14 ROYAL SOCIETY OF CANADA 



The needle was magnetized, placed at the centre of the circle and 

 the deflection of the magnetometer noted. When the discharge parsed 

 round the circle there was a fall of deflection due to the partial demag- 

 netization of the needle. The detector was then removed, magnetized 

 again to saturation and replaced. The direction of the discharge was 

 reversed and the fall of deflection again observed. One of these direc- 

 tions is such that the magnetizing force of the first, third, fifth, etc., 

 half-oscillations is in opposition to the magnetization of the needle, 

 while in the other direction it is the second, fourth, sixth, etc., half- 

 c«cillations which demagnetize the needle. Since the magnetic force 

 on the needle is proportional to the length of arc traversed by the dis- 

 charge, if the arc in one case is adjusted so that it will give the same 

 deflection as that given in the other, then the ratio of the smaller arc 

 to the greater, will be the ratio the second half-oscillation bears to the 

 first. 



The circuit on which most of the observations were made was 

 rectangular in shape, measuring 63 cm. by 35 cm. The discharge wires 

 were of copper Oo mm. in diameter, and the air-break varied in length 

 from one to ten millimetres. A large Wimshurst supplied the current. 



The damping of the oscillations was found to increase rapidly and 

 steadily with the length of the spark-gap. The ohmic resistance cor- 

 responding to the absorption of energy by the air-break was found by 

 inserting in the circuit a known electrolytic resistance. A solution of 

 sulphate of ziuc with zinc electrodes proved most suitable for the pur- 

 pose. The damping due to the resistance in the air-break and the 

 known resistance was observed and a comparison of this with the damp- 

 ing due to the air-break alone, gave the resistance of the latter. By 

 varying this electrolytic resistance, a curve was obtained, giving the 

 resistance for all values of the damping. 



When the capacity of the circuit was altered by var}'ing the number 

 of Leyden jars, the inductance and spark-gap remaining the same, the 

 damping was observed to vary as the square root of the capacity. 



When the capacity and spark-gap were kept constant and the 

 inductance varied by changing the length of the circuit, the damping 

 was found to decrease steadily as the inductance increased. 



The copper wires of the circuit were replaced by iron wires of the 

 same diameter and the damping considerably increased. That this 

 was not due only to the greater resistance of the iron wire was proved 

 by the fact that copper wire of smaller diameter than that usually 

 employed gave no increase whatever in the damping. The currents of 

 high frequency magnetize the iron, they thus penetrate less deeply into 

 the metal and meet with greater resistance. 



