1886.] Passage of an Electric Discharge through Nitrogen. 337 



produced by the gas we are testing. We found in this way that our 

 gas certainly did not contain one part of oxygen in 500, and probably 

 not one part in 1000. 



The Experiments. 



A tube carefully prepared and sealed off in the way we have already 

 described was taken, and after the cathetometer readings had shown 

 that the pressure had remained constant for several hours, it was 

 sparked through, generally with an induction coil. In order to get 

 the effect we are describing, it is necessary to introduce a large resis- 

 tance into the circuit, for this purpose we used a piece of wetted 

 string, this makes the discharge through the tube much less intense 

 and the heat produced comparatively small ; we were not able to get 

 the effect when the discharge passed straight through the tube 

 without any resistance beyond that of the tube and the connecting 

 wires. The fact that heat restores the gas to its original condition is 

 sufficient to account for this, for when there is only a small resistance 

 in the circuit, the heat developed in the tube is much greater and 

 the tube becomes very hot. We noticed a similar thing when we 

 used a Holtz machine instead of a coil : if we charged up five large 

 Leyden jars with the Holtz and then discharged the jars through the 

 tube, no permanent alteration in the pressure was observed ; if, how- 

 ever, we never allowed the jars to get fully charged, but sent a 

 succession of small sparks through the tube, then a permanent 

 diminution in the volume of the gas took place. 



When the discharge from the coil with a piece of wet string in the 

 circuit went through the tube, a slow diminution in the volume of the 

 gas took place ; the rate of diminution diminished as the sparking went 

 on and ultimately the permanent volume of the gas remained un- 

 affected by the passage of the sparks. It took, however, a considerable 

 time to reach this state ; as a rule each tube was sparked through for 

 between three and four hours on each of three consecutive days, the 

 diminution in the volume at the end of the first day was about two- 

 thirds of the maximum diminution, there was a diminution of about 

 one-half of this on the following day, and no appreciable diminution on 

 the third day. 



The gauges of the tubes used at first were filled with sulphuric acid, 

 and after the discharge had passed through until the pressure had 

 become steady, it was always found that the sulphuric acid in the 

 limb of the gauge next the tube had risen towards the tube, showing 

 that the pressure exerted by the gas in the tube had diminished. 

 The difference of level between the legs of the gauge increased in the 

 case of five different tubes by from 4*5 to 7 mm. ; now the pressure in 

 the tube was originally 8 mm. of mercury, or about 58 mm. of H 2 S0 4 , 

 so that the diminution in 'the pressure exerted by the gas is from 



2 A 2 



