INDUCTION METHOD. 599 



The error may be greatly diminished by searching by trial for 

 the initial position of the frame in the second case, so that the 

 discharge is the same as in the first* More generally, the frame 

 being made movable about an axis perpendicular to the meridian, 

 two initial positions are determined, such that the deflection of the 

 ballistic galvanometer is the same as for a rotation of 180; the 

 bisection of this gives the direction of the inclinations. 



Any measurement of the discharge is eliminated, as well as the 

 regulation of the initial and final positions of the movable frame, 

 by determining the right line about which rotation should take 

 place so that no current is induced, f The apparatus is quite 

 similar to a dip needle ; the axis of rotation of the frame may 

 be displaced on a vertical circle which itself is movable about a 

 horizontal circle. 



By means of a milled head the frame may be turned 180 to the 

 right or left. The deflections observed in a ballistic galvanometer 

 enable us by a series of methodical trials to put the axis of rotation 

 first in the meridian, and then in the direction of the dip needle. 

 The accuracy of the observations is much increased, for when the 

 current is near null, we can combine the movements of the frame 

 with those of the ballistic galvanometer, and multiply the angles of 

 swing. 



The adjustment for the magnetic meridian is so accurate that 

 the apparatus can even serve as declination compass. 



The sensitiveness may moreover be increased by putting soft iron 

 in the frame, since the changes of magnetisation are null for any 

 given rotation about an axis parallel to the field. 



We may finally replace the ballistic galvanometer by a telephone, 

 giving the frame a continuous motion. The periodic currents pro- 

 duced in these conditions do not vary suddenly, and can with 

 difficulty be perceived in a telephone^ but they become more 

 perceptible when a make-and-break is introduced in the circuit 

 for instance, a toothed wheel. In this way, if the frame turns con- 

 tinuously, the telephone is only silent when the rotation is parallel 

 to the dipping needle, but without giving a sufficient approximation. 

 Better results are obtained by putting soft iron in the frame. 



* H. WILD. Bulletin de FAcad. des Sc. de St. Petersburg, Vol. xxvn., 

 p. 320. 1 88 1. 



t MASCART. Comptes rendzis, Vol. xcvu., p. 1191. 1883. 

 J J. STEPHAN. Wiener Sitzungsberichte, p. 262. 1880. 

 W. SCHAPER. Meteorolog.'Zeitschrift,vT*- 1886. 



