19 1 2. No. 12. THE SEA WEST OF SPITSBERGEN. 25 



The values thus obtained b}- comparing some few vertical temperature- 

 curves cannot, however, be considered very accurate, also for this reason, 

 that the conditions in the sea may evidently change much in a short time, 

 as we have previously pointed out; and there are probabl}' great periodical 

 or unperiodical vertical movements of the water-strata, by which a stratum 

 with comparatively low temperatures, ma}' for instance at one time be 

 observed in 500 metres, and at another time in 450 metres or higher. 

 Such a vertical movement may possibl}- explain wh}- at Stat. 29, the 

 curve of September 6th shows lower temperatures, in depths below 500 

 metres, than the curve of July 22nd (see Fig. 29). If the water-strata of 

 the September-curve be lowered about 40 metres, the two curves would 

 nearly coincide below the depth of 500 metres, while the difterence between 

 them would be somewhat increased above that level. 



B}- taking the mean increase of temperature from June and Julv to 

 September at the three Stations mentioned, 3, 4 and 29, we find an in- 

 crease of o.io^ C. in 10 days in the mean temperature of the water of 

 the Spitsbergen Atlantic Current between 50 and 450 metres of depth. 

 This value does not give the increase in the temperature of the same 

 water that is continually moving northwards; but it indicates the increase in 

 the temperature in the locality where the water is continually being renewed. 



In the vertical sections of the Norwegian Atlantic Current northwest- 

 wards from the Sognefjord, we have found [see 1909, p. 230] an increase 

 in the mean temperature of the Atlantic water (except the surface-strata) 

 with salinity above 3500 ^' 00. h'om 7.23° C. on May 23rd — 24th, 1903, to 

 7.66*^ C. on August loth, 1903; /. e. an increase of 0.43'^ C. in 78 davs, 

 or 0.055^' C. in 10 days. 



From August loth, 1903, to November 17th — i8th, 1903, the corre- 

 sponding increase was from 7.66° C. to 8.13" C: i.e. 0.47° C. in 99 days, 

 or 0.047 C. in 10 days. 



The increase in the temperature of the water of the Atlantic Current 

 should consequently be much slower in the region of the Sognetjord- 

 sections, or in about 62*^ N. Lat. than in the region of Isachsen's Sects. 

 I and I\', or in about 75*^ and 78'' N. Lat. The values are not, however, 

 quite comparable, as Isachsen's Stats. 3, 4, and 29, were in the warmest 

 part ot the current, /. c. near its right side, while the mean temperatures 

 of the Sognefjord-section were the means of all observations in the Atlantic 

 water of the whole section. 



The mean temperatures of the water between 50 metres and 450 metres 

 (computed from the temperature at 100, 200, 300, and 400 metres as above) 

 are the following in Isachsen'sfour sections of the Spitsbergen Atlantic Current: 



