MAGNETIC SUEVEY OF JAPAN FOR THE EPOCH 1895.0. 163 

 >■= /4 -^-— -?-____= .844 K 



or a depth of about 91>2 kilom. if the point be an epicentei'. 



^^' '' — a/ '•> • , - 1 or? — o,^ o-, =1.453 E 



^ 3 sm (.j4.°7— 20.°.)) 



or a height of about 2890 kih^m. if the point be a subcenter. 



Taking tlie latter vahie, Ave are wholly within the nodal 

 circle of no vertical force which will be about 46.°5 from the 

 pericenter, and therefore the horizontal force H' must be of the 

 same sign as its vertical variation, that is, the upward decrease 

 of the west components found in the variational anomalies must 

 be looked upon as upward increase of eastward forces diverging 

 from the pericenter, and consequently the image must be positive. 

 This contradicts however the observed positive values of the varia- 

 tions of the vertical component on the west and its negative value 

 on the east. The image must therefore be below the surface and 

 negative in sign, and the horizontal force H' must be considered 

 as converging toward the point, that is negative calculated in 

 the sense of increasing 0, and its vertical variation positive ; or 

 practically there nuist be an upward increase of eastward force, 

 or decrease of westward force which is just what is found. 



We may next find the zenith distances c of the forces from 

 (o) by eliminating 7n and /', thus, 



2 (. p èVl + i/ 3 



2) being the ratio — g— /-^7-, the sign of the second term and 

 the value of n depending upon r and 6. Confining the inverse 

 sines within the first quadrant, these are 



