HAUNTS AND HABITATS 37 



introduction from some other part of the earth could the 

 islands be populated ; but on such a supposition we are unable 

 to explain the harmonious distribution, we cannot explain why 

 every, or nearly every, island has its race or species. If some 

 animals could be carried hundreds of miles to the islands, why 

 are they not carried from one island to the other ? But, besides 

 this, how could we make plain the presence of such peculiar 

 forms as the gigantic land-tortoises, for instance? According 

 to the elevation theory, we can only think of an accidental im- 

 portation of these tortoises by some current, because they are 

 unable to swim. After the islands had been elevated out of 

 the sea, it happened once, by a peculiar accident, that a 

 land-tortoise was carried over. Alone it could not propagate. 1 

 This was only possible after a similar accident imported 

 another specimen of the same species, of the opposite sex, to 

 the same island. Or we could imagine that at the same time 

 animals of both sexes were thus accidentally introduced. By 

 this we could at least explain the population of a single island. 

 But how did all the other islands become populated ? To 

 explain this we should have to invoke a thousand accidents. 



" The most simple solution is given by the theory of sub- 

 sidence. All the islands were formerly connected with each 

 other, forming a single large island ; subsidence kept on, and 

 the single island was divided into several islands. Every 

 island developed, in the course of long periods, its peculiar 

 races, because the conditions on these different islands were not 

 absolutely identical." 



In other words, in place of being "oceanic," the Galapagos 

 group is " continental," and was once joined to the American 

 mainland. And there is little doubt that a similar explanation 

 must hold good with the other " tortoise-islands ". In Tertiary 

 times giant tortoises of the genus Testudo inhabited all the 

 great continents ; the earliest known form being T. amnion of 

 the Lower Eocene of Egypt, which may have been the parent 

 stock of all. It is from these continental tortoises that the 

 island forms of to-day are derived. 



It may be mentioned that the great antiquity of giant 

 tortoises of the genus Testudo affords an excellent example of 

 the early date, as compared with mammals, of the dispersal 

 1 Unless it happened to be a female with eggs (R.L.). 



