﻿Ionization by Collision in Helium, 841 



The helium upon which the experiments were eventually 

 made was kept in the purification chamber for a considerable 

 time. There is no doubt that the helium was rendered very 

 pure by this process, a fact which was confirmed by spectro- 

 scopic observation. 



The impurities due to the form of the apparatus will now 

 be discussed : — In the first place a reference to the figure will 

 show that mercury vapour from the pumps and gauge must 

 be present in the bell-jar; but the amount, owing to the low 

 vapour-pressure of mercury, is certainly small, and an experi- 

 ment was made in which the gas entered the bell-jar through 

 an additional charcoal tube in liquid air of the U-form inserted 

 between the bell-jar and the parts of the apparatus containing 

 mercury. The vapour was thus unable to reach the bell-jar, 

 but no variation in the results obtained previously without 

 this precaution could be detected. The effect of the mercury 

 vapour seems therefore inappreciable. 



Secondly, the apparatus was not quite (though very nearly) 

 air-tight. The leak, which was tested over about a month, 

 was uniform and caused a rise in pressure of g 1 ^ mm. 

 of mercury a day. An experiment to find a and /3 took 

 usually a little less than an hour, and therefore the pressure 

 of the impurity present at the end of such a set was about 

 ^oo mm * 5 ^he lowest pressure of helium in an experiment 

 was 1*5 mm., and therefore the maximum impurity present 

 was not more than *03 per cent. 



In addition there were present vapours given off by the 

 cement and vaseline used in the bell-jar. From the account 

 given of the electrical arrangements it appears impossible to 

 eliminate these ; for instance, a quartz window had to be 

 affixed and the various connecting wires had to be taken out 

 of the jar. It is probable that these impurities are very 

 small ; the apparatus was exhausted for about six weeks 

 before the helium was admitted, so that the more volatile 

 gases would probably all have been given off ; and certainly 

 any residue was not enough to show on the McLeod gauge, 

 which was read to ¥ J-q mm. 



Finally it was necessary to eliminate impurities given oft* 

 from the plates and the glass by the electric discharge ; the 

 small currents in the gas of the order which affect an electro- 

 meter do not cause any trouble, but sparking at once brings 

 off gases from the electrodes and the glass, and after a spark 

 had been passed no observation was ever taken till the helium 

 had again been purified. The electrodes were prepared by 

 sparking between them for a very long time at a low pressure 

 before the helium was admitted to the apparatus at all. 



