Absorption-Spectrum of Potassium Vapour. 197 



n. Wave-length. 



10 2963-4 (2963-36 Kayser) 



11 2942-8 (2942-8 „ ) 



12 2928-0 



13 2916-6 

 J 4 2907*6 



15 2900-4 



16 2894*6 



17 2889*7 

 ] 8 2885*9 



19 2882-9 



20 2880-3 



21 2877*9 



22 2875-8 



23 2S74-1 



24 2872*5 

 -5 2871-1 

 2<S 2870*0 



The three formulae given by Kayser * give for the value 

 of the wave-lengths for n — 20 



2876*2 : 2876-4 ; 2879*4 : 



the departures from the observed values in the eases oE n = \0 

 are, however, 



+ 1*05 : -O'O : + 0*76 : 



€0 that the agreement is as close as could be expected, and a 

 recalculation of the constants i< necessary. 



The values for the wave-lengths are estimated to be correct 



to within -2 oi an Angstrom unit. With a larger spectro- 

 graph giving more dispersion and a wider aperture, there is 

 no doubt that many more lines of this series could be 

 photographed. 



There appears to be no trace of lines belonging to the 

 associated series for potassium in any photographs which the 

 author has taken. This fact i< of importance as it shows that 

 there must be a very great difference in the mechanism 

 corresponding to the principal series and the others, and that 

 sodium and potassium are exactly analogous in this way. 

 Another interesting point in connexion with these lines i- 

 the fact that though a largo number of lines of wave-lengths 

 less than 3000 have been observed in the spark spectrum j. 

 none of these lines after 2992*3 corresponding to /* = ( J 

 appear. 



* Handbuch der Spectroscopies ii. p. o:!4. 



t Eder & Valenta, Denksekr, Wien, lxi. 1894. 



