THE DISTURBANX'E OF ISOMAGNETICS. 



151 



with Mr. Imagawa made a survey, in which they tried to distrihute 

 the stations as uniformly as they could. They divided themselves 

 into two parties, indicated as 'south' and 'north,' and as the principal 

 work was done during the three months, 'Tuly, August, and Septem- 

 ber, the errors due to the secular and seasonal variations were but small. 

 The south party made observations of diurnal variations of declination 

 at every station except three. This, besides giving the mean value of 

 declination at the place, tells us the nature of the magnetic weather of 

 the da}'^ on which other magnetic elements were determined. They 

 made also three or four determinations of the horizontal intensity and 

 dip, the times of observation being distributed over the day. (For 

 details of this survey we refer to volume II of this Journal.) Had 

 we made another series of such observations, we could have allowed 

 for the secular variation in the results of the present survey, and 

 should have been able to tell more definitely the kind of disturbance 

 caused by the earthquake: but, as it is, our results will yet be of service 

 for determining the secular variation, when, in the future, another survey 

 shall be made, and then we may be able to draw inferences as to the effect 

 of the earthcjuake by allowing for the secular variation then ascertained. 



There was no particular division of work among the members of 

 the party ; generally speaking, however, the vibration and deflection 

 experiments were carried out b}^ the writers and either of the stu- 

 dents; the dip was determined almost exclusively by the students; and 

 the declination and time- observations, hoth for the chronometer rate 

 and for azimuth, were undertaken by the writers. 



The stations, as a simple matter of course, were those of the previous 

 survey in 1887. Although for guidance we had verbal descriptions of 

 the stations, and the memory of one of the party who had worked on 

 the previous occasion, the identification of places was not always 

 easy, as no fiducial m;irk of reference had been left. This was to be 



