DERIVATION OF THE FLORA AND FAUNA 385 



\Vith regard to the oljjection that oceanic islands are volcanic, and hence prob- 

 ably not the mountain-tops of sunk continents and that they contain no fossil mam- 

 mals, we have in the Malay Archipelago, vast areas of land which if submerged (and 

 they are exposed to constant subsidences and risings) would leave only isolated vol- 

 canic peaks, such as oceanic islands present. Were such an area to be submerged, 

 leaving exposed the volcanic peaks of Java and the Moluccas &:c, &c, should we expect 

 to find either recent or fossil terrestrial mammals upon them? Nor should it be over- 

 looked that, as a general rule, islands diminish in size and numbers toward the 

 centres of the great oceans, which, taken with the admission, that the great islands 

 adjacent to the continents were previously united to them, would favour the hypo- 

 thesis that all may have been so. And finally, we have instances of continental distri- 

 bution, presenting facts so analogous to oceanic, and hitherto so utterly unexplicable, 

 on any hypothesis of migration that does not embrace immense geological changes, 

 that we can scarcely avoid coupling the phenomena they present with those of oceanic 

 islands (p. 1 1 ). 



However, HoOKER did not stop here. When lie threw the pros and cons 

 into the balance, the scale of scruples sank: 



On the other hand, to my mind, the great objection to the continental extension 

 hypothesis is, that it may be said to account for everything, but to explain nothing; 

 it proves too much, whilst the hypothesis of trans-oceanic migration, though it leaves 

 a multitude of facts unexplained, offers a rational solution of many of the most puzz- 

 ling phenomena that oceanic islands present: phenomena which, under the hypothesis 

 of intermediate continents, are barren facts, literally of no scientific interest — are 

 curiosities of science, no doubt, but are not scientific curiosities. 



This was, I am afraid, to say too much — but are not Hooker's words a 

 good expression of the biogeographer's dilemma.' 



The latest brief outline of the history of the Andes is found in GoODSPEED's 

 monograph of Xicotiana [ii2]\ it was, partly at least, based on WEEKS' paper 

 {2'jg). The sequence of events especially refers to the central Andes, with which 

 the history of the southern Andes is said to agree. Four major periods of uplift 

 are recognized, the final one during the Pliocene-Pleistocene, with an average 

 elevation of 3500 m. Coincident with the uplift was a compensory coastal sink- 

 ing. The troughing along the coast ceased with the foundering of the Pacific 

 fore-lands at the close of the Miocene and early Pliocene. These lands are sup- 

 posed to have extended an unknown distance westward (p. 31). It is possible 

 that the border-lands extended all the way north along Central America, Mexico 

 and California, where are found islands in a position corresponding to that of 

 the Galapagos and Juan Fernandez Islands. 



From what has been said above I can onl)- find that there is good geo- 

 logical evidence for a former extension west of the continent, uniting southern 

 Chile with a "Juan Fernandez land '. Before attempting to trace the sequence 

 of immigration of the various floristic elements and of their fate we ought to 

 know something of the fossil floras of Chile and, further, to devote a chapter 

 to the supposed Antarctic migration routes. 



25 ~ 557857 The Nat. Hist, of Juan Fernandez and Easter Isl. Vol. I 



