up sand-banks on the coast of a shallow sea, and thus cutting off 

 the lakes from the sea. 



POSTSCRIPT. 

 In the discussion that followed the reading of this paper, the 

 opinion was advanced that an island cannot be thus formed in 

 the mouth of a river ; and that the bifurcation of a river must in 

 every case be due to the river having at some past epoch burst 

 its banks and formed a new channel without deserting the old 

 one. This no doubt is possible, and probably common, in 

 regions where the banks of the river, and the surface of the 

 river, when full, are higher than the neighbouring country. 

 But the possibility of an island being formed in the stream of a 

 river appears to be proved by the instance, to mention only one, 

 of the islands in the Rhine above Bingen. It is scarcely possible 

 for any one to steam past those well known islands without 

 feeling certain that they never were pari of the mainland, but 

 have been formed in the middle of the river by its alluvial 

 deposit. Every bar at the mouth of a river may also be regarded 

 as an incipient island so formed, or rather so forming, which 

 has not yet risen above the water. 



