116 BELL SYSTEM TECHNICAL JOURNAL 



care was taken to avoid capacity couplings. The measuring circuit 

 was shielded from the observer by a metal screen. In spite of these 

 precautions the results obtained on the same sample at different 

 times do not agree as well as might be desired although the individual 

 readings taken at the same time agree in most cases to within 5 per 

 cent. Other observers have found that measurements made on the 

 same sample at intervals of a few hours often differ by more than 

 the apparent error of the measurements and have attributed it to 

 actual changes in the properties of the material.'' Hence it is possible 

 that at least part of the apparent variation with frequency shown 

 above is due to unknown errors in the measurements or unknown 

 changes in the samples or both. 



As an illustration of the error involved in taking only phase differ- 

 ence as a measure of power loss, suppose we wish to compare hard 

 rubber having a phase difference of about 0.5 degree and a dielectric 

 constant of 3., with a certain grade of glass having a phase difference 

 of about 0.3 degree and a dielectric constant of 7. On the basis of 

 phase difference alone the hard rubber appears very much worse than 

 the glass, but when the dielectric constant is taken into account, the 

 glass is found to give a 40 per cent higher power loss. Similarly, some 

 untreated woods were found to have considerably lower losses than 

 the phenol fibres although their phase differences are nearly the same. 



Conclusion 



In the case of ordinary insulation where the object is to provide a 

 mechanical separator or support, the product of phase difference and 

 dielectric constant is a true measure of the energy loss per unit volume 

 as shown by the equation (1). In the case of a condenser where the 

 object is to obtain a given capacity, the phase difference alone de- 

 termines the power loss since in this case the effect of the increased 

 dielectric constant is exactly balanced by the smaller volume of 

 dielectric required. 



^ R. Mesing — L'Onde Electrique, April, 1922, p. 235 and Augustin Frigon — Comptes 

 Renins, May 22, 1922, p. 1339. 



