POLAR ICE. — INFLUENCE 'ON ATMOSPHERE. 299 



the advance of the wind must in consequence be 

 farther and farther before it be reduced to the tem- 

 perature of the ice ; and, therefore, some snow would 

 continue to be precipitated to an increasing and un- 

 limited extent. 



Hence, as winds blowing from the Pole must be 

 replaced by air that is both warmer and damper, it 

 is not possible that they should carry away more 

 moisture from the circumpolar regions than the 

 air which replaces them brings : but, on the con- 

 trary, as the snow deposited on the interior ice by 

 southerly storms, (from the nature of the circum- 

 stances), must be derived from evaporations out of 

 the sea ; it is evident, that there must be an in- 

 crease of snow in the icy latitudes, and that we can- 

 not possibly determine any limit beyond which it 

 may be affirmed that no snow can be deposited. 



4. On approaching a pack, field, or other com- 

 pact aggregation of ice, the phenomenon of the 

 ice-blink is seen whenever the horizon is tolerably 

 free from clouds, and in some cases even under a 

 thick sky. The ice-blink consists in a stratum of a 

 lucid whiteness, which appears over ice in that part 

 of the atmosphere adjoining the horizon. It appears 

 to be occasioned thus : Those rays of light which 

 strike on the snowy surface of the ice, are reflected 

 into the superincumbent air, where they are rendered 

 visible, either by the reflective property of the air, 

 simply, or by a light haze, which, on such occasions, 



