1922.] S. L. Hora : Fish of Mountain Torrents. 61 



has been shown elsewhere, 1 the forms from various localities are 

 not specifically identical. 



Methods of propagation. — All the hill-stream fishes with which 

 I am acquainted are oviparous. The rapid current in these 

 streams makes it impossible for them to lay their eggs loose in the 

 rocky beds, as they are liable to be carried down stream and 

 destroyed. In my tours to the hill-streams I have always found 

 that the pools in the courses of these streams were full of young 

 specimens of the genera Nemachilus and Bariiius and of a. few 

 other small fish that are usually found in pools. In no case was 

 I able to find young of Garra or of any other highly modified 

 hill-stream forms. Dr. Annandale and Major Sewell, however, 

 found Garra and Psilorhynchus in pools of hill-torrents in the 

 Western Ghats. There appear to be two possibilities, (i) that the 

 hill-stream fishes migrate to slow running waters to lay their eggs 

 and that every generation has later on to ascend up-stream to its 

 natural habitat or, (ii) that the eggs are tightly fixed to stones. 

 I am unable to say at present which hypothesis can be accepted 

 as correct. 



' Hora, Rec. Ind. Mas. XXII, pp. 633-687 (1921). 



