Discharge from an Electrified Point. 589 



by a plate can be explained as follows : N is giving positive 

 ions to P at the ordinary high values of field at N. P starts 

 discharging : this means that a sudden supply of negative 

 ions enters the high field at N, the value of the field there 

 being considerably greater than that necessary to cause the 

 arriving ions to ionize. When the flow from P to N is con- 

 centrated by a plate behind P, this may mean a very sadden 

 increase in current from N to P which may possibly result 

 in a spark. Throughout the work with a negative P the 

 point was very unsteady, no doubt owing to this tendency to 

 spark, and attempts to find the effect of varying r have given 

 exceedingly irregular results. With each point investigated, 

 however, a similar drop in f has been observed for small 

 values of y, and it appears not unlikely that, as in the case 

 of a positive P, y is also independent of the size of P. For 

 negative discharge y is about 1*5 cms. 



Sometimes the discharge takes the "streamer" form, in 

 which there is a continuous glow stretching across between 

 the two points even when they are a centimetre or two apart. 

 The field at P is in this case greatly reduced, especially when 

 P is large. An isolated case of this effect^ which has since 

 been frequently observed, was given in a previous paper 

 (loc. cit.). 



When P is positive, sparks also pass between P and N at 

 short distances, but it has been found that the value of y at 

 which this occurs becomes smaller as the radius of N is 

 decreased. In fact the results obtained when N was a piece 

 of the finest platinum wire cut obliquely with scissors show 

 that the values of the fields f c given in the previous work 

 are not correct. These were obtained by extrapolating the 

 f Q -y curve to y = 0, N being throughout a sewing-needle. 

 The dotted curve in fig. 1 shows similar results which were 

 obtained when P was of radius *062 cm. and N '039 cm. 

 But using the above very fine N the values shown in full 

 lines in Curves 1 were obtained, and it appears not unlikely 

 from these that / and y become zero almost simultaneously. 

 The same applies to the f -y curve for a negative P with a 

 very fine N in Curves 2. 



The reason for this sparking when P is positive is also 

 clear. When y is quite small the positive ions from P at 

 first find a field at N which is larger than necessary to give 

 ionization there, so that a sudden increase in current results. 

 The phenomenon is, therefore, similar to that occurring at a 

 negative P, except that at a positive P it only occurs at small 

 values of ?/. For a very sharp point, the lines of force spread 

 rapidly and the distance y must be decreased, so that the field 



