130 



(de Muizon 1982a). Our analysis indicates that the auricular foramen first appears 

 ancestrally in the phocids (again, assuming that it is absent in the remaining caniforms). 

 and is separate from the stylomastoid foramen. This condition is retained throughout both 

 subfamilies, with their confluence into a single auriculostylomastoid foramen (state 0) 

 arising in parallel in Mirounga spp. and Pusa sibirica. This latter observation is another 

 a posteriori assessment, based on the assumption that these phocids still possess the 

 auricular foramen and have not reverted to the primitive condition for the caniforms, in 

 which it is absent. 



110) relationship of tympanohyal and stylomastoid foramen: 0 = separated; 1 = closely 

 associated (Wozencraft 1989). 



This and the following character deal with the relationship of the tympanohyal to the 

 stylomastoid foramen. As the hyoid apparatus is rarely referred to in the literature, and 

 even more rarely preserved in museum collections (due, in part, to the tympanohyal being 

 cartilaginous), we were forced to rely on the observations of Wozencraft (1989) for both 

 this and the following character. Among the caniforms, a close association between the 

 tympanohyal and the stylomastoid foramen is plesiomorphic. The apomorphic condition, 

 where the two are separated, is a lutrine-pinniped synapomorphy, with an independent 

 origin in the ursids. Observations by Burns & Fay (1970) may contradict this for the 

 phocids; however, the degree of the contradiction depends on the definitions of "closely 

 associated" versus "separated" employed by Wozencraft (1989). 



111) location of tympanohyal relative to stylomastoid foramen: 0 = anterior: 1 = posterior 

 (Wozencraft 1989). 



An anterior placement of the tympanohyal relative to the stylomastoid foramen is a 

 synapomorphy of the phocids (Wozencraft 1989). This morphology has been corroborated 

 in the phocids by Burns & Fay ( 1970). 



112) position of petrosal relative to intracranial ridges of basioccipital continuous 

 anteriorly with the dorsum sellae: 0 = widely separate; 1 = intermediate; 2 = closely 

 adjacent (Wozencraft 1989) (Fig.21). 



As originally coded by Wozencraft (1989), this character dealt with the nature of the 

 petrosal-basioccipital suture, with a note that it was usually not visible from the ventral 

 side of the skull. A wide separation was held to define the clade of the ursids plus the 

 otarioids (Wozencraft 1989). However, the exact nature of this character is elusive, as this 

 character is not apparent from an examination of Wozencraft's (1989) citation for it (van 

 der Klaauw 1931), which appears to refer to either the petrobasilar fissure (see characters 

 #95 and 96), or something analogous to character #84. In any case, in keeping with 

 Wozencraft's (1989) apparent intention (i.e., determining the intracranial approach of the 

 petrosal to the basioccipital), we modified the character slightly to how it now appears 

 above. 



The condition where the petrosal and intracranial basioccipital ridges are closely adjacent 

 is plesiomorphic among the caniforms, with any apomorphic separation of the two 

 structures generally characterizing the pinnipeds only (Ursus is polymorphic between all 

 three states). This is largely manifested by a wide separation (the ancestral pinniped 



