20 Messrs. E. B. Rosa and A. W. Smith on a Resonance 



electric capacity, it also had only one -twelfth the capacity 

 for heat of the other condenser ; and since the temperature 

 rose five times more slowly, therefore that the percentage loss 

 of energy was only one-sixtieth as great, that is, £$ of 3 %, 

 or 0*05 °/ , giving his condenser an efficiency of 99*95 °/ ! 

 The reasoning is, however, quite inconclusive, for nothing is 

 said concerning the thickness of the paraffin-paper dielectric 

 and the volume of the condenser. Suppose the dielectric of 

 the smaller condenser to be 0'0129 inch thick, which is three 

 times the thickness of the other. (It might have been as 

 thick as that, seeing it sustained for a considerable period an 

 alternating voltage of 3000.) Then the volume per unit of 

 capacity would be nine times as great as the other, and the 

 heat required to raise its temperature as rapidly would be 

 nine times as great, assuming the capacity for heat and rate 

 of radiation the same for both. The loss would then have 

 been 0*45 % instead of 005 %. This illustrates how entirely 

 inconclusive any determination merely of rise of temperature 

 is, unless all such circumstances as heat capacity, rate of 

 cooling, thickness of dielectric (or intensity of the electro- 

 static induction) are specified. 



The enormous discrepancy between the results above re- 

 ferred to as to the quantity of the heating effect in condensers, 

 and the almost entire lack of precise statements as to its 

 numerical value, led us to undertake, more than a year ago, 

 to measure this energy loss in such a way that it could be 

 expressed absolutely. 



We proposed to measure by means of a wattmeter the 

 energy dissipated in a condenser when it is subjected to an 

 alternating electromotive force. In order that the frequency 

 of charge and discharge be perfectly definite, the electromotive 

 force should be a simple harmonic one, that is, the upper har- 

 monics of the fundamental should be absent. This is most 

 easily effected by inserting in the circuit in series with the 

 condenser a coil of wire having large self-induction, but 

 without an iron core. The variable permeability of the iron 

 will give rise to upper harmonics, especially if the magnetic 

 induction of the core attains large values ; and hence a coil 

 without an iron core is necessary. Moreover, if the self- 

 induction is not large enough it will reinforce some of the 

 upper harmonics instead of quenching them; and the presence 

 of the coil will be detrimental rather than beneficial. The 

 best value of the self-induction of the coil is such that the 

 fundamental is reinforced to a maximum degree; in other 

 words, it is that value of L given by the equation 



£=2ttx/LC; 



