32 FRIDTJOF NANSEN. M.-N. Kl. 
As this heaviest water close to the bottom, has sometimes salinities 
higher than were ever observed in the vicinity in higher strata, it might 
seem somewhat difficult to believe that the salinity at some place in the 
same region, could have been, during the winter, so high in all strata 
between the surface and the bottom, at least for any length of time. If it be 
supposed however, that the vertical circulation during winter, had etab- 
lished a nearly uniform density between surface and bottom; further that 
the temperature had been lowered towards freezing-point, and that the 
salinity had been increased by formation of ice; and if then, on extremely 
cold days, great open lanes had been suddenly formed in the thick ice, and 
thus a great open water-surface suddenly exposed to very rapid cooling, 
there would have been a very intense formation of ice. The upper 
water-stratum might in this way be cooled down and its salinity increased 
more rapidly than it could be replaced by vertical circulation. The very 
heavy water thus formed might then sink through the underlying water 
and might perhaps in some cases, especially where the sea is shallow, 
reach the bottom before it was too much intermixed with intermediate 
strata. If such a lane be covered with new ice, and reopened several 
times, as is very frequently the casel, the effect of the local cooling 
and rapid formation of ice might be still more increased, and a column 
of very cold and saline water might actually be formed for a while be- 
tween the surface and the bottom. But as soon as the sea-surface is 
again covered by thicker ice, the very heavy water, will sink below the 
upper strata, and regular, more uniform conditions will be reestablished, 
while the heavy layer will remain on the bottom*. 
! The ice-pressure is generally repeated very regularly with the tidal currents, along such 
channels and lanes in the ice, where a line of weakness has been formed. In the 
intervals between the pressures the channels are videly opened. This happens especially 
at the time of spring-tide. 
Professor Edlund’s theory (cf. Övers. Kongl. Vet. Akad. Handlingar, Stockholm, 1863) 
that, under certain conditions, ice might be formed near the bottom of the sea, if 
nw 
correct, offers another possible explanation of the above fact. Edlund’s idea was that 
the sea-water might be cooled below freezing-point, even though in slight motion, and 
such super-cooled water might therefore sink towards the bottom without freezing, 
until suddenly, by some accident, the freezing process be started, and quantities of ice 
formed which would suddenly rise to the surface. Edlund based his theory upon ob- 
servations made by E. A. Nordenskiöld, who with a maximum and minimum 
thermometer actually found temperatures below freezing-point in the sea near Gotland; 
but with the imperfect instruments of those days too low temperatures were frequently 
observed in the sea. During Nordenskiöld’s Vega-Expedition temperature-readings as low 
as —2'3 and —2'4° C. were repeatedly taken from the deep water of the Kara Sea, 
in August 1878, when there was no possibility of super-cooling. And at Vega’s winter- 
harbour near Bering Straits, temperature-readings as low as —2'8° and even —3’0 C, 
were taken in water from 2, 4 and 6 metres below the ice-covered surface (cf. O. 
