﻿514 
  The 
  R6le 
  of 
  Water 
  Vapour 
  in 
  Gaseous 
  < 
  induction. 
  

  

  the 
  capillary 
  where 
  sodium 
  was 
  most 
  abundant, 
  the 
  hydrogen 
  

   spectra, 
  both 
  simple 
  and 
  compound, 
  were 
  weak 
  ; 
  on 
  cooling 
  

   the 
  tube, 
  they 
  both 
  increased 
  in 
  intensity 
  and 
  showed 
  no 
  signs 
  

   of 
  approaching 
  dissolution. 
  The 
  tube 
  was 
  finally 
  accidentally 
  

   broken, 
  and 
  on 
  being 
  immediately 
  examined 
  it 
  was 
  found 
  

   that 
  there 
  were 
  large 
  patches 
  of 
  sodium 
  with 
  clean 
  metallic 
  

   lustre. 
  Whatever 
  other 
  impurities 
  might 
  have 
  been 
  there, 
  

   it 
  is 
  difficult 
  to 
  see 
  how 
  any 
  free 
  water-vapour 
  or 
  oxygen 
  

   could 
  have 
  been 
  present. 
  Moreover, 
  as 
  the 
  hydrogen 
  

   spectrum 
  had 
  its 
  full 
  intensity 
  at 
  the 
  first 
  flash 
  through 
  the 
  

   tube, 
  it 
  seems 
  improbable 
  that 
  the 
  current, 
  first 
  passing- 
  

   through 
  the 
  sodium-vapour, 
  could 
  have 
  set 
  free 
  oxygen 
  from 
  

   any 
  of 
  its 
  compounds 
  present, 
  thus 
  enabling 
  hydrogen 
  to 
  take 
  

   subsequent 
  part 
  in 
  the 
  conduction. 
  

  

  Some 
  experiments 
  by 
  Warburg* 
  appear 
  to 
  have 
  some 
  

   bearing 
  on 
  this 
  matter. 
  He 
  found 
  that 
  in 
  a 
  hydrogen 
  tube 
  

   containing 
  some 
  pure 
  electrolytically 
  prepared 
  metallic 
  sodium, 
  

   the 
  cathode 
  fall 
  of 
  potential 
  had 
  a 
  minimum 
  value 
  of 
  168 
  

   volts. 
  With 
  measurable 
  traces 
  of 
  oxygen 
  present 
  the 
  fall 
  

   was 
  240 
  volts, 
  gradually 
  falling 
  to 
  a 
  minimum 
  of 
  about 
  200 
  

   volts 
  as 
  the 
  oxygen, 
  after 
  combination 
  with 
  the 
  hydrogen, 
  

   was 
  removed 
  by 
  the 
  ordinary 
  drying-tubes. 
  Only 
  after 
  the 
  

   more 
  nearly 
  perfect 
  drying 
  produced 
  by 
  metallic 
  sodium 
  was 
  

   the 
  lower 
  minimum 
  reached. 
  This 
  indicates 
  that, 
  so 
  far 
  as 
  the 
  

   processes 
  at 
  the 
  cathode 
  are 
  concerned 
  (and 
  here 
  the 
  current 
  

   meets 
  with 
  the 
  greatest 
  opposition), 
  the 
  current 
  can 
  pass 
  with 
  

   more 
  facility 
  through 
  perfectly 
  dry 
  hydrogen 
  than 
  through 
  

   that 
  containing 
  traces 
  of 
  oxygen 
  or 
  water-vapour. 
  In 
  the 
  

   case 
  of 
  nitrogen, 
  Warburg 
  found 
  that 
  small 
  traces 
  of 
  these 
  

   impurities 
  lowered 
  the 
  cathode 
  fall, 
  but 
  in 
  dry 
  nitrogen 
  it 
  had 
  

   a 
  definite 
  upper 
  limit. 
  The 
  writer 
  has 
  used 
  one 
  of 
  Professor 
  

   Warburg's 
  nitrogen 
  tubes, 
  and 
  although 
  the 
  gas 
  had 
  been 
  in 
  

   contact 
  with 
  sodium 
  for 
  about 
  ten 
  years, 
  the 
  current 
  passed 
  

   with 
  ease 
  and 
  gave 
  the 
  ordinary 
  spectrum 
  of 
  nitrogen. 
  

  

  It 
  certainly 
  seems 
  imposs'ble 
  to 
  eliminate 
  all 
  traces 
  of 
  

   impurity 
  from 
  vacuum-tubes, 
  and 
  it 
  may 
  be 
  that 
  one 
  or 
  more 
  

   such 
  impurities 
  may 
  take 
  a 
  very 
  active 
  and 
  essential 
  part 
  in 
  

   the 
  processes 
  which 
  transform 
  a 
  gas 
  into 
  a 
  conductor 
  ; 
  but 
  

   if, 
  under 
  the 
  conditions 
  referred 
  to, 
  comparatively 
  small 
  

   electromotive 
  forces 
  can 
  drive 
  a 
  current 
  through 
  hydrogen 
  

   or 
  nitrogen, 
  it 
  seems 
  at 
  least 
  doubtful 
  whether 
  water-vapour 
  

   is 
  necessary 
  to 
  gaseous 
  conduction, 
  t 
  

  

  University 
  of 
  California, 
  Berkeley, 
  

   January 
  1902. 
  

  

  * 
  Warburg, 
  Wied. 
  Ann. 
  xl. 
  p. 
  1 
  (1890). 
  

  

  