Discharge from cm Electrified Point. G03 



number by clustering with oxygen before hydrogen can be 

 suitably encountered. The curve of combination per coulomb 

 and percentage of oxygen will thus pass through a maximum 

 value ; this was found to occur experimentally, the combina- 

 tion per coulomb being greatest at 0'003 per cent, oxygen, 

 and being then three times greater than at 14 per cent. 

 oxygen. 



There was apparently a much smaller maximum in positive 

 discharge ; this also is reasonable because the main oxygen 

 layer is then at the negative plate, where the number of free 

 corpuscles is very much less, if not nil. 



Summary. 



1. Further measurements of the field at the surface of a 



discharging point have been made, and it is suggested 

 that some of the observed effects, originally attributed 

 to the action of unclustered ions, are due to an 

 emission of uncharged doublets during ionization by 

 positive ions. 



2. When extrapolated, the curve of wind pressure on a 



plate (pj and the distance between point and plate (z) 

 cuts the axis of z at a positive value z . z is inde- 

 pendent of the field at the point surface and of the 

 current. It is in general about 4 millimetres in 

 length, but is much greater in negative discharge in 

 pure hydrogen. The effects in air are explained by 

 the doublet theory, and in hydrogen by the view tbat 

 the negative ions in that gas are, in the main, cor- 

 puscular in nature. 



3. The apparent velocities of the ions in air, as measured 



by the wind-pressure method, become greater as the 

 current between point and plate increases. This is 

 explained by the presence of a back discharge from 

 the plate, which prevents the method from being 

 applied with accuracy except when the current is 

 small. 



4. Evidence is adduced in support of the view that the 



positive ion in pure hydrogen is a charged molecule. 



5. The amount of combination which point discharge pro- 



duces in hydrogen containing different percentages 

 of oxygen and nitrogen, and the form of the wind- 

 pressure curves in these mixtures are discussed and 

 explanations are offered. 



2 R 2 



