80 



TABLE III (Continued). 



1 X 847, 2 X 848, 3 X 847. From tliis ensues that Table III is not 

 nneqnivocallj determined, like II, because when e.g. only 7^ and Q 

 occur in a row, they can now equally well be placed in another 

 row in the Q and R or R and S columns. 



It makes the impression that the number 847 = Z) has a physical 

 meaning, as also a value 846,2 occurs in the argon spectrum, which 

 may possibly be a more accurate value for D. 



One more detail of the experiments deserves to be mentioned. 

 In some cases the argon spectrum was observed in the potassium 

 tul)e at the same time with the first spark spectrum of potassium. 

 We have not to do here with a case of transmutation ofpotassiimi 

 into argon, but with the penetration of atmospheric air, of which 

 the argon has been (inally left. When, however, all precautions are 

 taken, the spark spectrum of potassium is emilted without argon lines. 



