58 VOYAGE TO THE POLAR SEA. July 



matioii was apparently the same as that of the original 

 snow layer. Unfortunately a sample was not obtained, 

 and I cannot say whether the ice was brackish or not, 

 but I think that it was so and that the change had 

 taken place by the percolating downwards of the salt 

 brine from the ice above, as noticed in other floebergs. 



4 On the whole I conclude that the Polar ice in- 

 creases in thickness below and not superficially, and by 

 the natural freezing of the water at its lower surface 

 during the winter. As before mentioned, a Polar 

 floe only one year old is composed, not of ordinary 

 ice frozen on the surface of a space of water, but 

 of a quantity of conglomerate ice pressed together by 

 the general movement of the pack and then frozen into 

 a floe ten or twelve feet and upwards in thickness ; and 

 to whatever thickness such a formation is continued 

 the freezing and consolidation of the whole into a 

 compact mass of ice stakes place at all depths, for it is 

 remarkable that only once have we found a cavity 

 denoting where a hollow, left when the pieces com- 

 posing the floe were first pressed together, has not 

 become filled up. In more southern latitudes, where 

 such cold water is not found at the same depth, similar 

 cavities remain unchanged. 



' In the middle of each of the large shore lakes we 

 find very thick ancient ice ; wdiether it is frozen to the 

 bottom or not we have no means of ascertaining ; but 

 it is apparently immovable by the wind. Near the 

 shore the inpour of heated water during the summer is 

 sufficient to prevent its growth at a less depth than 

 eight feet. During the winter the ice newly formed 

 near the border of the lake by the natural freezing of 



