Phylogenetic Position of the Giant Panda HISTORICAL CONSENSUS THROUGH SUPERTREE ANALYSIS ERHAPS NO QUESTION inmammalian Ailuropoda "sufficiently distinct" to justify being systematics has engendered such long-term placed in its own family?),whereas a study of controversy and uncertainty as the phylogenetic its phylogenetic or sister-group relationships is placement of the giant panda (Ailuropoda much more objective and concrete. Second, mdanoleuca). Although its formal introduction through the use of supertree construction (sensu to Western science placed it as a member of the Sanderson et al. 1998), I am able to infer the bear family (Ursidae) (David 1869), similarities consensus estimate of the affinity of Ailuropoda to the lesser or red panda (Aihsfulgens) and, for any given time period. Thus, I can demon- by extension, raccoons and allies (Procyonidae) strate how consensus opinion has shifted over were quickly noted (Milne-Edwards1870). Since time and by how much. This resembles work that time, a variety of evidence has been used to done previously with Ailurus and the pinnipeds ally Ailuropoda with ursids, procyonids, or Ailu- (Bininda-Emonds 2oooa). Because Ailurus has rus (either within the previous two families or as played a critical historical role in the controversy the separate family Ailuridae), or to place it in a surrounding the origin of Ailuropoda, I also per- family by itself (Ailuropodidae).Fueled largely form similar analyses for it here. by molecular evidence, there is perhaps finally a growing consensus that Ailuropoda represents ISSUES OF EVIDENCE the sister group to the remaining ursids. AND CONVERGENCE In this chapter, I approach the question of panda phylogeny from a historical perspective to The uncertainty surrounding the placement of examine trends in the placement of Ailuropoda Ailuropoda derives from the numerous similar- through time. My work follows on from that of ities that it shows to each of ursids, procyonids, O'Brien et al. (1991),although it differs in two and especially to Ailurus. In fact, were it not for key respects. First, I employ a phylogenetic rather the existence of the procyonid-like Ailurus, the than a taxonomic perspective. The taxonomic acceptance of Ailuropoda as an ursid would likely status of any species is highly subjective (e.g.,is be considerably less-if at all-in dispute. The Giant Pandas: Biology and Conservation (Lindburg, D.G. and K. Baragona, eds.) University of California Press, Berkeley, CA. 2004. TABLE 1.1 Shared Features between Ailuropoda and Ursids, Procyonids, and Ailurus FEATURE OF AILUROPODA URS1DAE PROCYONIDAE AIL URUS Morphology Size and gross morphology + Skull robustness - Brain morphology + Auditory region and ossicles + Epipharyngeal bursa Ridges on hard palate + Dentition (especially massive size) - + Skeletal robustness - Specialized sesamoid on forepaw ? (unique) Respiratory tract + Intestines (shorter and less complex) - External (soft)morphology ? (unique) Genitalia Coloration Hair structure + Molecular Karyotype + Serology/immunology + Sequence data + Behavioral Vocalizations - Life history traits + Scent marking (and glands) - Feeding behavior Mating behavior Other Fossil affinities + NOTES: A plus sign indicates that the feature is similar between the two taxa. A minus sign indicates that the feature has been used to argue against a relationship between the two taxa, even ifAilumpoda does not share the feature with another taxon. key lines of evidence used to infer the ancestry any species, Davis (1964: 322) declaredthat "every of Ailuropoda historically are given in table 1.1. morphological feature examined indicates that The majority ofevidence, both morphological and the giant panda is nothing more than a highly molecular, allies Ailuropoda with ursids. Fossil specialized bear." However, a few morphological evidence, particularly the inferred close relation- features have been used to argue against such a ship with the Pliocene fossil ursid Hyaenarctos relationship, even if they do not indicate a rela- (now included in Agriotherium, a member of tionship with either procyonids or Ailurns ("neg- the extinct sister group to ursids) (McKenna and ative evidence"; see below and table 1.1). These Bell 1997; Hunt, chapter 3), also unanimously include the greater than expected robustness of indicates an ursid origin. the skull, skeleton, and dentition of Ailuropoda In one of the most thorough and impressive for an animal of its size (Morris and Morris comparative morphological investigations for 1981). 