167 



Mirounga angustirostris 



Mirounga leonina 



Cystophora 



Ommatophoca 



Erignathus 



Halichoerus 



Pusa caspica 



Histriophoca 



Pagophilus 



Pusa hispida 



Phoca largha 



Pusa sibirica 



Phoca vitulina 



Monachus schauinslandi 



Monachus tropicalis 



Leptonychotes 



Lobodon 



Monachus monachus 



Mart es 



Procyon 



Enhydra 



Lutra 



Odobenus 



Zalophus 



Urs us 



Hydrurga 



Canis 



Mirounga angustirostris 



Mirounga leonina 



Cystophora 



Ommatophoca 



Phoca vitulina 



Erignathus 



Pusa sibirica 



Histriophoca 



Pagophilus 



Halichoerus 



Pusa caspica 



Pusa hispida 



Phoca largha 



Monachus schauinslandi 



Monachus tropicalis 



Leptonychotes 



Lobodon 



Monachus monachus 



Mart es 



Enhydra 



Lutra 



Procyon 



Odobenus 



Zalophus 



Ursus 



Hydrurga 



Canis 



Fig.25A: Consensus solutions (top - strict algorithm; bottom - majority rule algorithm) resulting from 

 a parsimony analysis of the inversely weighted data matrix subdivided according to the region of 

 character origin: (A) snout (18 characters, n = 66 trees, length = 5,667 steps, CI = 0.486, HI = 0.710, 

 RI = 0.701, RC = 0.474). Lengths and goodness-of-fit statistics apply only to the majority rule 

 consensus solution, where, unless otherwise indicated, all nodes were found in 100% of the equally 

 most parsimonious solutions. 



