CH. Il] BEST GROUP FOR DEFINING REGIONS. 87 



about the scheme of classification, shows the breaks that 

 exist between the various families. Though it is true that 

 we have a fair anatomical knowledge of the Mammalia, 

 yet there are plenty of lacunae; the broad facts are 

 perhaps ascertained, but there are many details of im- 

 portance which require further elucidation. For instance 

 there is but little knowledge, and that purely osteological, 

 of the structural differences between the different kinds of 

 bears and cats. These two families have a wide distribu- 

 tion and it would be of great interest to know if for 

 example the muscular anatomy of the new world forms 

 is distinctive of them and different from that of the old 

 world genera or species. Besides much of our knowledge 

 of the Mammalia is limited to skins and horns, in many 

 cases necessarily. Unsuspected differences so constantly 

 turn up between superficially very similar forms, that 

 further knowledge is desirable. A noteworthy instance 

 are the discoveries by the late Prof. Garrod in the anatomy 

 of the soft parts of the rhinoceroses. 



On the whole, however, it is impossible to avoid agree- 

 ing with Mr Wallace that the Mammalia are the most 

 satisfactory group. And moreover the adoption of the 

 regions necessitated by the distribution of this group is 

 in harmony with the distribution of some other groups 

 and does no great violence to distributional fact any- 

 where. 



