MAGNETIC SURVEY OF JAPAN FOE THE EPOCH 1895.0. lo9 



If tlie poles be of unequal strength, the excess of the 

 stronger over the weaker only will be felt at great distances. 

 Such a magnet of an apparently odd pair of poles may be pro- 

 duced by local heterogeneity of magnetization in a body like 

 the earth, one pole being concentrated at a place and the other 

 widely spread out over the rest, when the effect of the former 

 can be represented by a single image for approximate purpose. 

 Tlie same will also be the case when the length of the magnet 

 is excessively long, compared with the region through which the 

 disturbance is considered. 



Turning to Table XXI p. 135, we observe that in Japan 

 there is a close coincidence of the vertical variations of the 

 north and upward components with — oh/R times the respective 

 forces, but we are struck by a remarkable discrepancy in those 

 of the west component. It stands so peculiar among those of the 

 other components that some arithmetical error was suspected, and 

 the calculations were examined repeatedly through, till a thought 

 occurj'ed that the effect might be due to the presence of the con- 

 tinent on the west ; the extent of that effect throughout the 

 whole country showing that the origin of the disturbance is likely 

 to be found at some distance. 



Assuming then that a large part of the disturbance can be 



represented by a simple magnetic image, we are much helped 



by the presence of a few points of demarkations in the country 



in locating its position. Takino- for the first approximation, 



. . / ,. , , . . 9X' OY' rW 



those variational anomalies to be the variations ^y^, ~ös~' ~dz~ 



given above for a simple source, we notice tliat the sign 



OX' 

 of -^c — chano'es in middle of the countrv between III and IV, 



OY' 

 while that of —^ — remains the same. Hence it seems that we 



oz 



are crossed ])y one of the nodal ellipses ; the fact is further 



