160 THE ORIGIN OF SPECIES 



ing oceanic islands. Several facts in distribution — such 

 as the great difference in the marine faunas on the oppo- 

 site sides of almost every continent — the close relation of 

 the tertiary inhabitants of several lands and even seas to 

 their present inhabitants — the degree of affinity between 

 the mammals inhabiting islands with those of the nearest 

 continent, being in part determined (as we shall hereafter 

 see) by the depth of the intervening ocean — these and 

 other such facts are opposed to the admission of such 

 prodigious geographical revolutions within the recent 

 period as are necessary on the view advanced by 

 Forbes and admitted by his followers. The nature and 

 relative proportions of the inhabitants of oceanic islands 

 are likewise opposed to the belief of their former con- 

 tinuity with continents. Nor does the almost universally 

 volcanic composition of such islands favor the admission 

 that they are the wrecks of sunken continents; — if they 

 had originally existed as continental mountain ranges, 

 some at least of the islands would have been formed, 

 like other mountain summits, of granite, metamorphic 

 schists, old fossiliferous and other rocks, instead of con- 

 sisting of mere piles of volcanic matter. 



I must now say a few words on what are called acci 

 dental means, but which more properly should be called 

 occasional means of distribution. I shall here confine 

 myself to plants. In botanical' works, this or that plant 

 is often stated to be ill adapted for wide dissemination; 

 but the greater or less facilities for transport across the 

 sea may be said to be almost wholly unknown. Until 

 I tried, with Mr. Berkeley's aid, a few experiments, it 

 was not even known how far seeds could resist the inju- 

 rious action of sea-water. To my surprise I found that, 



