MAGNETIC SURVEY OF JAPAN. 



217 



For all the Stations. 



AYhere two ditlerent values for tlie mean probable error are given, 

 the second is calculated after those stations, which are characterised by 

 veri/ large differences between the observed and calculated values, 

 have been thrown out. Thus in group I, Nos. 39 and 41 are neg- 

 lected in the calculation of the second values for the I)i[) and 

 Horizontal Force ; No. 38 for the Total Force ; and Nos. 41 and 64 

 for the Declination. In Group XL there were no stations prominent 

 by the relative hugeness of their differences. In Group III, No. 73 is 

 neglected for the Dip and Horizontal Force ; Nos. 11 and 18 for the 

 Total Force: atid No. 10 for the Declination. 



For the Fifty Stations only. 



The general conclusion to be deduced from both these tables is 

 that the mean probable error is greater the more volcanic the region. 



