Interbreeding of Birds. 549 



tinually lost by interbreeding ; so that the similarity of 

 individuals in a species is retained, Avhilst the sterility pro- 

 duced by a specific variation prevents the universal mongreli- 

 zation of species which might otherwise take place. 



Interbreeding is a check upon the indefinite multiplication 

 of species ; whilst the narrow limit in which it is possible 

 provides against the extinction of specific differences. 



A very interesting case of the interbreeding of forms 

 hitherto supposed to be specifically distinct has just come to 

 my knowledge. 



Cinclus cashmiriensis is a well-known species of Dipper, in 

 which the dark-brown and white on the underparts are dis- 

 tributed in the same manner as they are in our bird, the white 

 throat and breast being divided abruptly from the dark-brown 

 belly and flanks. Its range extends from Lake Baical to the 

 Altai Mountains. In its more northern locality it meets with 

 C. leucogaster, with which it apparently interbreeds ; for, as 

 is well known, every intermediate form, as well as both 

 extreme forms, are found in the district. I have lately had 

 an opportunity of examining a large series of Dippers, sent 

 from the Altai Mountains by the Siberian collector of Herr 

 Tanere of Anclam ; and I find that in the southern extremity 

 of its range C. cashmiriejisis comes into contact with C. sor- 

 didus, with whom it also apparently interbreeds ; for here again 

 we have every intermediate form as well as both extreme 

 forms. It is impossible to say whether C. leucogaster would 

 interbreed with C. sordidus or not, because we do not know of 

 any locality where both are found; but it is difficult to avoid 

 coming to the conclusion that both of them interbreed with 

 C. cashmiriensis. By obtaining a series of examples from 

 both localities, a complete series of Dippers may be obtained, 

 beginning with birds having nearly the whole of the under- 

 parts white, and ending with birds having the whole of the 

 underparts brown, the throat and breast being only a shade 

 paler than the belly and flanks. 



In all these forms of Cinclus there is little or no diff'erence 

 in the colour of the upper parts, which makes it very diflScult 

 to suppose that the difference in colour can be a {)rotective 



SER. IV. VOL. VI. 2 P 



