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



it has a clearlymarked maximum, and for the stations to the south a gre- 

 ater and greater bend in the curve at noon. 



The arc of the great circle, b, between Zi-Ka-Wei and the point M 

 will be the difference between the latitude of these two places, viz. 30° 44'. 

 If we compare this with the corresponding value, calculated according to 

 formula (2), for each of the other 4 points through whose zenith the cur- 

 rent must be assumed to pass, we-- obtain 



b 



Mean 30° 59' 



This mean value is only 13' less than the latitude of Zi-Ka-Wei, In 

 other words, if the current be imagined as a complete circle lying in the 

 plane that is tangent to the earth in Zi-Ka- Wei's parallel, the southern- 

 most part of the circle, that part of the current that sets from W to E, 

 will at noon almost pass over the equator, and will thus approximately 

 intersect the line of connection between the sun and the earth at an 

 altitude above the latter of 1060 km., this being the value of /«, 'which, 

 by lormula (i) answers to <^ ^ 30° 59'. The mean radius of the current- 

 circle, Ç = /? tan b, will be 3825 km. 



A.s mean . time is used throughout this paper, and the equation of time 

 for the two days treated of is about + S"" , all the hours must be re- 

 duced by o-i hr., in order to obtain apparent time. 



The arrow to the sun in Figs. 4 and 5 points therefore in reality about 

 2° west of the apparent sun. 



VII. 



The results of the above special investigation may be summed up in 

 the following main points: 

 1. The normal diurnal variation of terrestrial magnetism at a place on 

 the northern hemisphere is dependent not only on the position of 

 the sun in relation to the horizon of that place, but also on the situ- 

 ation of the place in relation to the magnetic north pole. Correspon- 

 ding conditions probably also exist with regard to the southern hemis- 

 phere. 



