SECT. 7] THE PHYSICS OF SEA-ICE 827 



addition, no large body of water is ever still and currents frequently lead to 

 stratification in which a denser but warmer layer of higher salinity underlies 

 the cold surface water. In spite of these effects, cases have been observed in 

 which the temperature of sea-water below an ice cover remained constant at 

 about — 2°C to depths of a few hundred metres. The term thermocline is also 

 used with sea-water to describe a boundary layer with a large temperature 

 gradient separating two distinct layers of water, even though the method of 

 formation of the thermocline is different from that in fresh water. 



The mechanism of sea-ice formation is complex. Studies of ice formation, 

 and of the crystal structure of ice generally, are made easier by the fact that 

 ice consists of optically uniaxial, birefringent crystals. The degree of bire- 

 fringence is small but sufficient to display the structure readily when a thin 

 section is viewed by polarized light. The orientation of the principal optic axis, 

 called the c-axis, can be found using a polariscope with a universal stage. Some 

 difficulty is experienced in defining a single crystal of ice. The term will 

 here be used to refer to a piece of ice showing uniform colour under crossed 

 polaroids. Such a single crystal has an internal structure consisting of many 

 parallel layers or platelets. 



When the surface layer of sea-water is at the freezing point, further cooling 

 results in the formation of small discoids of pure ice in the top few cm of water. 

 These discoids or platelets are sometimes called frazil ice. Their average size is 

 about 2.5 cm x 0.5 mm and they may be of various shapes, ranging from 

 almost square plates to hexagonal dendrites (Weeks, 1958). The c-axis is 

 always perpendicular to the plane of the discoid. These platelets float to the 

 surface and must initially lie with their planes parallel to the surface of the 

 water, that is with their c-axes vertical. Wind and wave action will compact 

 them, forcing some towards a vertical position. When the cover solidifies it 

 will consist of platelets with a variety of orientations, although for still condi- 

 tions the majority of the platelets will have vertical c-axes. As freezing con- 

 tinues these platelets act as seed crystals and it is found experimentally that 

 the growth of crystals with c-axes inclined away from the vertical is favoured. 

 Two explanations have been offered : larger effective thermal conductivity 

 perpendicular to the c-axis than parallel to it (Anderson and Weeks, 1958; 

 Anderson, 1958), and greater availability of lattice positions in the 0001 planes 

 (Percy and Pounder, 1958). In any event the result is that the closer the c-axis 

 is to the horizontal the larger the crystal grows, at the expense of crystals with 

 less favoured orientations. By a depth of about 5 cm all the ice crystals remaining 

 have essentially horizontal c-axes, and this structure persists to the bottom of 

 the ice sheet with the individual crystals increasing in size with depth. (Note, 

 however, that Shumskii (1955) reported observations on Arctic sea-ice in- 

 dicating uniformly fine-grained crystals.) The regular structure just described 

 will show slight modifications at depths where there were significant increases 

 in the freezing rate (caused by abrupt drops in air temperature). At these 

 depths a number of small, randomly oriented crystals are found, but only the 

 horizontally oriented ones continue to grow with depth. 



