454 The Movement of Gas in " Vacuum Discharges." [Mar. 30, 



pressure, although neither was observed except when the pressure was 

 such as to approach the stage when Crookes' phosphorescence was 

 produced. 



These phenomena also reproduced themselves in another way. Some 

 tubes, after having been completed and taken off the pump, showed a 

 decreased pressure after a prolonged passage of a strong current, 

 others an increased pressure, but among both classes tubes were not 

 unfrequently found which recovered their original pressure after a 

 period of rest or cessation of discharge. 



Matters remained in this rather confused state until we observed 

 with more care than before a tube of which the exhaustion was near 

 the phosphorescent state, and of which both terminals were metallic 

 cones, and consequently presented large surfaces for any action which 

 might take place upon them. 



In what may be considered to have been its normal condition, this tube 

 showed three or four large white striae with a dark space of con- 

 siderable size round the negative terminal. On passing the discharge 

 through the tube for some minutes the dark space increased, the 

 striae became fewer and feebler in illumination, the green phosphores- 

 cence began to show itself, and the discharge showed the usual signs 

 of reduced pressure. On suddenly reversing the current the striae 

 became again more numerous and more brightly illuminated, precisely 

 as they would be by an increase of pressure, while the other features 

 of the discharge in a great measure resumed their original character ; 

 and not only so, but by a comparatively slow process, occupying many 

 seconds in duration, the indications of increasing pressure continued 

 still further, until they implied a pressure even beyond that at which 

 the tube stood when the experiments began, after which the appear- 

 ance slowly changed as before in a manner indicating reduced 

 pressure. This reversal of the discharge was repeated many times 

 with the same result in every case. The amount of change in 

 pressure indicated by the appearance on each reversal was found to 

 depend within wide limits upon the duration of the previous discharge, 

 or, what is the same thing, upon the amount of depression below the 

 normal pressure indicated by the previous discharge. 



The most probable explanation of these phenomena appears to be this, 

 that the effect of the discharge is actually to alter the pressure in the 

 tube, not by any modification in the chemical composition of the gas, still 

 less by anything that could be represented as a destruction of matter, 

 but simply by driving occluded gas out of one terminal, and by drawing 

 it in, or occluding it, at the other. On reversing the discharge, the 

 operation is reversed, and the occluded contents of one terminal are 

 thrown along the tube to be occluded at the other. This view of the 

 mechanism whereby the observed phenomena are produced is supported 

 by the absence of these appearances when the terminals are compara- 



