VIL] METHODS AND RESULTS OF ETHNOLOGY. 165 



if a continent like Asia had once occupied the area of 

 the Pacific, the peaks of its mountains would now show 

 not more numerous than the islands of the Polynesian 

 Archipelago. 



What lands may have been thickly populated for 

 untold ages, and subsequently have disappeared and 

 left no sign above the waters, it is of course impossible 

 for us to say ; but unless we are to make the wholly 

 unjustifiable assumption that no dry land rose elsewhere 

 when our present dry land sank, there must be half-a- 

 dozen Atlantises beneath the waves of the various oceans 

 of the world. But if the regions which have undergone 

 these slow and gradual, but immense alterations, were 

 wholly or in part inhabited before the changes I have 

 indicated began and it is more probable that they 

 were, than that they were not what a wonderfulty 

 efficient "Emigration Board" must have been at work 

 all over the world long before canoes, or even rafts, were 

 invented ; and before men were impelled to wander by 

 any desire nobler or stronger than hunger. And as 

 these rude and primitive families were thrust, in the 

 course of long series of generations, from land to land, 

 impelled by encroachments of sea or of marsh, or by 

 severity of summer heat or winter cold, to change their 

 positions, what opportunities must have been offered for 

 the play of natural selection, in preserving one family 

 variation and destroying another ! 



Suppose, for example, that some families of a horde 

 which had reached a land charged with the seeds of 

 yellow fever, varied in the direction of woolliness of 

 hair and darkness of skin. Then, if it be true that 

 these physical characters are accompanied by compara- 

 tive or absolute exemptions from that scourge, the 

 inevitable tendency would be to the preservation and 

 multiplication of the darker and woollier families, and 



