1922.] N. Annandale & S. h. Hora : Parallel Evolution. 507 



dense and uniform black colour. Similar conditions are found as 

 regards structure in the fishes of large pools of hill-streams, which 

 are usually species with strong swimming-powers but not highly 

 specialized, belonging to such genera as Barbus and Barilius. 



The general analogy between the structure and form of these 

 less specialized members of the fauna of hill-streams becomes in 

 many instances particular as between fish and tadpoles when the 

 more specialized members, living in more peculiar conditions, are 

 critically examined. We know of no exact parallel between the 

 oral floats of the larvae of Megalophrys parva and its allies and 

 any similar structure in a mountain fish ; but when we come to 

 the production of adhesive organs many specific parallels occur. 

 Between the enlarged lips and ventral mouth of several fish of 

 the genera Glyptothorax and Glyptosternum and of tadpoles like 

 those of Bufo penangensis there is a close analogy, and just as we 

 find the oral suckers in different stages of evolution in different 

 species of tadpoles, so also do we find them in different fish of the 

 suborder Siluroidea. In the larvae, for example, of Rana assamensis 

 and its close allies the lips, though ventral and enlarged, are not 

 greatly enlarged and the organ produced is not conspicuous. 

 Almost every stage between this condition and that of Bufo penan- 

 gensis has been observed. Similarly in such species as Glyptoster- 

 num andersoni and G. feae the lips are comparatively small, while 

 in other species of the same genus (e.g. G. labiatum and G. blythi) 

 they are much more highly developed. In both groups the 

 evolution can be correlated with life in waters of stronger and 

 stronger current. 



It is, however, in the ventral suckers of certain tadpoles of 

 the section Ranae Formosae 1 of the subgenus Hylorana on the one 

 hand and similar structures in fish of the genera Garra (or Dis- 

 cognathus) of the family Cyprinidae and Glyptothorax of the 

 family Sisoridae on the other that the closest analogy is to be 

 sought, especially, so far as the fish are concerned, in the former 

 genus. 



We have thought it worth while not only to give a brief 

 general account of the convergence that occurs between these fish 

 and tadpoles but. also to consider the minute structure of the 

 adhesive disc of such Ranid tadpoles as Rana afghana (Gunther) 

 [=R. lalopalmala, Blgr.] and R. livida (Blyth) and to compare it in 

 detail with that observed in the fishes of the genus Garra, in 

 which the modifications are of a similar nature and occupy a 

 similar position on the ventral surface just behind the mouth. 



The disc of Rana afghana 1 is a well-marked structure; it is 

 almost as broad as the bod}' and a little more than half its 

 length. The disc is provided with a free border except at the 

 anterior end, where the border is replaced by the posterior lip. 

 The comparatively thin central portion of the disc in preserved 



1 Boulenger, Rec. Ind. Mus. XX, pp. 123, 130(1920). 

 * Annandale, Rec. Ind. Mus. VIII, pi. iv, fig. 3a. 



