Oscillations in the Discharge of a Leyden-ji 



95 



zation ; therefore the ratio 72/71, i. e. the damping, will appear 

 greater than it really is. The effect decreases steadily with 

 the decrease in the diameter of the wire, and with very fine 

 wires it is almost negligible. The detector used in all the 

 experiments was composed of 55 hard steel wires '0015 cm. 

 in diameter, very carefully insulated and made up into a 

 compound magnet 1*5 cm. in length. With such a needle, 

 in a circuit where from theoretical considerations the damping- 

 could be shown to be very small, a damping of about 2 per cent, 

 could be measured, so that any correction due to magnetic 

 shielding in such a case must have been trifling. 



The following arrangement, which is theoretically equivalent 

 to that of the two solenoids described above, was adopted in 

 practice. 



A strip of brass was taken and bent into an almost complete 

 circle (see fig. 1) which was fixed on a block of ebonite. At 

 the centre of the circle an ebonite tube projected which served 



Fiar. 1. 



as the axis of a metal arm which pressed against the circum- 

 ference of the circle, and could be moved about it. The 

 detector was fixed in the end of a glass tube which could be 

 easily slipped in and out of the central ebonite tube. A scale 

 dividing the circle into 120 divisions was placed about its 

 circumference, and the whole fixed in position before a mirror- 

 magnetometer. The circle was placed in series with the 

 discharge-circuit, one wire being connected to one extremity 

 of the circle and the other to the movable arm, so that any 

 desired portion of the circumference might be included in the 

 circuit (fig. 1). This arrangement is practically a form of 

 the tangent-galvanometer with a single coil, the intensity of 



