56 



Table 2: Bootstrap frequencies indicating support within the inversely weighted data matrix for a 

 monophyletic grouping of various outgroup taxa and either the phocids alone or the pinnipeds as a 

 whole. Numbers given as the total number of trees (out of 1000) / percentage of all trees. 



Alternative outgroup For phocids alone For all pinnipeds 



arctoid 



< 0.08 / 



« 1 



1000* / 100* 



procyonid 



1.00 / 



< 1 



1.00 / < 1 



ursid 



< 0.08 / 



« 1 



21.00 / 2 



mustelid (including lutrines) 



115.50 



/ 12 



680.50 / 68 



musteline 



4.00 / 



< 1 



43.67 / 4 



lutrine (both Enhydra and Lutra) 



219.67 



/ 22 



729.33 / 73 



Enhydra alone 



11.25 



/ 1 



41.00 / 4 



Lutra alone 



119.93 



/ 12 



428.67 / 43 



otarioid 



687.00 



/ 69 



n/a 



Odobenus alone 



59.00 



/ 6 



n/a 



Zalophus alone 



24.50 



/ 2 



n/a 



* Due to Canis being the ultimate outgroup, this arrangement was necessarily found in all bootstrap 

 replicates. 



a consensus solution; see Methods and Materials), we should look instead to the overall 

 solution. 



When bootstrap frequencies are determined for the nodes present in the overall solution 

 (Fig.8B), a clear dichotomy in support can be observed. Outgroup relationships tend to 

 be moderately to strongly supported, with only the node for Lutra plus the pinnipeds 

 falling below a bootstrap frequency of 50%. However, this merely reflects an equally 

 strong tendency for the two lutrines to form a monophyletic sister group to the pinnipeds 

 (bootstrap frequency = 39%), in essence a minor alteration. A monophyletic Otarioidea is 

 particularly strongly indicated. Bootstrap frequencies for alternative outgroup arrange- 

 ments are noticeably smaller, especially for those postulating a diphyletic Pinnipedia 

 (Tab.2). 



The monophyly of both the phocids as a whole, and of each of its two subfamilies, show 

 comparable bootstrap frequencies to the outgroup nodes. Beyond this, support for the 

 relationships within each subfamily was distinctly weaker. Only the species clusters of 

 Histriophoca plus Pagophilus, Mirounga spp., and Monachus spp. (and M. schauinslandi 

 plus M. tropicalis within this) display bootstrap frequencies greater than 50%. A 

 monophyletic Phocini (plus Erignathus) is also relatively strongly indicated (bootstrap 

 frequency of 62%), giving further support to the basal position of Cystophora within the 

 phocines. In fact. Cystophora displays an unusually strong tendency to cluster with the 

 monachines (bootstrap frequency of 31%). something that might be expected more of the 

 supposedly more monachine-like Erignathus. but was not supported here (bootstrap 

 frequency of <1%). 



