﻿the Electric Discharge in Liquid Oxygen, Air, fyc. 237 



through the liquid and partly through the gas immediately 

 above it, the discharge took place more easily, and the con- 

 tinuous spectrum became, relatively to the lines, less bright. 

 Without the jar, however, the spectrum had generally the 

 same character as before. The three lines above mentioned 

 were still the most conspicuous bright lines. On putting the 

 jar into the circuit, however, many more bright lines came 

 out. The well-known orange line of oxygen, X6171, ap- 

 peared expanded into a band with its sharper edge on the 

 more refrangible side at a wave-length of about 615, and 

 fading towards the red but traceable as far as X618. Schuster 

 and Piazzi Smyth give a compound line at about wave-length 

 X6156; and it is possible that this band may represent this 

 compound line. It will be noticed that, as to the shading, it 

 follows the character of A and B, but with the dispersion we 

 used it was not possible to resolve it into lines. Probably, 

 under the circumstances, the lines of which we may suppose 

 the band composed would be so much expanded as to overlap 

 one another. Besides this band, blue lines were conspicuous 

 at wave-lengths about 435, 441, 459, 465, 470, all corre- 

 sponding to known lines of oxygen. The green line of 

 platinum at wave-length about 530 came out brightly, as 

 well as platinum lines at about X583 and X580. Another 

 line, less bright, appeared at wave-length about 557, and at 

 times a second line near it at about 555. These have not 

 been described as oxygen lines by other observers, but they 

 fall within one of the green bands described by Schuster as 

 seen in the negative glow in a vacuous oxygen tube. 



We next proceeded to exhaust the gas above the liquid in 

 the vessel, until the pressure was reduced to about 1 centim. 

 of mercury. The liquid, of course, boiled away fast until the 

 temperature had fallen to something like —200° ; and the 

 gas at the reduced pressure otfered comparatively little resist- 

 ance to the passage of the discharge. 



So long as both electrodes were immersed in the liquid, the 

 reduction of pressure and of temperature did not make any 

 marked difference in the appearance of the spectrum. But 

 as the liquid evaporated and only the lower electrode was 

 immersed, so that the discharge was partly through the gas, 

 the continuous spectrum was much weakened, and two bright 

 green bands came out extending from about X 521 to X 531, 

 and from about A, 553 to X 561. This was without a jar in 

 the circuit. The bands were nearly uniformly bright with 

 both edges diffuse. They were much better seen when both 

 electrodes w T ere out of the liquid, and were brightest in the 

 glow which surrounded the poles. They were equally well 



