Measurement of Electromagnetic Radiation. 51 



intervals. When, however, a torrent of sparks was produced, 

 the glass surface would take up a screening charge equivalent 

 to the mean charge on the oscillator-ends ; then, when the 

 oscillator-ends had charges either above or below the mean 

 charge, the statical action on the resonators would be the same 

 as that due to the excess or defect on the oscillator, so that the 

 steady repulsion observed was due to induction alone, and had 

 nothing whatever to do with electrical oscillations proper. 



In order to perfectly screen off this induction effect, and 

 yet leave the action of the true electrical oscillations, it was 

 necessary to so increase the conductivity of the surface of the 

 glass jar that it would have time to acquire a screening charge 

 which should follow variations of potential accompanying the 

 sparks at the rate of, say, 10 or 20 a second, and yet leave it 

 so imperfect as to be practically an insulator to oscillations at 

 the rate of 500,000,000 a second, which corresponded to the 

 apparatus in use. 



Prof. J. J. Thomson has shown* that the conductivity of 

 liquids may be determined by finding what thickness of a sheet 

 of liquid is just sufficient to prevent oscillations of a known 

 rate from exciting a resonator on the other side, and that the 

 thickness must be proportional to the specific resistance. 

 For instance, he found that a layer of dilute sulphuric acid 

 three or four millimetres thick was just opaque to oscillations 

 at the rate of 100,000,000 a second. It would therefore 

 require a greater thickness to be opaque to oscillations of 

 500,000,000 a second, and a very much greater thickness for 

 water to be opaque to oscillations of this rate. A duster, 

 therefore, soaked in water, though it might reflect some 

 fraction of this radiation, should be practically transparent to 

 it. Nevertheless a wet duster would be amply sufficient to 

 screen the action of true harmonic variation of potential of 

 the frequency of the sparks, i.e. 10 or 20 a second. As a 

 matter of fact, these variations of potential are not simply 

 harmonic, but since by Fourier's principle the non-harmonic 

 periodic change is equivalent to a series of true harmonic 

 changes at successively higher rates all superposed, the wet 

 duster should completely screen all the lower of these com- 

 ponents up to a certain point, and probably far enough. On 

 wrapping up the jar with a wet duster all effect was found to 

 cease, and the spurious character of the deflexions was de- 

 monstrated. It is possible that these are not the only experi- 

 ments with electrical oscillations in which a wet duster or its 

 equivalent might be found of service. 



The apparatus was then remade with the view of obtaining 



* Proc. Roy. Soc. vol. xlv. p. 269. 

 E2 



