SAUNDERS. — ARC SPECTRA OF THE ALKALI METALS. 451 



results of the measurements in the deep red differ from those of Lehmann 

 to a marked degree. As a check on the present metliods, the writer took 

 a set of measuremeuts of several other and better known lines in this 

 spectrum. The results, which are included in the table, agree reasonably 

 well with those of Kayser and Runge. As a further check, the wave- 

 number differences for the pairs were calculated, using the writer's values 

 for the wave-lengths. The results were : — 



1' serk'S. 



237.8 



The constancy of these numbers is satisfactory, with the exception of 

 the short wave-length ends of the series, where the measurements do not 

 pretend to be accurate. A similar table calculated from Lehmann's 

 results shows a much wider divergence. 



Ramage has noted a line at 6306 ; nothing was seen at this place 

 except a ghost of the strong line at 6299. He has also given a sharp 

 line at 5165; at this point on the writer's photographs appeared a sharp 

 line which was the head of a carbon band. 



Caesium. 



In tlie caesium spectrum, five new lines were found. The one at 

 7944.7 was predicted by Ritz and is a member of the second subordinate 

 series, along with 7609.7. 



Two new series can be arranged from the odd lines in this spectrum, 

 which might be called the third and fourth subordinate series, as they 

 evidently belong to this class. 



III. IV. 



8080 ? 



8019 9171 ? 



6872.6 7280.5 



6826.9 7228.8 



6475 6630.5 



6434 6588.0 



6359 

 6325 



