121 



(DELTRAN optimization), or a synapomorphy necessitating a reversal to the 

 plesiomorphic condition ancestrally for the pinnipeds (ACCTRAN optimization). 



85) relationship between auditory bulla and paroccipital process: 0 = does not reach 

 process; 1 = reaches (or very closely approaches) process (Wyss 1988a). 



In acknowledgement of the limitations of the previous character mentioned by Ray 

 (1976b), Wyss (1988a) proposed this related feature. Wyss (1988a) noted that the condition 

 where the auditory bulla covers the petrosal is coincidental with the bulla extending 

 posteriorly to nearly contact the exoccipital. Through our observations, we have modified 

 this character still further, asking if the bulla contacts (or at least closely approaches) the 

 major process of the exoccipital, the paroccipital process. Naturally, contact with the 

 paroccipital process, or lack thereof, is tied in with inflation of the auditory bulla, 

 particularly the posterior region (Flower 1869). Most fissipeds are noted for a relatively 

 inflated bulla (Segall 1943; Repenning 1972; Hunt 1974), and this is sufficient in Cards 

 and Procyon to cause contact between it and the paroccipital process (Davis 1964). This 

 condition is also obtained in feloids (Turner 1848; Wozencraft 1989), but not in ursids, 

 where the bulla is relatively flat (Turner 1848; Segall 1943; Davis 1964). However, contact 

 is maintained in the ursids by a bony ridge running between the bulla and the paroccipital 

 process (Segall 1943). In contrast, Flower (1869) holds that the paroccipital process is 

 generally separate from the auditory bulla in most arctoids. 



The plesiomorphic state for the Caniformia is inferred to be one where the auditory bulla 

 and paroccipital process are in contact. Together with its occurrence in felids, this suggests 

 that this state is primitive at the level of the carnivores. The derived condition, where 

 contact is lost, describes a synapomorphy linking the lutrines with the pinnipeds. However, 

 this distribution is contingent on our equating of the bony ridge possessed by ursids with 

 the primitive condition, where the auditory bulla and the paroccipital process are directly 

 in contact. 



86) groove separating mastoid bulla and petrosal: 0 = absent; 1 = present (King 1972; de 

 Muizon 1982a). 



There is some uncertainty on our part as to what feature de Muizon (1982a) was attempting 

 to diagnose with this character. The region around the posterolateral edge of the auditory 

 bulla is punctuated by a number of grooves, pits, and foramina in phocids [see characters 

 #87, 88, 108, and 109; Figs.6 and 7 in de Muizon (1982a)], and de Muizon's description 

 makes it unclear as to which exact feature he is referring. As de Muizon (1982a) describes 

 the transition of a pit unique to Histriophoca and Pagophilus into a groove in Cystophora, 

 the likely candidate is the "digastric pit" of Burns & Fay (1970: 374). However, as we 

 were unable to find any evidence of either the pit or the groove in the above taxa, we 

 settled instead for a definition synonymous with the stylomastoid groove of King (1972). 

 In other words, the definition we employed is the literal one: is there a groove running 

 between the posterolateral edge of the auditory bulla and the petrosal? 

 The groove, as we have defined it. is an apomorphic trait found only in Odobenus and 

 all phocids. [However, this is dependent upon the assessment of state 0 by PAUP for non- 

 lutrine fissipeds. In these taxa, the plesiomorphic contact between the auditory bulla and 

 the paroccipital process (see previous character) made it impossible to determine the 



