( 443 ) 

 The constants have been calculated from the lines 4, 5 and 6. 



ALUMINIUM. FIRST ASSOCIATED SERIES. 



The agreement with the first lines (1, 2 and 3), leaves much to be 

 desired. The value of the constant h is here 1,03806, greater than 

 the value of a in that formula; this does nof, occur with any of the 

 other series. 



With 4 constants, so with : 



n =z A 



109675 



b c 



m -\~ a -\ [- — 



m m 



a better result is most likely reached. When the constants h and here 

 probabl}' also the c, are not small with respect to a, then the influ- 

 ence of those constants is very great, particularly for small values 

 of in. The deviation for the first line of the above series (3082,27), 

 however, is so great, that I doubt if this is really the first line of 

 this series. 



The behaviour of this Aluminium series is certainly peculiar, and 

 a further investigation is desirable. 



For the way in which the constants in the formula were calculated, 

 and for the spectra of Potassium, Rubidium and Calcium, of Magne- 

 sium, Calcium, Cadmium and of Helium and Oxygen, I refer to my 

 thesis for the doctorate, which will shortly be published. 



29 



1'iccceding.s lloyal Acad. Amslei'dam. Vol. IX. 



