48 Brush Discharges in Gases, 



We have also experimented with other gases, but owing to 

 difficulties in obtaining them in a state of purity, on account 

 of limited space and want of general conveniences, we have 

 not obtained reliable results. 



The results given above, however, may be depended upon, 

 as they were all repeated some hundreds of times under 

 varying conditions, and have always been satisfactory and 

 consistent. 



We may therefore sum up the results which we have 

 established as follows : — 



In a brush-discharge in air positive electricity passes more 

 readily than negative from a point onto any neighbouring 

 conductor. 



In oxygen the same is the case. 



In hydrogen the reverse holds, negative electricity passing 

 more readily. 



To a certain extent these results are anticipated by previous 

 experiments. It has long been known that positive brushes 

 were somewhat stronger than negative. Again, Guthrie has 

 shown that a red-hot conductor in air will retain a negative 

 but not a positive charge (Phil. Mag. vol. xlvi. p. 257, 1873). 

 This experiment, as having a bearing on our subject, we have 

 repeated and confirmed. 



In conclusion we may mention a few precautions which 

 must be taken in conducting these experiments. 



(1) The arrangement of two condensers in series was 

 suggested by Prof J. J. Thomson in his paper on " Discharges 

 in Gases without Electrodes/'' We have found it a very 

 necessary precaution : if the discharge of only one condenser 

 is used the unsymmetrical arrangement of the apparatus leads 

 to very confusing results. 



(2) The greatest care must be taken with the insulation of 

 the transformer, both in insulating the primary from the 

 secondary and also in insulating the turns of the secondary 

 itself. 



(3) It is advisable to prevent all brush-discharges except 

 at the point where they are to be observed. This, as far as 

 we know, can only be done by immersion of every part of the 

 conductor in oil. 



(4) It is necessary to use the gases in large bulk or to keep 

 up a continuous flow, as otherwise they acquire a charge 

 which confuses the results. 



(5) The distance between the point and the conductor which 

 is to be charged must be so adjusted that no spark passes. 

 If a spark does pass, the conductor is entirely discharged. 



