﻿914 
  Dr. 
  Norman 
  Campbell 
  on 
  a 
  

  

  Therefore 
  

  

  b 
  = 
  J 
  -=i 
  . 
  . 
  . 
  . 
  . 
  . 
  . 
  . 
  (14) 
  

  

  and 
  

  

  • 
  i 
  2X 
  htC\ 
  

  

  I 
  =1 
  -y 
  5 
  ( 
  15 
  ) 
  

  

  where 
  

  

  * 
  = 
  ~ 
  k 
  (16) 
  

  

  It 
  is 
  to 
  be 
  observed 
  that 
  the 
  assumptions 
  made 
  are 
  not 
  

   mutually 
  inconsistent, 
  for 
  (14) 
  shows 
  that 
  in 
  fact 
  b 
  is 
  small 
  

   when 
  the 
  current 
  is 
  nearly 
  saturated. 
  It 
  is 
  not 
  pretended, 
  

   however, 
  that 
  the 
  calculation 
  is 
  rigidly 
  justifiable, 
  but 
  it 
  

   seems 
  to 
  lead 
  to 
  results 
  in 
  accordance 
  with 
  observation. 
  It 
  

   must 
  be 
  remarked 
  that 
  two 
  conditions 
  must 
  be 
  fulfilled 
  if 
  

   (15) 
  is 
  to 
  be 
  applicable. 
  In 
  the 
  first 
  place 
  the 
  effects 
  of 
  

   recombination 
  must 
  be 
  negligible, 
  that 
  is 
  to 
  say, 
  an 
  1 
  n 
  2 
  must 
  

   always 
  be 
  small 
  compared 
  with 
  q. 
  If 
  we 
  take 
  the 
  values 
  of 
  

   n^n 
  2 
  from 
  (7) 
  and 
  ^8) 
  and 
  observe 
  that 
  n 
  Y 
  n 
  2 
  is 
  a 
  maximum 
  

   when 
  a 
  = 
  0, 
  the 
  necessary 
  condition 
  is 
  found 
  to 
  be 
  that 
  

  

  j(~) 
  — 
  is 
  small 
  compared 
  with 
  unity. 
  . 
  (A) 
  

  

  In 
  the 
  second 
  place 
  2irbl 
  2 
  must 
  be 
  small 
  compared 
  with 
  X, 
  

   that 
  is 
  to 
  say, 
  the 
  necessary 
  condition 
  is 
  that 
  

  

  rysj~ 
  ' 
  1S 
  smal 
  l 
  compared 
  with 
  unity. 
  . 
  (B) 
  

  

  (15), 
  if 
  it 
  is 
  true 
  in 
  any 
  possible 
  experimental 
  conditions, 
  

   represents 
  rather 
  a 
  remarkable 
  result. 
  For 
  a 
  well-known 
  

   argument 
  due 
  to 
  Townsend 
  shows 
  that 
  X 
  is 
  a 
  universal 
  

   function 
  of 
  the 
  temperature 
  of 
  the 
  gas 
  and 
  the 
  ionic 
  charge. 
  

  

  p 
  

   It 
  is 
  equal 
  to 
  <==- 
  3 
  where 
  P 
  is 
  the 
  pressure 
  of 
  a 
  gas 
  containing 
  

  

  N 
  molecules 
  per 
  cm. 
  3 
  ; 
  it 
  has 
  the 
  dimensions 
  of 
  a 
  potential 
  

   and 
  the 
  numerical 
  value 
  O0248 
  volt 
  at 
  16° 
  C. 
  Accordingly, 
  

   in 
  the 
  circumstances 
  to 
  which 
  (15) 
  applies, 
  i/I 
  should 
  depend 
  

   only 
  on 
  the 
  potential 
  between 
  the 
  plates 
  and 
  the 
  temperature 
  

   and 
  should 
  be 
  independent 
  of 
  the 
  pressure 
  or 
  nature 
  of 
  the 
  

   gas, 
  of 
  intensity 
  of 
  the 
  ionization 
  and 
  of 
  the 
  distance 
  between 
  

   the 
  plates. 
  

  

  