320 Dr. A. Schuster. Experiments on the 



made. I have varied the conditions of the experiments in many ways, 

 but always with the same result. When the air was sufficiently 

 oxpelled, the spark does not break through the tube at the ordinary 

 temperature ; when the mercury is heated the discharge passes, but 

 always in a perfectly continuous stream, joining the positive and 

 negative electrode. It is difficult to prevent small bubbles of mercury 

 from condensing on the surface of the glass, and the absence of dark 

 space and glow might be explained by the irregularities thus intro- 

 duced, but I think the following observations conclusive in this 

 respect. 



The mercury in the tube would always at first contain small bubbles 

 of air between it and the glass, and these could be purposely left, say, 

 on one side, by volatilising the mercnry always on the other. One 

 tube could in this way be prepared which, with a small but visible 

 bubble on one side kept down by the mercury, did not allow the 

 discharge to pass. If the mercury was heated without disturbing the 

 bubble, the current passed, illuminating the tube with a perfectly 

 continuous stream of light. As soon, however, as the bubble was 

 driven into the tube, a perfectly distinct and clear dark space became 

 visible at once. 



The following observation was also repeatedly made. As the 

 electrodes generally contain a large quantity of air dissolved in 

 them, I have been accustomed to send through my various tubes a 

 very strong discharge, using a Wilde dynamo machine as a primary 

 current. The currents thus obtained I have found very useful in 

 driving out the air from the electrodes. While the mercury was kept 

 hot a strong current was thus sent through the tubes. Instanta- 

 neously a dark space appeared surrounding the negative electrode ; 

 the appearance presented was that of a discharge through mercury 

 superposed on one through air, and whether these discharges passed 

 through the tube simultaneously or one after the other, the result 

 proved that the dark space and glow can and generally do appear 

 under the exact conditions under which they are absent in mercury 

 vapour. The strong current of the dynamo machine has generally 

 been fatal to the life of the tube, and I am not sure that I have ever 

 been able to examine the discharge from electrodes really free of air. 

 To that small remnant of air it is due, perhaps, that the discharge, 

 though continuous, was seldom quite symmetrical with respect to the two 

 poles. There generally seemed to be a tendency to surround the nega- 

 tive, and to set out only from the point of the positive electrode. But 

 the appearance was much disturbed by globules of condensed mercnry 

 on the electrodes ; sometimes both electrodes were equally surrounded 

 by the discharge, and sometimes the latter would pass only from the 

 points. 



[There are, however, other causes at work which tend to produce 



