﻿Interferometer 
  for 
  the 
  Study 
  of 
  Band 
  Spectra. 
  177 
  

  

  spacing 
  of 
  the 
  lines 
  in 
  the 
  spectrum 
  is 
  fortuitous, 
  the 
  fringes, 
  

   which 
  break 
  each 
  line 
  up 
  into 
  a 
  vertical 
  row 
  of 
  dots, 
  cannot 
  

   possibly 
  form 
  collectively 
  a 
  symmetrical 
  pattern. 
  Curved 
  

   patterns 
  were, 
  however, 
  seen 
  in 
  portions 
  of 
  the 
  spectrum, 
  

   which 
  indicated 
  a 
  definite 
  law 
  of 
  spacing 
  in 
  the 
  case 
  of 
  the 
  

   line 
  groups 
  upon 
  which 
  the 
  patterns 
  appeared. 
  It 
  has 
  been 
  

   my 
  intention, 
  ever 
  since 
  this 
  observation 
  was 
  made, 
  to 
  make 
  

   a 
  further 
  study 
  of 
  the 
  subject, 
  but 
  pressure 
  of 
  other 
  work, 
  in 
  

   combination 
  with 
  the 
  hope 
  that 
  someone 
  else 
  would 
  clear 
  the 
  

   matter 
  up, 
  has 
  caused 
  me 
  repeatedly 
  to 
  postpone 
  a 
  return 
  to 
  

   the 
  problem. 
  

  

  During 
  the 
  past 
  winter 
  I 
  have, 
  however, 
  taken 
  up 
  the 
  

   matter 
  again, 
  and 
  by 
  studying 
  the 
  phenomenon 
  in 
  the 
  case 
  

   of 
  band 
  spectra, 
  where 
  there 
  is 
  present 
  a 
  known 
  law 
  of 
  

   spacing, 
  have 
  reached 
  a 
  very 
  definite 
  conclusion 
  as 
  to 
  the 
  

   manner 
  in 
  which 
  the 
  symmetrical 
  patterns 
  are 
  produced, 
  and 
  

   to 
  what 
  extent 
  they 
  may 
  be 
  used 
  for 
  the 
  study 
  of 
  spectrum 
  

  

  regularities. 
  

  

  The 
  bright 
  and 
  dark 
  bands 
  which 
  form 
  the 
  patterns 
  appear 
  

   to 
  be 
  of 
  a 
  type 
  of 
  interference 
  fringes 
  hitherto 
  undescribed, 
  

   though 
  they 
  are 
  of 
  the 
  same 
  nature 
  as 
  the 
  shadow 
  bands 
  

   which 
  appear 
  when 
  any 
  simple 
  periodic 
  structure 
  is 
  viewed 
  

   through 
  a 
  similar 
  superposed 
  structure. 
  By 
  their 
  aid 
  we 
  can 
  

   show 
  periodic 
  discontinuities 
  in 
  a 
  spectrum 
  which 
  are 
  beyond 
  

   the 
  resolving 
  power 
  of 
  the 
  spectroscope, 
  just 
  as 
  we 
  can 
  be 
  

   sure 
  that 
  a 
  distant 
  roof 
  is 
  tiled 
  if 
  shadow 
  bands 
  appear 
  when 
  it 
  

   is 
  viewed 
  through 
  the 
  meshes 
  of 
  the 
  fiy-screen 
  in 
  the 
  window,, 
  

   even 
  though 
  we 
  cannot 
  see 
  the 
  individual 
  tiles. 
  

  

  The 
  manner 
  in 
  which 
  the 
  bands 
  are 
  formed 
  is 
  as 
  follows. 
  

  

  If 
  we 
  employ 
  the 
  crater 
  of 
  the 
  carbon 
  arc 
  as 
  our 
  source 
  of 
  

   light 
  and 
  locus 
  the 
  image 
  of 
  the 
  circular 
  ring 
  system 
  formed 
  

   by 
  passing 
  tiie 
  light 
  through 
  a 
  Fabry 
  and 
  Perot 
  interfero- 
  

   meter, 
  upon 
  the 
  slit 
  of 
  a 
  spectroscope, 
  we 
  find 
  the 
  continuous 
  

   spectrum 
  traversed 
  by 
  vertical 
  dark 
  arcs, 
  convex 
  towards 
  the 
  

   red 
  end, 
  which 
  are 
  the 
  loci 
  of 
  wave-lengths 
  refused 
  trans- 
  

   mission 
  through 
  the 
  half-silvered 
  plates 
  of 
  the 
  interferometer. 
  

   If 
  the 
  centre 
  of 
  the 
  ring 
  system 
  is 
  at 
  the 
  centre 
  of 
  the 
  slit, 
  

   the 
  portions 
  of 
  the 
  slit 
  above 
  and 
  below 
  the 
  centre 
  are 
  illumi- 
  

   minated 
  by 
  rays 
  which 
  have 
  traversed 
  the 
  air-film 
  between 
  

   the 
  plates 
  in 
  an 
  oblique 
  direction, 
  and 
  have 
  in 
  consequence 
  a 
  

   smaller 
  path 
  difference, 
  the 
  formula 
  for 
  the 
  path 
  difference 
  

   being 
  e 
  = 
  2T 
  cos 
  0. 
  in 
  which 
  T 
  is 
  the 
  thickness 
  of 
  the 
  film, 
  

   and 
  6 
  the 
  angle 
  which 
  the 
  rays 
  make 
  with 
  the 
  normal. 
  

  

  A- 
  the 
  plates 
  of 
  the 
  interferometer 
  are 
  separated, 
  the 
  dark 
  

   arcs 
  become 
  narrower 
  and 
  more 
  numerous, 
  until 
  finallv 
  the 
  

  

  Phil. 
  Mag. 
  S. 
  6. 
  Vol. 
  2G. 
  No. 
  151. 
  July 
  1913. 
  X 
  

  

  