Chap. XII.] MEANS OF IDISPERSAL. 141 



connected with Europe or Africa, and Europe likewise 

 with America. Other authors haye thus hypothetically 

 bridged over every ocean, and united almost every is- 

 land with some mainland. If indeed the arguments 

 used by Forbes are to be trusted, it must be admitted 

 that scarcely a single island exists which has not re- 

 cently been united to some continent. This view cuts 

 the Gordian knot of the dispersal of the same species to 

 the more distant points, and removes many a difficulty; 

 but to the best of my judgment we are not authorised in 

 admitting such enormous geographical changes within 

 the period of existing species. It seems to me that we 

 have abundant evidence of great oscillations in the level 

 of the land or sea; but not of such vast changes in the 

 position and extension of our continents, as to have 

 united them within the recent period to each other and 

 to the several intervening oceanic islands. I freely 'ad- 

 mit the former existence of many islands, now buried 

 beneath the sea, which may have served as halting-places 

 for plants and for many animals during their migration. 

 In the coral-producing oceans such sunken islands are 

 now marked by rings of coral or atolls standing over 

 them. Whenever it is fully admitted, as it will some 

 day be, that each species has proceeded from a single 

 birthplace, and when in the course of time we know 

 something definite about the means of distribution, we 

 shall be enabled to speculate with security on the for- 

 mer extension of the land. But I do not believe that it 

 will ever be proved that within the recent period most 

 of our continents which now stand quite separate have 

 been continuously, or almost continuously united with 

 each other, and with the many existing oceanic islands. 

 Several facts in distribution, — such as the great difference 



