12 INTRODUCTORY. [CHAP. 



These orders are further subdivided into families and genera. 

 In regard to the number of the latter of these, there is still a con- 

 siderable difference of opinion among naturalists ; but in the 

 present work those adopted in Flower and Lydekker's Study of 

 Mammals 1 will be in the main adhered to, with such corrections 

 and additions as are necessary owing to recent emendations in 

 nomenclature or to the discovery of new forms. 



Omitting from consideration the purely aquatic and volant 



members of the class, the most effectual barriers to 



Dispersal of the dispersal of mammals are formed by channels 



Mammals. 



their being crossed by swimming. And it is this inability to 

 traverse any extent of water that renders what are known as 

 oceanic islands practically devoid of all mammalian life, with 

 the exception of a few bats and small rodents ; the latter animals 

 having apparently some means of dispersal not common to other 

 members of this class. Oceanic islands, it may be explained, are 

 such as rise from great depths in the ocean, and are composed, 

 almost invariably, either of volcanic rocks or of coral. They show, 

 for the most part, no decisive evidence of having been connected 

 with any continental land, and thus have never been enabled to 

 receive a mammalian fauna 2 . In marked distinction to these are 

 the so-called continental islands, such as Madagascar and Great 

 Britain, which, both from the evidence of their mammalian fauna 

 and their geological conformation, have indubitably been in direct 

 communication with the adjacent continent at no very distant 

 epoch. As a rule, the channels between such islands and the 

 mainland are comparatively shallow, so that a moderate degree of 

 upheaval would place the two in direct connection. 



The relative depth of the channel between two islands, or 

 between an island and a continent is indeed of much more im- 

 portance in regard to the dispersal of mammals than is its width. 

 This is best exemplified by the well-known case of "Wallace's 

 line " in the Malayan archipelago ; that name being applied to 



1 London, 1891. 



2 There is a possibility that some oceanic islands may have been connected 

 with continents, and that their original mammalian fauna has been destroyed 

 by submergence. 



