424 Trowbridge — Spectra of Gases. 



had been changed. This holding of the mercury meniscus was 

 large and had to be overcome by vigorous tapping of the tube. 

 An analogous effect was obtained when the Geissler tube was 

 filled with rarified air, and also when it was filled with nitro- 

 gen. When, however, it was filled with dry hydrogen, the 

 holding effect was comparatively inappreciable. The oxygen 

 produced by the dissociative effect of the electric discharge 

 combined with the hydrogen and no longer oxidized the sur- 

 face of the mercury. In this connection it may be observed 

 that the mercury meniscus in the Lippman electrometer is 

 affected principally when it is made the positive pole, and, 

 therefore, oxygen is liberated. 



Perhaps the most striking experiment in this connection can 

 be made with the steady current from a large storage battery. 

 When Geissler tubes, preferably of half a centimeter internal 

 diameter, are provided with copper terminals, and are filled 

 with dry hydrogen at pressures of one millimeter to one-tenth 

 of a millimeter, a steady diminution in the pressure of the gas 

 results from the application of the discharge ; the light of the 

 spectrum grows dimmer and dimmer, then the cathode rays 

 appear, finally the X-rays, and then no discharge can be forced 

 through the tube until a much higher electromotive force is 

 employed, or heat is applied to the tube. This heat evidently 

 drives off water-vapor from the walls of the tube together 

 with air ; a fresh application of the steady current again dimin- 

 ishes the pressure in the tube to an apparent vacuum. Thus 

 one can exhaust, so to speak, a Geissler tube by employing a 

 steady current of electricity to dissociate the ever present 

 water-vapor. With copper electrodes, the oxidization pro- 

 duced by this dissociation is more evident than with the other 

 metals ; although I have observed it with magnesium termi- 

 nals, with iron terminals and with other metals. 



These experiments lead me to believe that, just as in chem- 

 ical reactions, a certain amount of water- vapor or humidity is 

 essential to conduction in gases whether brought about by 

 what is called chemical affinity or electrolytic action. 



I have dwelt upon the broadening of the lines of metals in 

 capillary tubes. This phenomenon is also observed with hydro- 

 gen lines, and was first noticed by Liveing and Dewar, Chem. 

 News, xlvii, p. 122, 1883. These authors attributed the 

 broadening to compression of the gas in the narrow capillary 

 under the effect of a powerful condenser discharge. Their 

 method of experiment was as follows : The tube was exhausted 

 only to perhaps five or six centimeters pressure, so that a white 

 discharge of a spark nature passed through the capillary and 

 then spread out to electrodes placed in the large ends of the 

 tube. When the tube was viewed end-on, a continuous spec- 



