320 BIRKELAND. THK NORWEGIAN AURORA POLARIS EXPEDITION, igO2 1903. 



The method employed in the working-up, is exactly analogous to that used with the observations 

 from 1902 and 1903, except that here, instead of registerings we have readings of the magnetic elements 

 for every fifth minute. 



The variation in these elements, in the case of a number of stations, is represented graphically in 

 the respective publications, and for stations where this is not already done, we have ourselves drawn 

 the magnetic variation curves. The same hour-length is employed throughout, namely, 15 mm. per hour, 

 whereas the scale for the deviations varies somewhat from place to place, according to the amplitude of 

 the oscillations. 



These curves are placed under one another in plates, thereby affording a clear view of the course 

 of the perturbations from station to station. These plates in reduced size will be found at the end of 

 this section. Further, the perturbing forces are calculated for a series of points of time, these beinj. 

 represented on a polar chart by current-arrows in precisely the same manner as before. In this con- 

 nection, however, it should be remarked that in calculating P d , it is the value of H existing at the 

 moment, that has to be employed but during the powerful storms this value may vary so considerably 

 that the same value of H cannot be used for the whole perturbation, and a correction must be intrc 

 duced. This correction is always evident during the powerful storms that take place in the regions here 

 studied. Z./, is given in the plates for the value of H, which answers to the normal line. During 

 powerful storms in H, therefore, direct use cannot be made of this, if fairly great accuracy is desired; 

 but as a rule the error will not be very great. For this reason, the values of Pd that we find at Fort 

 Conger during powerful storms will be somewhat uncertain, as we there have only absolute observations 

 of H to go upon. This is of little significance, however, in our studies. 



The scale of the arrows on the charts is given, and, as will be seen, is generally about five times 

 that employed in the previous observations. By this means the current-arrows at the polar stations are 

 of a suitable size, while at the southern stations another scale must be used. This is indicated by thtre 

 adding |, f , and so on, which means that the scale employed is f , f , etc. of the general given one. 

 This is a reversal of our former plan of introducing the factors , f , etc. at the polar stations in order 

 to indicate the local scale there in each case. 



On most of the charts here, moreover, there are several sets of current-arrows for one series of 

 generally as many as three different points of time. Instead of vertical arrows, which are found upon 

 the charts on which only one point of time is marked, a little table is here placed beside the station, 

 giving the corresponding values of P,. A similar table is given for P^ at Fort Conger, where only 

 term-days observations of the declination were carried out. Further, the magnetic meridian of that place 

 is indicated, and an idea is thus obtained of the direction, and to some extent of the magnitude of the 

 perturbing force at the various times. A powerful westerly-directed perturbing force in D thus corres- 

 ponds to a current-arrow directed westwards, more or less NW or SW, according as the perturbing 

 force in H might be more or less powerful, positive or negative. It will be seen that the magnetic 

 meridian and the geographical meridian at this place are nearly at right angles to one another, so that 

 a westerly-directed perturbing force as regards the magnetic meridian, answers to a perturbing force 

 directed southwards. If, on the other hand, Pj is only small, there is either, if PI, is also small, only a 

 small current-arrow, or, if PI, is fairly large, a current-arrow directed northwards or southwards. In this 

 way it is possible to make use of these observations. 



For the calculation of the perturbing forces, it is necessary to have a more or less accurate know- 

 ledge of the diurnal variation. By the diurnal variation must be understood the variation that there 

 would have been in the magnetic elements, if there had been no perturbations, in other words, if the 

 day had been a 'quiet day'. The diurnal variation, however, in the case of certain stations, has been 

 calculated as the mean of all the observations in a certain space of time. The results found therefrom, 



