CONTEMPORARY ADVANCES IN PHYSICS 603 



resonance. He did not measure A^, but something probably (though 

 not certainly) proportional to A^, the factor of proportionality not 

 known: the direct current i flowing across the gas between two probes 

 inserted on opposite sides of the tube and maintained at a constant 

 P.D. Plotting y and P against i, he observed a crinkle in the former 

 curve, a minimum in the latter — evidence of a natural frequency at a 

 particular value of electron-concentration. 



In his earlier work (which I mention because the curves are often 

 reproduced) Gutton connected the plates of the ionization-condenser 

 to one another through a thermocouple, and to this circuit coupled 

 an oscillator of which the frequency could be varied. For each of a 

 number of values of the ionizing-current in the discharge-tube he 

 varied the frequency until resonance was declared by a maximum of 

 the mean square of the current in the thermocouple; he measured 

 this maximum P and the wave-length X of the oscillations. Not 

 having any quantitative measure of the ionization against which to 

 plot P and X, he plotted one against the other — P as ordinate, X as 

 abscissa — and obtained curves of the curious appearance shown in 

 Fig. 10, in which the arrow indicates the sense in which the ionization 

 increases along each curve. The start is made from the wave-length 

 (408 cm.) at which the system is in resonance when the gas is not 

 ionized. At low pressure, the curve bends first to the left and down- 

 ward; this signifies that as the ionization increases the dielectric 

 constant of the gas is falling and the conductivity rising, as by the 

 theory they should. Then at an unknown but seemingly sharply- 

 marked value of ionization, the curve bends sharply to the right; 

 and this signifies the same as the crinkle and the minimum in the 

 other more-fully-comprehended curves which I have been discussing. 



It is still dubious whether the natural frequency so revealed is 

 that which the foregoing theory predicts. Neither in his earlier nor 

 in his later experiments did Gutton measure N (the estimate of its 

 value which he once makes is derived in an indirect and fallible way). 

 Measuring the values of i at which the resonance appeared in his 

 later work, and comparing them with the corresponding values of the 

 frequency, he found i^^ proportional to i^'*; the range of frequencies 

 was comparatively small (not quite 4 : 1) but if the result is certain 

 and i is truly proportional to N, equation (24) is contradicted. Apple- 

 ton and Chapman testify that they found the resonance-effect observed 

 by Gutton, but at values of N (which they measured by the Langmuir 

 method) entirely too great to permit of regarding it as the plasma- 

 electron frequency, from which they believe it to be distinct. 



As for the search for spontaneous oscillations probably having the 

 proper plasma-electron frequency, it is at present even less advanced. 



