84 



BJ0RN HELLAND-HANSEN 



[rep. of the "MICHAEL SARS" NORTH 



ing mean variations. In so doing we must use observa- 

 tions from several years and neglect the changes from 

 one year to another. The result of such a computation 

 is given in the table above. The stations used are 

 grouped about the middle of April, the middle of May, the 

 first half of June and the middle of July (cf. Fig. 15). 

 The mean dates are noted in the table, as well as the 

 number of observations. The temperature at the upper 

 level minus that at the lower has been computed for each 

 station, the average of such differences being given in the 

 columns headed A t. Numbers showing the mean varia- 

 tions are printed in brackets. 



The vertical variation of temperature between the 

 surface and 25 metres is very irregular. The observations 

 in May were taken in 1911 and 1922; in the latter year 

 especially they showed a considerably higher temperature 

 at the surface than at 25 metres. Observations at 12 stations, 

 between the 5th and 12th of June, 1911, showed, on the 

 other hand, a lower temperature at the surface than at 

 25 metres. Alter a gale the temperature difference between 

 the two levels is generally very small even in summer. 



The difference of temperature and the mean varia- 

 tions below 25 metres progress very regularly in space 

 and time. The vertical variations of temperature have a 

 maximum between 25 and 50 metres in spring and sum- 

 mer. The above computations show a mean difference of 

 014" C. per metre in the middle of July in area "B". 

 This difference is evidently still greater at the end of August. 



The conditions alter from region to region, but the 

 chief characteristics of the upper water-layers, shown in 

 the example from area "B", are valid for most of the 

 North Atlantic: there are small values for the mean varia- 

 tion from the surface down to 20 or 25 metres even in 

 summer; at a short distance below this depth there are 

 values that increase numerically all the time during spring 

 and summer and reach an absolute maximum (in the warm 

 season they are higher than anywhere else along the verti- 

 cal from surface to bottom); while below some 50 to 75 

 metres the values decrease numerically downwards to a 

 minimum which is met with at varying depths. 



The vertical convection in winter, resulting in uniform 

 temperatures from the surface downwards, attains very dif- 

 ferent depths in the different parts of the sea. We have 



in winter very few observations. On the "Challenger" 

 Expedition serial observations were made in the latter 

 half of February, 1873, at Stats. 4, 5, 8 an 10 across the 

 deep eastern part of the North Atlantic in 25-23°N., and 

 in the first half of March at Stats. 15, 17, 18 and 20 across 

 the deep western part in 21 — 19°N. At all these stations 

 observations were made at the surface, 100, 200, 300 fathoms 

 and so forth. When interpolating for metres instead of 

 fathoms we get the following averages of A r (in °C.) 

 and mean variation: 



The vertical variation of temperature between the 

 surface and 200 metres is much greater here than was 

 shown above for the area "B" in April (when the heating 

 of the surface had commenced) and even greater than in 

 May. At the first 4 "Challenger" stations mentioned 

 observations were not made between the surface and 100 

 fathoms except at Stat. 10, where the following tempera- 

 tures were observed: 



depth: 10 20 30 40 50 60 70 80 90 100 fathoms 

 I : 21-7 221 221 221 220 2i-9 21-8 21-4 20 7 201 19-4° C. 



Even if the surface is further cooled during several 

 weeks, it is not likely that the winter convection will go 

 deeper than 70 — 80 fathoms or about 150 metres. There 

 are no determinations of salinity so the stability cannot 

 be examined. 



At Stat. 18 in the western part of the North Atlantic 

 the temperature was observed at 50 fathoms, being here 

 23-0° C. while it was 23-3^ C. at the surface and 210'' at 

 100 fathoms. At this stations the vertical convection in 

 winter would scarcely extend to as much as 100 metres 

 below the surface. 



In the North Atlantic near the equator the annual 

 range of temperature has a minimum (Fig. 9). In a region 

 where the mean range is less than 2°C. and, in part, less 

 than 1°, observations were made in the middle of July 



