igzi.] S. L. HoRA : Fishes of the genus Garni. 643 



(vii) The eyes in the adult are usually invisible from below 

 though visible in the first few stages on the under surface. 



It Heems quite probable that the various changes enumerated 

 above have been brought about by the rolling of the skin cover- 

 ing the snout towards the under surface and that the mouth 

 changes its position and is being gradually shifted backwards. By 

 this process the anterior labial fold is formed and the divergence 

 and reduction of the branchiostegal membranes and the rays 

 brought about. How the true lips are replaced by secondary folds 

 is a matter of detail The disc develops from the papillate difc 

 rudiment. 



In adults of certain less specialised species of the genus, the struc- 

 ture of the disc and the associated organs resembles a stage in the 

 development of typical species of the genus. Thus G chaudhurii 

 and G. quadriiiiaciiLatus are similar to stage 4, while G. vinciguerraz 

 shows a considerable resemblance to stage 5. In almost all the 

 species that occur in Persia, Syria and Africa the mental disc is 

 less specialised and the true lips are usually present. The new 

 species from Darjiling is represented hy three specimens two of 

 which are mature males and the other a ripe female. The mental 

 disc in the three specimens shows progressive specialisation begin- 

 ning with stage 4 of the developmental series described above. 

 It is not uncommon to find one or two specimens in a big collec- 

 tion of typical Garra in wliich the branchiostegal rays meet at an 

 acute angle behind the disc and the branchial isthmus is narrow. 

 It is still more common to meet with examples in which the poste- 

 rior and the anterior borders of the disc are poorly developed, but it 

 is always possible, after examining a large number of individuals 

 from the same localit}-, to refer them along with normal speci- 

 mens to their proper species. 



In the collection of the Zoological Survey of India there is a 

 series of young specimens from Madras collected by Major Sewell. 

 Of these I have been able to determine only four stages which are 

 separated from each other by fairly wide gaps. 



SKELETON OF THE MOUTH-PARTS. 



Boulenger (1907), when defining the genus Discognathus, gave 

 the following short account of its skeleton : — ' ' The skeleton is 

 very similar to that of Labeo, but the premaxillaries emit short 

 ascending processes, the posterior edge of the mandible is raised 

 into a process at the symphysis, and the clavicles do not form a 

 diaphragm." So far as the skeleton of the mouth-parts is con- 

 cerned, I find great dissimilarity between that of Labeo and that of 

 Garra. The nearest approach to Garra is made by Cirrhina and 

 Crossochilus. The following are some of the salient points in 

 which the mouth-parts of Labeo (fig. 2, 3, 5a) differ from those of 

 the other genera enumerated above : — 



(i) The bones are distinct and separate and have not coalesced 

 to form a rigid structure. 



