[ 253 ] 



XXI. 



ON THE VACUUM TUBE SPECTRA OF SOME METALS AND 

 METALLIC CHLORIDES. Paut II.— LEAD, IRON, MAN- 

 GANESE, NICKEL, COBALT, CHROMIUM, BARIUM, CAL- 

 CIUM, STRONTIUM, MAGNESIUM, POTASSIUM, SODIUM, 

 AND LITHIUM. 



By JAMES H. POLLOK, D.Sc, 

 Royal College of Science for Ireland. 



[Plates XXIL anb XXIIL] 



[Published May 7, 1912.] 



The experiments were conducted precisely in the manner described in the 

 first part of this paper, the same apparatus and method being used through- 

 out. The quartz vacuum tube has already been illustrated and described in 

 detail. The illustration in Plate XXIII. shows the whole apparatus, with 

 tlie spectrograph, coil, condenser, pump, dryers, gauge, vacuum tube, and 

 Meker burner, in working order. 



The generalizations arrived at in previous observations were confirmed by 

 the observations on this new series of metals and their chlorides. Bands 

 were not so prominent a feature of the spectra of the compounds examined 

 in this part of the paper, except in the cases of manganese and magnesium ; 

 with them a greater number of lines show when the condenser or Leyden jar 

 is used, and the bands are less conspicuous. Apart from the bands, the 

 spectra of the vapours of metallic chlorides appear to consist exclusively of 

 the lines of the spark spectra of the metals of which they are composed, with 

 or without the lines of the spark spectrum of chlorine, together with the 

 ultimate lines of any impurities that may be present. There is, so far, no 

 reason to believe that compounds give rise to extra lines in the spectrum ; 

 and in all oases where lines have been observed that were not due to 

 the metal and chlorine, and have been carefully measured and investigated, 

 they have invariably proved to be the ultimate lines of the vacuum tube 

 spectrum of some other metal or non-metal present as au impurity. It 

 would be premature to assert that no lines due to the compounds of the 

 metals with chlorine existed until every line in each spectrum was 

 accurately measured and identified ; but it may safely be asserted that 

 such lines are not a prominent feature of spectra, and, if they exist at all, 



SOIENT. PROC. B.D.S., VOL. XIII., NO. XXI, 2 B 



