Discharge from an Electrified Point. 591 



When P was negative the presence of N, whether dis- 

 charging or not, made practically no difference to the field 

 at P, as shown in the field-current curves 46. This is also 

 what one would expect on the above theories, if N at short 

 distances is only supplying a small proportion of the current 

 between the points. At greater distances, whatever the size 

 of N, a large supply of positive ions from it to P has no 

 effect on the field at P because that must still be the field in 

 which positive ions ionize. 



Lastly, an interesting" effect was obtained when P was 

 made a white hot platinum loop. When P was positively 

 charged f was greatly reduced, being from 40 to 70 per cent, 

 lower than that when P was cold and other conditions re- 

 mained the same. The greatest reduction occurred when y was 

 small. A much smaller reduction of from 5 to 10 per cent, 

 occurred when P was negative. This is what might be 

 expected from the well-known fact that the mobility of a 

 negative ion increases much more rapidly with temperature 

 than that of a positive ion. The fields, therefore, in which 

 these ionize will be correspondingly lower, and the decrease 

 in / will be most marked when N is supplying negative 

 ions. 



(2) Experiments on the pressure of the Electric Wind. 



It has been shown by Chattock* for various gases, and 

 later by Chattock and the author (loc. cit.), for mixtures of 

 hydrogen and oxygen, that if a point is placed opposite a 

 plate at a distance z from it, the average pressure on the 

 plate p for a given current density is a linear function of z 

 when the values of the latter are of the order of a few centi- 

 metres. If, however, these values of p and z are plotted and 

 the curve is produced to p = 0, it cuts the axis of z at some 

 positive value, referred to in what follows as z . 



For negative discharge in " pure " hydrogen z was as 

 much as 4 centimetres, but it rapidly decreased to 3 or 4 

 millimetres on the addition of small percentages of oxygen 

 or air. For positive discharge under the same conditions it 

 was 3 or 4 millimetres throughout. In air z appeared to 

 have about the same value, 3 or 4 millimetres, for both signs 

 of discharge. 



It was suggested that ~ was the average distance that an 

 ion travelled before clustering. Support for this view was 

 obtained from some experiments of Franckf when discharge 



* Chattock, Phil. Map-. [5] xlviii. p. 401 (1890). 

 t J. Frauck, Ann. d. Thys. [4] Bd. xxi. p. 084. 



