n;oi>rn:i> i;v nor IM.ATINTM IN I>II-TI;I;KNT GASKS. 



oxygen and nitrogen act quite independently of each other in producing the positive 

 ionisation reasons will be adduced later for showing that this is improbable the 

 proportion of oxygen to nitrogen required to account for the whole of the leak in 

 nitrogen at atmospheric pressure would be about 1 to 10,000. Against this it is 

 necessary to set the following facts : 



(1) The ions in nitrogen appear to be considerably bigger than those in oxygen, as 

 shown by the current E.M.F. curvea 



(2) The experiments on air indicate that nitrogen does not simply act as a diluent 

 to oxygen, but exerts a marked retarding effect on its power of producing positive 

 ionisation, so that a trace of oxygen present would exert a much smaller effect than 

 that obtained on the basis of the above calculation. 



(3) The positive ionisation in nitrogen varies more rapidly with temperature than 

 that in oxygen. The numbers indicate that the energy necessary to liberate a 

 positive ion from hot platinum in nitrogen is rather more than twice the value for 

 oxygen. 



On the whole, the balance of evidence seems to be in favour of the view that 

 nitrogen produces a genuine positive ionisation which is considerably smaller than 

 that in oxygen at alxmt 900 C. 



V. 12. THE IONISATION IN AIR. 



The air used in these experiments was purified by passing slowly through two 

 tubes, one packed with cotton wool and the other with fused caustic potash. It was 

 suKsequently dried over phosphorus pentoxide inside the apparatus. The general 

 characters of the results in air are similar to those in oxygen, but the current was 

 found to be more difficult to saturate under otherwise like conditions. 



The relation between the current and the electromotive force at atmospheric 

 pressure is given by the lower curve in fig. 12. The observations to which both the 

 curves in fig. 12 refer were made with the water-cooled apparatus already described, 

 The temperature was 976 C., the effective area of the wire 0'170 sq. centitna, 

 .-iml the unit of current 7xlO~ ls ampere. The upper curve gives the value of the 

 saturation current at different pressures. In this curve the leak at the four highest 

 pressures \\.is measured with +704 volts, the next with + 5 12 volts, and the two 

 lowest with +120 volts. This method of gradually decreasing the voltage as the 

 pressure was ln\\ered, so as to ensure approximate saturation and at the same time 



