CONTEMPORARY ADVANCES IN PHYSICS 595 



symbols because they are used in some of the original articles, not 

 because they add anything to the fundamental ideas.) 



The value of co may be determined by the Langmuir probe-method; 

 values for /, the mean-free-path of the electrons, are supplied by 

 various methods of "varying reliability. Childs, in his experiment 

 which I have quoted, obtained with the probe the three values 1.3, 

 1.5, 1.7 times 10^ (cm. /sec) for w, these corresponding to the three 

 cited values of the current maintaining the ionization; combining 

 these with estimates ^ of /, he obtained for mco// the values 3.2, 3.8, 

 4.3-10-^8. Comparing these with the values found for g, and re- 

 membering the manifold chances of faults in the assumptions, one is 

 favorably impressed with the agreement. 



Could we not evaluate g directly, by measuring the drift-speed u 

 of the electrons exposed to a certain constant fieldstrength F, and 

 forming the ratio of eF to n to which, according to equation (12), 

 g is equal? Here we must be careful. According to equation (13), 

 g depends on the vivacity of the random motions of the molecules, 

 whereof co is the mean speed; now, co depends on the fieldstrength; 

 we must select such a value of F, that the random agitation of the 

 electrons shall be the same as it is in the actual gas on w^hich the 

 high-frequency field is imposed. Now this actual gas is subjected to 

 a constant field, that which maintains the ionization. It is natural 

 and simple to assume, that this field controls the value of co, the 

 effect of the high-frequency field on the mean speed of random agitation 

 of the electrons being presumably slight. The measurement of the 

 drift-speed should therefore be made in the very gas under the very 

 same constant field, in w^hich the high-frequency phenomena are 

 observed. 



Let me denote by i the steady current along the length of the tube, 

 due to the constant field; by A^, the number of electrons per unit 

 volume; by u, their drift-speed; by A, the cross-sectional area of the 

 tube. It can then be readily seen that i is equal to NeuA} The 

 method is consequently simple in principle; in practice, the chief 

 difficulty apparently is that N is not the same near the w^alls of the 

 tube as along its axis, so that probe-measurements should be made at 

 a number of distances from the axis and the results averaged. Apple- 

 ton and Chapman determined iV at a single point of the tube; the 

 value of the ratio eFju came out equal to 2.3- 10"^* — again, a remark- 



7 Apparently he multiplied by 4V2 the value of the mean-free-path of molecules 

 of nitrogen, this being the simple gas presumably most like air; at any rate he used 

 the value .036 cm. 



^Electrical Phenomena in Gases, p. 207, pp. 232-233; on the two latter pages I 

 describe Killian's application of the method to mercury vapor. 



