126 B1RKELAND. THE NORWEGIAN AURORA POLARIS EXPEDITION, 1902 1903. 



to the south-east of Dyrafjord, almost in the south-east of Iceland. The point of convergence must lie 

 somewhere in the regions between Iceland and Stonyhurst, probably nearer the latter station, as the 

 force there is so small. 



Taking for granted the point of convergence, the horizontal force should first increase along 

 the transverse axis of the system from 0, and then slowly decrease; and we do indeed find that the 

 force increases from Stonyhurst towards Munich, Val Joyeux and San Fernando, and then becomes 

 smaller towards Tiflis and Dehra Dun. The change, moreover, corresponds fairly well with that which 

 we find in the two calculated systems in Table XV, where the horizontal part of the current lies at 

 a height of 300 km. 



The vertical components at Val Joyeux, Wilhelmshaven, Potsdam and Pawlowsk are all directed 

 upwards, just as we should expect. At Pola there was earlier a fairly considerable vertical component 

 directed upwards; but it now about equals 0. 



Chart IV. Time /'' o m . 



The arrows in Europe and Asia have continued to turn. In the United States also, the arrows 

 have now turned a little. The alteration in the field is fully explained by the assumption that our 

 current-system has moved a little farther in the direction from Dyrafjord to Axeleen. 



The directions of the arrows in Europe show that the point of convergence of the horizontal 

 components ought now to be found a little to the north of Pawlowsk. At Pawlowsk, as we should 

 expect from its lying near the point of convergence, PI, is exceedingly small, only 7.5 y; but on the 

 other hand P v = 14 y, and is directed upwards. 



At Tiflis and Potsdam, P, is directed upwards, but is rather small. At Wilhelmshaven, P t = o. 

 At Pola, a small force is directed downwards. 



It is in harmony with our assumption that we also find a larger horizontal force south of Pawlowsk 

 than at that station itself. It is greater even at Dehra Dun, Irkutsk and Zi-ka-wei. At the last-named 

 station, P t = o. It appears from the vertical forces at our stations, that the principal axis of our 

 system should lie to the south-east of Dyrafjord and Axeleen, as P, is there directed downwards. At 

 Kaafjord, however, we find P v directed upwards, which also indicates that the axis lies between the two 

 first-named stations and the latter. 



Chart V. Time i h 15. 



The current-arrows in the United States are turned so that their direction is now about west, 

 answering to a southward direction of PI,. 



In Europe, PI, is turned farther in the same direction, and is now directed eastwards. The field 

 during this period resembles that at the conclusion of the perturbation of the isth December, or those 

 of the 22nd March and loth February. 



At Pawlowsk there is still a considerable vertical component directed upwards. The point of 

 convergence should now have moved farther east. 



Chart VI. Time i k jo m . 



The distribution of force is as in the preceding chart, but the forces are much smaller. In the 

 case of the European stations, the turning is continued a little. 



During this great but gradual alteration in the outer field, the conditions at Dyrafjord and Axel- 

 een, notwithstanding small local irregularities, have remained very constant. At both stations the current- 

 arrows have been directed all the time south-west; and the vertical component all the time has been 

 directed downwards. At Kaafjord, on the contrary, the vertical force has been directed upwards all the 

 time, with the exception of a short time at about o h 28, and attains a magnitude of 209 y. 



