328 Profs. J. A. Fleming and J. Dewar. On the Dielectric 



observers who have examined the dielectric constant of this sub- 

 stance at temperatures above 0° C. have found for glycerine a negative 

 coefficient of temperature variation. Thus W. Cassie (' Roy. Soc. 

 Proc.,' vol. 46, 1889 ; also ' Phil. Trans.,' A, 1890, vol. 181) finds for 

 glycerine between 18° C. and 41° C, a negative temperature coefficient 

 equal to —0*006. Hence within this range, increase of temperature 

 decreases the dielectric constant of glycerine. C. B. Thwing ('Zeit. 

 f iir Phys. Chem.,' vol. 14, p. 293) gives the value 56*2 for the dielectric 

 constant of glycerine at ordinary temperatures. Since at low tem- 

 peratures we find a positive coefficient, and much smaller values, it 

 is clear that glycerine, like water, must have a maximum value for 

 its dielectric constant at a certain temperature. It is very important 

 that this point should be examined more fully. Up to the present 

 time electrical theory has offered no complete explanation of how it is 

 that change of temperature has such profound effect in modifying the 

 dielectric quality of certain kinds of matter. Ice, for instance, taken 

 at — 200° has about 2*4 times the dielectric power of the electro- 

 magnetic medium alone. At — 60° from 50 to 100 times the power. 

 At the absolute zero of temperature perhaps about twice the power. 

 On the other hand, we may carry such a bod}'' as ethylene dibromide 

 over the same range of temperature, without making any very 

 sensible change in the degree to which the substance affects the 

 electrical qualities of the space in which it exists. 



It is evidently, however, very closely a chemical question. Com- 

 paring the variation of many other physical qualities, such as density, 

 refractive index, thermal conductivity, &c, with dielectric constants, 

 we see that the variations which occur in the latter by changes of 

 temperature and changes of substance are much more akin to the 

 wide differences which occur between bodies in chemical activity. 

 The presence of certain chemical radicles, such as OH, CO, COH, 

 JTO 2 , has been shown by C. B. Thwing (' Zeit. fur Phys. Chem.,' vol. 

 14) to be a most potent factor in determining high values for the 

 dielectric constant at ordinary temperature, but the remarkable 

 behaviour of an aqueous 5 per cent, solution of caustic potash has 

 shown us that questions of concentration and the nature of the 

 solvent have to be considered as well, whilst the behaviour of all 

 electrolytes at the very low temperature indicates that decrease in 

 temperature destroys the influence of the radicle in this respect. 



One of the most interesting points is the evidence which now 

 oxists that water, glycerine, ethyl alcohol, and no doubt many other 

 bodies, have a maximum value for their dielectric constants at certain 

 temperatures. 



Thus W. Cassie ('Phil. Trans.,' A., 1890, vol. 181, p. 16) has 

 shown that the curves representing the variation of the dielectric 

 constant of glycerine, carbon disulphide, olive oil, benzol and tur- 



