544 PROF. C. G. KNOTT ON 



(1) The magnetic change rate of resistance of nickel increases steadily with 



increase of field, but at a somewhat slower rate as the field increases ; 



(2) The magnetic change rate increases unmistakably, with rise of temperature up 



to 100° C. and probably to higher temperatures ; 



(3) The thermal change rate is fairly constant throughout, tending to diminish a 



little in the lower fields. 



In the earlier set of experiments, in which the apparatus was heated in an air-bath, 

 the temperature was pushed up to 170° C, bej^ond which it was dangerous to go in case 

 the silk insulation should break down. As a matter of fact it was because of this 

 heating that the insulation did ultimately break down and prevented the experiments 

 in their later and improved form being carried up to even such a perfectly safe 

 temperature as 130°. These earlier experiments, however, gave the same general results 

 as the later, and indicate that up to 170° in the lower fields at least the magnetic change 

 goes on increasing. There was a good deal of uncertainty in these first experiments as 

 to the temperature of the nickel wires, especially in the higher fields, and the numbers 

 are not deemed sufficiently accurate to be placed alongside of the numbers obtained in 

 the later form of experiment. 



In the lower fields there was evidence of hysteresis, the change of resistance at break 

 of the magnetising current being somewhat less than at make ; but in fields above 

 thirty or forty there was no measurable hysteresis. As a general rule the field was 

 applied cyclically through several cycles in alternate directions before the observations 

 were taken. Occasionally, however, the magnetising current was applied and removed 

 and applied again in the same direction instead of being reversed ; that is to say, 

 instead of the cycle being ( + , 0, — ,0, + ) it was either ( + , 0, + ) or ( — ,0,—). In 

 such cases it was invariably found that the accompanying change of resistance was 

 greater under the cycle ( + , 0, -f ) or ( — , 0, — ) than, under the complete reversing 

 cycle ( + , 0, - , 0, + ). 



If, following Professor J. J. Thomson's theory of electric corpuscles, we try to get 

 at the significance of these results, we arrive at some interesting conclusions. These 

 negatively charged corpuscles of mass, very small compared to the mass of the neutral 

 molecule or of the positively charged cormolecule, are regarded as the agents which 

 carry the charge through the conducting material. The process is analogous to the 

 process of conduction of heat in gases according to the generally accepted kinetic theory ; 

 and Professor J. J. Thomson has given reasons for believing that the process is due to 

 the impacts of the corpuscles on the neutral molecules, the result of such a collision in 

 any particular case being the entanglement of the impinging corpuscle with the neutral 

 molecule and the instantaneous setting free of another corpuscle which moves on with 

 its charge to the next collision. The effect of an externally applied electric force is to 

 give direction to the motion of the momentarily dissociated corpuscles. The rate at 

 which the charges are carried, that is, the amount of electricity conducted, will depend 

 upon the velocities of the corpuscles and their free paths, and also upon their state of 





