462 University of California Publications in Zoology [Vol. 12 



cannot be said, on grounds of compared characteristics, to be closer to 

 belugae than it is to eastern canadensis. Belugae is more closely 

 allied to canadensis of eastern Canada than is either Castor c. leuco- 

 donta or Castor c. pacificus. It is impossible to decide to which of 

 the two mainland subspecies (belugae or pacificus) leucodonta is most 

 closely related. Frondator is more closely allied to canadensis of 

 eastern Canada than it is to the California species of beavers or to 

 the belugae-leucodonta-pacificus series. Castor c. micliiganensis finds 

 its closest relative in canadensis. 



The California beavers stand by themselves, having undergone 

 considerably more divergence than the other subspecies, and their 

 immediate ancestry is decidedly uncertain. Mr. F. H. Holden, of 

 the staff of the Museum of Vertebrate Zoology of the University of 

 California, who has carefully compared such skeletons of beavers as 

 are available, has called the writer's attention to the fact that the 

 majority of skeletal characters of Castor subauratus subauratus would 

 relate it rather to the Eurasiatic Castor fiber than to the North Ameri- 

 can Castor canadensis. This testimony, however, is not borne out by 

 the skull characteristics, which would seem definitely to relate subau- 

 ratus to the canadensis series. Perhaps the most striking differences 

 between the subauratus series and the canadensis series are the dif- 

 ferent shapes of the foramen magnum and of the median process in 

 the interpterygoid fossa. Castor c. pacificus and Castor c. belugae 

 are somewhat more variable respecting these characters than are any 

 other subspecies of canadensis, and individual skulls of these show a 

 more decided bent in the direction of subauratus than do the other 

 forms, which suggests the possibility that subauratus has been derived 

 from the form which was ancestral also to the subspecies occupying 

 the mainland of Oregon, Washington, British Columbia and Alaska. 

 If this should be the case, the relationship of the California forms with 

 canadensis would be close, and the comparatively large amount of 

 divergence would be noteworthy as possibly indicating a more rapid 

 evolution, or a more complete isolation from the parent stock (possi- 

 bly both together) than is exemplified by the other forms. 



Remarks on Isolation and Its Relation to Speciation 

 consideration of some op the evidence 



Evidence from the Study of Beavers 

 We may now profitably undertake a review of some of the facts 



