234 Barss — Ion ization by Collision in Gases and Vapors. 









Table I. 











Pressure, 4-17""". 





X 



a (I 



, 2"") 



a (L 1="') 



a (L 0-5 cm ) 



120 



•13 



(•13) 







160 



•30 



(•28) 



•26 ( -30) 





200 



•50 



(•50) 



•48 ( -51) 





240 



1-08 





1-14 ( '99) 



1-16 



320 







2-06 (2-1 ) 



2-20 (2-2) 



400 







Table II. 

 Pressure, 2-09 mm . 



3-80 (3-6) 



X 



a (L 2<™) 



a (L 1"") 



a (L 0-5 tm ) 



80 



•15 



(•13) 







120 



•38 



(•42) 



•42 ( -40) 





160 



•87 



(•90) 



•89 ( -90) 



•92 



240 



1-63 





1-61 (1-60) 



1-55 



320 







2-50 (2-35) 



2-52 (2-35 



for the same field v strength, the same distances between the 

 plates and pressures of 4-13 mm and 2-12 mm . 



It will be noticed that the curve corresponding to a distance 

 of 2 centimeters between the plates rises more abruptly than 

 the curve corresponding to 1 centimeter ; and in the curve cor- 

 responding to half a centimeter the current rises more slowly 

 with the field strength than in either of the other cases. In 

 other words, there is a divergence of the curves for different 

 distances when the pressure is constant. When the plates are 

 close together the mass of gas to be ionized is less, and we 

 should expect, according to Townsend's theory, a smaller col- 

 lision effect than when the plates are separated to a greater 

 distance. 



The next experimental results were obtained with hydrogen. 

 The hydrogen employed was prepared by the chemical action 

 of hydrochloric acid on zinc. It was then dried by being 

 passed through tubes containing calcium chloride and phos- 

 phorous pentoxide, and further purified by the action of liquid 

 air and charcoal. A series of curves was obtained at relatively 

 high pressures (19-6 cm , 15-5 cm , 10 em , and 4-2 cm ), which agreed 

 very well with those obtained by Bishop, who used ultra-violet 

 light as the ionizing agent. The collision effect appears to set 

 in at a smaller field strength ; this probably points to the fact 

 that in the present case a slightly higher degree of purity was 

 obtained. 



Curves were also obtained with different distances between 

 the plates, the pressure being constant. The values of a calcu- 

 lated from these curves agreed very well among themselves, 

 the greatest deviation being about 5 per cent. 



