82 



is the disruption of the two subfamilies as they are commonly recognized. Two distinct 

 clades are still present, but Cystophora is now seen to cluster with the monachines, forming 

 the sister taxon to this subfamily. This result is fairly labile, however, with cladograms of 

 only one additional step finding Cystophora back as the sister taxon to the remaining 

 phocines (results not shown). However, as indicated by the bootstrap, Cystophora displays 

 a strong tendency to join the monachines even in the inversely weighted data set (bootstrap 

 frequency of 31%). 



This tendency on the part of Cystophora likely reflects its similarities, be they convergent 

 or symplesiomorphic (with respect to all phocids). with Mirounga. The failure of Mirounga 

 spp. to show an analogous predisposition (bootstrap frequency of only 2%) relates in turn 

 to a third taxon, Ommatophoca. Together, these three genera are uniquely characterized 

 among phocids by nasal processes of the premaxilla that distinctly fail to reach the nasal 

 bones (see Character Analysis, character #12). Although this morphology is not as 

 developed in Ommatophoca, together with the large number of characters used to describe 

 the nasal region (see Character Analysis, characters #5-12), it is apparently sufficient to 

 have all three taxa occupying the basal positions within the one clade. (These same 

 characters, being predominantly multistate, exhibit less of an influence in the inversely 

 weighted data matrix. However, note that even under such ameliorating conditions that 

 Cystophora still displays a marked tendency to cluster with the monachines.) In fact, the 

 tendency for all three genera to group together is demonstrated by the inclination for both 

 Mirounga spp. and Ommatophoca to join the phocines (bootstrap frequency of 9%) being 

 marginally higher than of Mirounga spp. alone. This also reflects the surprisingly high 

 tendency of Ommatophoca to join the phocines (bootstrap frequency of 14%), presumably 

 to cluster basally with Cystophora. 



Within the monachines proper, the paraphyly of the lobodontines is even more pronounced 

 with the shift of Ommatophoca to its basal position between Cystophora and Mirounga 

 spp. As well, the traditional lobodontine relationships (i.e., paired Hydrurga-Lobodon and 

 Leptonychotes-Ommatophoca clades) are contradicted further with Leptonychotes and 

 Lobodon forming a monophyletic clade. Monophyly of Monachus is still indicated, 

 however. 



The loss of Cystophora to the monachines results in Erignathus resuming its traditional 

 sister taxon status to the remaining phocines. Resolution within the now monophyletic 

 Phocini is again limited, but paraphyly of Phoca (sensu Burns & Fay 1970), Phoca (sensu 

 stricto), and Pusa are all indicated, with some novel relationships presented between Phoca 

 spp. and Pusa spp. Histriophoca and Pagophilus continue to constitute a clade although, 

 again, it is shifted basally with the loss of Erignathus. 



This analysis is mentioned primarily as a curiosity into the effects of character "weighting" 

 on this data matrix. As mentioned previously, the uncorrected use of binary and multistate 

 characters will not result in equally weighted characters (as desired here) due to the 

 algorithms used in PAUP (Swofford 1993). However, the results from this analysis do 

 mirror the general conclusions from the analysis of the inversely weighted data set: 

 strongly supported relations for the outgroup taxa and at the level of the phocid 

 subfamilies, and weakly supported/resolved relations within the subfamilies (albeit 



