72 



Monachus schauinslandi 



Monachus tropicalis 



Monachus monachus 



Lobodon 



Ommatophoca 



Leptonychotes 



Hydrurga 



Mirounga angustirostris 



Mirounga leonina 



Odobenus 



Zalophus 



Histriophoca 



Pagophilus 



Erignathus 



Pusa hispida 



Pusa sibirica 



Pusa caspica 



Phoca vitulina 



Phoca largha 



Halichoerus 



Cystophora 



Lutra 



Enhydra 



Mart es 



Procyon 



Urs us 



Canis 



Monachus schauinslandi 



Monachus tropicalis 



Monachus monachus 



Ommatophoca 



Leptonychotes 



Lobodon 



Hydrurga 



Mirounga angustirostris 



Mirounga leonina 



Cystophora 



Pusa hispida 



Pusa sibirica 



Pusa caspica 



Phoca largha 



Phoca vitulina 



Histriophoca 



Pagophilus 



Halichoerus 



Erignathus 



Odobenus 



Zalophus 



Lutra 



Enhydra 



Mart es 



Procyon 



Ursus 



Canis 



D 



Pusa hispida 

 Pusa sibirica 

 Phoca vitulina 

 Pusa caspica 

 Phoca largha 

 Erignathus 

 Histriophoca 

 Pagophilus 

 Halichoerus 

 Cystophora 



Monachus schauinslandi 



Monachus tropicalis 



Monachus monachus 



Ommatophoca 



Leptonychotes 



Lobodon 



Hydrurga 



Mirounga angustirostris 



Mirounga leonina 



Odobenus 



Zalophus 



Lutra 



Enhydra 



Mart es 



Procyon 



Ursus 



Canis 



Monachus schauinslandi 



Monachus tropicalis 



Monachus monachus 



Ommatophoca 



Leptonychotes 



Lobodon 



Hydrurga 



Mirounga angustirostris 



Mirounga leonina 



Cystophora 



Pusa hispida 



Pusa sibirica 



Pusa caspica 



Phoca largha 



Phoca vitulina 



Histriophoca 



Pagophilus 



Halichoerus 



Erignathus 



Odobenus 



Zalophus 



Lutra 



Enhydra 



Mart es 



Procyon 



Ursus 



Canis 



Fig.l4A-D: Cladograms resulting from a constraint analysis examining various alternative hypotheses 

 of ingroup relationships: (A) (not) phocidae *, (B) (not) two subfamilies, (C) three subfamilies *, 

 and (D) cystophorinae *. An asterisk indicates a majority rule consensus solution, where, unless 

 otherwise indicated, all nodes were found in 100% of the equally most parsimonious solutions. See 

 also Fig.4 and Tab.4. 



