42 



multiple independent speciation events from a polytomy) as a computational aid 

 (Maddison & Maddison 1992; Swofford 1993). Although a consensus tree does not 

 necessarily represent an optimal solution for a given data set (Swofford 1993), and does 

 not here, we will refer to the consensus solution as the overall parsimony solution as we 

 believe it to present the best summary of the data set. The consensus solution reflects the 

 conflict present within the phocines, and its single cladogram provides a more efficient 

 discourse. All tree descriptions (see Appendices D through F) are derived from the 

 consensus solution. However, the optimal tree length will be taken to be that of the two 

 equally most parsimonious solutions (69,834 steps) and the various goodness-of-fit 

 statistics will also refer to the optimal solutions (as will be the case for all other analyses 

 as well), unless specified otherwise. 



Outgroup relations (Fig.5C) 



This analysis confirms a monophyletic Pinnipedia, with the otarioids forming a 

 monophyletic sister group to the phocids. This is in accordance with the recent upswing 

 in support of a monophyletic Pinnipedia among morphological studies (e.g., Wyss 1987; 

 Flynn et al. 1988; Wolsan 1993; Wyss & Flynn 1993), but contradicts the recent contention 

 of an Odobemts-phocid clade (Wyss 1987; Wyss & Flynn 1993; Berta 1991; Berta & Wyss 

 1994; Vrana et al. 1994). However, a monophyletic Otarioidea is among the most strongly 

 supported of all clades, ranging in support from 16 to 21 unweighted steps, depending 

 upon the optimization criterion employed (see Appendix D for weighted branch lengths; 

 the identities of the synapomorphies supporting each node can be found in Appendix E, 

 and are discussed in the Character Analysis section). 



A somewhat unexpected result was that of a lutrine affinity for the pinnipeds, with Lutra 

 being the immediate sister group and Enhydra being the sister group to Lutra and the 

 pinnipeds. Although many early workers allied the phocids and the lutrines based on 

 superficial similarities [see references in Taylor (1914)], a lutrine affinity for the phocids 

 based on more robust characters has only been suggested by four workers: Flower (1869), 

 Mivart (1885), McLaren (1960b), and de Muizon (1982a). To our knowledge, however, 

 the equivalent scenario has never been postulated for the pinnipeds as a whole, as the 

 otarioids are typically allied with the ursids under this otherwise diphyletic scenario. Most 

 recent studies advocating a monophyletic Pinnipedia (and including both ursid and 

 mustelid outgroups) conclusively indicate an ursid affinity for the pinnipeds (e.g., Vrana 

 et al. 1994; Lento et al. 1995). Only Wolsan (1993) allies a monophyletic Pinnipedia with 

 the lutrine-like fossil Potamotherium (which he considers to be a pinniped) within his 

 Mustelida, but his exclusion of any undisputed lutrines [a lutrine affinity for 

 Potamotherium has recently come into question (C.A. Repenning pers. comm.; A.R. Wyss 

 pers. comm.)], plus the lack of an ursid outgroup, precludes any definitive statement on 

 lutrine affinities. However, a Lwrra-pinniped pairing is generally supported here no less 

 strongly than any other outgroup node and is minimally indicated by seven unequivocal 

 synapomorphies. As well, a lutrine affinity also implies a mustelid affinity for a mono- 

 phyletic Pinnipedia, which, although still rare, is a somewhat more commonly held 

 hypothesis [e.g., Arnason & Widegren 1986; Miyamoto & Goodman 1986 (albeit as the 

 sister group to a mustelid-procyonid clade); Wolsan 1993]. 



