8 Mr. H. Eamage. Relations of Spectra, Densities, [Nov. 7, 



selected in the same way, but our knowledge of these series is not as 

 complete as our knowledge of the principal series. Kayser and 

 Eunge did not observe any lines in the sharp subordinate series of 

 caesium : no lines, therefore, of the sharp series appear on the 

 diagrams. 



Copper, Silver, Gold. — There are only two strong lines in the flame 

 spectra of copper and silver, but some weak lines occur in the less 

 refrangible region. The wave-lengths of the strong lines are : — 

 Copper, 3274-1 and 3247*6 ; silver, 3380-8 and 3283*0. 



A line with wave-length 2675 occurs in the oxy hydrogen flame 

 spectrum of gold. It is much more refrangible, therefore, than the 

 above lines of copper and silver, and is on this account much weaker ; 

 the same* source of energy does not produce these more rapid mo- 

 tions with the same amplitude as the slower motions. This line and 

 another of wave-length 2428 occur in the arc and spark spectra of 

 gold : the two lines are easily reversed, and appear in all respects to 

 correspond to the above doublets of copper and silver. Eydberg 

 selected these six lines as the first members of the principal series 

 of these metals. He gave the wave-lengths, on Angstrom's scale, 

 as : — 



Copper. Silver. G-old. 



P 2 3273-2 3382-3 2675-4 



Pi 3246-9 3280-1 2427-5 



There are also several remarkable! bands in the flame spectra of 

 these three metals. The bands in the silver spectrum differ in 

 character from those in the copper and gold spectra. Every band 

 terminates in a sharp line on the more refrangible side. The lines 

 forming the bands are sharper and more widely separated in the 

 spectra of copper and gold than in the spectrum of silver. 



Magnesium, Zinc, Cadmium, Mercury. — The flame spectra of the first 

 three consist of lines and bands. The bands are very complicated in 

 structure, and are degraded on the more refrangible side. Eder and 

 Valentaj and Huff§ have described the production of a banded spark 

 spectrum of mercury, which is similar in character to those of zinc and 

 cadmium. No bands degraded in the same direction have been 

 discovered in the spectra of other metals. 



The following lines occur in the flame spectra of these metals : — 



* Hartley and Ramage, 'Trans. Roy. Dublin Soc.,' N.S., vol. 7, part xii, p. 341. 



f Hartley and Ramage, ibid. 



% 'Denkschr. K. Akad. Wien,' vol. 01, 1891. 



§ ' Astrophys. Jour.,' Sept. 1900. 



I 



