640 REPORT — 1900. 



work with the following observations as to the cause of the phenomenon. The 

 experiment just referred to consisted in usino; a pivoted disc of alumimum as the 

 negative electrode in a bulb containing rarefied air, so that when green patches 

 appeared a rotational movement of the disc (actuated by means of an external 

 magnet) would show whether the patches of green moved in a corresponding 

 manner or not. It was seen that they did so move. 



A distinct feature of the experiment, however, consisted in ascertammg 

 whether those flecks which persisted after the discbarge had ceased were still 

 sensitive to movements of the cathode. This also was found to be the case. The 

 proof, therefore, that the green patches were associated with an emission from the 

 cathode appeared complete. j, ,. , , 



But it was further noticed that subsequently to the passage ot a discharge, and 

 even when either or both of the electrodes were connected to earth, still the green 

 flecks were easily visible upon the glass walls of the bulb, and continued to move 

 as before when the cathode was rotated. 



This was apparently a case in which cathode rays were emitted from an 



electrode at zero potential. „ , . ^ , , , n- i 



I do not know that any explanation of this effect has so far been otiered, and 

 I therefore venture to bring forward the following suggestions, supported by 

 farther experiments. , ^ , 



When a piece of metal is placed in a rarefied atmosphere and made the nega- 

 tive pole for an electrical discharge passing across the attenuated gas, innumerable 

 small bricrht specks of light appear over the surface of the metal. It was found 

 convenient in my particular case to use an iron electrode for _ observations of this 

 eifect, because it had the advantage of being readily magnetisable from without. 

 The addition of a similar iron electrode to act as the anode determined the shape 

 of the magnetic field. With such an apparatus it could be seen that the bright 

 spots appeared principally upon the cathode while the discharge passed, and that 

 tlie creation of a magnetic field between the electrodes made visible the individual 

 luminous streams of gas emanating from those tiny points of light. The paths of 

 these luminous streams, becoming bent by the action of the magnetic field, followed 

 the direction of the lines of magnetic force and exhibited a tendency to become 

 spiral in accordance with well-known laws. In this way a fine layer of sodium 

 upon the anode was caused to fluoresce through the action of the negatively 

 electrified particles beating down upon it. One was able, in fact, by this means 

 to cast shadows of objects placed in the paths of the bent streams, and at a pres- 

 sure considerably higher than that necessary for the production of the well-known 

 cathode shadow effects. In the above case, however, the method served to clearly 

 establish the fact that the bright points of light upon the surface of the cathode 

 indicated the places from which the jets of gas originated.^ 



The number of these jets became less as the fxhauftion was continued, and 

 individual streams were very clearly seen owing to the action of tlie magnetic 

 field. When the discharge began to cause fluorescence in the glass of the bulb 

 some green patches made their appearance, and, in some cases, when the cathode 

 was magnetised a bright spot, which was previously judged to indicate the origin 

 of a o-reen patch upon the glass, would shift to a new position upon the electrode. 

 In all such cases the corresponding patch also moved in a similar manner upon 



the glass. , , . , 



At sufficiently high exhaustions the bright spots upon the electrode disappeared 

 euti'relv, although green flecks were still visible upon the glass. But a movemen'-. 

 as before of the "electrode as well as its magnetisation gave results consistent with 

 previous observations. , , , ,, , ;, 



Under these conditions, when the discharge through the bulb was stopped 

 ■-'roen flecks were still visible for about ten seconds. 



° A positively charged body was brought up to the outside of the bulb and the 

 patches brightened considerably. A negative charge similarly placed extin- 

 guished the flecks completely. 



Finally, vacuum bulbs, such as X-ray lamps, &c., exhibiting green flecks while 

 in operation, were generally found to deteriorate if laid aside for a mouth or two, 



