1878.] 



On the Viscosity of Dielectrics. 



241 



curve expressing the time increase of strain in a substance subjected 

 to a constant stress, and careful examination snows that all our results 

 up to the present time bear so close an analogy with the stress and 

 strain phenomena in viscous substances, that we feel that this analogy 

 means a physical connexion. We therefore propose, even before the 

 completion of the experiments, to show in the present paper the bear- 

 ing of this analogy; since not only does it constitute an extension of 

 Faraday's stress theory of induction, but, in addition, it forms a veri- 

 fication of the extension of our knowledge regarding stress and strain 

 phenomena, as afforded by our recent experiments on the viscosity of 

 substances. Besides this, the theory that we are about to sketch roughly, 

 and which we hope to establish in all its details by experiment, is 

 valuable in furnishing an easily conceived image of the internal action 

 taking place in a dielectric. 



A complete series of experiments will soon be ready for publication, 

 showing that almost every body which can be examined exhibits a 

 time increase of strain under a constant stress. Thus, when a small 

 couple, much less than that which is required to produce any perma- 

 nent set, is applied to twist a prism, there is a rapid production of 

 torsion due to a property of the material measured usually by a 

 number called the modulus of rigidity; the rapidity depending mainly 

 on the mass which has to be moved, and also depending on the viscous 

 yielding (or on something which is related to the viscous yielding) of 

 the material. If we could eliminate the effect of inertia, then not an 

 instantaneous but a very rapid production of torsion would be ob- 

 served, the production of strain along the prism not being instan- 

 taneous even at the beginning, but proceeding with a velocity com- 

 parable with the velocity of electric induction. After a very short 

 interval the torsion no longer increases rapidly, and only its amount 

 (D) at the end of the interval is what has usually been taken into con- 

 sideration by physicists and engineers. But it should be known that 

 it is only after a long time that the production of strain ceases, and 

 the additional torsion (d) in glass and steel fibres after some hours or 

 days in some of our experiments became equal to (^D) . If when the 

 production of strain appears to have ceased the twisting couple be 

 suddenly removed, then there is a rapid but not instantaneous reduc- 

 tion of strain (D), and an additional reduction (d) effected after a con- 

 siderable time. 



Thus for all strained substances that we have experimented on, the 

 equilibriated state is one in which there is a free strain or a strain that 

 is almost instantly removable, and there is also an absorbed strain 

 which is only slowly produced, and which is only slowly removable. 



Exactly in the same way, when a condenser is charged by a battery 

 of electromotive force (V), the charge at first increases rapidly, due to 

 a property of the condenser called its capacity (S) ; this rapidity 



