540 Prof. J. J. Thomson on the Relation between the 



most pronounced the more electropositive the metal, an 

 electrolyte, on the other hand, when exposed to ultra-violet 

 light does not lose its charge, whether this be positive or 

 negative. 



Another example of the discharge of electricity by the 

 second process, is that in the glow-discharge from a metal 

 where the temperature is low the negative discharge occurs 

 more easily than the positive, i. e. the potential difference 

 between the point from which the electricity escapes and the 

 nearest large conductor is less when the point is negatively 

 electrified than when it is positively. It would be interesting 

 to see whether this difference would still persist if the point 

 from which the discharge takes place were made of an electro- 

 lyte and not of metal. On the hypothesis we have just been 

 describing, we should expect the difference to be much reduced 

 by the substitution of an electrolyte for the metal. Some 

 recent experiments by Prof. S. P. Thompson show that the 

 conditions of discharge are modified when lead peroxide is 

 substituted for metal. 



When a glow-discharge takes place from a metal point 

 connected directlv with a Ruhmkorff coil or a Wimshurst 

 electrical machine the surrounding gas gets negatively elec- 

 trified. This, however, is not always the case if we connect 

 the point from which the discharge takes place to an instru- 

 ment producing potential differences greatly exceeding those 

 given by a Ruhmkorff coil. A very convenient form of such 

 an instrument is the one described by Prof. Elihu Thompson, 

 which is a transformer whose primary is a coil connecting the 

 outside coatings of two Leyden jars, while the secondary is a 

 coil with a large number of turns carefully insulated from the 

 primary. When the oscillating current produced by dis- 

 charging the jars passes through the primary, the electro- 

 motive force induced in the secondary is able to induce 

 sparks of great length. Using a transformer of this kind 

 instead of the Ruhmkorff coil, Harvey and Hird (Phil. Mag. 

 vol, xxxvi. p. 45) and Himstedt (Wied. Ann. Hi. p. 473; 

 found that when the point discharged in oxygen, the gas 

 was positively electrified, while when it discharged in hy- 

 drogen, the hydrogen was negatively electrified, just as 

 when the Ruhmkorff is used. The very intense field pro- 

 duced by the transformer would split up the molecules of the 

 gas and promote direct combination between the metal and 

 the gas ; this, as we have seen, would cause the gas to be 

 positively, the metal negatively electrified, and this, as we 

 have seen, is what happens in the case of oxygen, though not 

 in that of hydrogen. In the case of the electrification of the 



