Spectra of Potassium, etc., and their Mutual Relations. 303 



'The Spectra of Potassium, Rubidium, and Caesium, and their 

 Mutual Relations." By Hugh Ramage, B.A., St. John's 

 College, Cambridge. Communicated by Professor G. D. 

 Liveing, F.RS. Received May 14 —Read June 5, 1902. 



The spectra of this group of metals have already been considered 

 in a paper by the author, on " A Comparative Study of Spectra," &c* 

 It was there shown that the differences between the principal series of 

 lines in these metals depended on the atomic mass alone ; and also 

 that there was a close connection between the subordinate series and 

 the atomic mass. A further study of the latter series was impossible at 

 the time of writing the paper owing to the fewness of the lines which 

 had been observed and measured in them • practically no lines were 

 known in the second subordinate series of rubidium and caesium. 



Some lines belonging to the subordinate series have been measured 

 in Bunsen-flame and spark spectra by Lecoq de Boisbaudran, and in 

 the arc spectrum by Liveing and Dewar, and by Kayser and Runge. 

 Lehmann has measured some lines in the arc spectra in the red 

 region. Lines recorded by these observers were found by the writer, 

 with considerable intensity, in the oxyhydrogen flame spectra of the 

 metals ; and other lines, weaker than the above, were present which 

 had never been recorded. Photographs of these high-temperature 

 flame spectra were taken with a spectrometer designed by Professor 

 Liveing, fitted with a Rowland plane grating ruled with 14,438 lines to 

 the inch. The quartz lenses were plano-convex with a focal length for 

 the D lines of about 778 mm. The spectra in the first and second 

 orders were photographed, and some measurements were made in the 

 red region by eye observations. Spark spectra were photographed, 

 superimposed on the flame spectra, of iron and titanium principally, 

 but other metals were also employed. These furnished the numerous 

 fiducial lines required for the accurate determination of the wave- 

 lengths. 



The lines in the subordinate series are generally more diffuse than 

 those in the principal series. Some of the weaker lines, notably those 

 of caesium, are very broad with diffuse edges ; very accurate measure- 

 ment of these is impossible. 



Particulars of the spectra are recorded below; the oscillation fre- 

 quencies are reduced to their values in a vacuum. The lines have 

 been sorted into the principal and the first and second subordinate 

 series, and marked P, I, or II, with the number of the line, according 

 to Rydberg's formula, in the sixth column. The wave-lengths of the 

 lines which have been observed before are given in the fourth column. 



* < Koy. Soc. Proc.,' vol. 70, p. 1, 1902, 



