Oscillations in the Discharge of a Let/den-jar. 93- 



cases ; but so complicated are the arrangements for these 

 photographic experiments that no great range of observations 

 could be made ; and this is very necessary in order to make- 

 any reliable generalizations from so variable a phenomenon 

 as the spark under consideration*. 



If two small oppositely-wound solenoids A and B are placed 

 in the discharge-circuit of a leyden-jar and two exactly similar 

 steel needles, both magnetized to saturation, be placed one in 

 each, with their north poles facing in the same direction, them 

 on the passage of a discharge it will be found that the re- 

 duction of the magnetic moment is not the same in the two 

 cases. Let a 1? 2 ? <*3> & c -> De the half-oscillations of the 

 discharge in one direction, and ft u ft 2 , ft 3 , &c, the half- 

 oscillations in the other direction. Suppose that a ± tends to 

 magnetize the needle in A still further, no effect is produced, 

 since it is magnetized to saturation, 8 X demagnetizes the 

 surface-layer, a 2 tends to remove the effects of ft l , and so on. 

 In the solenoid B, a x demagnetizes the needle, ft x tends 

 to remagnetize it, a 2 to undo the effects of ft, and so on. 

 Since the maximum value of the current a Y is greater than 

 that of fti, the needle in B will be more demagnetized than 

 that in A. If, however, the number of turns per centimetre 

 on the solenoid A is increased until the effects on the two 

 needles are exactly the same : then, assuming that the value 

 of the current decreases in geometrical progression, the 

 maximum value of the magnetic force due to the oscillation 

 ft x acting on A is equal to the maximum value due to a x acting 

 on B. Then if y u y 2 are the maximum values of the first 

 and second half-oscillations respectively, and n x and n 2 the 

 number of turns on A and B ; then, since 



47m 1 72=47m 2 7 1 , 



72 _ "2 



The ratio of the second to the first half-oscillation is therefore 

 known, and the damping determined. 



The method of two solenoids was not adopted in practice; 

 but one theoretically equivalent was employed, in which it 



* Since finishing this investigation my attention has been drawn 

 to a paper by Cardacci (Physikalische Zeitschrift, i. p. 23) on the resist- 

 ance of the electric spark. The value there given for the ohmic resistance 

 of the spark corresponds very closely, as nearly as I can judge by the 

 data given, with that which I have obtained; but his results show an 

 increase of resistance with decrease of pressure in air: while all my 

 results in air, hydrogen, and carbonic acid gas show a decrease of resist- 

 ance with decrease of pressure. 



