bEPORt ON FISHES. PART 1: COBITIDAfi 305 



lagoons. No doubt tlie latter are far richer in food stuffs than the lake itself. 

 In the case of the third species ( N. panguri), from Tso Nyak and Pangur Tso, 

 there is no doubt that the s])ecies actually lives at present in the lakes which are ' 

 fairly rich in vegetation, though it also enters the streams running into them. To 

 my mind there can be no doubt that all three species developed in the late glacial 

 lake but only the third one has remained common in lacustrine environments, the 

 others hanging on as best they can chieHy in small pools in the basin." 



It is clear from the above that at the present time the Panggong Lake acts as an effective 

 barrier for the distribution of various species. Further it seems probable that this hal)itudi- 

 nal segregation may have induced the development of different species. It is likely that in 

 the late glacial lake there was only one species derived from a torrential stock and that when 

 the environments became restricted, it developed along different lines in different localities 

 and resulted in the production of several new species. ** This supposition supports the 

 hypothesis of Regan^ "that as a rule the first step in the origin of a new species is the for- 

 mation of a community with a new and restricted environment, or with new habits; in other 

 words, that some form of isolation, either localization or habitudinal segregation, is the 

 condition of the development of a new species." 



From the modifications of the bladder described above, it does not follow that the lake 

 species cannot enter into brooks or vice versa. All species of Nemachilus are flattened and 

 adapted for clinging to foreign objects and if ponds, pools and lakes are in communica- 

 tion with brooks it is possible that the species uf one habitat may enter the habitat of the 

 other set of species. For instance, it often happens that torrential forms enter springs and 

 pools for breeding purposes. The occurrence of a species (A^. panguri) in both types of 

 habitats, therefore, does not in any way help to fix its evolutionary status. 



A'^. x'ittatus seems to have developed in the Kashmir lakes and is endemic in them. 

 Neiiicichiliis sttilicckae, N. iiiicrtipx. N. touiicauda and N. gracilis are widely distributed 

 species. The first three are, however, restricted to the high altituiles, whereas A', gracilis 

 is found as far down in the Indus as Attock.^" Almost all the species were obtained by the 

 Netherland Karakorum Expedition not only from the Nubra Valley but also from the Kara- 

 kash river, which now drains into the Tarim river system. Mukerji and I referred to this 

 discontinuous distributidu of the species, but the difficulty has now disappeared for "On the 

 basis cjf his geomorphological studies. Dr. de Terra has reconstructed the Tertiary drainage 

 pattern of the western part of the Tibetan plateau. A nunilx-r of rivers ran from west to 

 east, one of them occupying the present valley of the Upper Indus."'' The close similarity 

 between the torrential fish fauna of the Karakash river and of western Tibet suggests, at 

 any rate, a common drainage for the waters of these two areas at no great distant date and 

 lends great support to the hypothesis advanced by de Terra. Reference may also i)e made 

 to the occurrence of A', ladacensis in Ladakh and the Karakash \'alley, but it has to be 

 remembered that only a few specimens of this species are known so far, and, in consequence, 

 its specific limits have not yet been precisely defined. 



' Dr. A. W. C. T. Herre liad a similar problem in the evolution of the seventeen species of fishes of Lake 

 Lanao in the Philippines {Amer. Nat. XLVII, pp. l.S4-l()2, 1933). 

 •Regan, Nature. CXIII, p. 569 (1924). 

 '"Hora, Rcc. hid. Mus. XXXV, p. 189 (1933). 

 "Hutchinson, Naliire, CXXXIV, p. 87 (1934). 



