DISPERSAL OF MAMMALS. 43 



dence that in the former periods of the earth's history the surface of 

 the globe must have undergone such vicissitudes as to have at vari- 

 ous times disturbed the general relations existing between land and 

 water. In other words, much of the surface that at one tune was 

 occupied by water must have been replaced by land, and, per contra, 

 what was at one time land must at another have been water. And 

 evidences of such variations in terrestrial equilibrium are abundantly 

 afforded by geological landmarks. Had the greater portion of the 

 surface of the globe at one time since the introduction of the Mam- 

 malia consisted principally of dry land, or had there been since 

 that period a general alternation in the relative positions of the land 

 and water areas, the geographical distribution of the Mammalia 

 would have been very different from what we actually find it to be. 

 Hence, it must be assumed that a land and water alternation, such 

 as could have brought about the present result, must have taken 

 place in certain parts of the earth's surface only, and without affect- 

 ing others. There would seem to be very strong grounds for con- 

 cluding that the most recent connection uniting the principal land- 

 areas of the globe was formed in the Northern Hemisphere, as a belt 

 closing off the Arctic Sea (if it then existed) from the Pacific and 

 Atlantic oceans. 



The only class of terrestrial mammals to which a broad arm 

 of water offers no impediment in the way of migration or disper- 

 sion is that of the bats ; and, singularly enough, just in the case of 

 these animals, as has already been remarked, are we furnished with 

 an example of universal distribution, there being but very few of 

 the habitable oceanic islands which are not tenanted by one or more 

 representatives of the order. But even among the habitually ter- 

 restrial Mammalia there are certain exceptional methods by which 

 dispersion to very considerable distances from the mainland can be 

 effected. In the northern regions the frozen sea constitutes a con- 

 necting bridge between distantly-removed land-masses which is 

 constantly taken advantage of by various forms of Arctic animals. 

 By the breaking up and drifting away of fragments of the northern 

 ice-masses animals that might be temporanly wandering over them 

 could readily be transported to very considerable distances from 

 their true homes ; and, indeed, it is through such means that polar 

 bears are periodically stranded upon the coast of Iceland. In one 

 year alone twelve of such wandering animals made their appearance 



