Uil Mr. J. N. Lockyerow Recent 



line of sodium. Every turn of the screw which raises or 

 lowers the upper pole, enables me to vary to a very consider- 

 able extent indeed the thickness of that absorption- line. 



Now the way in which that has been managed is very simple. 

 The only arrangement required is one which shall enable me 

 at will to vary the density of the sodium-vapour. When I 

 make the sodium-vapour as dense as possible, then the line is 

 very thick. When I make it much less dense, the line becomes 

 thinner. If the spectrum on the screen had been a gas-spec- 

 trum (supposing it were possible to exhibit a gas-spectrum to 

 an audience), the exact equivalent of that experiment would 

 have been, that the gaseous spectrum at atmosphere pressure 

 would have given us most of the lines as thick as the sodium- 

 line was at its thickest ; while if by any possibility we could 

 have rendered the phenomena visible while the pressure was 

 being reduced, as the pressure of the gas was reduced the 

 line would thin. Now there are very great objections to the 

 using of Geissler tubes. One very valid objection is that the 

 gas becomes much less luminous as its pressure is reduced. 



Here is a method which is excellent in this way, that it en- 

 ables all the work connected with gaseous spectra to be done 

 at atmospheric pressure, and we get the line down as thin as we 

 choose, not by reducing the pressure, but by reducing the quantity 

 of gas in a mixture. If we take, for instance, a spark in ordi- 

 nary atmospheric air and observe its spectrum, we find the lines 

 of the constituents of atmospheric air considerably thick ; but 

 if I wish to reduce the lines, say of oxygen, down to a consi- 

 derable fineness so that I can photograph its lines (these should 

 be fine, in order to enable me to determine their absolute posi- 

 tion ; to accomplish this) the spark is made to pass in a glass 

 vessel with two adits and one exit tube. If I wish to observe 

 the oxygen -lines fine, I flood the vessel with nitrogen so that, 

 say, there is only 1 per cent, of oxygen present, and observe the 

 current between the enclosed electrodes. If I wish to observe 

 nitrogen-lines fine, I flood it with oxygen, so that there is 

 only 1 per cent, of nitrogen present. In this way, by merely 

 making an admixture in which the gas to be observed is quan- 

 titatively reduced, so that the lines which we wish to investi- 

 gate are just visible in their thinnest state, we have a perfect- 

 means of doing this without any apparatus depending on the 

 use of low pressures ; and those who have worked most with 

 Geissler tubes will appreciate the very great simplicity of work 

 which is thus introduced. 



Another important application of spectroscopic theory re- 

 cently applied to the investigation of the chemistry of the sun 

 is this : — Assume that the spectrum of any substance is not a 



