Chap. XII. OCEANIC ISLANDS. 399 



by formerly continuous land, from America ; and the 

 Cape de Verde Islands from Africa; and that such 

 colonists would be liable to modification ; — the principle 

 of inheritance still betraying their original birthplace. 



Many analogous facts coidd be given : indeed it is an 

 almost universal rule that the endemic productions of 

 islands are related to those of the nearest continent, or 

 of other near islands. The exceptions are few, and 

 most of them can be explained. Thus the plants of 

 Kerguelen Land, though standing nearer to Africa than 

 to America, are related, and that very closely, as we 

 know from Dr. Hooker's account, to those of America : 

 but on the view that this island has been mainly stocked 

 by seeds brought with earth and stones on icebergs, 

 drifted by the prevailing currents, this anomaly dis- 

 appears. New Zealand in its endemic plants is much 

 more closely related to Australia, the nearest mainland, 

 than to any other region : and this is what might have 

 been expected ; but it is also plainly related to South 

 America, which, although the next nearest continent, 

 is so enormously remote, that the fact becomes an 

 anomaly. But this difficulty almost disappears on the 

 view that both New Zealand, South America, and 

 other southern lands were long ago partially stocked 

 from a nearly intermediate though distant point, namely 

 from the antarctic islands, when they were clothed 

 with vegetation, before the commencement of the Gla- 

 cial period. The affinity, which, though feeble, I am 

 assured by Dr. Hooker is real, between the flora of 

 the south-western corner of Australia and of the Cape 

 of Good Hope, is a far more remarkable case, and is 

 at present inexplicable : but this affinity is confined to 

 the plants, and will, I do not doubt, be some day ex- 

 plained. 



The law which causes the inhabitants of an archi- 



