Pinnipeds and Carnivores: Phyletic Relationships and Classification J. Scott Shannon

Shannon, J. S. (Department of Wildlife, Humboldt State University, Arcata, California 95521) 1987. Pinnipeds and carnivores: phyletic relationships and classification. Syst. Zool. xx:xxx-xxx. Evidence for monophyly and polyphyly of the pinnipeds is summarized and discussed with reference to the classification of the higher subtaxa of the Pinnipedia and the order Carnivora. On the basis of a consensus analysis of data supporting pinniped monophyly, an alternative classification of the Carnivora is proposed. If pinnipeds are demonstrated to be of monophyletic origin, they should be assigned no higher than superfamilial rank (superfamily Pinnipedea new) within the infraorder Arctoidea. If mustelid ancestry for the phocids is firmly established, the classification of Tedford (1976) should be retained. Determination of actual phyletic relationships between pinnipeds and carnivores awaits discovery of additional fossil evidence of proto-phocid and early arctoid forms. The phyletic relationships of the higher taxonomic groups comprised by the Pinnipedia (Chordata:Mammalia) have posed problems for mammal systematists for over a century. The question of whether the seals (superfamily Phocoidea Smirnov, 1908) and the sea lions and walruses (superfamily Otarioidea Gill, 1866) share an immediate common ancestor remains a topic for debate among taxonomists. At stake is the validity of the widely accepted classifications presently in use. Typically, the superfamilies Otarioidea and Phocoidea are united in the suborder Pinnipedia Illiger, 1811 of the order Carnivora Bowdich, 1821 (Simpson, 1945), although Pinnipedia has also been assigned full ordinal rank (Scheffer, 1958). For the two superfamilies to be legitimately united in their own distinct taxon, however, their descent from a common ancestor (monophyly) must be presumed. Conversely, if a taxon embraces animals which are descendants of different ancestors (polyphyly), then its members are falsely united, and as such, the taxon misrepresents actual phylogenetic patterns. If the taxon Pinnipedia is found to be invalid, then, this has direct implications for the classification of its higher subtaxa, and of the order Carnivora, as well. W. H. Flower (1869) subdivided the Carnivora into three superfamilies based upon the 2 morphology of the auditory bulla and associated basicranial structures. In a study of these basicranial characters in living pinnipeds, Mivart (1885) described similarities in the auditory bullae of pinnipeds with the carnivores of Flower's superfamily Arctoidea (the families Ursidae, Mustelidae, and Procyonidae). Mivart (1885) concluded that the pinniped family Otariidae was descended from an early ursid-like lineage, while the Phocidae was derived from an early lutrine (otter-like) mustelid lineage. This statement has subsequently become the classical thesis for pinniped polyphyly. Fossil evidence that could be used to test Mivart's (1885) hypothesis was slow to accumulate, however, because of the poor representation of early pinnipeds in the fossil record. Even today, no proto-phocid fossil form has been discovered. Still, there have been significant fossil discoveries that have helped to fill important gaps in our knowledge of pinniped origins. Much of this evidence has been interpreted as upholding Mivart's (1885) claim for a polyphyletic derivation of the Pinnipedia. Savage (1957) described the anatomy of an early arctoid lutrine, Potamotherium valletoni Geoffroy, 1833 (Figure 1), known from numerous well-preserved specimens collected from late Oligocene lacustrine deposits in central France. In his analysis, Savage (1957) noted a number of resemblances in the cranial, axial, and appendicular skeletons of Potamotherium and Phoca. Although Savage (1957) did not feel confident enough to propose Potamotherium as the ancestor of the phocids, he did believe that the resemblances between Potamotherium and Phoca were of possible phylogenetic significance. Savage (1957) also commented on the many similarities between Potamotherium and the Pliocene seal-otter Semantor macrurus described by Orlov (1931) (Figure 2). Semantor is known from a single specimen from the Neogene of western Siberia. Precise systematic placement of Semantor is not possible because only the post-lumbar portions of the axial and appendicular skeleton were preserved (Figure 3). Nevertheless, enough remains to reveal a picture of a large, facultatively terrestrial, lutromorphic pinniped. According to Savage (1957): ...while Potamotherium is a morphological link between Lutra and Phoca, Semantor is a morphological link between Potamotherium and Phoca. Because of its Pliocene appearance, however, Semantor could not be the evolutionary