10 MAEIOI^ EXPEDITION TO DAVIS STRAIT AND BAFFIN BAY 



The point of solidification of sea ice; i. e., the temperature at 

 which it becomes a true solid, is however, quite different from the 

 freezing point of sea water. The fact that sea ice is a conglomera- 

 tion of pure ice crystals and of particles of brine, combined in vary- 

 ing proportions, due to varying temperature and salinity, naturally 

 prevents one single definite point of solidification. Strictly speak- 

 ing, the solidification j^rocess in such ice can not be considered 

 completed until the last salt solidifies, namely CaCh, the eutectic 

 point of which is —55° C. ( — 67° F.). For general purposes of 

 calculation, therefore, the freezing point of sea ice is assumed to 

 occur at the eutectic point of its major salt NaCl (about 77 per cent 

 of the salt content of sea water), i. e., at approximately —22° C. 

 ( — 7.6° F.). The closer the solidification point of a sample of sea ice 

 lies to the freezing temperature of sea water, the more homogeneous 

 w411 be the composition of the ice, and the less the deposition of brine 

 and salt. The lower the tem])erature to which sea water is exposed, 

 the more rapidly will the net of ice crystals form. 



One of the most astonishing things about sea ice is the fact that it 

 is so fresh. Malmgren (1928, Tables 1 and 5), has made several 

 determinations of the salinity of the ice of the polar cap, showing it 

 to be subject to a relatively wide range — 0.05 0/00 to 11.59 0/00. A 

 salinity of 7 0/00, therefore, may be taken as the salt value of ordi- 

 nary sea ice.^ It would be interesting to learn what is the average 

 salinity of the Arctic water from which ice of 7 0/00 is formed. The 

 salinity of the surface layers of the north polar ocean, outside of the 

 continental slope, as Nansen (1928, p. 11) has pointed out, are greatly 

 diluted by the land drainage from the Eurasian side. Nansen esti- 

 mates that the inflow of fresh water from Siberian rivers alone is 

 sufficient to cover the ocean from the Asiatic coast to the pole each 

 year, with a surface film 3i/> feet thick. The distribution of salinity 

 of the water as found along the Frani's track was 28.40 0/00 to 29 0/00 

 to a depth of 30 meters, 90 miles north of the New Siberian Islands, 

 and only 31.7 0/00 to 33 0/00 at 30 meters, within 300 miles of the pole. 

 In view of these observations we may conclude that the surface layers 

 in the polar basin, i. e., the water subjected to freezing, vary little from 

 a mean salinity of 31 O/OO."' The average salinity of sea ice, approxi- 

 mately 7 0/00, shows, therefore, that only about one-fourth of the 

 total quantity of salts of sea water enters its ice. Nansen, on the 

 drift of the Fram, made no determinations of the salinity of sea ice, 

 but from our information on the distribution of the salinity in the 

 Arctic Ocean, the MaurPs ice experiments can, without criticism, be 

 conq)ared directly with the Frames hydrographical observations. 



The rate at which sea water freezes is a very important factor 

 in determining the percentage of salt imprisoned in the ice. The 

 more rapid the process, the more sudden the pure ice crj^stals im- 

 pregnate the water, catching within their meshes a large quantity 

 of brine drops. The saltiest piece of ice that Malmgren (1928, p. 



5 Weyprecht (1897, p. 58) during the drift of the Tcc/ctthoff found the salinity of a 

 piece of thin ice formed rapidly under low temperature in the Arctic to be 25 0/00, and 

 this is probably as salt as sea ice ever is. 



® No other ocean in the world can show such low salinities more than a few miles out 

 from land. The inflow of land water around the basin's border fails to explain a salinity 

 uniformly so low over the entire expanse. Probably the absence of evaporation and the 

 presence of ice are the two factors chiefly responsible, combined with the seasonal cycle 

 of freezing and melting, repeated throughout the years since the present cap became 

 permanent. 



