254 Hull — Initial Velocities of the Electrons. 



from A. Hence the electrometer deflections will be propor- 

 tional to the number of electrons whose energy per unit 

 charge* exceeds the corresponding potential of B. The 

 maximum energy possessed by any of the electrons will be 

 that potential which is just sufficient to prevent any of them 

 from escaping, and is given by the point where the curves in 

 figures 2, 3 and 4 become tangent to the straight line represent- 

 ing the constant negative leak due to reflected light. From the 

 values of the energy the velocity can at once be determined 



by substituting the known value of — . 



The experiment was first tried with hydrogen in the dis- 

 charge tube. When the screen cell C is exhausted, the hydro- 

 gen discharge throws upon A light of all wave-lengths down to 

 X 1230, the limit of transparency of white nuorite.f Filling 

 the cell with air at atmospheric pressure cuts off practically 

 everything below X 17104 In this way two definite spectral 

 regions are obtained, and since the effect without air in the 

 screen cell is 25 times as great as with it* that is, the electrons 

 produced by the hydrogen light of longer wave-length than 

 X1710 are only 4 per cent, of the total, this region of shorter 

 wave-lengths may be taken to extend from X 1230 to X 1710. 



The results are given in Table I and figure 2. The ordinates 

 of curve II are magnified so that their maximum, representing 

 the total number of electrons set free by the light, coincides 

 with that of curve I. While the two curves show a somewhat 

 different distribution of velocities, indicating a larger propor- 

 tion of high velocity electrons for the light of higher 

 frequency, the maximum velocity is the same for both. It 

 appeared from this that the proportionality, discovered by 

 Ladenburg, between the velocity of the electrons and the 

 frequency of the light producing them, did not extend to these 

 short wave-lengths. For the shortest wave-length in the 

 unscreened light was estimated to be X 1230, that in the 

 screened light X 1710, so that the maximum velocities should 

 have been very different in the two cases. 



As a variation of the experiment the discharge tube was 

 filled with carbon dioxide, and now a different result was 

 obtained. There was a very marked difference in the maxima, 

 as is shown in the second part of Table I and figure 3. With- 

 out the air screen the maximum energy of the electrons is 

 about 3*3 volts, with the screen only 2*5 volts. If the shortest 

 wave-length in the spectrum of carbon dioxide is taken as 



* For the sake of brevity, the energy per unit charge will hereafter be 

 designated simply as energy of the electrons, 

 f See T. Lyman, Astrophys. J., xxv, p. 45, 1907. 

 {See T. Lyman, ibid., xxvii, p. 87, 1908. 



