60 THE ROYAL SOCIETY OF CANADA 



Since 2504- 5+ 327 -3 = 2831 -8, we see groups I, II, and IV are 

 respectively Kayser and Runge's, I, II, and III groups, the III and V 

 groups being new. 



The lines 2332-96, 2204-18 and 1868-59 were obtained by 

 Saunders with a lead vacuum arc and a vacuum grating spectroscope. 

 The line 1904-2 was found by Fuller and Ainslie with a lead vacuum 

 arc and a fluorite vacuum spectroscope. Saunders has also pointed 

 out that the lines 5201,5005, 4340 form part of a group additional to 

 those of Kayser and Runge. This bears no relation to the grouping 

 here proposed. 



An attempt was made to see if the five groups are in reality five 

 series, the corrresponding members of which have constant frequency 

 difference. 



In looking for new series we are guided by three facts: — 



(1) the analogy which exists with other elements in the same group 

 of the Mendeleeff table. 



(2) All members of a series show the same Zeeman effect. 



(3) The heads of series are the lines which are fundamental for 

 that element. They are usually the only lines which appear in the 

 flame and absorption spectra. They alone are observed when the 

 metallic vapour is bombarded by electrons having the requisite 

 speed. When we examine the light from the electric spark in solu- 

 tions of the metal with increasing dilution, they remain long after 

 other lines in the arc have disappeared. 



In the case of lead we cannot derive aid from the first method, 

 for no series have been discovered in germanium, tin, lead, or the 

 other elements of this group. 



Purvis^ is the only one who has studied the Zeeman effect with 

 lead. He found two lines, 3740, 2873, were broken up into four 

 components, the remainder into triplets. 



We have thus only the third method available for this element. 

 In the flame it was formerly supposed that lead always gave a band 

 spectrum which was attributed to lead oxide. Later, under certain 

 conditions, a flame spectrum was observed consisting of three lines, 

 4057-97, 3683-60, and 3639-71. 



Spectrograms were taken of the flame of a Meeker burner, into 

 which was fed the vapour of lead from a small globule of lead sup- 

 ported by the gauze at the top of the burner. The above three lines 

 always appeared, but there was no trace of the band spectrum. 



1 J. E. Purvis, Camb. Phil. Soc. Proc, Vol. 14, p. 216 (1907). 



