392 BELL SYSTEM TECHNICAL JOURNAL 



obey the Fermi-Dirac statistics has been enlarged but little since my article 

 of 1929. 



As for the second question, I can give only the shadow of an answer. 

 The reason for adopting sometimes the Bose-Einstein and sometimes the 

 Fermi-Dirac statistics springs from wave-mechanics, and that requires an 

 article of its own. I can say, without proof, that the choice depends upon 

 the number of elementary particles in the atom. The gas is supposed to 

 conform to the Bose-Einstein or the Fermi-Dirac statistics, according as that 

 number is even or odd. An electron is an elementary particle all by itself, 

 wherefore the preceding paragraph. 



For material gases, the crucial number is obtained by adding up the 

 numbers of the protons and the neutrons in the nucleus, and the number of 

 orbital electrons which surround the nucleus and complete the atom. In 

 nature the atoms for which the crucial number is even vastly outnumber 

 those for which it is odd, and the Bose-Einstein statistics is therefore the 

 prevalent one. The principal isotope of nitrogen and the second isotope of 

 hydrogen do indeed belong to the rarer category, but in the gaseous state 

 their atoms always pair themselves into diatomic molecules, a circumstance 

 which restores these gases to the realm of Bose and Einstein. A detail in 

 the band spectrum of a diatomic molecule is available for telling which form 

 of statistics the individual atom would obey if free; it confirms what I have 

 just been saying — but this is an intricate story. 



