118 



BELL SYSTEM TECHNICAL JOURNAL 



to residual \'apors and partly to the electrons striking the walls of the 

 tube. The other curves are corrected for this ctTect, and then A is 

 calculated. Vov helium it is 25.10 '^ cm-; the values obtained by 

 modifications of the method agree well.* 



The helium atoms therefore behave as so many minute and yet 

 appreciab'le obstacles to the passage of the electron-stream, so long 

 as the electrons are not moving so rapidly that their energies of motion 

 do not surpass 19.75 volts. Klectrons as slow as these bounce off 

 from the atoms which they strike. When, however, an electron pos- 

 sessing kinetic energ\- greater than Id. 75 volts strikes a helium atom, 



c2 



06 



S-^ 04 



0.2 



0.0 



mmHg 



2.7 5.4 82 (0.9 (3.6 cm 



Fig. 2 — Curves illustrating the iiilircupliDii of electrons by nitrogen niuleculcs which 

 they strike. (Mayer, Annalen der Pliysik) 



it is liable to lose 19.75 volts of its energy to the atom, retaining only 



the remainder. This energ>- does not become kinetic energy of the 



atom, a process which would be incompatible with conservation of 



momentum; neither is the atom broken up; it receives the quota of 



energy into its internal economy, where some kind of a domestic 



change occurs with which we are not concerned for the moment, 



except in that it furnishes an exceedingly accurate indirect way of 



calculating the exact amount of energy taken from the electron. The 



atom is said to be put into an "excited" or sometimes into a "meta- 



* The niodifie<l methods are generally more accurate. Ranisauer's device, which 

 1 described in the first article of this scries, is probably the best. By a magnetic 

 field he swung a stream of electrons around through a narrow curving channel, and 

 those which were deviale<l even through a few degrees struck the limiting partitions 

 and were lost from the beam; he varied the numl>er of atoms in the channel by 

 varying the gas-pressure. In this way he discovered that A for argon atoms differs 

 very greatly for different sirixIs of the electrons; it was later found that other kinds 

 of atoms have a variable ,1, although happily the variations are not great. This 

 seems strange at first.'but it is probably stranger that A should have nearly the same 

 value for different 9pec<ls of the oncoming electrons, as for many atoms it does; and 

 alrangrr yet that it sliould have the s.itnc value for an oncoming atom as for an 

 oncoming electron, as is often tacitly assumed, and not too incorrectly. 



