828 POUNDER [sect. 7 



Close examination of a sea-ice crystal shows it to consist of a large number of 

 thin parallel layers, with inclusions of brine, air and possibly other impurities 

 between them. The average thickness of these layers is of the order of 0.5 mm. 

 This structure is not confined to sea-ice but is found in any ice formed from an 

 impure melt, and was observed as long ago as 1860 by Faraday. Examination 

 of the under surface of a growing ice sheet shows something of the method of 

 formation. The lower 1 to 2 cm consists of pure ice platelets with layers of brine 

 between them. The platelets in a single crystal are vertical and quite accurately 

 parallel to each other. This skeleton layer has no mechanical strength until 

 further cooling results in ice bridges growing between the platelets. The crystals 

 gradually become interconnected in this way and brine is trapped in pockets 

 or cells. These brine cells shrink in size as cooling continues, since the con- 

 centration of salt must increase so that it is at all times in thermal equilibrium 

 with the cooling ice. Ringer (1928) has investigated the composition and 

 concentration of sea-ice and sea-water at temperatures down to — 80°C. 



The brine trapped in sea-ice is thus found partly between crystals and partly 

 in intracrystalline layers, the latter sites probably accounting for a large 

 majority of the total quantity of brine. The brine cells are usually vertical 

 cylinders of circular or elliptical cross-sections. In any ice sheet with a tempera- 

 ture gradient across it (and this includes all natural sea-ice) the brine cells are 

 never in equilibrium. Diffusion will keep the concentration of brine within a 

 cell uniform, so that if it is in equilibrium with the ice at one depth it cannot 

 be at others. At the colder end, water will freeze out of the brine ; at the warmer 

 end, the brine will dissolve ice. The net effect is a migration of the cell along the 

 temperature gradient in the direction of higher temperatures. This migration 

 was demonstrated strikingly in the laboratory by Whitman (1926). The rate 

 of migration decreases rapidly with temperature, and the time for significant 

 movement is of the order of months for ice temperatures below — 15°C. It 

 should cease entirely when the brine is completely frozen. Investigations of the 

 freezing of sea-water by Nelson and Thompson (1953) show that this does not 

 occur for temperatures above — 50°C. Brine migration reduces considerably 

 the salinity of the upper layers of annual sea-ice in temperate latitudes, migra- 

 tion being aided by drainage under gravity in the spring as rising temperatures 

 cause the brine pockets to enlarge and become interconnected. These processes 

 are also of great importance in perennial sea-ice in high latitudes during the 

 summer months. Sea-ice which is over a year old can be melted to yield potable 

 water. 



Most sea-ice is thus honeycombed with a network of large numbers of brine 

 cells and air pockets. At least above temperatures of — 15°C, there are some 

 interconnections so that a piece of sea-ice removed from an ice cover and 

 stored at this or higher temperatures will gradually "bleed" brine. The rate of 

 brine seepage increases rapidly with temperature. It seems to be the general 

 opinion of most investigators of sea-ice that all of the physical properties of 

 this material are largely controlled by the brine content and the distribution 

 of the brine cells. 



