L., WUST-SAUCY, A.-G. & COS- D. M. (Eds). Baltimore, MD: The Johns Hopkins Uni- SON, J.-F. (1998): Comparative phylogeography and versity Press. postglacial colonization routes in Europe. Molecular YU, L., LI, Q., RYDER,O.A.& ZHANG, Y. (2004): Ecology 7: 453–464. Phylogeny of the bears (Ursidae) based on nuclear and TALBOT,S.L.& SHIELDS, G. F. (1996a): Phylogeography mitochondrial genes. Molecular Phylogenetics and Evo- of brown bears (Ursus arctos) of Alaska and paraphyly lution 32: 480–494. within the Ursidae. Molecular Phylogenetics and Evolu- YU, L., LI, Y.-W., RYDER,O.A.& ZHANG, Y.-P. (2007): tion 5: 477–494. Analysis of complete mitochondrial genome sequences TALBOT,S.L.& SHIELDS, G. F. (1996b): A phylogeny of increases phylogenetic resolution of bears (Ursidae), a the bears (Ursidae) inferred from complete sequences of mammalian family that experienced rapid speciation. three mitochondrial genes. Molecular Phylogenetics and BMC Evolutionary Biology 7: 198–209. Evolution 5: 567–575. ZHANG,Y.& RYDER, O. (1994): Phylogenetic relation- THENIUS, E. (1989): Molekulare and ‘‘adaptive’’ Evolu- ships of bears (the Ursidae) inferred from mitochondrial tion, Kladistik und Stammesgeschichte. Erga¨nzungen zu DNA sequences. Molecular Phylogenetics and Evolution einer Arbeitshypothese. Zeitschrift fu¨r Zoologische Evo- 3: 351–359. lutionforschung 27: 94–105. UPHYRKINA, O., MIQUELLE, D., QUIGLEY, H., DRISCOLL,C. Manuscript submitted 15 April 2009; revised & O’BRIEN, S. J. (2002): Conservation genetics of the Far Eastern leopard (Panthera pardus orientalis). The Jour- 23 June 2009; accepted 15 September 2009 nal of Heredity 93: 303–311.

Int. Zoo Yb. (2010) 44: 33–46. c 2010 The Authors. Journal compilation c 2010 The Zoological Society of London