1879.] on the Reversal of the Lines of Metallic Vapours. 357 



When lithium carbonate was introduced the, red line was at once 

 seen reversed, but not much expanded. When some aluminium was 

 added, the lithium blue line (wave-length 4,604) was seen with a dark 

 line in the middle for a short time only. The green line of lithium 

 was very bright indeed, and appeared somewhat expanded on the 

 addition of aluminium, but showed no reversal. 



On putting some baryta into the crucible the line with wave-length 

 5,535 was reversed, appearing very black but narrow. No other 

 barium line could be seen reversed in that crucible, but in another 

 crucible into which magnesia had been introduced, a dark line, with 

 wave-length about 4,930, was observed, which may probably be ascribed 

 to barium. 



With magnesia and aluminium the least refrangible of the b group 

 was seen reversed, all the b group being expanded. 



When silver was introduced, on drawing the poles apart, both the 

 brightest green lines (wave-lengths 5,464 and 5,209) were seen for a 

 short time with a black line down the middle. 



Frequently on parting the poles, whatever might be the substance 

 in the crucible, the whole of the brightest part of the spectrum, from 

 the orange to the blue, appeared filled with dark lines, all equidistant 

 and equally dark, like a fine grating. With a high dispersion these 

 lines are seen to be ill-defined at the edges. We can only suppose 

 them to be a banded spectrum of some compound of carbon. 



The lime crucibles are very quickly destroyed, but we hope to get 

 some more compact lime than we have hitherto had, and to employ a 

 more powerful electric current. The use of carbon or magnesia for 

 crucibles will, we anticipate, enable experiments of this kind to be 

 extended much further, and applied to various reactions taking place 

 at the temperature of the arc. In the case of carbon crucibles the 

 block of carbon itself will form one electrode, the other electrode 



Fig. 3. 



