180 



transition of energy between electrons and atonas, and supplement 

 it by a method wliicli admits of a clear distinction between light 

 emission and ionization also in the case of unelastic collisions of 

 small efficiency. 



The methods applied up to now foi- the study of the quantum emis- 

 sion of energy consist in this that either the radiation or ionization 

 that take place starting from a definite potential, or the phenomenon 

 that the impinging electrons lose energy, is used as a proof of the 

 occurrence of unelastic collisions. In this way a curve is obtained 

 in which the different steps of energy appear as breaks ; an accurate 

 measurement of them is often difficult, especially for the higher 

 steps of energy. It seemed, therefore, desirable to me to use as 

 criterion for the quantum transition of eneigy a characteristic that 

 immediately disappears again when the critical potential is exceeded, 

 and consequently causes the separate steps of energy to stand forth 

 as sharp maxima. Such a characteristic is the occurrence of electrons 

 with the velocity zero. For as soon as an electron possesses exactly 

 the energy required for the excitation of a definite quantum transition, 

 it may lose all its energy at the collision, and be left behind as 

 an electron with the velocity zero. If, however, it possesses a greater 

 energy, it retains the rest after the collision, and remains behind as 

 an electron with a velocity which, though smaller, is yet different 

 from zero. If, therefore, electrons of a definite velocity, are admitted 

 into a space in which they collide with atoms of a noble gas, 

 electrons of the velocity zero will only occur when the energy of 

 the electrons is precisely equal to the work required for the excitation 

 of a quantum transition. When, therefore, the number of electrons 

 which are left behind with the velocity zero, is plotted as function 

 of the accelerating potential, a sharp niaximum must be obtained for 

 every potential corresponding with an energy-quantum that can be 

 transferred at a collision of electrons. In consequence of the inevitable 

 distribution of velocity of the electrons it is not possible to determine 

 the number of electrons which have rigorously a velocity zero. 

 Therefore the number of those electrons the velocity of which lies 

 below a definite small value (in our measurements mostly 0.2 Volt) 

 will be plotted as function of the tension. 



The measurements according to this principle are carried out in 

 the following way : 



The electrons emitted by a short incandescent wire Z) of tungsten 

 (Fig. 1) enter the field-free space R through the gauze ISf^ after 

 acceleration through an electric tield, in which space they collide 

 with the atoms of a rare gas. Part of the electrons passes through 



