156 A. TAXAKADATE AND H. NAGAOKA. "^ 



those regions, so that a Daniel 1 element prepared in a small bottle 

 proves suflicient to give the re({iiisite field. Its internal resistance 

 varies between 20 and 60 ohms., according to the strength of the 

 solution in the Daniell cell. In order to adjust the strength of the 

 current for use, a small resistance box consisting of 1, 2, 4, 8, 16, 32 

 ohms was put in the circuit. This arrangement of resistances gives 

 the fewest number of plugs with which the resistance can be adjusted, 

 within one ohm, from 1 up to the total sum of resisümces in the box. 



The method of observing the declination was the same as before 

 described ; that is, by finding the two positions of the coil where its 

 mao-netic axis coincides with the direction of the earth's field, in one 

 of which the pivots are reversed after the usual fashion of collimating 

 telescopes. The mean of the two azimuths is taken to be the 

 position of the Y'« of the theodolite in magnetic E.-W., corresponding 

 to the instant at the middle of two observations ; and the difference 

 as the magnetic collimation of tlie coil. This magnetic collimation 

 should be constant if the instrument remains perfectly unaltered in 

 shape. The collimation varied from J' to 1' on different occasions. 

 We first thought this difference to be personal, but after some num- 

 ber of triîds, it was found that the same collimation was obtained 

 by different observers when they worked in close succession, and that 

 the change was gradual and in the instrument. It might be supposed 

 that this fluctuation in the collimation is due to variation of the 

 declination during the course of observation. ]>ut the process of 

 observation takes only tvvo or three minutes, in no case more than 

 five minutes, and since the greatest change of declination is only 

 about two minutes per hour, this change cannot be the cause. The 

 fluctuation is due probably to temperature variation of the coil. 



The position of the circle corresponding to the astronomical 

 meridian was observed, together with the time; and was calculated, 



