198 Prof. Millikan and Mr. Winchester on Influence of 



due to the imperfection in the method of eliminating the leak 

 due to the heat. 



The conclusion which seems to be justified by these experi- 

 ments, considered in connexion with the previous experiments 

 with aluminium in which the temperature was carried to 

 343° C, is that tlie photo-electric effect, like radio-activity, in a 

 phenomenon which is completely independent of temperature. 

 Elster and Geitel's opposite result in the case of potassium 

 must be attributed to secondary causes of some sort. It is to 

 be hoped that further investigation will bring to light the 

 nature of this cause in the case of potassium and of the alkali 

 metals which according to Thomson behave like potassium 

 in this respect. 



The independence of the photo-electric effect upon tem- 

 perature constitutes very conclusive evidence that, if free 

 electrons exist at all within metals, it is not these electrons 

 which escape under the influence of the ultra-violet light 

 Nor does it seem possible, in view of these results, that the 

 suggestion made by Ramsay and Spencer * to account for a 

 discontinuity in fatigue effects which they observe in air, 

 that the discharge at first consists of metallic electrons and 

 later of electrons which are constituents of the atoms, can be 

 correct ; for the rate of escape of any electrons which are not 

 such constituents would certainly be expected to depend upon 

 temperature (cf. also §10). 



6. Effect of Temperature upon the Velocity of. Projection 

 of Electrons from Different Metals. — If the electrons which 

 escape under the influence of the light are not free electrons 

 of the metal, then they must be regarded, as Leuardhas held, 

 as constituents of the atom which become detached from it 

 through the influence of the light, and in this case the maxi- 

 mum velocity with which they are projected from the surface 

 of the metal, after they get outside the range of molecular 

 forces, can be determined by measuring the positive potential 

 which the metal assumes in a perfect vacuum when the light 

 acts upon it for an indefinite period ; for, obviously, if e is 

 the charge and m the mass of the electron, and if v is its 

 velocity of escape from the metal, and P the positive poten- 

 tial which the latter must have in order to overcome this 

 velocity and cause the electron to return again to the metal, 

 then we have 



or 



,= A / 2 p^, 



V m 



Whether or nut this represents the actual velocity of the elec- 

 tron within the atom, depends upon whether or not the 

 electron, after being set free, meets with any resistance to its 

 * Ramsay & Spencer, Phil. Mag. xii. p. 417 (1906). 



