Gases of Atmospheric Air, and their Spectra. 211 



and 669*2; so that for these substances and many others 



A — A 1 is always a small quantity, while — ~ — is considerable 



and increases as T diminishes. Hence the ratio of p to p 1 

 increases rapidly as T diminishes, and by evaporating all the 

 gases from the solid state and keeping the solid at as low a 

 temperature as possible, the gas first coming off consists in by far 

 the greatest part of that which has the lowest boiling point, which 

 in this case is nitrogen, and is succeeded, with comparative abrupt- 

 ness, by the gas which has the next higher boiling point. The 

 change from one gas to another is easily detected by examining 

 the spectrum in the sparking tube, and the reservoirs into which 

 the gases are pumped can be changed when the spectrum 

 changes, and the fractions separately stored. Or, if several 

 sparking tubes are interposed in such a way as to form parallel 

 communications between the tubes b and e, any one of them 

 can be sealed off at any desired stage of the fractionation. 



The variation of the spectra of both xenon and krypton 

 with variation in the character of the electric discharge is very 

 striking, and has already been the subject of remark, in the 

 case of krypton, by Runge, who has compared krypton with 

 argon in its sensitiveness to changes in the electric discharge. 

 Runge distinguishes krypton rays which are visible without a 

 jar and those which are only visible with a jar discharge. The 

 difference in the intensity of certain rays, according as the dis- 

 charge is continuous or oscillatory, is no doubt very marked, 

 but, with rare exceptions, we have found that the rays which 

 are intensified by the oscillatory discharge can be seen with a 

 continuous discharge when the slit of the spectroscope is wide. 

 Runge used a grating, whereas we have, for the sake of more 

 light, used a prism spectroscope throughout, and were there- 

 fore able to observe many more rays. 



There is one very remarkable change in the xenon spectrum 

 produced by the introduction of a jar .into the circuit. With- 

 out the jar xenon gives two bright green rays at about X4917 

 and X 4924, but on putting a jar into the circuit they are 

 replaced by a single still stronger ray at about X 4922. In no 

 other case have we noticed a change so striking as this on 

 merely changing the character of the discharge. Changes of 

 the spectrum by the introduction of a jar into the circuit are, 

 however, the rule rather than the exception, and there are 

 changes in the spectrum of krypton which seem to depend on 

 other circumstances. In the course of our examination of 

 many tubes filled with krypton in the manner above indicated, 

 we have found some of them to give with no jar the green ray 

 X5571, the yellow ray X 5871, and the red ray X 7600 very 



