84 



Contrary to most other analyses, the changes were the most severe in the monachines 

 where paraphyly is indicated for the subfamily. This is likely attributable to the improper 

 presumption of a monophyletic Lobodontini, forcing Monachus to become the sister taxon 

 to the remaining phocids. This latter scenario has been postulated to be conducive towards 

 obtaining a paraphyletic Monachinae (Berta & Wyss 1994). Moreover, under these 

 conditions of imposed monophyly, the lobodontines and Mirounga do form a clade, as 

 suggested by Hendey (1972) and King (1983) (but see Sarich 1976). Within the phocines, 

 the only alteration was the exclusion of Erignathus from the Phocini, where it was 

 normally firmly entrenched. Cystophora, however, maintains its position as the most basal 

 of the phocines. Overall, the internal phylogeny of the phocids in the condensed solution 

 strongly resembles the solution of Wyss (1988a). which, in effect, included some 

 constrained monophyletic higher taxa. The more traditional appearance of the ingroup 

 relationships (see below also) hints that previous studies into phocid phylogeny probably 

 assumed the monophyly of some higher level taxa to varying degrees as well. 

 The visualization of the internal phylogeny of the condensed higher level taxa by a 

 constraint analysis of the inversely weighted data matrix (Fig.l4P and Tab. 4) revealed 

 branching patterns that largely agree with those of the overall solution. Despite their altered 

 placement within the monachines, the internal topology of both Monachus and the 

 lobodontines were identical with those of the overall solution. Within Phoca (sensu 

 Burns & Fay 1970), a polytomy was again found, indicating uncertainty regarding the 

 relationships in this region. However, Pusa hispida and Pusa sibirica do continue to form 

 a clade in most solutions. As well, the enforced removal of Erignathus again resulted in 

 a more basal shift of the clade of Histriophoca plus Pagophilus relative to Phoca spp. 

 and Pusa spp. 



The condensed solution is also typified by an increased amount of homoplasy (relative to 

 the overall solution) presumably to account for the constrained monophyly of the otherwise 

 normally paraphyletic higher taxa. This is most clearly indicated by the ensemble CI for 

 the condensed solution (0.518), which although higher than the analogous value of the 

 overall solution, actually falls below the value expected for a study of 15 taxa [0.618; 

 Sanderson & Donoghue (1989)], while that of the overall solution was about on a par 

 with its expected value (see Overall Parsimony Analysis). The increased levels of 

 homoplasy present in the condensed solution were also indicated by the constraint analysis, 

 where the constraint of the inversely weighted data matrix to the basic pattern of the 

 condensed solution (see Fig.4P) resulted in one of the larger increases in length of all the 

 constraint analyses, 619 extra steps (nine extra corrected steps) (Tab.4). 

 The increase in homoplasy for the condensed solution is argued against by several factors, 

 however. Firstly, the ensemble HI is slightly lower (which might arise from the effective 

 removal of nine taxa), while both the ensemble RI and RC are higher than their equivalent 

 values in the overall solution. Another line of evidence originates from the somewhat 

 surprising result that all the bootstrap frequencies for the condensed solution were roughly 

 equivalent to those of the overall solution (compare Figs.8B and 17). While this could 

 have been expected for the outgroup nodes (as their taxa, and thus hopefully their 

 interrelationships, were unchanged), the failure of the bootstrap to reflect the presumably 



