30 



ML'. O. \V. IIICIIAUDSON ON TIIK IOXISATION 



tubes of aluminium. The two outside tubes acted as a guard ring, so that the leak 

 was only measured from the uniformly heated central part of the wire to the 

 surrounding middle cylinder. In addition to this, all the insulation was surrounded 

 by earthed tubes, so that there was no possibility of it becoming charged up by strong 

 ionisation. The effects previously observed were fully confirmed with this apparatus. 

 A new wire, not specially cleaned, placed in this apparatus gave with + 40 volts a 

 leak of 100 divisions which remained constant for 1 hour 40 minutes. On raising the 

 potential to +760 volts the following values of the current were obtained at the 

 times stated : 



On returning to +40 volts, the leaks at successive six-minute intervals were 

 80, 84, 90, 94. The temperature in this experiment was 925 C. A similar experi- 

 ment with a hot platinum tube 2 millims. in external diameter showed the same 

 kind of effect, and also an increase in the positive leak, when the tube had been left 

 negatively charged. 



The obvious conclusion from these experiments is that part of the leak from a new 

 wire is emitted by something which is itself positively charged, and is therefore 

 removed when the wire is positive, but not when it is negative. Dust given off by 

 the platinum might be expected to become positively electrified by virtue 'of point 

 discharge action in the strong field in the neighbourhood of the hot wire whatever 

 the sign of its electrification when it left the wire. It would thus be retained by a 

 negatively, but not by a positively-charged wire. On the other hand, there appears 

 to be no obvious reason either why dust should produce the observed effects or why a 

 new wire should give off more dust than an old one. Blowing a current of air past 

 the wire had no marked effect on the positive leak under a high voltage. The general 

 complexity of these initial phenomena would seem, however, to indicate that there is 

 more in them than the above simple explanation demands. 



What the process of ageing a wire for the purpose of this investigation consists in 

 is not quite clear. Mere heating alone will not do it. A new wire, which was heated 

 lor about a week in the air of the laboratory before being tested, was found to behave 

 like a new wire when the leak was tested. The above experiments would seem to 

 indicate that tlie presence of a big positive charge on the wire is instrumental in 

 Bwifitatifig the decay of the initial leak, whilst, other things being e|iial. the rate of 

 decay is greater at low than at high pressures. It looks as though the initial leak 



