MEMORIAL TO CLASSICAL STATISTICS 135 



Sackur-Tetrode formula. By "desirable and welcome" I mean that it is 

 harmonious with the idea that entropy is a measure of disorder, an idea 

 plausible in itself and fruitful in its applications. A chemical element per- 

 fectly cr^'stallized at the absolute zero is supposed to be the exemplar of 

 supreme order, and therefore its entropy ought to be nil. But this is an 

 enormous subject requiring at least one other article, and I am glad that my 

 attempt at writing such an article stands already in print in the June( 1942) 

 issue of this Journal. 



Here then is the astonishing history of the Classical Statistics. By a 

 strangely artificial device, the numbering of atoms deemed identical, it 

 arrived at the proper distributions — that is, the distributions ratified by 

 experiment — in ordinary space and in momentum-space. It then proposed 

 a picture of entropy partially right, yet wrong in its dependence on the 

 number of atoms, and therefore fatally wrong. With another artificial and 

 dubious device, it corrected itself by adopting a new picture of entropy, this 

 time depending in the right way upon the number of atoms. With a third 

 artificial device (the introduction of Planck's constant in a peculiar way) it 

 completed the formula for entropy in a manner leading to the consequence 

 that the entropies of solidified cr^^ stallized elements are zero at absolute zero. 

 All of these feats and more were subsequently achieved by the New Statistics, 

 in a manner which I hope to explore on a later occasion. 



