— 241 — 



been selected, which have well developed diagrams, clia- 

 racteristic for each of the different types. 



Accordingly Kingua Fjonl, Jan Mai/en^ Sodankifla, Greenwich, 

 Tipis, Zi Ka Wei, Batavia and Cai)e Hoorn have been chosen; 

 (Godfhaah has been left out the vertical component for that 

 station not being available). 



Firstly the hourly values of the components for these 

 stations have l)een smoothed graphically, because it was clear, 

 that the irregularities in those series had been caused chiefly 

 by want of material. From the smoothed curves representing 

 the values of Xs, y.s and Zs for the selected stations, the values 

 for the fuU hours of Göttingen mean time have been read. 



They are tabulated in Table 2. 



Now these values have been computed in most cases for 

 for the same couple of months, viz. .Tune — July, but not for the 

 same years. As it is probable that the diurnal inequality of 

 disturbing force increases and decreases together with that 

 foi-ce itself in the eleven year period, the figures are not 

 comparable directly. A factor redncing them to the same year 

 may be found as follows. 



In the "Observations made at the Royal Magiietical and 

 Meteorological Observatory at Batavia" 1809 part IT, table 69, 

 we find the mean horizontal disturl)ance (sh) per day and 

 from these values we may derive the formula 



Sh = 22 + 1.00 R 

 R being Wolf's sunspot-number. 



With this formula we find for the reducing factors: 



Greeturich 1.27 



Tiflis 1.27 



Batavia 1.17 



As these factoi-s differ not much from the unity and are 

 moreover approximations, and also as the components of these 

 stations themselves are small in comparison with arctic values, 

 no use has been made of these factois. 



The construct ion of the tield has been performed in the 

 following manner. 



LXIII. 16k 



