410 ISLAND LIFE 



Now, when the division of the a,rea leaves one portion of 

 the species in ,an island, a similar modification of the 

 species, either in the island or in the continent, occurs, 

 resulting in closely-allied but distinct forms ; and such 

 forms are, as we have seen, highly characteristic of island- 

 faunas. But islands also favour the occasional preservation 

 of the unchanged species — a phenomenon which very 

 rarely occurs in continents. This is probably due to the 

 absence of competition in islands, so that the parent 

 species there maintains itself unchanged, while the con- 

 tinental portion, by the force of that competition, is driven 

 back to some remote mountain area, where it also obtains 

 a comparative freedom from competition. Thus may be 

 explained the curious fact, that the species common to 

 Formosa and India are generally confined to limited areas 

 in the Himalayas, or in other cases are found only in 

 remote islands, as Japan or Hainan. 



The distribution and affinities of the animals of con- 

 tinental islands thus throws much light on that obscure 

 subject — the decay and extinction of species ; while the 

 numerous and delicate gradations in the modification of 

 the continental species, from perfect identity, through 

 slight varieties, local forms, and insular races, to well- 

 detined species and even distinct genera, afford an over- 

 whelming mass of evidence in favour of the theory of 

 " descent with modification." 



We shall now pass on to another class of islands, which, 

 though originally forming parts of continents, were 

 separated from them at very remote epochs. This 

 antiquity is clearly manifested in their existing faunas, 

 which present many joeculiarities, and ofier some most 

 curious problems to the student of distribution. 



