MAGNETIC SURVEY OF JAPAN FOR THE EPOCH 1895.0. KU 



Meau 116°^ W. 34?o N. 



Those coordinates have very different weights depending npon 

 tlic angles at which the circles cross each other, and upon the 

 probable errors of the azimuths. Supposing the latter to increase 

 with the distance from the middle of the country, as already 

 discussed under the mean isomagnetics, it is taken to be 1 for 

 the point III, 2 for II and IV, and 3 for I and V ; and the 

 weight of the point of intersection of any two circles is taken 

 inversely proportional to the sum of squares of their azimuth 

 errors and directly as the square of the sine of the angle at 

 which they cross each other. The weights of the longitude and 

 latitude are resolved parts of the weights so found. 



There was no apprecial)le improvements by restoring the last 

 figures of those coefficients which were cut off in the table above 



* To draw tliose circles on the gh)be through the points, tlie rotation axis was inclined 

 to the horizon circle at an angle sin-i (cos -^ sin A) and liy rotating the globe the point 

 )s brought to tlie plane of the horizon circle which is then the circle required, 



