of the Ionization Potential of Helium. 573 



introduced by Franck and Hertz to explain the results of 

 certain of their experiments, and seems in general to be in 

 accordance with the facts. The ordinary interpretation of 

 this hypothesis leads to the statement that an electron cannot 

 exist in mercury vapour with a velocity above 5 volts, or 

 in helium with a velocity above about 20 volts, provided that 

 the dimensions of the receptacle are sufficiently large compared 

 with the mean free path of: an electron at the given pressure. 

 A check on this hypothesis can be obtained from a con- 

 sideration of the successive current maxima at multiple 

 values of the critical voltage. With an apparatus like that 

 used by Franck and Hertz, in which the positive ions are 

 dealt with, no conclusion can, however, be drawn from the 

 successive values of the current on account of the retarding 

 effect of the accumulated positive ions. Tn mercury the suc- 

 cessive maxima appear to be, as a matter of fact, nearly equal 

 one to the other. In studying this point with the apparatus 

 used in the present investigation, the pressure and the 

 filament temperature were necessarily much reduced in 

 order to get that approach to saturation in each successive 

 current which was needed to allow the current increase at 

 the succeeding critical value to be apparent. The pressure 

 worked with (*2 mm.) was unfortunately one for which the 

 mean free path of the electrons is not sufficiently small, 

 compared with the radius of the receiving cylinder, to 

 render it safe to draw definite conclusions from the 

 observations. The results obtained are illustrated by the 

 curves of fig. 4, in which the electron current arriving; at 

 the copper cylinder is plotted against the accelerating 

 voltage. A is a preliminary curve in which particular 

 attention was paid to the points around 40 volts. In B 

 the points are carried to 85 volts. It will be observed that 

 a definite increase takes place at multiple values of the 

 ionization potential, but that the successive currents are not 

 related to the original current in a 1.2.4.8 ratio, as 

 would be the case if every electron collided inelastically 

 with an atom and there were no disturbances due to positive 

 layers. The only effects which can be produced by the 

 positive ions in this experiment are, first, their very marked 

 effect on the distribution of potential — a matter not yet 

 sufficiently investigated — and, second, their action in remov- 

 ing a certain number of free electrons by recombination. 

 The current ratios taken from curve B are, roughly, 

 1 . 1*25 . 1*7G . 2*18. On the assumption that three tenths 

 of the electrons undergo collisions at the specified ranges 

 these numbers would be 1 . 1'30 . 1*69 and 2*20. This 



