﻿236 Profs. Liveing and Dewar on the Spectrum of 



Liquid oxygen, air, and nitrogen, like non-electrolysable 

 compound liquids, offer very great resistance to the passage 

 of an electric discharge, so that with a powerful induction-coil 

 we could hardly make the spark pass through a thickness of 

 one millimetre of liquid. When the thickness of liquid tra- 

 versed was less than this, a succession of sparks could be 

 maintained ; but the resistance appeared to be very great, and 

 the disruptive effect on the electrodes sufficient to discolour 

 the liquid by the particles thrown off their surfaces. 



The discharge through the liquid in all cases gave a con- 

 tinuous spectrum and some bright lines traceable to the 

 electrodes, while the rays which we suppose to have been 

 emitted by the molecules of the liquid were less conspicuous. 

 It seems not unlikely, therefore, that the continuous spectrum 

 was due to the particles thrown off the electrodes. 



Liquid Oxygen. 



When both electrodes were immersed in liquid oxygen 

 boiling at atmospheric pressure, and therefore at about 

 — 180° 0., and the distance between the electrodes about 

 one millimetre or less, the spectrum of the spark, without jar, 

 was chiefly a continuous one, brightest in the yellowish-green 

 but extending for some distance both on the red and on the 

 blue side. The absorption-bands of oxygen were conspicuous 

 on this bright background. A few bright lines were also 

 seen, of which the most prominent were three in the green 

 and yellowish-green, with the approximate wave-lengths 

 505, 533, and 547. These lines were not much brighter than 

 the continuous spectrum. Glimpses of other lines were seen, 

 but too faint and intermittent to be measured. Of these lines 

 that with wave-length about 533 is no doubt due to oxygen, 

 for it is described as a triple line of the discharge in vacuous 

 oxygen tubes by Piazzi Smyth, and as a double line by 

 Schuster. The other two lines appear to be lines of platinum, 

 5059 and two lines 5475, 5478, according to Thalen. 



The introduction of a Leyden jar into the circuit increased 

 the brilliance both of the continuous spectrum and of the 

 lines, or of some of them, and brought out some additional 

 lines ; but the intermittent character of the sparks made it 

 almost impossible to measure the places of any of the lines. 



When the discharge from a Wimshurst machine was used 

 instead of that from the coil, only a continuous spectrum, 

 with the usual oxygen absorptions, was seen. No bright 

 lines could be detected on the continuous spectrum. 



By keeping only the lower electrode (the convex disk) 

 immersed in the liquid, so that the spark passed partly 



