soMF. coMi.\H'(U<.ih:y .tnr.ixcis- i\ riiysn\- riii 4,u 



riMson whaU'viT for thinking llial tlii- atoms arr siiliji-cl to tiiuisii.il 

 strt'sscs." The fxci'ptions, howi-vt-r, aro iint mmuToiis «-iioii^;h to 

 jeopardize the rule. 



Two other features of tlu- roliimns slioiiKI l)e pointed oiil; first, 

 that tlie suiTessive levels in each column are not scattered at random, 

 l)ut form a converging series approaching the lop of the column as 

 limit (their energy-values form a secpience converging to zero); and 

 secontl, that there is nothing arbitrary about the order of the coiumrts, 

 since the cohnnn at the extreme left admits of transitions to only 

 one other column and therefore is unmistakable, and all the others 

 follow after it in an immutable order. 



E 0. A Digression About Xotation 



The symbol for a transition between two stationary states, and for 

 the spectrum line which corresponds to that transition, consists of 

 the symlx)ls for the two states separated by an arrow, or a dash, or a 

 semicolon, or any convenient mark. The final state is commonly 

 written first. Thus the line due to the transition from a state B to 

 a state A is designated thus: {A) — {B). Chess-players will be re- 

 minded of the "Continental" system of describing moves at chess, in 

 which symbols for the squares from which and to which the piece is 

 moved are written down one before the other. 



The notation for spectrum lines thus flows easily and naturall\' 

 from the notation for stationary states. This notation is not in 

 principle ver\' difficult, but it has become confused and confusing, 

 largely because of the alterations which have been wrought upon it 

 to make it express not the facts, but divers theoretical interpretations 

 of the facts. Alterations in names and notations generally produce 

 an e\il effect in physics even when justified in the highest degree, 

 for the old systems and the new persist side by side and cause in- 

 terminable trouble; all the more is this so when the alterations are 

 based on uncertain grounds and impermanent. The notation for 

 stationar\- states has already suffered much in this manner, and 

 probably the worst is yet to come. 



The classification of levels which I ha\e just (lescrit)ed enables and 

 ref|uires us to give a twofold symbol to each le\el; the symbol must 

 designate the column in which the le%el stands, and its order-number 

 or serial number in that column. The columns are generalh' desig- 



" Footc, Meggers antl Mohler observed a line corresponding to a change of two 

 units in k (the line i\,s) — {^,d), in the notation to lie explained in section K5) under 

 circumstances in which it seemed im|X)ssililc to t)elieve in an abnormally large 

 electric field. 



