paa u CO st 
1906. No.3. AMUNDSEN'S OCEANOGRAPHIC OBSERVATIONS IN 1901. 31 
bulk of water, and the formation of ice on the surface will gradually 
increase the salinity, especially whereever the sea be shallow. This 
increase of salinity by ice-formation must obviously be the explanation 
of the remarkable fact, that the heaviest bottom-water, ¢. g. on the coast 
banks of Novaya Zemlya, has very often a higher salinity than was ever 
found in the comparatively warm Atlantic water of the eastern Barents 
Sea. If for instance, it be supposed that by the vertical circulation of 
the water during the winter, the salinity has become almost uniform, 
about 34°70 %0 (cf. Russian Stat. 10, Aug. 6, 1904, in Table below) from 
the surface to the bottom, in a region where the sea during the winter 
and spring be covered by a layer of new ice, with a mean thickness of 
2 metresl, the salinity of the underlying 198 metres of water will be 
increased by about 0'33 %/00, and instead of 34°70 %0 the mean salinity 
will become about 35°03 %0. If the sea be shallower, the formation of 
2 metres of ice during the winter would increase the salinity by still more. 
If the sea be only 100 metres deep it would in this way be possible to 
start with a water of mean salinity 3440/00, and still get water with a 
salinity above 35 °/oo after 2 metres of ice had been formed. The effect 
of ice-formation may naturally be somewhat reduced locally by horizon- 
tal circulation in the sea, carrying a new supply of water in under the 
ice; but still the effect upon the whole mass of the water will be the same, 
and whereever ice is formed, it may consequently be expected that the 
salinity will be higher in the winter than in the summer, as is also 
actually borne out by the observations. 
The cold, heavy water thus formed, will sink and spread out on the 
banks, and if it be heavier (either on account of lower temperature or 
higher salinity) than the bottom water on the same level or at lower 
levels in the neighbourhood, it will gradually creep along the bottom in 
under such waters. This is probably the explanation of the very frequent 
and sudden occurrence at a great many stations, of a thin bottom-layer 
of higher salinity and density, and a temperature generally lower, but 
sometimes higher, than the overlying water. 
! The formation of ice on the surface during winter, is much helped by ice-pressures, 
which breaks the ice, and piles the floes up into high hummocks. When the pressures 
ceases again, open lanes and channels are formed which are rapidly covered by new 
ice, again to be broken and piled up. Thus ice-pressure greatly increases the forma- 
tion of ice as welll as the cooling of the sea. If the ice were not continually being 
broken, but formed a continuous and permanent sheet, it would increase extremely 
slowly after having attained a certain thickness, and after being covered with much 
snow, and the underlying water would be protected against cooling. 
