NATURAL HISTOEY OP LAKE TJEMI. 353 



former lake-basins by an examination of terraces and of sinter 

 and travertine formations ; but in the case of Lake Urmi neither 

 were the investigations of Loftus or Eodler rewarded by the 

 discovery of lake-terraces, nor do the travertine deposits of 

 Dashkiesen appear to be o£ lacustrine origin. The travertine 

 deposits round Lake Urmi seem to be the result of mineral 

 springs. Pohlig mentions Pleistocene flood-zones of Neritina 

 and Dreissena on the islands and on the Shahi peninsula; but 

 although I saw a single specimen of a Dreisse^za-traverfcine mass 

 in Urmi, I was not able to discover where it had been found, and 

 consequently I was not able to confirm Pohlig's theory. 



Lahe of Urmi. — The lower parts of the depression of the 

 Azerbaijan plateau are covered by the great salt-lake of Urmi. 

 The length of this sheet of water, as measured from north to 

 south, does not fall short of 80 miles, and its breadth is about 

 25 miles at the wider parts, but is far less where the Shahi 

 peninsula, jutting out, lessens the distance to ten miles. In 

 former days the mountain of Shahi seems to hf?ve been entirely 

 surrounded by water, but now it is counected with the eastern 

 shore by dry land except when the spring floods convert the 

 isthmus into a marsh. The area covered by the lake is about 

 1750 square miles. 



The shores of the lake, in some few places where the hills come 

 down to the water's edge, are rocky, and falling away abruptly 

 exhibit low cliffs of erosion ; but for the most part the land 

 slopes so gradually that the bather may have to wade for a mile 

 -or two before he reaches water which is out of his depth. The 

 more notable elevations on the coast-line besides Mfc. Shahi are 

 at Gruverchin Kala, a promontory of granite with pink felspars 

 and of Miocene limestone with shells and echiuoderms, near 

 G-avilan ; at St. George's Hill, Superghan ; and at the Bezau 

 Daghi, which are at any rate partly of volcanic origin, since 

 their lower flanks are composed of pumiceous hornblende-biotite- 

 andesite. 



The islands will be described in greater detail below. There 

 are a few small islets off Gruverchin Kala, at the northern end 

 of the lake, which I was unable to visit ; but the more important 

 group of islands is situated in the southern half of the lake, 

 though its exact geographical position has yet to be defined. 



The depth of the lake is inconsiderable. At no place does 

 this huge expanse of water appear to be more than 40 feet in 



