192 ANNUAL. REPORT SMITHSONIAN INSTITUTION, 1924 



eease to play with a nucleus about which the electron orbits are 

 largely sj^mmetrical. In other words, if we retain the relativity 

 explanation of the spectroscopic-doublet formula, we are obliged 

 to suppose that two orbits which have the same shape but different 

 orientations with respect to the nucleus may exhibit widely different 

 screening constants — which is only another way of saying that 

 these orbits may possess widely different energies. 



To this extent, then, I am able to help out the chemist in his 

 attack upon the electronic orbits of the physicists. I am able to 

 enable him to say with a good deal of certainty that these orbits 

 can not be of precisely the type which we physicists have been 

 playing with so assiduously for the past five years. If we retain 

 the explanation w^hich has heretofore been given to the relativity 

 doublet formula, an explanation which requires entirely different 

 shapes for the two orbits corresponding to these doublets, then 

 we must begin to work with an atom which is very much less sym- 

 metrical with respect to the differently oriented orbits than we have 

 hitherto been imagining. 



