392 Alf Trolle. 



of the water being then read off on the rod. This measurement was 

 dependent, however, on how much ice had been pressed down by the 

 weight of snow etc. on the surface of the ice, and it would have been 

 more exact to measure the distance between the upper and under 

 surface of the ice. This could not be done, however, as several 

 layers of frozen snow were formed аЬол'е the ice in the course of the 

 winter. 



When the sample of water was taken, great care had to be used 

 to prevent snowilakes and powder in the water from accompanying 

 the latter into the bottle. This could not be prevented always and 

 irregularities in the values of the salinity are certainly often due to 

 the presence of this line snow in the bottles. 



For the sake of control, samples from 8 and 16 m. were collected 

 with Ekmann's water-bottle a few times during the winter. 



In spring the samples were taken from a depth which was equal 

 to the thickness of the ice, as the salinity at the surface itself was 

 then far too irregular, owing to the presence of water from the mel- 

 ting snow and ice. The observations are shown in curves in PI. XXIII. 



The curves display a distinct connection between increase in 

 the surface salinity and the increasing thickness of the ice in the 

 course of the winter. The following was the sequence of events. 



On September 21st 1906 ice formed round about the ship and 

 then rapidly increased in thickness. The autumn was clear with but 

 little snow. The new ice was, as usual, very plastic and gradually 

 became harder, the salt crystals being separated out on the upper 

 side of the ice. 



The ice increased most in thickness in the autumn and tlie tirst 

 part of winter, and at the same time the increase in salinity was great- 

 est (see PI. XXIII). 



From middle of February till end of May both the thickness of 

 the ice and the salinity remained almost unchanged. The air temper- 

 ature reached its minimum in the middle of February. At the end 

 of May the air-temperature became higher than the freezing point 

 of the sea-water ( — 1.7°) and the ice must then have begun to melt. 

 Nevertheless, the thickness of the ice increased just at this time, new 

 layers being formed on the under side of the winter ice, as the w-ater 

 from the melting snow and ice became frozen where it came into con- 

 tact with the sea-water, which had a temperature of about — 1.7°. 



On June 28th the rivers began to break up on land and let the 

 warm water flow out (the river water had a temperature of + 10 



1- 14° C). The ice now rapidly decreased in thickness. On July 



6th 1907 the ship was free of ice. In the fjords, however, the ice still 

 remained and the thickness of the ice measured in the mouth of the 

 fjord at Maroussia in the middle of July was 188 cm., at the end of 

 July 155 cm. 



