596 BELL SYSTEM TECHNICAL JOURNAL 



able agreement with the values deduced from the high-frequency 

 phenomena ! ^ 



(It is rather surprising to realize that in such discharges, we must 

 not conceive the electrons as moving with a steady velocity along the 

 axis of the tube, on which a sinusoidal oscillation parallel to the field 

 is superposed. If we could follow the wanderings of an individual 

 electron, the sinusoidal oscillation would be as little obvious as the 

 steady drift. Only the rapid zigzag motions of the corpuscles would 

 be conspicuous; what I have been calling a vibratory motion is, in 

 truth, only a slight bias of these random flights, just as the apparent 

 steady drift is itself a slight bias.) 



Another experiment, capable in principle of testing the expressions 

 for dielectric constant and conductivity and of giving the values of N 

 and g, consists in sending electromagnetic waves of the frequency 

 desired across a stratum of ionized gas, and measuring their index of 

 refraction and their absorption in the stratum. The former of these 

 two has not, so far as I know, been measured; but apparatus for 

 determining the second (it is that of Hasselbeck's experiment) is 

 shown in Fig. 7 : one sees the parafifin lenses which convert a diverging 

 beam of Hertzian waves into a parallel beam which is sent through 

 the ionized gas, and others which reconvert the parallel beam into 

 one which converges upon a bolometer. Part of the beam is reflected 

 from a semi-transparent mirror onto another bolometer, so that the 

 ratio of the intensities of the waves before and after the passage 

 through the gas can be determined without regard to fluctuations. 



The formulae for refractive index Wo (I add the subscript because 

 of having already used n, the conventional symbol, for another 

 purpose) and absorption-coefficient k are familiar to everyone who 

 has studied the theory of absorption and reflection of light by metals, 

 for in the classical theory of metals such a substance is conceived 

 exactly as we are now conceiving an electron-populated gas. We 

 have: ^^ 



Wo' - k^ = e, nok = iTa/n, (15) 



and putting for e and a the expressions (7, 8), we get: 



solving which equations for k, and putting m/er for g (according to 

 equation 14) and wi' for the combination AirNe^/m (we shall meet it 



9 Similar observations by Jonescu and Mihul on air and on hydrogen, subjected 

 in some cases to magnetic field, have recently been published, but with scant detail. 



"See for instance P. Drude, "Treatise on Optics" (page 361 of the English 

 translation). 



