43 



Interestingly, despite Taylor (1914) noting a high degree of convergence between Enhydra 

 (but not Lutra) and the phocids due to the constraints of their largely aquatic habitats, it 

 is the apparently less aquatically adapted Lutra that forms the immediate sister group to 

 the pinnipeds here. The overall implication of this result is that the relationships advocated 

 herein (and especially the historically unusual lutrine-pinniped pairing) are based more on 

 phylogenetically informative characters than on convergent aquatic features. 

 If the pinnipeds are momentarily ignored, the fissiped outgroups otherwise fall out as is 

 commonly held for the caniforms. The lutrines form a monophyletic clade within the 

 remaining mustelids (Martes), with the procyonids (Procyon), ursids (Ursus), and canids 

 (Canis) forming successive sister taxa to the mustelids (see Tedford 1976; Miyamoto & 

 Goodman 1986; Flynn et al. 1988; Wyss & Flynn 1993). Support for the constituent nodes 

 appears reasonably robust, ranging from eight (Mustelidae under either ACCTRAN or 

 DELTRAN optimization) to 27 (Arctoidea under ACCTRAN optimization) unweighted 

 steps (Fig.5C; Appendices D and E). It should be noted, however, that this arrangement 

 is dependent on the somewhat subjective placement of the root of the cladogram (which 

 has Canis as the ultimate outgroup). This placement was chosen in accordance with the 

 strong agreement for the canids being the most primitive of the extant Caniformia (Tedford 

 1976; Flynn et al. 1988; Wyss & Flynn 1993; Vrana et al. 1994; but see Wozencraft 1989), 

 but any alternative placement of the root will disrupt the above pattern of sister taxa re- 

 lationships. Fortunately, however, these alternative placements will not affect the ancestral 

 state assessment used to determine the pattern of ingroup relationships (Maddison et al. 

 1984). 



Ingroup relations (Fig.5C) 



A monophyletic Phocidae enjoys the strongest support of any node in the overall solution, 

 being supported by 30 or 31 synapomorphies (21 of which are unequivocal). However, in 

 contrast to the findings of Wyss (1988a) and Arnason et al. (1995), the overall solution 

 indicates the division of the phocids into two major monophyletic clades: the Monachinae 

 and Phocinae. Support for this arrangement is strong, with the phocines having moderately 

 better support at 13 to 28 unweighted steps (10 of which are unequivocal) versus 10 to 

 17 (6 of which are unequivocal) for the monachines, depending on the optimization 

 criterion employed (Fig.5C; Appendices D and E). Although the membership of these 

 subfamilies is as expected (see King 1966), some novel internal relations are indicated in 

 both cases. 



Relationships within the Monachinae 



With respect to the monachines, the elephant seals (Mirounga spp.) are held to be basal, 

 a position traditionally accorded to Monachus spp. (Hendey 1972; Repenning & Ray 1977; 

 Repenning et al. 1979; de Muizon 1982a; King 1983; Wyss 1988a; Arnason et al. 1995; 

 Lento et al. 1995). Therefore, coupled with the view that Monachus spp. is the most 

 primitive of all extant phocids, as well as a good number of fossil forms (Repenning & 

 Ray 1977; de Muizon 1982a), it is surprising to find the monk seals occupying a terminal 

 position deep within the lobodontines. Overall, the indicated topology for the monachines 



