Light from Mercury Vapour. 63 



as to deflect any ions which come into this region. When 

 these electrodes are uncharged, the luminosity spreads out 

 from the opening b in a symmetrical manner, becoming 

 fainter toward the left of the drawing. When an E.M.F. 

 of 160 volts was applied between g and h, the position of 

 the luminosity was shifted. 



In all of the photographs shown by Strutt there was an 

 increase of the luminosity in the neighbourhood of the positive 

 electrode. In the majority of the photographs there was also 

 an increase in the light near the negative electrode. According 

 to Strutt, the positive and negative ions are separated by the 

 field, and the luminosity near the positive electrode is caused 

 by the negative ions alone, and that near the negative electrode 

 by the positive ions alone. 



As has often been shown, it is impossible to separate 

 positive and negative ions by this method unless there is a 

 very high vacuum on account of the formation of new r ions. 

 Thus in the stream of gas coming from A there are electrons, 

 and since the potential difference between g and h is 160 volts, 

 the electrons will be drawn toward the positive electrode with 

 sufficient velocity to ionize by impact any molecules which 

 may be present. The recombination of the ions thus formed 

 will then give out light. That is, a second discharge is pro- 

 duced between g and h } giving out light as any other discharge 

 through a rarefied gas gives out light. In fact, the potential 

 difference is sufficiently large here to make it possible for the 

 positive ions as well as the negative ones to ionize by impact, 

 so that new ions may be formed both in the region of the 

 positive and in that of the negative electrode. 



The only wa}^ to avoid such action is to obtain a gas so 

 rarefied that there are no molecules present. Unfortunately, 

 no measurements were made of the pressure of the gas in 

 Strutt' s experiment; but it is evident that there was a sufficient 

 number of molecules present to allow of ionization by impact, 

 since an arc was apt to start between g and h when the 

 voltage was raised above 160 volts (p. 911). It is impossible 

 to have an arc in a region where there are no molecules, for 

 no ions can be formed without molecules from which to form 

 them. This is shown, for example, in the Coolidge X-ray 

 tube, where the voltage may be several thousand volts without 

 starting an arc. 



Furthermore, it was found that there was the greatest 

 tendency to start an arc when the gas was least dense. If 

 there had been very few molecules present, the lower the 

 density of the gas the less possibility there would have been 

 of the formation of an arc. 



Again, to start an arc there must be ionization by the 



