MAGNETIC SURVEY OF JAPAN FOE, THE EPOCH 1895.0. 15 



vertical. The reading of the azimuth thus found was compared 

 witli that found when the vibration experiment was carried. 

 The difference between those two kept nearly constant ; hence 

 when vibration experiment was done before the dip, the magnetic 

 meridian was obtained by applying this small difference. The 

 metliod was especially convenient when experiment was done in 

 open in " SyuUyö.^^ 



In reversing magnetism by means of the coils the current 

 was kept for ten seconds, at first the magnet was reversed with 

 regard to the coils and the current reversed to eliminate any 

 unsymmetry. It was found however in the course of the survey 

 that this double operation was quite superfluous. 



The Declination was observed in the same way as in the previ- 

 ous survey, the only difference consisted in taking four positions 

 of the coil in each set of observations, reversing the coil with 

 regard to north and south and again with regard to east and 

 west, although the mean of the last pair is usually sufficient. 



The Astro7iom(cal 3[eridimi was usually found by observation 

 of the Polaris, and then transit of six stars were taken with 

 the telescope reversed in Y's at the middle, each three being 

 north, zenith and south stars. From these observations azimuth 

 coUimation and clock errors were found by Mayer's method. 

 In setting the theodolite into the meridian, the graphical method 

 given by the writer (Sfigaku-Buturigakkwai Kizi Vol. VI p. 21) 

 was very handy as the azimuth could be found at a glance to 

 fraction of a minute. 



The attached specimen pages of field notes which were 

 entered in printed schedule will illustrate the process. 



Local Time and Chronometer Bate were determined at nearly 

 every station by taking observations in the evening and either 