12 01.AF R. P. BININDA-EMONDS Virtually all molecular evidence points to first holds that Ailuropoda is an ursid that has an ursid affinity for Ailuropoda (Waits, brief re- shifted to an almost exclusively herbivorous diet port 1.1). Only the karyotype is equivocal. In its (the so-called "bear school"). This scenario ac- diploid number, Ailuropoda (2N=42) resembles counts for such features as the more robust procyonids (e.g., 2N = 42 for Procyon lotor) and dentition as being obvious (convergent)adapta- Ailurus (2N = 44) more so than ursine bears tions for herbivory. The second holds that Ail- (2N=74). Although this observation is correctly uropoda is instead derived from a small herbi- attributed to Newnham and Davidson (1966), vore, typically with procyonid affinities, that has the attendant implication of procyonid ancestry converged secondarily on a larger bearlike body is not. Newnham and Davidson (1966: 161) ex- plan (the "raccoon school"). The greater than plicitly pointed out that large differences in karyo- expected robustness of the skull and skeleton types and diploid numbers occur within such has been used as evidence of rapid growth in the families as Canidae (dogs) and that the evidence lineage leading to Ailuropoda (Morris and Mor- merely indicates Ailuropoda to be a different ris 1981). The raccoon school in particular has species from ursine bears. They added that the relied on the negative evidence found in table 1.1 number of chromosomal arms ("nombrefonda- in combination with behavioral information. mental") might be a more informative measure Proponents of the raccoon school concede in this regard. Along these lines, O'Brien et al. that the majority of evidence places Ailuropoda (1985) discovered that, although Ailuropoda pos- with ursids (e.g., Ewer 1973; Morris and Morris sesses a procyonid-like diploid number, the band- 1981).In supporting a nonursid origin, they in- ing patterns of its chromosomes are virtually stead argue that the fewer features that cluster identical to those of ursine bears. Together with Ailuropoda with either procyonids or Ailurus rep- other molecular evidence, they persuasively ar- resent evolutionary novelties that are more diffi- gued that Ailuropoda is related to ursids and that cult to envisage evolving on multiple occasions its reduced diploid number is a result of exten- (Ewer 1973; Morris and Morris 1981). Similar sive chromosomal fusion in the past (see also reasons have been used to cluster megachirop- Nash et al. 1998). It is noteworthy that similar, teran bats with primates on the basis of a shared albeit independent, fusions have been inferred optic network and other neural features, in spite to explain the karyotype of the spectacled bear of an overwhelming number of similarities, (Tremarctos omatus) (2N=52) (Nash and O'Brien mostly related to flying, with microchiropteran 1987; Nash et al. 1998). bats (Pettigrew 1986,1991).An extension ofthis Finally, behavioral characteristics strongly general argument is that organisms with a sim- group the two panda species together. Only life ilar body plan are more likely to develop conver- history traits are similar between Ailuropoda gently evolved features, given similar selective and ursids (Garshelis, chapter 4). Otherwise, Ail- regimes. This argument was used to explain uropoda resembles Ailurus in its unusual scent- why the two main groups of pinnipeds (true seals marking behavior and in its mating and feeding versus sea lions and walruses) were so similar behaviors. The latter is the most remarkable, morphologically, despite formerly being believed with both pandas being renowned for the ability to have separate ancestors (McLaren 1960; to manipulate precisely their herbivorous food Mitchell 1967; Repenning 1990). This conclu- items, although only Ailuropoda possesses an sion is now held to be false (Wyss 1987; Vrana enlarged sesamoid that acts as an analog of an et al. 1994; Bininda-Emonds et al. 1999).With opposable thumb. respect to Ailuropoda, the full argument is that Researchers have sought to make sense of its overall similarity with ursids applies only to the conflicting signals within the phenotypic features that are phenotypically plastic or that evidence (i.e.,morphological and behavioral data) are expected to show a greater degree of conver- through one of two evolutionary scenarios. The gence in response to similar selection pressures, PHYLOGENETIC POSITION OF THE GIANT PANDA 13 possibly due to the inheritance of a common, spect to whether panda relationships were ex- primitive, arctoid body plan (Ewer 1973; Morris amined in isolation or not. The first method and Morris 1981). assessed simple statements of phylogenetic affin- The issue is not easily resolved. Molecular ity of the form "Ailuropodais most closely related evidence has played a valuable role, because con- to. " To summarize this information quanti- vergence at this level is unlikely to mirror that tatively, I derived a simple affinity metric. State- at the morphological