Discharge from an Electrified Point. 597 



field, therefore, at a point would be that in which negative 

 ions will ionize, and this does not seem to be the case. 



This doublet theory is thus consistent with all the experi- 

 mental facts above stated. 



Changes of Wind Pressure icith Current. 

 There is still the fact to be explained that in the wind- 

 pressure work the values of -~- per unit current decrease 

 with increasing values of current. 



The fact that -f- is constant for values of z included in 



dz 



the centre portion of Curve 5 implies that the carriers of the 

 current are not generated in the main body of the gas, but 

 travel a distance which does not differ very appreciably from 

 the whole distance z between the point and the plate. * 



The observed fall in -^ per unit current might occur if 



the current is carried by two kinds of ions, clustered and 

 unclustered, both travelling from point to plate and varyino- 

 in relative numbers with the current. Although, as shown 

 below, such a view is possible for negative discharge in 

 hydrogen containing traces of oxygen, it does not seem 

 probable in air*. 



The simplest explanation is that the current between the 

 point and the plate is not wholly carried by ions of one si on 

 only, but that there is a certain amount of back discharge 

 from the plate which at very small currents is negligible in 

 amount, but which increases as the current and the field at 

 the plate increase. 



That such a back discharge can exist in air, even at low 

 currents, is shown by the part BC of Curve 5. In this 



region there is a rapid fall in -+- . At a current of 1 micro- 

 ampere the value of z at which the curve leaves the straight 

 is about ti'O centimetres, and at 10 microamperes it is about 



* It is true that in a recent paper (Phil. Mag-. Feb. 1911) Sir J. J. 

 Thomson has shown that positive ions 3 and 6 6 are present in the 

 positive rays at low pressures. It is, therefore, conceivable that these 

 and similar clusters with negative charges may be present at atmospheric 

 pressure in varying amounts depending on" the value of the current 

 flowing from the point ; if they carry an appreciable percentage of that 



current, the observed changes in -^ may thus be produced. 



