UPON ELECTPvICAL DISCHARGE PHENOMENA IN RAREFIED GASES. 
145 
It is obvious, however, that the above observations can only refer to a particular 
bulb and a certain set of conditions. The principal difficulty lay in determining tlie 
conditions. Ileadings taken at the beginning of the week could not be accurately 
repeated at the end of it, although the curves plotted from such records were in the 
main parallel. Tliose shown in fig. 7 were obtained with oxygen as the residual gas 
within the bulb, and are typical of the residts obtained with a particular bulb, 
irrespective of whether the residual gas was air, hydrogen, carbonic dioxide, or 
oxygen ; the only difierence in all the various cases being a shifting, parallel to itself 
of the complete curve. 
Reference has already been made to the Interval which occasionally elapsed between 
the excitation of the magnet and the appearance of a luminous ring within the hull). 
A similar pause was noticed in connection with experiments upon the diselectrification 
effects. It should also he pointed out that the sign of the electrification of the bulb, 
both in the case of the luminous-iing experiments and those relating to diselectrifica¬ 
tion, was invariably negative through tlie action of the magnetic field. Moreover it 
has already l)een seen that luminous flashes usually accompanied diselectrification. 
The rpiestion as to tlie origin of the luminous ring appeared therefore to he intimately 
involved in any explanation that might be found to account for the diselectrification 
phenomenon, and it became important to see finally whether conditions favourable to 
the one hut unfavourable to tlie other could he found to exist simultaneously. 
A general survey of the results up to this stage pointed to a concentration of 
negative ions at the centre of the bulb, and a series of independent experiments were 
tlien carried out in order to ascertain what was the distribution of charged gas 
within the bulb at the moment the induction coil discharge ceased to pass through it. 
For instance, the action of the magnetic field upon the residual gas particles was 
examined not only suhseipiently hut also during the passage of the discharge. A 
screw thread was cut upon each electrode, so that liy its means the uniform green 
fluorescence of tlie glass might appear as a luminous sjiiral and show, by widening 
out or twisting, the paths of the electrified particles sliot olf from the metal. The 
action of the magnetic field upon jets of gas"^ which, at tlie lowest pressures and even 
after continued exhaustion, were found to he given off from metal exposed within the 
bulb, was also examined. The results of tliese experiments are embodied in fig. 8, 
which shows the probable distribution of the electrified gas to he such that the central 
portion of the bulb was fairly uniformly filled with negative ions, while a layer of 
positively electrified gas resided upon the interior surface of the glass : the former 
being indicated by the small dots and the latter by the larger circles. The diagram 
represents the state of aftairs at the moment an oscillatory discharge had passed 
through the bulb, and I desire to call attention to the accumulation of negative ions 
at either end. 
With the object of ascertaining whether these clouds of negative ions were aftected 
* ‘ Electrician,’vol. 41, p. 425; ‘ B. A. Report,’ 1900, p. G39. 
VOL. CXCVIT.—A. 
U 
