1878.] 



On the Viscosity of Dielectrics, 



243 



really increase by electrification. Besides this there is, of course, 

 some of the absorbed energy which is recoverable, and which, there- 

 fore, must not be confounded with conduction, just as the energy 

 recoverable from a deflected beam must be distinguished from that 

 lost through conversion into heat on account of internal friction. 



True conduction in a condenser, that is, the amount flowing in after 

 much time has elapsed, and which is entirely lost as electric energy, 

 however difficult it may be to measure with accuracy, is perhaps the 

 only property of dielectrics having no reference to exact time, and, 

 therefore, enabling comparison to be made between different dielectrics ; 

 that is, when the flow into a dielectric is steady this flow may possibly 

 be proportional to the electromotive force. We say advisedly " may 

 possibly," because we have reasons for thinking our ideas of stress and 

 strain will not only throw light on the questions of absorption pre- 

 viously referred to, but will in addition suggest an explanation of the 

 hitherto apparent inconsistency between the laws for change of re- 

 sistance by heat in conductors and in insulators. If that be so, then 

 we think it may be possible that we shall find out that Ohm's law is 

 always but an approximation, and that this approximation is the 

 nearer and nearer the truth the less the resistance of the conductor, 

 and that for insulators the current, after it has become steady, may 

 obey some such law as that experimentally determined by Mr. C. 

 Varley for conduction through rarefied gases. (See " Proceedings 

 Royal Society," October 5th, 1870.) 



True conduction is found to be extremely small after some weeks 

 for the flint glass in Sir W. Thomson's electrometers, care being taken 

 that the charge in the jar is left quite undisturbed by induction, from 

 electrified moving neighbouring bodies for example, whereas there is 

 much loss of charge due to conductivity as well as that due to absorp- 

 tion in the first few days after charging. Thus two electrometer jars 

 in our possession when left untouched for some weeks were found, if 

 the first measurement were made not much less than one week after 

 charging, to retain nearly the whole of their charge, the electrometers 

 being in constant use, but the replenisher not touched during the 

 entire time, to avoid alteration of internal distribution. Now when 

 the conductivity of glass is measured in any ordinary way a few hours 

 after charging, it seems to be very great even when we are able to 

 eliminate true absorption, which is all recoverable as electric energy. 

 Inattention to these considerations, we think, led some of the students 

 of Sir W. Thomson to infer that the true conductivity of .flint glass 

 was incomparably greater than it really is. 



In fact, we believe that up to the present time no experiments have 

 ever been made which determine the amount of the true conductivity 

 of flint glass, that is, the conductivity a long time after charging. 

 Since very hot solid glass is like pitch, a " truly viscous fluid," that is 



