Discharge from an Electrified Point. 595 



opposite a plate is the true ionizing field for unclustered 

 positives, and that the effects of external ions at short dis- 

 tances must be explained in some other way. Then when N 

 supplies negative ions from a distance greater than y , the 

 field at P is /— . One would expect, however (though not 

 with certainty), that f — would be greater than the ionizing 

 field for unclustered positives, whereas experiment shows it 

 to be considerably less. 



Suppose, on the other hand, that the positive ion at birth is 

 already of molecular magnitude, and that the negative ion 

 is the only one to go through a clustering stage. This would 

 explain the effects at a positive P for all values of y. In fact, 

 in the previous paper (Phil. Mag. xx. 1910) the possibility of a 

 positive clustering ion was ignored from lack of experimental 

 data. The theory, however, would not explain (1) why the 

 field at a negative P decreases as N is brought within y , and 

 (2) why there is a z wind effect in positive discharge. 



On these grounds the author considers that, at any rate 

 for air, the above growing ion theory must be abandoned. 

 If in air the negative ion is at birth a corpuscle, it must 

 immediately cluster by taking on an oxygen molecule or 

 molecules. Further evidence for this view is given below. 

 With regard to possible alternatives the following considera- 

 tions present themselves : — 



1. A z effect will be present if there is a large amount of 

 ionization occurring within a few millimetres of a discharging 

 point. Such ionization may occur if the point is emitting 

 radiations with powerful ionizing properties, the resultant 

 positive and negative ions giving mutually neutralizing wind 

 pressures. On this view, it is not clear why the field at P 

 should fall when brought close to a discharging N, because 

 although ions are being produced close to P by the radiations 

 from N, the field at P before it can discharge must be such 

 as to make these ions produce more. It is true that radiations 

 falling on the surface of a negatively charged P may give 

 photoelectric effects, and a supply of negative ions may thus 

 be obtained at a low field, but to fit the facts there must be 

 similar effects at P when it is positively charged. Also 

 it is probable that any such photoelectric currents will be 

 negligibly small compared with the currents used in the 

 above experiments. 



2. The author tentatively suggests the following hypothesis 

 as one which seems to fit most of the experimental facts. 



It is well known that electrically neutral bodies — possibly 

 doublets formed by the union of positive and negative ions — 

 are present in discharge-tubes at low pressures : evidence 



