AKSEL S. STEEN. M.-N. Kl. 



The basis, then, or uniting link in a connection such as this, should 

 be electric currents in the atmosphere, whose presence would be recog- 

 nised by the aid of magnetic observations in the application of Ampere's 

 law, and which, on the other hand, might also be imagined to have an 

 influence, by change of direction and force, on the meteorological con- 

 ditions. 



If in the mean time we take for granted that such currents do actu- 

 ally exist, it is, in m}^ opinion, necessary first of all to examine more 

 closel}^ their connection with the terrestrial-magnetic phenomena. Here, 

 however, we are immediately met by the difficulty presented by the 

 rather complicated nature of the magnetic observations at our disposal. 

 A simultaneous observation of declination, horizontal intensity, and vertical 

 intensity (or inclination), gives us, it is true, an expression for the total 

 magnetic force at the moment of observation at the place in question; 

 but this expression^ as we know, is made up of four separate constituents, 

 namely : 



1. The normal distribution of magnetism over the earth's surface at a 

 certain epoch, 



2. The effect of the secular variation, 



3. The effect of the diurnal variation, and 



4. The effect of a possible magnetic perturbation. 



Of these four constituents, however, the first two together may be 

 regarded as a constant in the discussion of the present question; and it 

 will be the diurnal variation and the perturbations that must be treated 

 separately, the diurnal variation first, as its careful examination ought to 

 form the foundation for all more extensive investigations in this subject. 



The considerations here put forward by way of preface, have led 

 me into studies of which the first results are recorded in the present 

 paper. It forms a link in the chain of special operations organised by 

 the Norwegian Meteorological Institute, which have in view the possible 

 improvement of weather-forecasts. 



