CHAP, xviri JAPAN" AND FORMOSA 409 



species die out and are replaced by others, which quite 

 agrees with what the theory of evohition assures us must 

 have occurred. On a continent, the process of extinction will 

 generally take effect on the circumference of the area of 

 distribution, because it is there that the sjDecies comes into 

 contact with such adverse conditions or competing forms 

 as prevent it from advancing further. A very slight change 

 Avill evidently turn the scale and cause the species to 

 contract its range, and this usually goes on till it is reduced 

 to a very restricted area, and finally becomes extinct. It 

 may conceivably haj^pen (and almost certainly has some- 

 times happened) that the process of restriction of range by 

 adverse conditions may act in one direction only, and over 

 a limited district, so as ultimately to divide the specific 

 area into two separated parts, in each of which a portion 

 of the species will continue to maintain itself. We have 

 seen that there is reason to believe that this has occurred 

 in a very few cases both in North America and in Northern 

 Asia. (See pp. 65-68.) But the same thing has certainly 

 occurred in a considerable number of cases, only it has 

 resulted in the divided areas being occupied by representa- 

 tive forms instead of by the very same species. The cause 

 of this is very easy to understand. We have already shown 

 that there is a large amount of local variation in a 

 considerable number of s]3ecies, and we may be sure that 

 were it not for the constant intermino^lino^ and inter- 

 crossmg of the individuals inhabiting adjacent localities 

 this tendency to local variation in adaptation to slightly 

 different conditions, w^ould soon form distinct races. But 

 as soon as the area is divided into two portions the inter- 

 crossing is stojDped, and the usual result is that two closely 

 allied races, classed as representative species, become 

 formed. Such pairs of allied species on the two sides of a 

 continent, or in two detached areas, are very numerous ; 

 and their existence is only explicable on the supposition that 

 they are descendants of a parent form which once occupied an 

 area comprising that of both of them, — that this area then 

 became discontinuous, — and, lastly, that, as a consequence 

 of the discontinuity, the two sections of the parent species 

 became seerreo-ated into distinct races 



