540 Royal Society . 



The electric spark in passing through it exhibited a violet colour, 

 and gave the spectrum of fluted bands : 



5129 4436 



4981 4390 



4649 4318 



4556 4237 



The sodium was next heated until it presented a clean metaUic 

 surface. The light which the tube now emitted was bluish white, and 

 much fainter than before ; and the whole appearance of the spectrum 

 had changed to that of the second order with its characteristic 

 green line. It was, however, found that the pressure in the tube 

 had slightly increased, owing most likely to the vapour of the sodium 

 present ; and on bringing the mercury to its former level, the spec- 

 trum became brighter, but remained the same in character. New 

 nitrogen was then led into the tube, and after exhaustion the old fluted 

 spectrum again appeared ; this was, however, at once changed into 

 that of hues by heating the sodium. This process was repeated several 

 times in succession, but invariably with the same result. I have in my 

 possession two tubes sealed off under 2 millims. pressure, one without 

 sodium, showing the fluted bands, the other containing sodium, 

 showing the spectrum of lines. Two other tubes, sealed off under 15 

 millims. pressure, show the same thing. I have repeatedly convinced 

 myself that, from the highest pressure under which the spark of the 

 induction-coil passes to the lowest pressure which I could obtain with 

 an ordinary air-pump, pure nitrogen invariably gave one and the same 

 line-spectrum. Once, when I intended to seal a tube off under 

 higher pressures, it was found that the sodium was not sufficient to 

 absorb all the oxygen present, so that a sort of mixture of the two 

 spectra was seen. Such a mixture was often observed by Pliicker and 

 Wullner at the point where one spectrum changed into the other ; 

 it is characterized by the green line of nitrogen and the fluted violet 

 bands at the same time. 



i* The tube showing the mixture at 15 millims. pressure was gradu- 

 ally exhausted, but the spectrum remained exactly the same. If the 

 formation of the two spectra depends merely upon the pressure or 

 temperature to which the gas is subjected, how can a mixture of the 

 two spectra, indicating a state of transition, exist under so entirely 

 different pressures and different temperature? 



In order to ascertain whether nitrogen even carefully prepared 

 contains oxygen, a drop of a solution of iodide of potassium and 

 starch was introduced into the tube ; after the spark had passed for 

 a few seconds only, the liquid was coloured blue — showing either 

 the formation of oxides of nitrogen or of ozone, but at any rate the 

 presence of oxygen. 



7. Spectrum of oxides of nitrogen. — I tried to obtain the spectra 

 of the different oxides of nitrogen ; they all give the same fluted 

 spectrum, and t could get no information as to which particular oxide 

 the fluted spectrum is due : this is, however, easily understood if we 

 remember that it is just as difficult to prepare the oxides of nitrogen 



