Mr. J. N. Lockyer on Spectrum Analysis. 



147 



It scarcely needs remark that trie difference in breadth of the line 

 H (5 in different stars affords us information of the difference of den- 

 sity of the gas by which the lines of absorption are produced. A 

 discussion of the observations in reference to this point, and other 

 considerations on the physical condition of the stars and nebulae, I 

 prefer to reserve for the present. 



Table II. — Stars approaching the Sun. 



Star. 



Arctunis 



Vega 



a Cygni , 



Pollux 



a Ursce majoris 



y Leonis 



e Bootis 



7 Cygni 



«Pegasi 



y Pegasi? 



a Andromedee , 



Compared 

 with 



Mg 



H 



H 



Mg 



Mg 



Mg 



Mg 



H 



H 



H 



H 



Apparent 

 motion. 



Earth's 

 motion. 



50 



40 to 50 



SO 



32 



35 to 50 



+ ^ 

 + 3-9 

 + 9 

 + 17 

 +11 



Motion to- 

 wards sun. 



55 



44 to 54 



39 



49 



46 to 60 



December 12.— -"William Spottiswoode, M.A., Treasurer and Vice- 

 President, in the Chair. 



The following communication was read : — 



" Researches in Spectrum Analysis in connexion with the Spec- 

 trum of the Sun."— No. I. By J. Norman Lockyer, P.R.S. 



The author, after referring to the researches in which he has 

 been engaged since January 1869 in conjunction with Dr. Prank- 

 land, refers to the evidence obtained by them as to the thickening 

 and thinning of spectral lines by variations of pressure, and to the 

 disappearance of certain lines when the method employed by them 

 since 1869 is used. This method consists of throwing an image of 

 the light-source to be examined on to the slit of the spectroscope. 



It is pointed out that the phenomena observed are of the same 

 nature as those already described by Stokes, W. A. Miller, Eobinson, 

 and Thalen, but that the application of this method enables them to 

 be better studied, the metallic spectra being clearly separated from 

 that of the gaseous medium through which the spark passes. Pho- 

 tographs of the spark, taken in air between zinc and cadmium and. 

 zinc and tin, accompany the paper, showing that when spectra of 

 the vapours given off by electrodes are studied in this manner, the 

 vapours close to the electrode give lines which disappear from the 

 spectrum of the vapour at a greater distance from the electrode, 

 so that there appear to be long and short lines in the spectrum. 



The following elements have been mapped on this method : — « 

 Na, Li, Mg, Al, Mn, Co, Ni, Zn, Sr, Cd, Sn, Sb, Ba, and Pb, 

 the lines being laid down from Thalen's maps, and the various 

 characters and lengths of the lines shown. 



In some cases the spectra of the metals, enclosed in tubes and sub- 



L2 



