PART III. EARTH CURRFXTS AND KARTH MAGNKTISM. CHAP. 



747 



We know that in a number of cases the magnetometer oscillations are direct effects from earth- 

 currents underneath the magnets, and for these oscillations, at any rate when their beginning is abrupt 

 rind well marked, we should expect to find simultaneity within the limits of experimental errors, because 

 the delay caused by the periods of the apparatuses can only be a question of seconds. 



To be clear of this question, we must have recourse to our rapid registerings. Of these we have 

 i gvi -at number, but as, for this purpose, the occurrence of especially characteristic serrations is required, 

 .here are not very many that are of use to us. We find a number of these reproduced in Series III. 

 We have taken a number of the most characteristic notches on these curves, and the time-differences 

 "mind between deflections in the earth-currents and the declination are given in the following table. 



TABLE CX1V. 



