v] ANCIENT GEOGRAPHY 77 



Land tortoises (genus Testudo\ have reached the 

 Galapagos by land, not astride a log, just as certain 

 as lemurs in Madagascar and elephant remains in 

 the Andes are evidence of former land-connexions. 

 Oceanic islands have at best a scanty terrestrial fauna, 

 but it is a mistake to refer this entirely to accidental 

 spreading. It would first have to be shown whether 

 these islands arose out of the ocean, or whether they 

 are remnants of sunken continents. If Africa sub- 

 sided two miles there would remain a few isolated 

 volcanic peaks, like Cameroon and Kenia, with all 

 the appearance of oceanic islands, but with a derelict 

 indigenous fauna. It depends upon the age of such 

 islands, whether they could have only reptiles, or also 

 mammals, and even then absence of, say, mammals 

 on the Galapagos proves but little since so much 

 depends upon the physical conditions of the land 

 (cf. p. 59). 



Of course enormous subsidences and elevations of 

 extensive areas are required, although many authori- 

 ties shrink from their application. The permanence 

 of the great ocean basins had become a dogma since 

 it was found that a universal elevation of the land to 

 the extent of 100 fathoms would produce but little 

 changes, and when it was shown that even the 1000 

 fathom-line followed the great masses of land rather 

 closely, and still leaving the great basins, by general 

 consent this quite arbitrary unit of about one mile 



