XX 



section of Hopkinson's original work. It was shown that a dielectric 

 which had been subjected to successive electromotive forces of opposite 

 polarity gave a corresponding inverted succession of residual dis- 

 charges, and these were treated in the manner of Boltzmann's theory 

 of the after-effects of mechanical strain. The specific inductive 

 capacity of many substances was measured in the earlier experiments 

 for periods ranging from 1/2 to 1/20,000 of a second, and was com- 

 pared, on Maxwell's theory, with the refractive index for long waves. 

 It was shown that in hydrocarbon oils the results were in agree- 

 ment with Maxwell's theory, but with other substances there was 

 generally disagreement. The last paper is of particular interest as 

 supplying a key to this discrepancy. It describes experiments on ice, 

 and on glass at various temperatures under high as well as low 

 frequencies of charge and discharge, the high frequency ranging from 

 2,500,000 down to 8,000. The capacity of ice was found to be of the 

 order 80, when measured by charges and discharges with a frequency 

 of 1/10 or 1/100 second, but to have a value less than 3 when the 

 frequency was such as one-millionth of a second. This showed that 

 the apparently excessive capacity was to be ascribed to residual 

 charge. Further, in the case of glass, a high temperature was found 

 to increase the apparent capacity for comparatively slow frequency of 

 discharge, but not for high frequency. Here, again, the difference is 

 due to residual charge. Again, the insulation of heated glass was 

 observed to be less after 1/50,000 second electrification than after 

 1/10,000 second, but to be sensibly constant for longer times of electri- 

 fication. The current which flows when electromotive force is applied 

 to a condenser is ordinarily treated as consisting of three parts, the 

 charge proper, the polarisation or residual charge, and the conduction 

 current due to imperfect insulation. Hopkinson pointed out the 

 arbitrary nature of this distinction and the real continuity of the 

 phenomena. These three terms, though separated for convenience, are 

 really parts of one continuous magnitude. In a dielectric which 

 exhibits residual charge and deviates from Maxwell's law, the action is 

 essentially the same in kind as that which is found in an ordinary 

 electrolyte. 



Before noticing the large section of Hopkinson's work which relates 

 to magnetism, mention should be made in passing of a short but 

 suggestive paper on the Hall effect (1880), where it is suggested that 

 the effect is completely expressed by Maxwell's " Rotatory Coefficient " 

 of resistance, and of another " On the Seat of Electromotive Force in 

 the Voltaic Cell " (1885), where it is pointed out that the controversy 

 between those who held the difference of potential between metals in 

 contact to have the value deduced from electrostatic experiments, and 

 those who held it to be measured by the Peltier effect was one of 

 definition and of hypotheses used for the expression of admitted facts. 



