Physical Geography. 149 



still less numerous fauna of the Caspian Sea, as far as 

 that fauna is known, which sea, or brackish lake, was, it 

 is believed, once connected with the northern ocean, as 

 the fauna seems to testify. My belief is, that these 

 Permian waters were also of an inland, unhealthy 

 nature, and, like those of the Caspian, had previously 

 been connected with the open ocean. 



Besides the poverty in number and small size of 

 the mollusca, the chemical composition and lithological 

 structure of the Magnesian Limestone, seem to me to 

 afford strong hints that it was originally deposited in a 

 large inland salt lake, and not that it was entirely de- 

 rived from calcareous organisms, and subsequently 

 altered into dolomite by chemical changes. I am well 

 aware that there are such masses, occasionally, for ex- 

 ample, in the Carboniferous Limestone which was formed 

 in an open sea. Some modern atolls are known to be- 

 come dolomitised, as described by Dana, but in the 

 Magnesian Limestone corals are chiefly conspicuous by 

 their absence. I repeat that the Permian Magnesian 

 Limestone was not, as used to be supposed, formed 

 in the sea, but in an inland salt lake, under such 

 circumstances that carbonates of lime and magnesia 

 were deposited simultaneously, probably, by concentra- 

 tion of solutions due to evaporation. In an open sea, 

 lime and magnesia only exist in solution in very small 

 quantities, and limestone rocks there are formed, as in 

 coral reefs, by organic agency. 



In some of the lower strata of the Magnesian Lime- 

 stone, when weathered, it is observable that they consist 

 of many curious thin layers, bent into a number of very 

 small convolutions, approximately fitting into each 

 other, like sheets of paper crumpled together. These 

 dolomitic layers convey the impression that they are 



