CONTEMPORARY ADVANCES IN PHYSICS 



675 



tendency to spread other than that for which the gas itself is account- 

 able. As early as 1894, Lenard did such experiments with 30-kilovolt 

 electrons (cathode-rays which had emerged from a discharge-tube 

 through a window of thin metal foil). A few years later, he added 

 data obtained with slower electrons, and the work was continued 

 by Becker and by Silbermann.^ For the 30-kilovolt corpuscles, the 

 cross-section o- is very much smaller than the gas-kinetic co — only a 

 few per cent as great. As the energy of the electrons is decreased, a 

 rises towards ao. AH this is illustrated in Fig. 1. 



Apqioo 



660 



T T — r 



500 1000 2000 



4000 



30000 

 VOLT 



Fig. 1 — Cross-section for interception of electrons, plotted for various gases over a 

 wide range of electron-speeds. (P. Lenard, Annalen der Physik.) 



Obviously, these are not experiments in which the corpuscles are un- 

 able to ionize or to excite ; quite the contrary. One might be tempted 

 to rush to the other extreme, and guess that all of the electrons missing 

 from the beam have effected ionizations, that the quantity a of Lenard's 

 experiments is a measure of likelihood of ionization; but in the present 

 state of knowledge, this would be going too far. It is important, 

 however, that for the/a5/ electrons a turns out to be proportional to the 

 atomic number of the gas, if this is monatomic; and to the sum of the 

 atomic numbers of the atoms constituting the molecule, if the gas is 

 diatomic or triatomic. This sort of rule suggests that interception of 

 fast-flying corpuscles is due either to the nuclei of the atoms, or to some 

 action of the bound electrons of the atoms in which they all are equally 

 potent, however tightly or loosely they are bound.* 



^ References are given at the end of the paper. 



* Silbermann (dissertation, Heidelberg, 1910) and A. Becker investigated a large 

 number of organic compounds, using fast electrons; and they found that the a of each 

 of these molecules (some of them composed of five to twenty atoms) could be predicted 

 accurately, simply by adding together the values of <r for the constituent atoms as 

 determined by experiments on simpler gases. 



