204 



KNOTT AND TAXAKADATE 



eiices between the observed and calculated valuer? of the Total Force. 



Running our eye down the difference colunnis in the Dip and 

 Horizontal Force Tables, we are struck by a general tendency tor 

 the dirterences at any one h^cality toha\e opposite signs. That is to 

 say, where the observed value of the dip is greater than the calculated 

 value, the observed value of the horizontal force is, in the majority 

 of cases, less than the calculated value. More particularly, (^f the 

 81 Stations only 2 G are characterised by having their dip and hori- 

 zontal force differences of the same sign ; or, if the selected stations 

 are ahjiie considered, of these 50 only 17 are similarly characterised. 

 It would thus appear that the magnetic disturbances in Japan are 

 of a nature to affect the direction rather than the amount of the total 

 force — a result cpiite in accordance with the usual laws of magnetic 

 action. In this connection there are two very interesting cases that 

 seem to call for special remark. To bring out their peculiarities the 

 m(jre distinctly, it is advisable to draw up in tabular form the 

 differences (jnly fjr the stations that are to be discussed. They are 

 as follows : — 



These form })airs of contigU(3Us })oints. In all four the Dip and 

 Horizontal Force differences are exceptionally large ; whereas the 

 Total Force Differences are all distinctly smaller than the mean proba- 

 ble error for the whole. Here we have evidence in both cases of a 



