niRKELAND. THE NORWEGIAN AURORA POLARIS EXPEDITION, IQO2 1903. 



S I = 9 P 4- S 1 " 



H ft \ W 



Further we shall introduce a quantity representing the difference between the positive and the 



negative storminess 



i 

 1 C 



', dt 



For Declination and Vertical Intensity similar expressions are defined. 



Finally, by the Total storminess in the same interval of time we are to understand the quantity. 



S T = i\S 



In accordance with the definition of storminess here given the positive and negative storminess 

 corresponds respectively to a positive and negative direction of the perturbing force. 



There are two problems which will be dealt with in the following pages and form the main ol>jm 

 of our investigation, viz: 



(1) The total storminess as a function of time. 



(2) The distribution of disturbances in magnitude and direction at the different hours of the day or 

 the possible diurnal variation of the storminess. 



For the practical carrying out of the calculation, the following mode of procedure has been adopted: 

 The storminess was calculated for each period of two hours, for all three components, and the positive 

 and negative storminess were taken out separately. The numbers for one day were placed in the same 

 horizontal line as shown in the first series of tables. 



For each five-day period, the mean was taken of all numbers in the vertical columns corresponding 

 to the same hour-interval. This gave a horizontal line containing the distribution of the storminess at the 

 various hours for a period of five days. Taking the mean of the positive and negative storminess in 

 this horizontal line, we obtain the positive and negative storminess S 1 ' and S" for a period of five days, 

 and their sum (S' 1 -)- 5") gives the absolute storminess for the same period. 



We think it of considerable importance that a continuous record should be given of the occurrence 

 of magnetic perturbations during the whole period of observation, believing that such a record will ^i v( 

 an idea of how far we have succeeded in the first part of this work in treating the most important of the 

 perturbations. 



For our present purpose, however, it is the average values, that mostly concern us; and we have 

 therefore decided on publishing the following separate tables: 



FIRST SERIES. 



Tables for the continuous two-hourly records. These tables will be divided into groups of five days, 

 corresponding to each five-clay period. The numbers will be expressed in arbitrary units, which will 

 differ for the three components, but will be the same for all four stations. The factors of transformation 

 into absolute units will be given for each component. 



SECOND SERIES. 



Tables giving the distribution of storminess at the different hours of the day for each five-day period. 

 The periods will be divided into groups of 6. The mean value for each vertical column is taken for each 

 group and placed in a horizontal line, thus giving the distribution of storminess for 30 days, which will 

 be taken as one month. 



Finally, we find the mean distribution for the whole period of our observations. For each station 

 there will be three tables, one for each component; and the numbers will be expressed in absolute 

 magnetic units. 



