364 ME. E. T. GTTNTHEE OTST THE 



have not been isolated from each other long enough for such 

 f aunistic differences to have appeared. But if, on the other hand,- 

 it can be proved that there are faunistic differences between the 

 rivers, then it must be inferred either that the rivers have been 

 populated from different stocks, or that they have been populated 

 from the same stock at so remote a period that their isolated 

 faunas have had time to become modified in different directions. 

 My collections and observations are by no means as complete 

 as could be desired for the solution of the problem, but still they 

 would seem to afford a strong indication of the existence of local 

 faunistic differences in the rivers. I think they show that there 

 are not only differences of species and varieties, but that the 

 relative abundance of the common species varies greatly in the 

 different rivers. 



As I have already pointed out, the rivers which flow into Lake 

 Urmi may be divided into three groups *. The Zola, Nazlu, 

 Shaher, Barenduz, and Gader rise in the mountains to the west 

 of the lake ; the Tatawa and the Djaghatu enter the lake from 

 the south ; and the Murdi, Safi, and Adji receive the eastern 

 drainage. 



Probably all the rivers contain the chub, Capoeta gracilis, 

 and axL Alburmis. The Gader, Tatawa, and Djaghatu are 

 noted for the Silurus glanis which is absent in all the western 

 series of rivers except the Gader Ohai, which opens near 

 the Tatawa, and which probably often mingles its flood-waters 

 with those of the latter stream on the flat land between their 

 mouths. Leuciscus ulanus seems to be peculiar to the waters of 

 the Plain of Salmas, and Leuciscus gaderanus to the southern 

 rivers. I succeeded in catching Sarhus caucasicus and Nema- 

 chihts persa only in the more northern streams, although the 

 former is common all over Persia. 



Most of the fish procured were taken by means of poison, a 

 method of fishing which, so far as its efficiency and unsportsman- 

 like character are concerned, is second only to the method of 

 fishing with dynamite. A mixture of flour and the pounded 

 berries of Coccuhis indicus is mixed with sufficient butter to 

 make a stiflBsh paste. A backwater or reach where the stream 

 does not flow too rapidly is then ground-baited with small pellets 

 of the paste. If the fish are feeding, the poison will begin to 

 work in about ten to fifteen minutes ; and the poisoned fish will 

 * ' Proc. Roy. Geographical Sociesty,' vol. xiv. pp. 504-523. 



