586 Mr. A. M. Tyndall on the 



ordinary discharge with glow sets in. For both signs of 

 discharge the supply of positive ions is kept up by ionization 

 due to negative ions, these having been produced by pre- 

 viously formed positive ions and so on. Both signs of ion 

 have, therefore, to be able to ionize as each produces the 

 other. Since positive ions require a stronger field for this 

 than negative, it is always the field required by the positive 

 ions which has to occur at the point P. 



Suppose now that ions of opposite sign to the charge on P 

 are supplied to it from an external point N in its neighbour- 

 hood. When P is negative the field necessary for glow at 

 its surface is UDaltered. 



When P is positive, the supply of negative ions, which is 

 in other cases kept up by ionization near P by positive ions, 

 is now continuously supplied from without. Glow therefore 

 appears at P when the field near it has reached that in which 

 negative ions will ionize. This field (/— ) is less than that 

 (/4-) in which positive ions will ionize. 



Experiments on the pressure of the electric wind further 

 suggested that the ions streaming away from a glowing point 

 are not at first fully formed, and travel a distance from the 

 point varying from 4 centimetres in the case of negative 

 ions in pure hydrogen to 3 or 4 millimetres in air, before 

 they completely cluster. 



Now unclustered ions ionize in a lower field than ions 

 fully grown, so that, as N is made to approach P, the field for 

 glow at the latter begins to drop when the distance between 

 N and P is decreased to values less than this critical distance. 

 Thus when P is positive and N very close to it, P is sub- 

 jected to a corpuscular bombardment and its field sinks to 

 that (/ c ) necessary for corpuscles to ionize. 



Lastly, when the current is small, ions oP one sign only 

 traverse the greater part of the distance between point and 

 plate, except in negative discharge in pure hydrogen. 



In the light of the experiments described below it appears 

 that certain parts of this theory must now be abandoned. 

 Thus it will be shown that except in the case of negative ions 

 in certain oxygen-free gases the growing ion theory presents 

 very serious difficulties, and that effects originally attributed 

 to unclustered ions are better explained on the view that, 

 accompanying the process of ionization by positive ions, 

 there is an emission of rapidly moving neutral bodies which 

 have a range of a centimetre or tw T o at atmospheric pressure. 



The statement that in air for small currents, ions of one 

 sign only are present between the glow and the point and 

 the plate is probably also not correct. It seems likely that 



