﻿Asymmetry 
  of 
  the 
  Zeeman 
  Effect* 
  251 
  

  

  and 
  the 
  fraction 
  of 
  the 
  separation 
  is 
  

  

  1 
  eTLim^ 
  + 
  m.y 
  

  

  4=m 
  1 
  m 
  2 
  (m 
  l 
  — 
  m2) 
  

  

  On 
  both 
  views, 
  then, 
  the 
  actual 
  difference 
  of 
  displacement 
  

   is 
  greater, 
  other 
  things 
  being 
  equal, 
  at 
  the 
  red 
  end. 
  But 
  the 
  

   fraction 
  which 
  the 
  asymmetry 
  is 
  of 
  the 
  separation 
  would 
  be 
  

   greater 
  at 
  the 
  violet 
  end 
  of 
  the 
  spectrum 
  on 
  Voigt's 
  theory, 
  

   and 
  greater 
  at 
  the 
  red 
  end 
  on 
  the 
  present 
  theory. 
  This 
  has 
  

   an 
  important 
  bearing 
  on 
  the 
  experimental 
  detection 
  of 
  the 
  

   effect. 
  

  

  We 
  may 
  further 
  consider 
  the 
  probable 
  numerical 
  values. 
  

   , 
  The 
  quantity 
  e 
  l 
  is 
  connected 
  with 
  the 
  molecule 
  in 
  such 
  a 
  

   w 
  r 
  ay 
  that 
  the 
  dielectric 
  constant 
  

  

  K-l+S^ 
  

   the 
  summation 
  referring 
  to 
  all 
  the 
  molecules 
  : 
  e 
  x 
  is 
  thus 
  a 
  

   quantity 
  of 
  order 
  about 
  10" 
  25 
  , 
  and 
  if 
  3 
  = 
  4 
  x 
  10~ 
  16 
  , 
  

  

  f 
  is 
  of 
  order 
  4x 
  I0" 
  41 
  . 
  

  

  If, 
  as 
  a 
  variety 
  of 
  experimental 
  work 
  indicates, 
  m 
  2 
  is 
  very 
  

   small 
  compared 
  with 
  m 
  l3 
  the 
  formula 
  (A) 
  becomes 
  

  

  2m 
  2 
  ° 
  8 
  m 
  2 
  2 
  ° 
  

  

  eB. 
  

  

  For 
  a 
  field 
  of 
  10 
  4 
  C.Gr.S. 
  units 
  we 
  may 
  take 
  — 
  about 
  10 
  n 
  ; 
  

  

  m 
  2 
  

  

  therefore 
  i— 
  A 
  2 
  is 
  about 
  8 
  X 
  10~ 
  21 
  , 
  

  

  2 
  m 
  2 
  

  

  1 
  e 
  2 
  H 
  2 
  

  

  and 
  ^ 
  —a- 
  V 
  is 
  about 
  8 
  X 
  10~ 
  26 
  . 
  

  

  8 
  m 
  2 
  2 
  

  

  Thus 
  f 
  appears 
  to 
  be 
  very 
  small 
  compared 
  with 
  

  

  le 
  2 
  H 
  2 
  

  

  m 
  2 
  2 
  ~* 
  ' 
  

  

  S 
  

  

  and 
  the 
  latter 
  term 
  might 
  just 
  come 
  within 
  measurable 
  

   amount 
  by 
  increasing 
  the 
  field. 
  

  

  One 
  more 
  point 
  in 
  connexion 
  with 
  the 
  proposed 
  view 
  may 
  

   be 
  noted, 
  which 
  is 
  that 
  it 
  provides 
  an 
  explanation 
  of 
  why 
  a 
  

   line 
  may 
  not 
  be 
  resolvable. 
  If 
  m-^ 
  — 
  m^ 
  there 
  would 
  be 
  no 
  

   doubling 
  of 
  the 
  line, 
  but 
  only 
  a 
  small 
  shift 
  towards 
  the 
  violet. 
  

   In 
  this 
  case 
  the 
  system 
  is 
  dynamically 
  symmetrical. 
  

  

  It 
  is 
  almost 
  unnecessary 
  to 
  remark 
  that 
  the 
  simple 
  system 
  

   selected 
  for 
  discussion 
  is 
  merely 
  illustrative 
  of 
  the 
  charac- 
  

   teristic 
  features 
  of 
  the 
  problem. 
  

  

  