— 13 — 



tangent screw until the reflected image is brought back to where it was when 

 no current was passiug. The current is then reversed by turning the twin- 

 plug half-way round and its strength is adjusted, if necessary, by means 

 of the resistance coils so that the resultant magnetic field is not reversed. The 

 image will again be displaced on account of inefficiency of the previous 

 adjustment. The position of the coil is readjusted to zero deflection of the 

 image and the angle reading on the azimuth circle taken. The coil is now 

 hfted up carefully from the Y's and replaced in the reversed position, after 

 the usual fashion of collimating a telescope. The observations are repeated 

 with this new position of the coil, and the mean of the two readings is 

 taken as the mean magnetic bearing between the two times of setting. This 

 subtracted from the meridian reading of the azimuth circle gives the declina- 

 tion required. 



It will be seen that if we take the mean of the two observations thus 

 made, the value obtained will be free from what may be called the error of 

 magnetic coUimation, that is, the error arising from the axis of the pivots not 

 being strictly perpendicular to the direction of the electro-magnetic field. This 

 error is half the difference of the two observations, provided the configuration 

 of the coil and the declination remain constant throughout the whole series of 

 operations. AYlien the instrument was in order, the observations could be 

 made in about three minutes, and the so called magnetic colliraation was pretty 

 nearly constant, being about 7". This obviously gives a check on any accidental 

 mistake on the angle reading. 



The following curve shows the variation' of magnetic declination at 

 Ogasawarajima as determined in this way. 



i. y «■ • 

 .' r 



S'o' 

 H lis' 



I is 



■^ 



i 





" 



■' 







1 j 











— 1 



1 1 



1 ' 

























' \ 





^ 



\ 

















--- 



— - 



























-^-=^ 



L^ 



\ 































^ -^ 



./„ 





Y 



\ *" 



PN 





































/ 





A 



"\=fej 







































■ 



\/. 





-\- 





































/ 





pT5- 





/ ' ^ 

































/ 



/ 





r 





r\ i 1^ 

















" "x^ 







.._ J J_L/_ 





4-- 



\ 



^-^-x^ 





^ 



V A 







~''':^^^ 



I — ,U 



r- 



JuJ,: 



\. -J 





-3s=» 



-^ 



-q 



■\\ 



-Vj 



\y \ 





^T \ 







^^- 



-H^/ 













>l 



V 



^ 



1 



/ 





1 ' i 







\ 



-1^ , 



I ' 





\^ 





"-^ 



y 





' 



- 



1 







\>- 



-' 1 ! 



/ 







^ J 





<^ 



V 











i 1 



















i 



: ' ; i 1 i 



1 



i 



i 1 

















i 



s 



1 I 



> 



> / 



1 



/ 1 



2 



i 



. 



r 4 



i 



« 



1 



i 



t 



t 



n 



jj o r. a / 



■r. „. 



• This dip in the ciin-c looks like nil aceiJciital error in an^'le reailing. 



t Thi» is proljuljly due to a iliaiilaceiueiit of the base since a ilowiiwaril sliift of Uic succecJiiip 

 portion fiU in well with the rest of the curve. 



