105 



otarioids with ursids and the phocids with mustelids (e.g., Mivart 1885; Tedford 1976; 

 King 1983; Wozencraft 1989). Although suggestive, this character does not resolve the 

 question of pinniped ancestry on its own. Howell (1928) has noted one otariid specimen 

 in which the canal was not bilaterally present, which fuels the contention of other workers 

 that the common absence of the canal in phocids and mustelids is due to convergence 

 (Wyss 1987; Wyss & Flynn 1993). 



Although our observations agree with those in the literature, the relationships advocated 

 here indicate a somewhat novel suggestion for the evolution of the alisphenoid canal. The 

 monophyly of the pinnipeds still necessitates homoplasy to explain the distribution of this 

 character; however, the homoplasy now arises from a reversal by the otarioids to re-obtain 

 the primitive condition of possessing the canal as found in Canis and Ursus. 



49) location of least interorbital width: 0 = distinctly anterior to middle of interorbital 

 region; 1 = approximately in the middle of interorbital region; 2 = distinctly posterior to 

 middle of interorbital region (Wyss 1988a). 



King (1972, 1983) has argued that the greater reliance of pinnipeds on sight rather than 

 smell has resulted in the lateral compression of the interorbital region (and the underlying 

 turbinal bones) to accommodate larger orbits. This has resulted in a proportionately 

 narrower interorbital region in pinnipeds as compared to fissiped carnivores (Howell 

 1928), and is the most pronounced in the smaller species (King 1972). The exact nature 

 of this compression is not constant within the pinnipeds, however, with the least 

 interorbital width varying in its location: anterior for the phocines minus Cystophora and 

 Erignathus, and posterior for all other pinnipeds (Burns & Fay 1970; King 1972; Wyss 

 1988a). 



Unfortunately, the distribution of this character does not lend itself to a simple description. 

 The general tendency is for the phocids to shift the least interorbital width anteriorly from 

 the plesiomorphic posterior placement to the middle of the interorbital region. This also 

 occurs in Martes and Procyon. The remaining outgroup taxa maintain the plesiomorphic 

 condition. ACCTRAN optimization indicates that state 1 is synapomorphic for both the 

 phocids and the clade of Martes and Procyon, with two intervening reversals accounting 

 for the lutrines and otarioids. DELTRAN optimization holds for independent origins in 

 Martes, Procyon, and in each phocid subfamily. However, both optimization schemes 

 indicate that a posterior placement is primitive in the pinnipeds, as Wyss (1988a) suggest- 

 ed. Within the phocids, the distributions mentioned above are only generally supported, 

 with most members retaining a middle placement. The diagnostic anterior and posterior 

 placements only occur for scattered phocines (and also Monachus tropicalis) and 

 monachines respectively. 



50) location of greatest zygomatic width: 0 = anterior to glenoid fossa (i.e., within 

 zygomatic arch proper); 1 = at level of glenoid fossa (i.e., at squamosal) (pers. obs.). 

 One mechanism to possibly accommodate the larger eyes and orbits of the phocids is for 

 the zygomatic arches to be bowed outwards and generally broadened (Howell 1928; see 

 character #51). However, we noted some variation in the location of the broadest point of 

 the zygomatic arches. In some forms, the broadest point was located at about the level of 



