503 

 TABLE IV {Continued). 



The constant differences seem soon to stop below X 3000. This 

 may be in connection with tlie appearance of the second spark 

 spectrnm of potassium. 



We have, however, also started an investigation of the lines that 

 satisfy formulae with fourfold and ninefold Rijdberg constans. By 

 this way the proof might be furnished that the observed spectrum 

 belongs to once ionized potassium; besides, a quantitative comparison 

 with the red argon spectrum may perhaps be possible. 



