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



In the case of the three other quantities, we have employed a somewhat different mode of procedure 

 in the calculation. For Dyrafjord they are found by the aid of the change in temperature that will 

 always take place during a determination of sensibility, and which can be determined by the temperature- 

 curves. In order to be sure that the displacement of the curve is due to the temperature, it must be 

 calm before and after. The diurnal variation must, moreover, be taken into consideration. The values 

 given are means of 10 determinations distributed over the various months. 



At the other three stations, a method has been employed by which we escape having to consider 

 the diurnal variation. Under normal conditions, the ordinates to the curve in points lying 24 hours from 

 one another -- provided the temperature is the same should be of the same length. The majority 

 of our magnetograms cover a period of 24 hours. 



We have now selected a series of registerings for the very calmest days with a great difference 

 in temperature between the beginning and the end. The required temperature coefficients are then 

 found from the differences between the ordinates to the terminal points and the difference in temperature. 

 This method is very suitable, as the temperature for Axeleen is read directly, at the beginning and end 

 of each magnetogram. At Matotchkin Schar, it is only at the beginning. At Kaafjord, on the other 

 hand, the temperature in the register-house was read only a few times in the course of the winter, and 

 there we have had to keep to the registered temperature-curve only. 



The values given in the table are in each case the mean of 12 such determinations. In the cal- 

 culation of the mean, we have assigned different weights to the determinations, according to the amount 

 of difference in temperature, and the calmness of the twenty-four hours. 



The temperature-coefficients of our registerings -- with the exception of those for Matotchkin 

 Schar -- are not inconsiderable; and as the temperature at these temporary stations undergoes great 

 changes, it has been necessary for us, in our calculations, in each case to direct our attention to its effect. 



EXPLANATION OF THE CHARTS. 



23. Our investigations of the distribution and course of the magnetic perturbations, divide, as already 

 mentioned, into two sections, the one embracing the whole earth, the other more especially confined to 

 the regions round the North Pole. 



We have here found it most practical to employ two different charts in the synoptic representation. 



For the universal part we have employed a map of the world on Mercator's projection. The 

 advantage of this projection is that it is orthomorphic, so that angles upon the earth can be marked 

 directly upon the chart. 



For the second section we have used a polar map in the equidistant zenithal-projection. This pro- 

 jection is not orthomorphic; but the angular deformation in the polar regions is very slight. For all 

 stations except that of Cape Thordsen we have, however, taken this deformation into account. As for 

 Cape Thordsen the deformation is less than the accuracy with which the angles can be determined. 



The previously explained current-arrows are marked on the maps, representing geometrically the 

 perturbing forces calculated for a particular point of time. The time is stated at the top of the map. 

 The length of the arrows is proportional in each chart to the perturbing forces. At the foot of the 

 chart a scale is marked, by means of which the magnitude of the perturbing force can easily be taken 

 directly from the chart. As the unit of magnetic force we have employed i y = io~ 5 absolute units. 



It has proved inexpedient to make all the arrows on one chart to the same scale, as the perturbing 

 forces at the northern stations are often more than 10 times as great as over the other parts of the 

 earth during the same period. 



