438 BIRKELAND. THE NORWEGIAN AURORA POLARIS EXPEDITION, IQO2 1903. 



If we look at these charts, the great accordance with the observed areas of perturbation is at once 

 apparent. 



We have finally made a direct comparison with one of the observed elementary storms (see pag. 437). 

 We have here placed our current-system with its storm-centre in the neighbourhood of Dyrafjord, 6 = j n 

 the point of intersection of the earth's magnetic axis with the surface of the earth, and we have employed 

 the system with the shortest horizontal piece of current, /\ = 75. With this arrangement this will come 

 very nearly along the auroral zone. The projection of the assumed current-system is indicated on the 

 chart by a dotted line. 



For this system the magnetic force-components are then calculated for the stations from which we 

 have observations. The agreement, as will be seen, is striking as regards the horizontal current-arrows, 

 except that the current-system employed seems to be a trifle too large. If we had taken ^tt a little 

 smaller, or if the storm-centre had been chosen somewhat more westerly, the agreement would unquestion- 

 ably have been still closer. In the vertical forces the arrow observed at Val Joyeux is considerably 

 smaller than might be expected from the calculations. The direction is the same, however, in both 

 cases. The cause of this is to be looked for partly in the fact that the constitution of the actual current- 

 system must only with a very rough approximation be assumed to be capable of being replaced by such 

 a system, and that the actual current-system might not possess such a strongly-marked horizontal com- 

 ponent as is here assumed. Perhaps the agreement would have been better also in the vertical intensity 

 if we had used a form of the current-system analogous to that given diagramatically in fig. 187. We 

 believe, moreover, that much of the disagreement may be due to earth-currents, 

 which would have the effect of increasing the horizontal magnetic force-components, 

 while reducing the vertical. It is possible that these currents played the most im- 

 portant part. We must further draw attention to the uncertainty that may be con- 

 nected with the observed values of P v . We see this with special distinctness in 

 Charts III and VII X for the I5th February, in which there seem to be powerful 

 perturbing forces in the vertical intensity at Uccle, while at the surrounding stations 



Val Joyeux, Wilhelmshaven and Munich no particularly noticeable effect is found. The uncertainty 

 in the determination of the normal line is, as will be understood, rather great. 



At Axeleen the observed vertical arrow is considerably greater than the calculated. This may 

 only be due to the great uncertainty which attatches to the statement of the scale value for Fat this place. 



