474 Long-Wave Limits of Normal Photoelectric Effect. 



known that the velocity of the photo-electrons increases with 

 the frequency of the light. Two laws have been proposed 

 for the relation between the velocity and the frequency, the 

 first being that the velocity squared, or the energy, of the 

 photo-electron is proportional to the frequency, that is 



V = /m-y , ...... (1) 



where V is the potential difference required to stop the 

 photo-electron, and therefore proportional to its energy, and 

 n the frequency. The second is that the velocity is directly 

 proportional to the frequency, or, 



VV = k'n-c (2) 



On account of the small ranges of wave-lengths available it 

 is difficult to decide which is the correct relation, but I 

 think that Richardson & Compton's investigations and mine 

 show undoubtedly that the results are well expressed by (1) 

 but not by (2). Drs. Pohl and Pringsheim, in support of 

 their contention that experiment has not decided between the 

 two laws suggested, refer to experiments by Kunz and his 

 pupil Cornelius, who conclude that their results are better 

 expressed by equation (2). The experimental arrangements 

 •of Kunz and Cornelius are open to considerable objections, 

 as has been pointed out very clearly by Compton (Pbys. 

 Rev. (2) i. p. 382, 1913)/ and it is therefore doubtful 

 whether their results are of any value in deciding the exact 

 nature of the relation between the velocity and the frequency. 

 The experiments of Richardson and Compton and my expe- 

 riments are the only experiments on this problem in which 

 the measurements were sufficiently accurate to distinguish 

 between the proposed relations (1) and (2), and both inves- 

 tigations lead to the conclusion that (1) is the true relation. 

 Drs. Pohl and Pringsheim offer several objections to these 

 researches, the first being that the intensity of the shortest 

 wave-length used (especially in my experiments) was so 

 feeble that the illuminated plate would never charge up 

 to its proper potential on account of insulation leaks. 

 The method used in my experiments was practically a null 

 method, so that this objection has but little weight. In 

 this connexion they state that errors may arise from light 

 being diffused to the electrode surrounding the illuminated 

 plate. The amount of light which gets through the mono- 

 -chromator, other than the isolated wave-length, is quite small, 

 and by the time a fraction of it is reflected by the illuminated 

 plate to the surrounding electrode, the effect is negligible. 

 This could be shown by a simple control experiment. 



