﻿produced 
  in 
  Gases 
  by 
  Ultra-Violet 
  Light. 
  561 
  

  

  observations 
  were 
  thus 
  made 
  with 
  different 
  distances 
  between 
  

   the 
  plates, 
  the 
  electric 
  force 
  and 
  the 
  pressure 
  being 
  kept 
  

   constant. 
  These 
  experiments 
  were 
  repeated 
  alternately, 
  and 
  

   the 
  mean 
  of 
  the 
  observations 
  was 
  taken. 
  As 
  a 
  rule 
  the 
  

   observations 
  did 
  not 
  vary 
  by 
  more 
  than 
  4 
  or 
  5 
  per 
  cent., 
  so 
  

   that 
  considerable 
  accuracy 
  could 
  be 
  obtained. 
  

  

  It 
  is 
  necessary 
  to 
  apply 
  a 
  small 
  correction 
  to 
  the 
  electro- 
  

   meter 
  readingsowing 
  to 
  the 
  increase 
  of 
  capacity 
  which 
  occurs 
  

   when 
  the 
  plate 
  B 
  is 
  moved 
  up 
  towards 
  A. 
  The 
  effect 
  of 
  

   moving 
  B 
  from 
  a 
  position 
  5 
  millimetres 
  from 
  A 
  to 
  1 
  milli- 
  

   metre 
  from 
  A 
  was 
  to 
  increase 
  the 
  capacity 
  of 
  the 
  apparatus 
  

   connected 
  to 
  the 
  quadrants 
  (including 
  the 
  quadrants) 
  in 
  the 
  

   ratio 
  of 
  82 
  : 
  80. 
  The 
  observations 
  at 
  the 
  latter 
  distance 
  

   were 
  increased 
  by 
  2J 
  per 
  cent., 
  and 
  intermediate 
  observations 
  

   by 
  a 
  smaller 
  factor, 
  in 
  order 
  that 
  the 
  observations 
  should 
  be 
  

   proportional 
  to 
  the 
  current 
  between 
  the 
  plates. 
  

  

  In 
  most 
  cases 
  three 
  conductivities, 
  n 
  1? 
  n 
  2y 
  and 
  w 
  3 
  , 
  were 
  

   found 
  for 
  three 
  distances, 
  d 
  h 
  d 
  2 
  , 
  and 
  d 
  3 
  , 
  between 
  the 
  plates, 
  

   the 
  electric 
  force 
  and 
  pressure 
  being 
  constant. 
  The 
  distances 
  

   were 
  such 
  that 
  d 
  2 
  —d 
  1 
  = 
  d^—d 
  2 
  . 
  

  

  According 
  to 
  the 
  formula, 
  n 
  = 
  n 
  Q 
  e 
  ad 
  , 
  the 
  corresponding 
  

  

  conductivities 
  satisfy 
  the 
  condition 
  — 
  = 
  — 
  when 
  the 
  increase 
  

  

  ni 
  n 
  2 
  

  

  in 
  conductivity 
  arises 
  from 
  the 
  action 
  of 
  the 
  negative 
  ions* 
  

  

  By 
  adopting 
  this 
  method 
  of 
  experimenting 
  it 
  is 
  possible 
  to 
  

  

  guard 
  against 
  using 
  forces 
  which 
  are 
  too 
  high, 
  and 
  which 
  

  

  give 
  rise 
  to 
  conductivities 
  so 
  large 
  that 
  they 
  cannot 
  be 
  

  

  attributed 
  to 
  this 
  process 
  of 
  ionization 
  alone. 
  

  

  In 
  the 
  ordinary 
  discharge 
  arising 
  from 
  a 
  high 
  potential- 
  

   difference 
  between 
  the 
  electrodes, 
  there 
  must 
  be 
  at 
  least 
  two 
  

   processes 
  of 
  ionization 
  at 
  work. 
  The 
  genesis 
  of 
  ions 
  by 
  the 
  

   motion 
  of 
  negative 
  ions 
  alone 
  cannot 
  explain 
  the 
  phenomenon 
  

   of 
  a 
  continuous 
  discharge. 
  It 
  is 
  evident, 
  therefore, 
  that 
  some 
  

   other 
  cause 
  is 
  at 
  work, 
  and 
  these 
  experiments 
  show 
  that 
  it 
  

   begins 
  to 
  take 
  effect 
  at 
  potentials 
  much 
  lower 
  than 
  the 
  poten- 
  

   tial 
  necessary 
  to 
  produce 
  a 
  discharge. 
  With 
  a 
  suitable 
  

   pressure, 
  it 
  was 
  found 
  that 
  the 
  ratios 
  of 
  the 
  conductivities 
  

   corresponding 
  to 
  equal 
  increments 
  of 
  the 
  distance 
  d 
  con- 
  

   tinually 
  rose 
  as 
  the 
  distance 
  between 
  the 
  plates 
  was 
  increased, 
  

   although 
  the 
  potential-difference 
  between 
  the 
  plates 
  was 
  less 
  

   than 
  300 
  volts. 
  With 
  hydrogen 
  this 
  effect 
  is 
  obtained 
  with 
  

   potentials 
  less 
  than 
  200 
  volts. 
  

  

  I 
  have 
  as 
  yet 
  made 
  only 
  a 
  very 
  incomplete 
  examination 
  of 
  

   this 
  second 
  process 
  of 
  ionization, 
  and 
  in 
  the 
  following 
  tables 
  

   only 
  a 
  few 
  examples 
  of 
  it 
  are 
  given. 
  

  

  