For studying the problem of the variation or non-variation in the annual water- 

 and heat-convection of the Atlantic current, the boundaries of its eastern ramifications, 

 such as the Barents Sea, Skager Rak and Kattegat, should be chosen. Hydrographically, 

 these seas are among the best 

 known. The projections of 

 the Atlantic current entering 

 there disappear at intervals, 

 being then replaced by water 

 of different origin. This too 

 is a point in our favour, as it 

 facilitates this kind of research. 

 Sections across the middle part 

 of the current, somewhere be- 

 tween theShetlands and Faeroes, 

 should be resorted to only as a 

 secondary means of evidence. 

 As long as this evidence can- 

 not be founded on rational 

 dynamic principles, or on exact 

 current measurements, an est- 

 imate of a current's intensity 

 from the diameter of the water- 

 layer is to be erroneous. In 

 one unit of time the smaller 

 and more rapid current may 

 carry a greater mass of water 

 through a given section than 

 a greater current of less veloc- 

 ity. Far more reliable results 

 can be expected from an in- 

 vestigation of the origin of 

 the current itself (in this case 

 the Gulfstream-circulation proper of the Trades), that is, if the pulsations of this area are 

 found to be in concordance with the variations of its remotest current-branches. 



The courses of the Atlantic current branches in the Barents Sea represented in the Barents Sea 

 current chart on this page, are determined by the troughs and deep channels of the sea bottom. 



Fi2 



Current chart of the Barents Sea [alter Knipowitsch a. Breitfu,ss] 



= Atlantic current branches 

 = = = = — — — as undercurrents 



— I^^ Cold arctic current 



