187 



excitation potentials (at 12.25 and 13.7 Volts), and a less distinct one 

 at 14.7 Volts, which is followed by a series of energy steps which 

 has not yet been dissolved. Here too the apparently sharp excitation 

 potentials will no doubt correspond with undissolved groups of terms 

 lying close together, on account of the complication of the argon 

 spectrum. The argon spectrum not yet having been split up into 

 series, a comparison is not yet possible. When a similar structui-e 

 is assumed for the spectrum of argon as for neon, then starting 

 from the measured values for excitation and ionization potentials, 

 the following mean values foi' the first groups of terms are to be 

 expected : 



0.5 s = 130000 db JOOO 

 1.5 s - 30400 

 2 p - 18600 

 higher terms <^ 10500 



The serial terms calculated by Nissen ^) do not fit in with this 

 scheme. Also the fact that according to Nisskn lines of the red and 

 the blue argon spectra are considered as members of the same 

 series, though the condition for the excitation of the two spectra 

 are different, pleads in my opinion against the validity of the terms 

 calculated by him. 



B'or the rest more complications may possibly be expected for 

 argon than for neon. The fact found by Paschkn *) that for part of 

 the neon series the limits are shifted by a constant amount in 

 comparison with the other series, was explained by Grotrian ') by 

 the aid of the L-doublet of neon. He has also already pointed out 

 that it must be expected for argon thad the multiple M-limits will 

 manifest themselves in an analogous way. 



Physical Laboratory of the ''N. V. Philips 

 Gloeilampenfa brie ken' ' . 

 {Philips Incandescent Lamp Works). 

 Eindhoven. 



^) K. A. Nissen, Phys. Zeitschr. 21, 25, 1920. 

 ») F. Paschen, Ann. d. l^hys. 63, 201, 1920. 

 ») W. Grotrian, Zeitschr. f. Phys. 8, 116, 1921. 



