igi2. No. 12. THE SEA WEST OF SPITSBERGEN. 23 



salinit}' in the region of Sect. VI than in the region of Sect. I and II. If 

 we were to take the decrease in the maximum salinities as an indicator, 

 it might, for instance, be assumed that the decrease in the salinity of the 

 current was about 0.05 ''00 from the region of Section I to that of Sect. 

 M, considering that the observed maximum salinities liave decreased from 

 between 35.06 and 35.09 '^co to between 35.01 and 35.04^00. If, there- 

 fore, one took, for instance, the isohaline of 35.0 '^ 00 as indicating the 

 boundary of the waters carried by the current through Section I, the 

 corresponding boundary-line of Section VI would be the isohaline of 

 34.95 "^' 00. To judge from our sections, Pis. I\'— \'I, this might also seem 

 a reasonable method; but as it would be somewhat complicated, without 

 warranting correspondingly accurate results, and as the boundar3'-line thus 

 found would at most stations lie approximately in 450 and 500 metres, 

 we consider it to be sufficiently accurate for our purpose simply to take 

 the mean of the temperatures of the water-strata down to 450 or 500 

 metres at all stations within the current in each section, which means, 

 in Sect. I, Stats. 2 to 5; in Sect. II, Stats. 7 — 11; in Sect. I\', Stats. 

 25 — 29, and in Sect. \'I, Stats. 17 — 19. We have here left out the sta- 

 tions near the coast, where the depths were less than 400 metres. As the 

 temperatures of the surface-layers are ver\- dependent on the radiation of 

 the sun, the temperature of the atmosphere and the melting of ice, etc., 

 the}' ought to be excluded, and the means should only be taken of the 

 strata between 500 metres and a certain depth below the surface. 



Another difficulty is involved in the circumstance that the dates of 

 the sections differ much. It is obvious, for instance, that the mean tem- 

 perature of Sect. M, taken on July 17th — i8th, 1910, should be compara- 

 tively warmer than that of Sect. I which was taken on June 25th — 26th, 

 or 22 days earlier in the summer. 



The observations repeated in September at Stats. 3, 4, 29, 30, and 

 31, afford, however, some means of finding out what the rise in the 

 temperature of the water-strata may be during the summer in the same 

 localities. By taking the mean of the temperatures for 50, 150, 250, 350, 

 and 450 metres, we ma}- get a fairly correct value for the mean tempera- 

 ture ot the whole volume of water between the surface and 500 metres. 

 The temperatures may easily be taken from vertical temperature-curves 

 (see Figs. 22 — 30). The following table gives the results for Stats. 3, 4, and 

 29. The water at Stats. 30 and 31 was too shallow (195 and 235 metres) 

 to give trustworthy results for such a comparison. 



