The Discharge of Electricity through Gases. 533 



I can see, they present no special features which are not easily 

 explained by well-known facts of vacuum discharges. Owing to the 

 incandescence of the filaments, the different potential at the ends is 

 sufficient to produce a discharge through the residual gas ; hence the 

 observed disintegration of the negative end of the filament. If a 

 separate metallic plate is inserted into the lamp, its potential will be 

 intermediate between that of the positive and that of the negative 

 pole. On being connected with the positive pole a current will tend 

 to pass in such a direction as will make the plate an anode. The 

 current actually will pass, because we know that a conductor offers 

 no difficulty to the discharge of positive electricity into a gas in which 

 the discharge is established. This phenomenon has been called the 

 "Edison effect." If the plate is connected with the negative elec- 

 trode, it will tend to become a kathode. The fall of potential 

 required is now much greater, and hence the current is much weaker, 

 and may escape detection.* 



According to the theory of electrolytic convection, we conclude 

 that when an electrode is raised to a red or white heat the molecules 

 partially dissociate and render conduction possible. It does not 

 necessarily follow that the temperature alone is sufficient to dissociate 

 the molecules; it maybe that the addition of an electric stress is 

 necessary. It will be a matter for future investigation to decide 

 whether the decomposition goes on at the same rate in contact with 

 a negative wire as it does in contact with a positive wire. 



2. Behaviour of Electrodes illuminated by Ultra-violet Light. — Hertz, f 

 during his celebrated experiments on electric oscillations, noticed that 

 a spark passed more easily between metallic points when these were 

 illuminated by a simultaneous strong spark from another source, and 

 traced the effect to the illumination of the secondary electrodes by 

 the ultra-violet rays sent out by the primary spark. Messrs. E. 

 Wiedemann and EbertJ found that the negative electrode only is 

 active in this case, and Hallwachs§ subsequently attacked the question 

 by a method which has yielded very interesting results. 



He connected a gold leaf electroscope with a clean plate of zinc, 

 upon which a beam of strong ultra-violet light was allowed to fall. 

 The zinc was found to be incapable of retaining a negative charge ; 

 the effect on a positive charge was so small that at first sight it 

 seemed totally absent. But the fact to which I wish to draw special 

 attention is the importance, in order to ensure the success of the 

 experiment, of taking a clean surface of zinc. Hallwachs found that 



* [All the facts described by Mr. Fleming in his recent communication (' supra, 

 p. 118) will be found to agree with this explanation. — May 30.] 

 f 'Wiedemann, Annalen/ vol. 31, p. 983 (1887). 

 J ' Wiedemann, Annalen,' vol. 33, p. 241 -(1888). 

 § 'Wiedemann, Annalen,' vol. 33, p. 301 (1888). 



