bjj the Motion of J\)sitici' and Negative Ions. 3(51 



the plates have shown that the properties of gases are the 

 same at one niilli metre from a plate as tliey are at two centi- 

 metres, we see that the eatalytio action of a metal extends 

 nndiminished from a millimeti-e to two centimetres at least. 

 Dr. Stark's view therefore leads to a most imi)robable con- 

 clusion. It is difficult to believe in the peculiar influence 

 which a metal is thus supposed to exercise, and as far as I 

 can see the ex[)erimental evidence is altogether opposed 

 to it. 



The one passnge in my paper which refers to Dr. Stark's 

 work is the followino; : '^Accordino- to Stark, a neo-ative ion 

 must travel between two points differing in potential by 

 50 volts in order to acquire sufficient velocity to produce new 

 ions on collision, because there is a fall of potential of 

 50 volts near the anode when a continuous discharge takes 

 place in air.''"' To what extent I have thus misrepresented 

 his work may be gathered from the following two passages 

 which occur in his paper, Annalen der Pliysik [4] A'oL viii. 

 (1902). On p- 81() the following statement occurs: "die 

 lonisirungsspannung des positiven Ions gegen Platin in Luft 

 betragt 340 Volt, diejenige des negativen im Luftinnern 

 etwa 50 Yolt.^'' And on p. 819: "In der erwahnten 

 Abhandluijg wurde dargelegt, dass die auf einer positiven 

 Schichtlange liegende Spannungsdiff'erenz eine obere Grenze 

 der lonisirungsspannung der negativen lonen im Gasinnern 

 (in Luft 50, in Stickstoff 45, in Wasserstoff 33 Volt) ist, dass 

 ferner der normale Kathodenfall die lonisirungsspannung der 

 positiven lonen gegen das Kathodenmetall darstellt.^^ 



I maintain that the voltages 340 and 50 are both too large, 

 and that the views w^hich Dr. Stark holds with regard to falls 

 of potential either at the electrodes or in layers near the 

 electrodes are inadequate proofs of his suppositions. 



With regard to the potential-difference required to produce 

 ions by collision, I have been led to my conclusions from expe- 

 riments which do not admit of interpretations in favour of any 

 particular value. The calculations which I have given show 

 that ionization by collision takes place in a small percentage 

 of cases when negative ions collide with molecules with a 

 velocity acquired by a fall of potential of 5 volts. It may 

 be seen from the paper in the ' Electrician ' to which I have 

 referred that the ionization produced by positive ions can 

 easily be detected with a potential-difference of 280 volts 

 between the electrodes. 



Yours very truly, 



John S. Townsend. 



